[libvirt] [PATCH v3 00/22] scripts: convert most perl scripts to python
This series is an effort to reduce the number of different languages we use by eliminating most use of perl in favour of python. This aligns with fact that the likely future build system we'll use (meson) is written in python, and that python is much more commonly used/understood by developers these days than perl. With this applied we use perl in a handful of places only: - src/rpc/gendispatch.pl - this is a horrendously large script and very hard to understand/follow. A straight syntax conversion to Python would still leave a hgue and hard to understand/follow script. It really needs a clean room rewrite from scratch, with better structure. - src/rpc/genprotocol.pl - fairly easy to convert, but might be obsolete depending on approach for rewriting gendispatch.pl, so ignored for now - tests/oomtrace.pl - will be purge by the patches that drop OOM handling anyway - tools/wireshark/util/genxdrstub.pl - a very large script, which I haven't got the courage to tackle yet. - cfg.mk/maint.mk - many syntax rules involve regexes which are fed to perl. Decision on what to do with syntax-check rules punted to another time. - build-aux/gitlog-to-changelog - build-aux/useless-if-before-free - Both pulled in from gnulib. Could be rewritten quite easily if desired, but given that we aren't maintainers of them right now, they're ignored as they don't really impact our developers. In v3: - All scripts comply with all flake8 style rules with exception of E129 visually indented line with same indent as next logical line In v2: - Pulled in patch to hacking file - Converted many more scripts - Forced UTF-8 character set to avoid ascii codec on py3 < 3.7 Daniel P. Berrangé (22): docs: document that C & Python are the preferred languages build-aux: rewrite augest test generator in Python build-aux: rewrite po file minimizer in Python build-aux: rewrite duplicate header checker in Python build-aux: rewrite whitespace checker in Python build-aux: rewrite mock inline checker in Python build-aux: rewrite header ifdef checker in Python src: rewrite ACL permissions checker in Python src: rewrite symfile sorting checker in Python src: rewrite symfile library checker in Python src: rewrite systemtap probe generator in Python src: rewrite systemtap function generator in Python src: rewrite driver name checker in Python src: rewrite driver impl checker in Python src: rewrite ACL rule checker in Python src: rewrite polkit ACL generator in Python src: rewrite remote protocol checker in Python tests: rewrite test argv line wrapper in Python tests: rewrite qemu capability grouper in Python tests: rewrite file access checker in Python docs: rewrite hvsupport.html page generator in python docs: rewrite polkit docs generator in Python Makefile.am | 12 +- build-aux/augeas-gentest.pl | 60 --- build-aux/augeas-gentest.py | 72 ++++ build-aux/check-spacing.pl | 198 ---------- build-aux/check-spacing.py | 229 +++++++++++ build-aux/header-ifdef.pl | 182 --------- build-aux/header-ifdef.py | 231 +++++++++++ build-aux/minimize-po.pl | 37 -- build-aux/minimize-po.py | 60 +++ build-aux/mock-noinline.pl | 75 ---- build-aux/mock-noinline.py | 89 +++++ build-aux/prohibit-duplicate-header.pl | 26 -- build-aux/prohibit-duplicate-header.py | 57 +++ cfg.mk | 30 +- docs/Makefile.am | 14 +- docs/genaclperms.pl | 125 ------ docs/genaclperms.py | 123 ++++++ docs/hacking.html.in | 30 ++ docs/hvsupport.pl | 458 ---------------------- docs/hvsupport.py | 516 +++++++++++++++++++++++++ po/Makefile.am | 2 +- src/Makefile.am | 156 +++----- src/access/Makefile.inc.am | 7 +- src/access/genpolkit.pl | 119 ------ src/access/genpolkit.py | 122 ++++++ src/bhyve/Makefile.inc.am | 4 +- src/check-aclperms.pl | 73 ---- src/check-aclperms.py | 78 ++++ src/check-aclrules.pl | 252 ------------ src/check-aclrules.py | 263 +++++++++++++ src/check-driverimpls.pl | 80 ---- src/check-driverimpls.py | 102 +++++ src/check-drivername.pl | 83 ---- src/check-drivername.py | 114 ++++++ src/check-remote-protocol.py | 137 +++++++ src/check-symfile.pl | 70 ---- src/check-symfile.py | 81 ++++ src/check-symsorting.pl | 106 ----- src/check-symsorting.py | 119 ++++++ src/dtrace2systemtap.pl | 130 ------- src/dtrace2systemtap.py | 143 +++++++ src/interface/Makefile.inc.am | 2 +- src/libxl/Makefile.inc.am | 4 +- src/locking/Makefile.inc.am | 6 +- src/logging/Makefile.inc.am | 2 +- src/lxc/Makefile.inc.am | 4 +- src/network/Makefile.inc.am | 2 +- src/node_device/Makefile.inc.am | 2 +- src/nwfilter/Makefile.inc.am | 2 +- src/qemu/Makefile.inc.am | 4 +- src/remote/Makefile.inc.am | 4 +- src/rpc/Makefile.inc.am | 2 +- src/rpc/gensystemtap.pl | 193 --------- src/rpc/gensystemtap.py | 184 +++++++++ src/secret/Makefile.inc.am | 2 +- src/storage/Makefile.inc.am | 2 +- src/vbox/Makefile.inc.am | 2 +- src/vz/Makefile.inc.am | 2 +- tests/Makefile.am | 4 +- tests/check-file-access.pl | 126 ------ tests/check-file-access.py | 123 ++++++ tests/file_access_whitelist.txt | 2 +- tests/group-qemu-caps.pl | 124 ------ tests/group-qemu-caps.py | 123 ++++++ tests/test-wrap-argv.pl | 174 --------- tests/test-wrap-argv.py | 170 ++++++++ tests/testutils.c | 16 +- 67 files changed, 3283 insertions(+), 2863 deletions(-) delete mode 100755 build-aux/augeas-gentest.pl create mode 100755 build-aux/augeas-gentest.py delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py delete mode 100644 build-aux/header-ifdef.pl create mode 100644 build-aux/header-ifdef.py delete mode 100755 build-aux/minimize-po.pl create mode 100755 build-aux/minimize-po.py delete mode 100644 build-aux/mock-noinline.pl create mode 100644 build-aux/mock-noinline.py delete mode 100644 build-aux/prohibit-duplicate-header.pl create mode 100644 build-aux/prohibit-duplicate-header.py delete mode 100755 docs/genaclperms.pl create mode 100755 docs/genaclperms.py delete mode 100755 docs/hvsupport.pl create mode 100755 docs/hvsupport.py delete mode 100755 src/access/genpolkit.pl create mode 100755 src/access/genpolkit.py delete mode 100755 src/check-aclperms.pl create mode 100755 src/check-aclperms.py delete mode 100755 src/check-aclrules.pl create mode 100755 src/check-aclrules.py delete mode 100755 src/check-driverimpls.pl create mode 100755 src/check-driverimpls.py delete mode 100755 src/check-drivername.pl create mode 100644 src/check-drivername.py create mode 100644 src/check-remote-protocol.py delete mode 100755 src/check-symfile.pl create mode 100755 src/check-symfile.py delete mode 100755 src/check-symsorting.pl create mode 100755 src/check-symsorting.py delete mode 100755 src/dtrace2systemtap.pl create mode 100755 src/dtrace2systemtap.py delete mode 100755 src/rpc/gensystemtap.pl create mode 100755 src/rpc/gensystemtap.py delete mode 100755 tests/check-file-access.pl create mode 100755 tests/check-file-access.py delete mode 100755 tests/group-qemu-caps.pl create mode 100755 tests/group-qemu-caps.py delete mode 100755 tests/test-wrap-argv.pl create mode 100755 tests/test-wrap-argv.py -- 2.21.0
Blacklist Perl and Shell code in favour of Python for sake of readability and portability. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- docs/hacking.html.in | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) diff --git a/docs/hacking.html.in b/docs/hacking.html.in index 90376968c3..edf2f54ce3 100644 --- a/docs/hacking.html.in +++ b/docs/hacking.html.in @@ -343,6 +343,36 @@ Richard Jones' guide to working with open source projects</a>. </p> + <h2><a id="lang">Language Usage</a></h2> + + <p> + The libvirt repository makes use of a large number of programming + languages. It is anticipated that in the future libvirt will adopt + use of other new languages. To reduce the overall burden on developers, + there is thus a general desire to phase out usage of some of the + existing languages. + </p> + + <p> + The preferred languages at this time are: + </p> + + <ul> + <li>C - for the main libvirt codebase. Dialect supported by + GCC/CLang only.</li> + <li>Python - for supporting build scripts / tools. Code must + run with both version 2.7 and 3.x at this time.</li> + </ul> + + <p> + Languages that should not be used for any new contributions: + </p> + + <ul> + <li>Perl - build scripts must be written in Python instead.</li> + <li>Shell - build scripts must be written in Python instead.</li> + </ul> + <h2><a id="tooling">Tooling</a></h2> <p> -- 2.21.0
On Tue, Sep 24, 2019 at 03:58:42PM +0100, Daniel P. Berrangé wrote:
Blacklist Perl and Shell code in favour of Python for sake of readability and portability.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- docs/hacking.html.in | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+)
Reviewed-by: Ján Tomko <jtomko@redhat.com> Jano
As part of an goal to eliminate Perl from libvirt build tools, rewrite the augeas-gentest.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. The use of $(AUG_GENTEST) as a dependancy in the makefiles needed to be fixed, because this was assumed to be the filename of the script, but is in fact a full shell command line. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/augeas-gentest.pl | 60 --------------------------- build-aux/augeas-gentest.py | 72 +++++++++++++++++++++++++++++++++ src/Makefile.am | 3 +- src/bhyve/Makefile.inc.am | 4 +- src/interface/Makefile.inc.am | 2 +- src/libxl/Makefile.inc.am | 4 +- src/locking/Makefile.inc.am | 6 +-- src/logging/Makefile.inc.am | 2 +- src/lxc/Makefile.inc.am | 4 +- src/network/Makefile.inc.am | 2 +- src/node_device/Makefile.inc.am | 2 +- src/nwfilter/Makefile.inc.am | 2 +- src/qemu/Makefile.inc.am | 4 +- src/remote/Makefile.inc.am | 4 +- src/secret/Makefile.inc.am | 2 +- src/storage/Makefile.inc.am | 2 +- src/vbox/Makefile.inc.am | 2 +- src/vz/Makefile.inc.am | 2 +- 19 files changed, 97 insertions(+), 84 deletions(-) delete mode 100755 build-aux/augeas-gentest.pl create mode 100755 build-aux/augeas-gentest.py diff --git a/Makefile.am b/Makefile.am index 711f365504..17448a914e 100644 --- a/Makefile.am +++ b/Makefile.am @@ -41,7 +41,7 @@ EXTRA_DIST = \ run.in \ README.md \ AUTHORS.in \ - build-aux/augeas-gentest.pl \ + build-aux/augeas-gentest.py \ build-aux/check-spacing.pl \ build-aux/gitlog-to-changelog \ build-aux/header-ifdef.pl \ diff --git a/build-aux/augeas-gentest.pl b/build-aux/augeas-gentest.pl deleted file mode 100755 index 65834b533b..0000000000 --- a/build-aux/augeas-gentest.pl +++ /dev/null @@ -1,60 +0,0 @@ -#!/usr/bin/env perl -# -# augeas-gentest.pl: Generate an augeas test file, from an -# example config file + test file template -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. - -use strict; -use warnings; - -die "syntax: $0 CONFIG TEMPLATE\n" unless @ARGV == 2; - -my $config = shift @ARGV; -my $template = shift @ARGV; - -open CONFIG, "<", $config or die "cannot read $config: $!"; -open TEMPLATE, "<", $template or die "cannot read $template: $!"; - -my $group = 0; -while (<TEMPLATE>) { - if (/\@CONFIG\@/) { - my $group = 0; - print " let conf = \""; - while (<CONFIG>) { - if (/^#\w/) { - s/^#//; - s/\"/\\\"/g; - print $_; - $group = /\[\s$/; - } elsif ($group) { - s/\"/\\\"/g; - if (/#\s*\]/) { - $group = 0; - } - if (/^#/) { - s/^#//; - print $_; - } - } - } - print "\"\n"; - } else { - print $_; - } -} - -close TEMPLATE; -close CONFIG; diff --git a/build-aux/augeas-gentest.py b/build-aux/augeas-gentest.py new file mode 100755 index 0000000000..18c6490a08 --- /dev/null +++ b/build-aux/augeas-gentest.py @@ -0,0 +1,72 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# augeas-gentest.py: Generate an augeas test file, from an +# example config file + test file template +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + +if len(sys.argv) != 3: + print("syntax: %s CONFIG TEMPLATE" % sys.argv[0], file=sys.stderr) + sys.exit(1) + +config = sys.argv[1] +template = sys.argv[2] + + +def expand_config(config): + with open(config) as fh: + optprog = re.compile(r'''^#\w.*''') + groupstartprog = re.compile(r'''.*\[\s$''') + groupendprog = re.compile(r'''#\s*\].*$''') + + group = False + for line in fh: + if optprog.match(line) is not None: + line = line[1:] + line = line.replace('"', '\\"') + print(line, end='') + if groupstartprog.match(line): + group = True + elif group: + line = line.replace('"', '\\"') + + if groupendprog.match(line): + group = False + + if line[0] == '#': + line = line[1:] + print(line, end='') + + +def expand_template(template, config): + with open(template) as fh: + markerprog = re.compile(r'''\s*@CONFIG@\s*''') + for line in fh: + if markerprog.match(line) is not None: + print(' let conf = "', end='') + expand_config(config) + print('"') + else: + print(line, end='') + + +expand_template(template, config) diff --git a/src/Makefile.am b/src/Makefile.am index cd955ee552..2956e4bf35 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -421,7 +421,8 @@ check-augeas: $(augeas_DATA) $(augeastest_DATA) fi .PHONY: check-augeas -AUG_GENTEST = $(PERL) $(top_srcdir)/build-aux/augeas-gentest.pl +AUG_GENTEST_SCRIPT = $(top_srcdir)/build-aux/augeas-gentest.py +AUG_GENTEST = $(RUNUTF8) $(PYTHON) $(AUG_GENTEST_SCRIPT) # diff --git a/src/bhyve/Makefile.inc.am b/src/bhyve/Makefile.inc.am index 195069872a..7e49a8bb5c 100644 --- a/src/bhyve/Makefile.inc.am +++ b/src/bhyve/Makefile.inc.am @@ -77,7 +77,7 @@ bhyve/virtbhyved.aug: remote/libvirtd.aug.in $< > $@ bhyve/test_virtbhyved.aug: remote/test_libvirtd.aug.in \ - bhyve/virtbhyved.conf $(AUG_GENTEST) + bhyve/virtbhyved.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) bhyve/virtbhyved.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ @@ -91,7 +91,7 @@ augeas_DATA += bhyve/libvirtd_bhyve.aug augeastest_DATA += bhyve/test_libvirtd_bhyve.aug bhyve/test_libvirtd_bhyve.aug: bhyve/test_libvirtd_bhyve.aug.in \ - $(srcdir)/bhyve/bhyve.conf $(AUG_GENTEST) + $(srcdir)/bhyve/bhyve.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/bhyve/bhyve.conf $< > $@ endif WITH_BHYVE diff --git a/src/interface/Makefile.inc.am b/src/interface/Makefile.inc.am index baa85b4ba9..7f161c487f 100644 --- a/src/interface/Makefile.inc.am +++ b/src/interface/Makefile.inc.am @@ -97,7 +97,7 @@ interface/virtinterfaced.aug: remote/libvirtd.aug.in $< > $@ interface/test_virtinterfaced.aug: remote/test_libvirtd.aug.in \ - interface/virtinterfaced.conf $(AUG_GENTEST) + interface/virtinterfaced.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) interface/virtinterfaced.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/libxl/Makefile.inc.am b/src/libxl/Makefile.inc.am index c53396b7f3..3f35d1f2eb 100644 --- a/src/libxl/Makefile.inc.am +++ b/src/libxl/Makefile.inc.am @@ -111,7 +111,7 @@ libxl/virtxend.aug: remote/libvirtd.aug.in $< > $@ libxl/test_virtxend.aug: remote/test_libvirtd.aug.in \ - libxl/virtxend.conf $(AUG_GENTEST) + libxl/virtxend.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) libxl/virtxend.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ @@ -125,7 +125,7 @@ augeas_DATA += libxl/libvirtd_libxl.aug augeastest_DATA += libxl/test_libvirtd_libxl.aug libxl/test_libvirtd_libxl.aug: libxl/test_libvirtd_libxl.aug.in \ - $(srcdir)/libxl/libxl.conf $(AUG_GENTEST) + $(srcdir)/libxl/libxl.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/libxl/libxl.conf $< > $@ INSTALL_DATA_DIRS += libxl diff --git a/src/locking/Makefile.inc.am b/src/locking/Makefile.inc.am index fae92a6e45..df0c51aaf6 100644 --- a/src/locking/Makefile.inc.am +++ b/src/locking/Makefile.inc.am @@ -219,7 +219,7 @@ endif WITH_SANLOCK if WITH_SANLOCK if WITH_QEMU locking/test_libvirt_sanlock.aug: locking/test_libvirt_sanlock.aug.in \ - locking/qemu-sanlock.conf $(AUG_GENTEST) + locking/qemu-sanlock.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) locking/qemu-sanlock.conf $< > $@ endif WITH_QEMU @@ -227,12 +227,12 @@ endif WITH_SANLOCK if WITH_QEMU locking/test_libvirt_lockd.aug: locking/test_libvirt_lockd.aug.in \ - locking/qemu-lockd.conf $(AUG_GENTEST) + locking/qemu-lockd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) locking/qemu-lockd.conf $< > $@ endif WITH_QEMU locking/test_virtlockd.aug: locking/test_virtlockd.aug.in \ - locking/virtlockd.conf $(AUG_GENTEST) + locking/virtlockd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/locking/virtlockd.conf $< > $@ endif WITH_LIBVIRTD diff --git a/src/logging/Makefile.inc.am b/src/logging/Makefile.inc.am index 7e441dbffb..ce777e278a 100644 --- a/src/logging/Makefile.inc.am +++ b/src/logging/Makefile.inc.am @@ -101,7 +101,7 @@ augeas_DATA += logging/virtlogd.aug augeastest_DATA += logging/test_virtlogd.aug logging/test_virtlogd.aug: logging/test_virtlogd.aug.in \ - logging/virtlogd.conf $(AUG_GENTEST) + logging/virtlogd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/logging/virtlogd.conf $< > $@ endif WITH_LIBVIRTD diff --git a/src/lxc/Makefile.inc.am b/src/lxc/Makefile.inc.am index b4d560702c..0e12800886 100644 --- a/src/lxc/Makefile.inc.am +++ b/src/lxc/Makefile.inc.am @@ -164,7 +164,7 @@ lxc/virtlxcd.aug: remote/libvirtd.aug.in $< > $@ lxc/test_virtlxcd.aug: remote/test_libvirtd.aug.in \ - lxc/virtlxcd.conf $(AUG_GENTEST) + lxc/virtlxcd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) lxc/virtlxcd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ @@ -222,7 +222,7 @@ augeas_DATA += lxc/libvirtd_lxc.aug augeastest_DATA += lxc/test_libvirtd_lxc.aug lxc/test_libvirtd_lxc.aug: lxc/test_libvirtd_lxc.aug.in \ - $(srcdir)/lxc/lxc.conf $(AUG_GENTEST) + $(srcdir)/lxc/lxc.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/lxc/lxc.conf $< > $@ INSTALL_DATA_DIRS += lxc diff --git a/src/network/Makefile.inc.am b/src/network/Makefile.inc.am index 17467a65ad..0ac0e03957 100644 --- a/src/network/Makefile.inc.am +++ b/src/network/Makefile.inc.am @@ -104,7 +104,7 @@ network/virtnetworkd.aug: remote/libvirtd.aug.in $< > $@ network/test_virtnetworkd.aug: remote/test_libvirtd.aug.in \ - network/virtnetworkd.conf $(AUG_GENTEST) + network/virtnetworkd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) network/virtnetworkd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/node_device/Makefile.inc.am b/src/node_device/Makefile.inc.am index eac7f92e88..a42cf8bc6a 100644 --- a/src/node_device/Makefile.inc.am +++ b/src/node_device/Makefile.inc.am @@ -120,7 +120,7 @@ node_device/virtnodedevd.aug: remote/libvirtd.aug.in $< > $@ node_device/test_virtnodedevd.aug: remote/test_libvirtd.aug.in \ - node_device/virtnodedevd.conf $(AUG_GENTEST) + node_device/virtnodedevd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) node_device/virtnodedevd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/nwfilter/Makefile.inc.am b/src/nwfilter/Makefile.inc.am index 6acb45705c..45be065f7c 100644 --- a/src/nwfilter/Makefile.inc.am +++ b/src/nwfilter/Makefile.inc.am @@ -105,7 +105,7 @@ nwfilter/virtnwfilterd.aug: remote/libvirtd.aug.in $< > $@ nwfilter/test_virtnwfilterd.aug: remote/test_libvirtd.aug.in \ - nwfilter/virtnwfilterd.conf $(AUG_GENTEST) + nwfilter/virtnwfilterd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) nwfilter/virtnwfilterd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/qemu/Makefile.inc.am b/src/qemu/Makefile.inc.am index 48fd0332ec..e9794520cf 100644 --- a/src/qemu/Makefile.inc.am +++ b/src/qemu/Makefile.inc.am @@ -169,7 +169,7 @@ qemu/virtqemud.aug: remote/libvirtd.aug.in $< > $@ qemu/test_virtqemud.aug: remote/test_libvirtd.aug.in \ - qemu/virtqemud.conf $(AUG_GENTEST) + qemu/virtqemud.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) qemu/virtqemud.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ @@ -184,7 +184,7 @@ augeas_DATA += qemu/libvirtd_qemu.aug augeastest_DATA += qemu/test_libvirtd_qemu.aug qemu/test_libvirtd_qemu.aug: qemu/test_libvirtd_qemu.aug.in \ - $(srcdir)/qemu/qemu.conf $(AUG_GENTEST) + $(srcdir)/qemu/qemu.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/qemu/qemu.conf $< > $@ INSTALL_DATA_DIRS += qemu diff --git a/src/remote/Makefile.inc.am b/src/remote/Makefile.inc.am index 5a5c90a922..b42b700488 100644 --- a/src/remote/Makefile.inc.am +++ b/src/remote/Makefile.inc.am @@ -285,7 +285,7 @@ remote/virtproxyd.aug: remote/libvirtd.aug.in $< > $@ remote/test_libvirtd.aug: remote/test_libvirtd.aug.in \ - remote/libvirtd.conf $(AUG_GENTEST) + remote/libvirtd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) remote/libvirtd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ @@ -298,7 +298,7 @@ remote/test_libvirtd.aug: remote/test_libvirtd.aug.in \ > $@ || rm -f $@ remote/test_virtproxyd.aug: remote/test_libvirtd.aug.in \ - remote/virtproxyd.conf $(AUG_GENTEST) + remote/virtproxyd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) remote/virtproxyd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/secret/Makefile.inc.am b/src/secret/Makefile.inc.am index 76bc67418c..1020fb64c8 100644 --- a/src/secret/Makefile.inc.am +++ b/src/secret/Makefile.inc.am @@ -93,7 +93,7 @@ secret/virtsecretd.aug: remote/libvirtd.aug.in $< > $@ secret/test_virtsecretd.aug: remote/test_libvirtd.aug.in \ - secret/virtsecretd.conf $(AUG_GENTEST) + secret/virtsecretd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) secret/virtsecretd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/storage/Makefile.inc.am b/src/storage/Makefile.inc.am index 4dccb14ac1..28ddfeec72 100644 --- a/src/storage/Makefile.inc.am +++ b/src/storage/Makefile.inc.am @@ -197,7 +197,7 @@ storage/virtstoraged.aug: remote/libvirtd.aug.in $< > $@ storage/test_virtstoraged.aug: remote/test_libvirtd.aug.in \ - storage/virtstoraged.conf $(AUG_GENTEST) + storage/virtstoraged.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) storage/virtstoraged.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/vbox/Makefile.inc.am b/src/vbox/Makefile.inc.am index 178c360b99..1f2bcb5488 100644 --- a/src/vbox/Makefile.inc.am +++ b/src/vbox/Makefile.inc.am @@ -117,7 +117,7 @@ vbox/virtvboxd.aug: remote/libvirtd.aug.in $< > $@ vbox/test_virtvboxd.aug: remote/test_libvirtd.aug.in \ - vbox/virtvboxd.conf $(AUG_GENTEST) + vbox/virtvboxd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) vbox/virtvboxd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ diff --git a/src/vz/Makefile.inc.am b/src/vz/Makefile.inc.am index f56fceb8f7..fa7b8b79dd 100644 --- a/src/vz/Makefile.inc.am +++ b/src/vz/Makefile.inc.am @@ -91,7 +91,7 @@ vz/virtvzd.aug: remote/libvirtd.aug.in $< > $@ vz/test_virtvzd.aug: remote/test_libvirtd.aug.in \ - vz/virtvzd.conf $(AUG_GENTEST) + vz/virtvzd.conf $(AUG_GENTEST_SCRIPT) $(AM_V_GEN)$(AUG_GENTEST) vz/virtvzd.conf \ $(srcdir)/remote/test_libvirtd.aug.in | \ $(SED) \ -- 2.21.0
On Tue, Sep 24, 2019 at 03:58:43PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the augeas-gentest.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
The use of $(AUG_GENTEST) as a dependancy in the makefiles needed
s/dependancy/dependency/
to be fixed, because this was assumed to be the filename of the script, but is in fact a full shell command line.
This is the case regardless of the Perl->Python conversion and can be done upfront to reduce the churn in this patch. Introduced by commit fb59cf7a5824b9c876737dcbf6aac97c29b1444a
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/augeas-gentest.pl | 60 --------------------------- build-aux/augeas-gentest.py | 72 +++++++++++++++++++++++++++++++++ src/Makefile.am | 3 +- src/bhyve/Makefile.inc.am | 4 +- src/interface/Makefile.inc.am | 2 +- src/libxl/Makefile.inc.am | 4 +- src/locking/Makefile.inc.am | 6 +-- src/logging/Makefile.inc.am | 2 +- src/lxc/Makefile.inc.am | 4 +- src/network/Makefile.inc.am | 2 +- src/node_device/Makefile.inc.am | 2 +- src/nwfilter/Makefile.inc.am | 2 +- src/qemu/Makefile.inc.am | 4 +- src/remote/Makefile.inc.am | 4 +- src/secret/Makefile.inc.am | 2 +- src/storage/Makefile.inc.am | 2 +- src/vbox/Makefile.inc.am | 2 +- src/vz/Makefile.inc.am | 2 +- 19 files changed, 97 insertions(+), 84 deletions(-) delete mode 100755 build-aux/augeas-gentest.pl create mode 100755 build-aux/augeas-gentest.py
Since this is a new file with clean history, it might actually deserve a better location than build-aux and we can leave this directory to Automake and gnulib to do whatever magic they do there. Also note that the directory is in .gitignore. (I added the exception for .pl files back when I added files here) Would 'scripts' be too vague? Could be a good place to put the helper scripts for generating QEMU caps files since I never seem to remember its name and tests/ is growing quite big. [...]
+def expand_config(config): + with open(config) as fh: + optprog = re.compile(r'''^#\w.*''') + groupstartprog = re.compile(r'''.*\[\s$''') + groupendprog = re.compile(r'''#\s*\].*$''') + + group = False + for line in fh: + if optprog.match(line) is not None: + line = line[1:] + line = line.replace('"', '\\"') + print(line, end='') + if groupstartprog.match(line): + group = True + elif group: + line = line.replace('"', '\\"') + + if groupendprog.match(line): + group = False + + if line[0] == '#': + line = line[1:] + print(line, end='') + + +def expand_template(template, config): + with open(template) as fh: + markerprog = re.compile(r'''\s*@CONFIG@\s*''') + for line in fh: + if markerprog.match(line) is not None:
Simpler as: if '@CONFIG@' in line: There's no need to use a regex here
+ print(' let conf = "', end='')
This uses three spaces instead of two like the original Perl script. Three match the rest of the .aug files, but by including that change here, this rewrite generates a different output.
+ expand_config(config) + print('"') + else: + print(line, end='') + + +expand_template(template, config)
Reviewed-by: Ján Tomko <jtomko@redhat.com> Jano
On Wed, Sep 25, 2019 at 03:25:39PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:43PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the augeas-gentest.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
The use of $(AUG_GENTEST) as a dependancy in the makefiles needed
s/dependancy/dependency/
to be fixed, because this was assumed to be the filename of the script, but is in fact a full shell command line.
This is the case regardless of the Perl->Python conversion and can be done upfront to reduce the churn in this patch. Introduced by commit fb59cf7a5824b9c876737dcbf6aac97c29b1444a
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/augeas-gentest.pl | 60 --------------------------- build-aux/augeas-gentest.py | 72 +++++++++++++++++++++++++++++++++ src/Makefile.am | 3 +- src/bhyve/Makefile.inc.am | 4 +- src/interface/Makefile.inc.am | 2 +- src/libxl/Makefile.inc.am | 4 +- src/locking/Makefile.inc.am | 6 +-- src/logging/Makefile.inc.am | 2 +- src/lxc/Makefile.inc.am | 4 +- src/network/Makefile.inc.am | 2 +- src/node_device/Makefile.inc.am | 2 +- src/nwfilter/Makefile.inc.am | 2 +- src/qemu/Makefile.inc.am | 4 +- src/remote/Makefile.inc.am | 4 +- src/secret/Makefile.inc.am | 2 +- src/storage/Makefile.inc.am | 2 +- src/vbox/Makefile.inc.am | 2 +- src/vz/Makefile.inc.am | 2 +- 19 files changed, 97 insertions(+), 84 deletions(-) delete mode 100755 build-aux/augeas-gentest.pl create mode 100755 build-aux/augeas-gentest.py
Since this is a new file with clean history, it might actually deserve a better location than build-aux and we can leave this directory to Automake and gnulib to do whatever magic they do there. Also note that the directory is in .gitignore. (I added the exception for .pl files back when I added files here)
Would 'scripts' be too vague? Could be a good place to put the helper scripts for generating QEMU caps files since I never seem to remember its name and tests/ is growing quite big.
That's a good question. As you see from this series, we've got random scripts scattered all over the sub-dirs. build-aux/ was in some sense to avoid polluting the top level dir. If we create 'scripts/' should we put everything in there, or just stuff that's related to the top level, and keep everything else in their current subdirs ? Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Wed, Sep 25, 2019 at 03:17:10PM +0100, Daniel P. Berrangé wrote:
On Wed, Sep 25, 2019 at 03:25:39PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:43PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the augeas-gentest.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
The use of $(AUG_GENTEST) as a dependancy in the makefiles needed
s/dependancy/dependency/
to be fixed, because this was assumed to be the filename of the script, but is in fact a full shell command line.
This is the case regardless of the Perl->Python conversion and can be done upfront to reduce the churn in this patch. Introduced by commit fb59cf7a5824b9c876737dcbf6aac97c29b1444a
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/augeas-gentest.pl | 60 --------------------------- build-aux/augeas-gentest.py | 72 +++++++++++++++++++++++++++++++++ src/Makefile.am | 3 +- src/bhyve/Makefile.inc.am | 4 +- src/interface/Makefile.inc.am | 2 +- src/libxl/Makefile.inc.am | 4 +- src/locking/Makefile.inc.am | 6 +-- src/logging/Makefile.inc.am | 2 +- src/lxc/Makefile.inc.am | 4 +- src/network/Makefile.inc.am | 2 +- src/node_device/Makefile.inc.am | 2 +- src/nwfilter/Makefile.inc.am | 2 +- src/qemu/Makefile.inc.am | 4 +- src/remote/Makefile.inc.am | 4 +- src/secret/Makefile.inc.am | 2 +- src/storage/Makefile.inc.am | 2 +- src/vbox/Makefile.inc.am | 2 +- src/vz/Makefile.inc.am | 2 +- 19 files changed, 97 insertions(+), 84 deletions(-) delete mode 100755 build-aux/augeas-gentest.pl create mode 100755 build-aux/augeas-gentest.py
Since this is a new file with clean history, it might actually deserve a better location than build-aux and we can leave this directory to Automake and gnulib to do whatever magic they do there. Also note that the directory is in .gitignore. (I added the exception for .pl files back when I added files here)
Would 'scripts' be too vague? Could be a good place to put the helper scripts for generating QEMU caps files since I never seem to remember its name and tests/ is growing quite big.
That's a good question. As you see from this series, we've got random scripts scattered all over the sub-dirs. build-aux/ was in some sense to avoid polluting the top level dir.
If we create 'scripts/' should we put everything in there, or just stuff that's related to the top level, and keep everything else in their current subdirs ?
We put all the driver-specific tests into one tests/ directory, I think doing it for scripts makes sense too. Jano
Regards, Daniel
On Thu, Sep 26, 2019 at 05:35:46PM +0200, Ján Tomko wrote:
On Wed, Sep 25, 2019 at 03:17:10PM +0100, Daniel P. Berrangé wrote:
On Wed, Sep 25, 2019 at 03:25:39PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:43PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the augeas-gentest.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
The use of $(AUG_GENTEST) as a dependancy in the makefiles needed
s/dependancy/dependency/
to be fixed, because this was assumed to be the filename of the script, but is in fact a full shell command line.
This is the case regardless of the Perl->Python conversion and can be done upfront to reduce the churn in this patch. Introduced by commit fb59cf7a5824b9c876737dcbf6aac97c29b1444a
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/augeas-gentest.pl | 60 --------------------------- build-aux/augeas-gentest.py | 72 +++++++++++++++++++++++++++++++++ src/Makefile.am | 3 +- src/bhyve/Makefile.inc.am | 4 +- src/interface/Makefile.inc.am | 2 +- src/libxl/Makefile.inc.am | 4 +- src/locking/Makefile.inc.am | 6 +-- src/logging/Makefile.inc.am | 2 +- src/lxc/Makefile.inc.am | 4 +- src/network/Makefile.inc.am | 2 +- src/node_device/Makefile.inc.am | 2 +- src/nwfilter/Makefile.inc.am | 2 +- src/qemu/Makefile.inc.am | 4 +- src/remote/Makefile.inc.am | 4 +- src/secret/Makefile.inc.am | 2 +- src/storage/Makefile.inc.am | 2 +- src/vbox/Makefile.inc.am | 2 +- src/vz/Makefile.inc.am | 2 +- 19 files changed, 97 insertions(+), 84 deletions(-) delete mode 100755 build-aux/augeas-gentest.pl create mode 100755 build-aux/augeas-gentest.py
Since this is a new file with clean history, it might actually deserve a better location than build-aux and we can leave this directory to Automake and gnulib to do whatever magic they do there. Also note that the directory is in .gitignore. (I added the exception for .pl files back when I added files here)
Would 'scripts' be too vague? Could be a good place to put the helper scripts for generating QEMU caps files since I never seem to remember its name and tests/ is growing quite big.
That's a good question. As you see from this series, we've got random scripts scattered all over the sub-dirs. build-aux/ was in some sense to avoid polluting the top level dir.
If we create 'scripts/' should we put everything in there, or just stuff that's related to the top level, and keep everything else in their current subdirs ?
We put all the driver-specific tests into one tests/ directory, I think doing it for scripts makes sense too.
I'd really like the move all the tests into their respective driver directories at some point :-) We don't have an enourmous number of scripts though, so having them all in one place is fine. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Thu, Sep 26, 2019 at 16:40:54 +0100, Daniel Berrange wrote:
On Thu, Sep 26, 2019 at 05:35:46PM +0200, Ján Tomko wrote:
On Wed, Sep 25, 2019 at 03:17:10PM +0100, Daniel P. Berrangé wrote:
On Wed, Sep 25, 2019 at 03:25:39PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:43PM +0100, Daniel P. Berrangé wrote:
[...]
If we create 'scripts/' should we put everything in there, or just stuff that's related to the top level, and keep everything else in their current subdirs ?
We put all the driver-specific tests into one tests/ directory, I think doing it for scripts makes sense too.
I'd really like the move all the tests into their respective driver directories at some point :-)
I hope you intend to stash them into a 'tests' subdirectory or something. Otherwise I'll object as the driver directories themselves are getting too clobbered with non-test stuff already.
On Thu, Sep 26, 2019 at 05:58:47PM +0200, Peter Krempa wrote:
On Thu, Sep 26, 2019 at 16:40:54 +0100, Daniel Berrange wrote:
On Thu, Sep 26, 2019 at 05:35:46PM +0200, Ján Tomko wrote:
On Wed, Sep 25, 2019 at 03:17:10PM +0100, Daniel P. Berrangé wrote:
On Wed, Sep 25, 2019 at 03:25:39PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:43PM +0100, Daniel P. Berrangé wrote:
[...]
If we create 'scripts/' should we put everything in there, or just stuff that's related to the top level, and keep everything else in their current subdirs ?
We put all the driver-specific tests into one tests/ directory, I think doing it for scripts makes sense too.
I'd really like the move all the tests into their respective driver directories at some point :-)
I hope you intend to stash them into a 'tests' subdirectory or something. Otherwise I'll object as the driver directories themselves are getting too clobbered with non-test stuff already.
I've not thought that far ahead to be honest - its just a random wishlist item at number 783 on the todo list. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
As part of an goal to eliminate Perl from libvirt build tools, rewrite the minimize-po.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/minimize-po.pl | 37 ------------------------- build-aux/minimize-po.py | 60 ++++++++++++++++++++++++++++++++++++++++ po/Makefile.am | 2 +- 4 files changed, 62 insertions(+), 39 deletions(-) delete mode 100755 build-aux/minimize-po.pl create mode 100755 build-aux/minimize-po.py diff --git a/Makefile.am b/Makefile.am index 17448a914e..8f688d40d0 100644 --- a/Makefile.am +++ b/Makefile.am @@ -45,7 +45,7 @@ EXTRA_DIST = \ build-aux/check-spacing.pl \ build-aux/gitlog-to-changelog \ build-aux/header-ifdef.pl \ - build-aux/minimize-po.pl \ + build-aux/minimize-po.py \ build-aux/mock-noinline.pl \ build-aux/prohibit-duplicate-header.pl \ build-aux/useless-if-before-free \ diff --git a/build-aux/minimize-po.pl b/build-aux/minimize-po.pl deleted file mode 100755 index 497533a836..0000000000 --- a/build-aux/minimize-po.pl +++ /dev/null @@ -1,37 +0,0 @@ -#!/usr/bin/perl - -my @block; -my $msgstr = 0; -my $empty = 0; -my $unused = 0; -my $fuzzy = 0; -while (<>) { - if (/^$/) { - if (!$empty && !$unused && !$fuzzy) { - print @block; - } - @block = (); - $msgstr = 0; - $fuzzy = 0; - push @block, $_; - } else { - if (/^msgstr/) { - $msgstr = 1; - $empty = 1; - } - if (/^#.*fuzzy/) { - $fuzzy = 1; - } - if (/^#~ msgstr/) { - $unused = 1; - } - if ($msgstr && /".+"/) { - $empty = 0; - } - push @block, $_; - } -} - -if (@block && !$empty && !$unused) { - print @block; -} diff --git a/build-aux/minimize-po.py b/build-aux/minimize-po.py new file mode 100755 index 0000000000..5046bacede --- /dev/null +++ b/build-aux/minimize-po.py @@ -0,0 +1,60 @@ +#!/usr/bin/env python +# +# Copyright (C) 2018-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + +block = [] +msgstr = False +empty = False +unused = False +fuzzy = False + +strprog = re.compile(r'''.*".+".*''') + +for line in sys.stdin: + line = line.rstrip("\n") + + if line == "": + if not empty and not unused and not fuzzy: + for b in block: + print(b) + + block = [] + msgstr = False + fuzzy = False + block.append(line) + else: + if line.startswith("msgstr"): + msgstr = True + empty = True + + if line[0] == '#' and line.find("fuzzy") != -1: + fuzzy = True + if line.startswith("#~ msgstr"): + unused = True + if msgstr and strprog.match(line): + empty = False + + block.append(line) + +if not empty and not unused and not fuzzy: + for b in block: + print(b) diff --git a/po/Makefile.am b/po/Makefile.am index da117edbd5..90025b15dd 100644 --- a/po/Makefile.am +++ b/po/Makefile.am @@ -58,7 +58,7 @@ update-mini-po: $(POTFILE) $(MSGMERGE) --no-location --no-fuzzy-matching --sort-output \ $$lang.po $(POTFILE) | \ $(SED) $(SED_PO_FIXUP_ARGS) | \ - $(PERL) $(top_srcdir)/build-aux/minimize-po.pl > \ + $(RUNUTF8) $(PYTHON) $(top_srcdir)/build-aux/minimize-po.py > \ $(srcdir)/$$lang.mini.po ; \ done -- 2.21.0
On Tue, Sep 24, 2019 at 03:58:44PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the minimize-po.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/minimize-po.pl | 37 ------------------------- build-aux/minimize-po.py | 60 ++++++++++++++++++++++++++++++++++++++++ po/Makefile.am | 2 +- 4 files changed, 62 insertions(+), 39 deletions(-) delete mode 100755 build-aux/minimize-po.pl create mode 100755 build-aux/minimize-po.py
diff --git a/Makefile.am b/Makefile.am index 17448a914e..8f688d40d0 100644 --- a/Makefile.am +++ b/Makefile.am @@ -45,7 +45,7 @@ EXTRA_DIST = \ build-aux/check-spacing.pl \ build-aux/gitlog-to-changelog \ build-aux/header-ifdef.pl \ - build-aux/minimize-po.pl \ + build-aux/minimize-po.py \ build-aux/mock-noinline.pl \ build-aux/prohibit-duplicate-header.pl \ build-aux/useless-if-before-free \ diff --git a/build-aux/minimize-po.pl b/build-aux/minimize-po.pl deleted file mode 100755 index 497533a836..0000000000 --- a/build-aux/minimize-po.pl +++ /dev/null @@ -1,37 +0,0 @@ -#!/usr/bin/perl - -my @block; -my $msgstr = 0; -my $empty = 0; -my $unused = 0; -my $fuzzy = 0; -while (<>) { - if (/^$/) { - if (!$empty && !$unused && !$fuzzy) { - print @block; - } - @block = (); - $msgstr = 0; - $fuzzy = 0; - push @block, $_; - } else { - if (/^msgstr/) { - $msgstr = 1; - $empty = 1; - } - if (/^#.*fuzzy/) { - $fuzzy = 1; - } - if (/^#~ msgstr/) { - $unused = 1; - } - if ($msgstr && /".+"/) { - $empty = 0; - } - push @block, $_; - } -} - -if (@block && !$empty && !$unused) {
I guess the fact !$fuzzy was missing in this condition was a bug that the new python code doesn't suffer from, right?
- print @block; -} diff --git a/build-aux/minimize-po.py b/build-aux/minimize-po.py new file mode 100755 index 0000000000..5046bacede --- /dev/null +++ b/build-aux/minimize-po.py @@ -0,0 +1,60 @@ +#!/usr/bin/env python +# +# Copyright (C) 2018-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + +block = [] +msgstr = False +empty = False +unused = False +fuzzy = False + +strprog = re.compile(r'''.*".+".*''')
question 1) what's the benefit of compiling a regex and using it only once? Btw python does cache every pattern passed to re.match (and friends) so compilation IMO hardly ever makes sense unless you're doing 1000s of searches for the same pattern in which case the latency would naturally accumulate. question 2) why do we need the '''.* and .*''' parts compared to the original perl regex? I'll go ahead and assume it's because re.match matches at the beginning of a string by default, in which case, shouldn't we use re.search which matches anywhere (that's what perl does by default) instead? Erik
+ +for line in sys.stdin: + line = line.rstrip("\n") + + if line == "": + if not empty and not unused and not fuzzy: + for b in block: + print(b) + + block = [] + msgstr = False + fuzzy = False + block.append(line) + else: + if line.startswith("msgstr"): + msgstr = True + empty = True + + if line[0] == '#' and line.find("fuzzy") != -1: + fuzzy = True + if line.startswith("#~ msgstr"): + unused = True + if msgstr and strprog.match(line): + empty = False + + block.append(line) + +if not empty and not unused and not fuzzy: + for b in block: + print(b) diff --git a/po/Makefile.am b/po/Makefile.am index da117edbd5..90025b15dd 100644 --- a/po/Makefile.am +++ b/po/Makefile.am @@ -58,7 +58,7 @@ update-mini-po: $(POTFILE) $(MSGMERGE) --no-location --no-fuzzy-matching --sort-output \ $$lang.po $(POTFILE) | \ $(SED) $(SED_PO_FIXUP_ARGS) | \ - $(PERL) $(top_srcdir)/build-aux/minimize-po.pl > \ + $(RUNUTF8) $(PYTHON) $(top_srcdir)/build-aux/minimize-po.py > \ $(srcdir)/$$lang.mini.po ; \ done
-- 2.21.0
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On Thu, Sep 26, 2019 at 12:39:39PM +0200, Erik Skultety wrote:
On Tue, Sep 24, 2019 at 03:58:44PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the minimize-po.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/minimize-po.pl | 37 ------------------------- build-aux/minimize-po.py | 60 ++++++++++++++++++++++++++++++++++++++++ po/Makefile.am | 2 +- 4 files changed, 62 insertions(+), 39 deletions(-) delete mode 100755 build-aux/minimize-po.pl create mode 100755 build-aux/minimize-po.py
diff --git a/Makefile.am b/Makefile.am index 17448a914e..8f688d40d0 100644 --- a/Makefile.am +++ b/Makefile.am @@ -45,7 +45,7 @@ EXTRA_DIST = \ build-aux/check-spacing.pl \ build-aux/gitlog-to-changelog \ build-aux/header-ifdef.pl \ - build-aux/minimize-po.pl \ + build-aux/minimize-po.py \ build-aux/mock-noinline.pl \ build-aux/prohibit-duplicate-header.pl \ build-aux/useless-if-before-free \ diff --git a/build-aux/minimize-po.pl b/build-aux/minimize-po.pl deleted file mode 100755 index 497533a836..0000000000 --- a/build-aux/minimize-po.pl +++ /dev/null @@ -1,37 +0,0 @@ -#!/usr/bin/perl - -my @block; -my $msgstr = 0; -my $empty = 0; -my $unused = 0; -my $fuzzy = 0; -while (<>) { - if (/^$/) { - if (!$empty && !$unused && !$fuzzy) { - print @block; - } - @block = (); - $msgstr = 0; - $fuzzy = 0; - push @block, $_; - } else { - if (/^msgstr/) { - $msgstr = 1; - $empty = 1; - } - if (/^#.*fuzzy/) { - $fuzzy = 1; - } - if (/^#~ msgstr/) { - $unused = 1; - } - if ($msgstr && /".+"/) { - $empty = 0; - } - push @block, $_; - } -} - -if (@block && !$empty && !$unused) {
I guess the fact !$fuzzy was missing in this condition was a bug that the new python code doesn't suffer from, right?
Yeah it was a pre-existing bug, but harmless.
diff --git a/build-aux/minimize-po.py b/build-aux/minimize-po.py new file mode 100755 index 0000000000..5046bacede --- /dev/null +++ b/build-aux/minimize-po.py @@ -0,0 +1,60 @@ +#!/usr/bin/env python +# +# Copyright (C) 2018-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + +block = [] +msgstr = False +empty = False +unused = False +fuzzy = False + +strprog = re.compile(r'''.*".+".*''')
question 1) what's the benefit of compiling a regex and using it only once? Btw python does cache every pattern passed to re.match (and friends) so compilation IMO hardly ever makes sense unless you're doing 1000s of searches for the same pattern in which case the latency would naturally accumulate.
Ah, I've just seen the docs now "The compiled versions of the most recent patterns passed to re.compile() and the module-level matching functions are cached, so programs that use only a few regular expressions at a time needn’t worry about compiling regular expressions." so with this in mind, I can probably just remove the 'compile' step from all the scripts in this series. I haven't used it consistently to start with.
question 2) why do we need the '''.* and .*''' parts compared to the original perl regex? I'll go ahead and assume it's because re.match matches at the beginning of a string by default, in which case, shouldn't we use re.search which matches anywhere (that's what perl does by default) instead?
Yeah, I didn't notice the 're.search' function existed & had the semantics closer to Perl. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Thu, Sep 26, 2019 at 02:16:04PM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 12:39:39PM +0200, Erik Skultety wrote:
On Tue, Sep 24, 2019 at 03:58:44PM +0100, Daniel P. Berrangé wrote: question 1) what's the benefit of compiling a regex and using it only once? Btw python does cache every pattern passed to re.match (and friends) so compilation IMO hardly ever makes sense unless you're doing 1000s of searches for the same
Some of the scripts here are run on the whole libvirt codebase so that is the case here. For example just removing the pre-compilation of regexes for comments from the spacing check script bumped the execution time from 6.5s to 7.4s Sadly, the one script where pre-compilation matters the most is the one where separating them puts them far away from the usage to not fit on one screen.
pattern in which case the latency would naturally accumulate.
Ah, I've just seen the docs now
"The compiled versions of the most recent patterns passed to re.compile() and the module-level matching functions are cached, so programs that use only a few regular expressions at a time needn’t worry about compiling regular expressions."
so with this in mind, I can probably just remove the 'compile' step from all the scripts in this series. I haven't used it consistently to start with.
For those regexes only executed on a fraction of lines this is neligible, but it would be nice to keep them for those that run on every line. Jano
question 2) why do we need the '''.* and .*''' parts compared to the original perl regex? I'll go ahead and assume it's because re.match matches at the beginning of a string by default, in which case, shouldn't we use re.search which matches anywhere (that's what perl does by default) instead?
Yeah, I didn't notice the 're.search' function existed & had the semantics closer to Perl.
Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On Thu, Sep 26, 2019 at 05:34:49PM +0200, Ján Tomko wrote:
On Thu, Sep 26, 2019 at 02:16:04PM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 12:39:39PM +0200, Erik Skultety wrote:
On Tue, Sep 24, 2019 at 03:58:44PM +0100, Daniel P. Berrangé wrote: question 1) what's the benefit of compiling a regex and using it only once? Btw python does cache every pattern passed to re.match (and friends) so compilation IMO hardly ever makes sense unless you're doing 1000s of searches for the same
Some of the scripts here are run on the whole libvirt codebase so that is the case here. For example just removing the pre-compilation of regexes for comments from the spacing check script bumped the execution time from 6.5s to 7.4s
Sadly, the one script where pre-compilation matters the most is the one where separating them puts them far away from the usage to not fit on one screen.
I could do a little custom function that caches all regexes recache = {} def research(regex, line): global recache if regex not in recache: recache[regex] = re.compile(regex) return recache[regex].search(line) then the loop we can do a normal research(r'''some regex''', line) so we can get readability and full caching together. Probably not worth repeating this trick for every script, but certainly the whitespace script and a few others probably benefit. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Thu, Sep 26, 2019 at 04:38:49PM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 05:34:49PM +0200, Ján Tomko wrote:
On Thu, Sep 26, 2019 at 02:16:04PM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 12:39:39PM +0200, Erik Skultety wrote:
On Tue, Sep 24, 2019 at 03:58:44PM +0100, Daniel P. Berrangé wrote: question 1) what's the benefit of compiling a regex and using it only once? Btw python does cache every pattern passed to re.match (and friends) so compilation IMO hardly ever makes sense unless you're doing 1000s of searches for the same
Some of the scripts here are run on the whole libvirt codebase so that is the case here. For example just removing the pre-compilation of regexes for comments from the spacing check script bumped the execution time from 6.5s to 7.4s
Sadly, the one script where pre-compilation matters the most is the one where separating them puts them far away from the usage to not fit on one screen.
I could do a little custom function that caches all regexes
recache = {}
def research(regex, line): global recache if regex not in recache: recache[regex] = re.compile(regex) return recache[regex].search(line)
I'm not sure how ^this would solve the slowdown Jano is seeing as this is exactly what python should already be doing internally, IOW the slowdown Jano reported is most likely caused by cache accesses which I don't think our own custom cache would solve, so we probably do want to keep the compilation in even though I personally don't mind the ~1 sec penalty here (compared to the 4x slowdown in the next patch which I think we need to do better to resolve). Erik
On Fri, Sep 27, 2019 at 09:22:13AM +0200, Erik Skultety wrote:
On Thu, Sep 26, 2019 at 04:38:49PM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 05:34:49PM +0200, Ján Tomko wrote:
On Thu, Sep 26, 2019 at 02:16:04PM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 12:39:39PM +0200, Erik Skultety wrote:
On Tue, Sep 24, 2019 at 03:58:44PM +0100, Daniel P. Berrangé wrote: question 1) what's the benefit of compiling a regex and using it only once? Btw python does cache every pattern passed to re.match (and friends) so compilation IMO hardly ever makes sense unless you're doing 1000s of searches for the same
Some of the scripts here are run on the whole libvirt codebase so that is the case here. For example just removing the pre-compilation of regexes for comments from the spacing check script bumped the execution time from 6.5s to 7.4s
Sadly, the one script where pre-compilation matters the most is the one where separating them puts them far away from the usage to not fit on one screen.
I could do a little custom function that caches all regexes
recache = {}
def research(regex, line): global recache if regex not in recache: recache[regex] = re.compile(regex) return recache[regex].search(line)
I'm not sure how ^this would solve the slowdown Jano is seeing as this is exactly what python should already be doing internally, IOW the slowdown Jano reported is most likely caused by cache accesses which I don't think our own custom cache would solve, so we probably do want to keep the compilation in even though I personally don't mind the ~1 sec penalty here (compared to the 4x slowdown in the next patch which I think we need to do better to resolve).
Yeah the slowdown Jano reports looks like a bigger problem to deal with. I think it could still be worth doing it for this patch, since although 1 sec doesn't sound like much, with the huge number of scripts we have, it all adds up. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
As part of an goal to eliminate Perl from libvirt build tools, rewrite the prohibit-duplicate-header.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/prohibit-duplicate-header.pl | 26 ------------ build-aux/prohibit-duplicate-header.py | 57 ++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 60 insertions(+), 29 deletions(-) delete mode 100644 build-aux/prohibit-duplicate-header.pl create mode 100644 build-aux/prohibit-duplicate-header.py diff --git a/Makefile.am b/Makefile.am index 8f688d40d0..a7d8da7146 100644 --- a/Makefile.am +++ b/Makefile.am @@ -47,7 +47,7 @@ EXTRA_DIST = \ build-aux/header-ifdef.pl \ build-aux/minimize-po.py \ build-aux/mock-noinline.pl \ - build-aux/prohibit-duplicate-header.pl \ + build-aux/prohibit-duplicate-header.py \ build-aux/useless-if-before-free \ build-aux/vc-list-files \ ci/Makefile \ diff --git a/build-aux/prohibit-duplicate-header.pl b/build-aux/prohibit-duplicate-header.pl deleted file mode 100644 index 4a2ea65665..0000000000 --- a/build-aux/prohibit-duplicate-header.pl +++ /dev/null @@ -1,26 +0,0 @@ -#!/usr/bin/env perl - -use strict; - -my $file = " "; -my $ret = 0; -my %includes = ( ); -my $lineno = 0; - -while (<>) { - if (not $file eq $ARGV) { - %includes = ( ); - $file = $ARGV; - $lineno = 0; - } - $lineno++; - if (/^# *include *[<"]([^>"]*\.h)[">]/) { - $includes{$1}++; - if ($includes{$1} == 2) { - $ret = 1; - print STDERR "$ARGV:$lineno: $_"; - print STDERR "Do not include a header more than once per file\n"; - } - } -} -exit $ret; diff --git a/build-aux/prohibit-duplicate-header.py b/build-aux/prohibit-duplicate-header.py new file mode 100644 index 0000000000..5ed3f82ba7 --- /dev/null +++ b/build-aux/prohibit-duplicate-header.py @@ -0,0 +1,57 @@ +#!/usr/bin/env python +# +# Copyright (C) 2016-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_file(filename): + includes = {} + lineno = 0 + errs = False + with open(filename, "r") as fh: + headerprog = re.compile(r'''^# *include *[<"]([^>"]*\.h)[">]\s*$''') + for line in fh: + lineno = lineno + 1 + + headermatch = headerprog.match(line) + if headermatch is not None: + inc = headermatch.group(1) + + if inc in includes: + print("%s:%d: %s" % (filename, lineno, inc), + file=sys.stderr) + errs = True + else: + includes[inc] = True + + return errs + + +ret = 0 + +for filename in sys.argv[1:]: + if check_file(filename): + ret = 1 + +if ret == 1: + print("Do not include a header more than once per file", file=sys.stderr) + +sys.exit(ret) diff --git a/cfg.mk b/cfg.mk index 3731a2830b..bf4a8270d8 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1152,8 +1152,8 @@ endif # Don't include duplicate header in the source (either *.c or *.h) prohibit-duplicate-header: - $(AM_V_GEN)$(VC_LIST_EXCEPT) | $(GREP) '\.[chx]$$' | xargs \ - $(PERL) -W $(top_srcdir)/build-aux/prohibit-duplicate-header.pl + $(AM_V_GEN)$(VC_LIST_EXCEPT) | $(GREP) '\.[chx]$$' | $(RUNUTF8) xargs \ + $(PYTHON) $(top_srcdir)/build-aux/prohibit-duplicate-header.py spacing-check: $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.c$$' | xargs \ -- 2.21.0
On Tue, Sep 24, 2019 at 03:58:45PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the prohibit-duplicate-header.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/prohibit-duplicate-header.pl | 26 ------------ build-aux/prohibit-duplicate-header.py | 57 ++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 60 insertions(+), 29 deletions(-) delete mode 100644 build-aux/prohibit-duplicate-header.pl create mode 100644 build-aux/prohibit-duplicate-header.py
diff --git a/Makefile.am b/Makefile.am index 8f688d40d0..a7d8da7146 100644 --- a/Makefile.am +++ b/Makefile.am @@ -47,7 +47,7 @@ EXTRA_DIST = \ build-aux/header-ifdef.pl \ build-aux/minimize-po.py \ build-aux/mock-noinline.pl \ - build-aux/prohibit-duplicate-header.pl \ + build-aux/prohibit-duplicate-header.py \ build-aux/useless-if-before-free \ build-aux/vc-list-files \ ci/Makefile \ diff --git a/build-aux/prohibit-duplicate-header.pl b/build-aux/prohibit-duplicate-header.pl deleted file mode 100644 index 4a2ea65665..0000000000 --- a/build-aux/prohibit-duplicate-header.pl +++ /dev/null @@ -1,26 +0,0 @@ -#!/usr/bin/env perl - -use strict; - -my $file = " "; -my $ret = 0; -my %includes = ( ); -my $lineno = 0; - -while (<>) { - if (not $file eq $ARGV) { - %includes = ( ); - $file = $ARGV; - $lineno = 0; - } - $lineno++; - if (/^# *include *[<"]([^>"]*\.h)[">]/) { - $includes{$1}++; - if ($includes{$1} == 2) { - $ret = 1; - print STDERR "$ARGV:$lineno: $_"; - print STDERR "Do not include a header more than once per file\n"; - } - } -} -exit $ret; diff --git a/build-aux/prohibit-duplicate-header.py b/build-aux/prohibit-duplicate-header.py new file mode 100644 index 0000000000..5ed3f82ba7 --- /dev/null +++ b/build-aux/prohibit-duplicate-header.py @@ -0,0 +1,57 @@ +#!/usr/bin/env python +# +# Copyright (C) 2016-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_file(filename): + includes = {} + lineno = 0 + errs = False + with open(filename, "r") as fh: + headerprog = re.compile(r'''^# *include *[<"]([^>"]*\.h)[">]\s*$''') + for line in fh: + lineno = lineno + 1 + + headermatch = headerprog.match(line) + if headermatch is not None: + inc = headermatch.group(1) + + if inc in includes: + print("%s:%d: %s" % (filename, lineno, inc), + file=sys.stderr) + errs = True + else: + includes[inc] = True + + return errs + + +ret = 0 + +for filename in sys.argv[1:]: + if check_file(filename): + ret = 1
so, how about instead of using @ret at all we make @errs global and then simply do: if err: print("Do not include a header...") It's just an alternative, maybe a tiny bit more intuitive. Other than that, same question about the re.compile as in the previous patch, but those are just nitpicks: Reviewed-by: Erik Skultety <eskultet@redhat.com>
+ +if ret == 1: + print("Do not include a header more than once per file", file=sys.stderr) + +sys.exit(ret)
On Thu, Sep 26, 2019 at 02:46:50PM +0200, Erik Skultety wrote:
On Tue, Sep 24, 2019 at 03:58:45PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the prohibit-duplicate-header.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/prohibit-duplicate-header.pl | 26 ------------ build-aux/prohibit-duplicate-header.py | 57 ++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 60 insertions(+), 29 deletions(-) delete mode 100644 build-aux/prohibit-duplicate-header.pl create mode 100644 build-aux/prohibit-duplicate-header.py
diff --git a/Makefile.am b/Makefile.am index 8f688d40d0..a7d8da7146 100644 --- a/Makefile.am +++ b/Makefile.am @@ -47,7 +47,7 @@ EXTRA_DIST = \ build-aux/header-ifdef.pl \ build-aux/minimize-po.py \ build-aux/mock-noinline.pl \ - build-aux/prohibit-duplicate-header.pl \ + build-aux/prohibit-duplicate-header.py \ build-aux/useless-if-before-free \ build-aux/vc-list-files \ ci/Makefile \ diff --git a/build-aux/prohibit-duplicate-header.pl b/build-aux/prohibit-duplicate-header.pl deleted file mode 100644 index 4a2ea65665..0000000000 --- a/build-aux/prohibit-duplicate-header.pl +++ /dev/null @@ -1,26 +0,0 @@ -#!/usr/bin/env perl - -use strict; - -my $file = " "; -my $ret = 0; -my %includes = ( ); -my $lineno = 0; - -while (<>) { - if (not $file eq $ARGV) { - %includes = ( ); - $file = $ARGV; - $lineno = 0; - } - $lineno++; - if (/^# *include *[<"]([^>"]*\.h)[">]/) { - $includes{$1}++; - if ($includes{$1} == 2) { - $ret = 1; - print STDERR "$ARGV:$lineno: $_"; - print STDERR "Do not include a header more than once per file\n"; - } - } -} -exit $ret; diff --git a/build-aux/prohibit-duplicate-header.py b/build-aux/prohibit-duplicate-header.py new file mode 100644 index 0000000000..5ed3f82ba7 --- /dev/null +++ b/build-aux/prohibit-duplicate-header.py @@ -0,0 +1,57 @@ +#!/usr/bin/env python +# +# Copyright (C) 2016-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_file(filename): + includes = {} + lineno = 0 + errs = False + with open(filename, "r") as fh: + headerprog = re.compile(r'''^# *include *[<"]([^>"]*\.h)[">]\s*$''') + for line in fh: + lineno = lineno + 1 + + headermatch = headerprog.match(line) + if headermatch is not None: + inc = headermatch.group(1) + + if inc in includes: + print("%s:%d: %s" % (filename, lineno, inc), + file=sys.stderr) + errs = True + else: + includes[inc] = True + + return errs + + +ret = 0 + +for filename in sys.argv[1:]: + if check_file(filename): + ret = 1
so, how about instead of using @ret at all we make @errs global and then simply do:
I'm biased against referencing global variables from within methods, because I feel it leads to harder to understand code when there's a mix of local & global vars used. It especially makes later refactoring of code more error prone as code being refactored interacts with global state in unpredictable ways. Now 'errs' is quite a simple usage pattern, but I'd still prefer to avoid mixing in globals.
if err: print("Do not include a header...")
It's just an alternative, maybe a tiny bit more intuitive.
Other than that, same question about the re.compile as in the previous patch, but those are just nitpicks:
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/check-spacing.pl | 198 -------------------------------- build-aux/check-spacing.py | 229 +++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 232 insertions(+), 201 deletions(-) delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py diff --git a/Makefile.am b/Makefile.am index a7d8da7146..03bf1beb78 100644 --- a/Makefile.am +++ b/Makefile.am @@ -42,7 +42,7 @@ EXTRA_DIST = \ README.md \ AUTHORS.in \ build-aux/augeas-gentest.py \ - build-aux/check-spacing.pl \ + build-aux/check-spacing.py \ build-aux/gitlog-to-changelog \ build-aux/header-ifdef.pl \ build-aux/minimize-po.py \ diff --git a/build-aux/check-spacing.pl b/build-aux/check-spacing.pl deleted file mode 100755 index 33377f3dd3..0000000000 --- a/build-aux/check-spacing.pl +++ /dev/null @@ -1,198 +0,0 @@ -#!/usr/bin/env perl -# -# check-spacing.pl: Report any usage of 'function (..args..)' -# Also check for other syntax issues, such as correct use of ';' -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. - -use strict; -use warnings; - -my $ret = 0; -my $incomment = 0; - -foreach my $file (@ARGV) { - # Per-file variables for multiline Curly Bracket (cb_) check - my $cb_linenum = 0; - my $cb_code = ""; - my $cb_scolon = 0; - - open FILE, $file; - - while (defined (my $line = <FILE>)) { - my $data = $line; - # For temporary modifications - my $tmpdata; - - # Kill any quoted , ; = or " - $data =~ s/'[";,=]'/'X'/g; - - # Kill any quoted strings - $data =~ s,"(?:[^\\\"]|\\.)*","XXX",g; - - next if $data =~ /^#/; - - # Kill contents of multi-line comments - # and detect end of multi-line comments - if ($incomment) { - if ($data =~ m,\*/,) { - $incomment = 0; - $data =~ s,^.*\*/,*/,; - } else { - $data = ""; - } - } - - # Kill single line comments, and detect - # start of multi-line comments - if ($data =~ m,/\*.*\*/,) { - $data =~ s,/\*.*\*/,/* */,; - } elsif ($data =~ m,/\*,) { - $incomment = 1; - $data =~ s,/\*.*,/*,; - } - - # We need to match things like - # - # int foo (int bar, bool wizz); - # foo (bar, wizz); - # - # but not match things like: - # - # typedef int (*foo)(bar wizz) - # - # we can't do this (efficiently) without - # missing things like - # - # foo (*bar, wizz); - # - # We also don't want to spoil the $data so it can be used - # later on. - $tmpdata = $data; - while ($tmpdata =~ /(\w+)\s\((?!\*)/) { - my $kw = $1; - - # Allow space after keywords only - if ($kw =~ /^(?:if|for|while|switch|return)$/) { - $tmpdata =~ s/(?:$kw\s\()/XXX(/; - } else { - print "Whitespace after non-keyword:\n"; - print "$file:$.: $line"; - $ret = 1; - last; - } - } - - # Require whitespace immediately after keywords - if ($data =~ /\b(?:if|for|while|switch|return)\(/) { - print "No whitespace after keyword:\n"; - print "$file:$.: $line"; - $ret = 1; - } - - # Forbid whitespace between )( of a function typedef - if ($data =~ /\(\*\w+\)\s+\(/) { - print "Whitespace between ')' and '(':\n"; - print "$file:$.: $line"; - $ret = 1; - } - - # Forbid whitespace following ( or prior to ) - # but allow whitespace before ) on a single line - # (optionally followed by a semicolon) - if (($data =~ /\s\)/ && not $data =~ /^\s+\);?$/) || - $data =~ /\((?!$)\s/) { - print "Whitespace after '(' or before ')':\n"; - print "$file:$.: $line"; - $ret = 1; - } - - # Forbid whitespace before ";" or ",". Things like below are allowed: - # - # 1) The expression is empty for "for" loop. E.g. - # for (i = 0; ; i++) - # - # 2) An empty statement. E.g. - # while (write(statuswrite, &status, 1) == -1 && - # errno == EINTR) - # ; - # - if ($data =~ /\s[;,]/) { - unless ($data =~ /\S; ; / || - $data =~ /^\s+;/) { - print "Whitespace before semicolon or comma:\n"; - print "$file:$.: $line"; - $ret = 1; - } - } - - # Require EOL, macro line continuation, or whitespace after ";". - # Allow "for (;;)" as an exception. - if ($data =~ /;[^ \\\n;)]/) { - print "Invalid character after semicolon:\n"; - print "$file:$.: $line"; - $ret = 1; - } - - # Require EOL, space, or enum/struct end after comma. - if ($data =~ /,[^ \\\n)}]/) { - print "Invalid character after comma:\n"; - print "$file:$.: $line"; - $ret = 1; - } - - # Require spaces around assignment '=', compounds and '==' - if ($data =~ /[^ ]\b[!<>&|\-+*\/%\^=]?=/ || - $data =~ /=[^= \\\n]/) { - print "Spacing around '=' or '==':\n"; - print "$file:$.: $line"; - $ret = 1; - } - - # One line conditional statements with one line bodies should - # not use curly brackets. - if ($data =~ /^\s*(if|while|for)\b.*\{$/) { - $cb_linenum = $.; - $cb_code = $line; - $cb_scolon = 0; - } - - # We need to check for exactly one semicolon inside the body, - # because empty statements (e.g. with comment only) are - # allowed - if ($cb_linenum == $. - 1 && $data =~ /^[^;]*;[^;]*$/) { - $cb_code .= $line; - $cb_scolon = 1; - } - - if ($data =~ /^\s*}\s*$/ && - $cb_linenum == $. - 2 && - $cb_scolon) { - - print "Curly brackets around single-line body:\n"; - print "$file:$cb_linenum-$.:\n$cb_code$line"; - $ret = 1; - - # There _should_ be no need to reset the values; but to - # keep my inner peace... - $cb_linenum = 0; - $cb_scolon = 0; - $cb_code = ""; - } - } - close FILE; -} - -exit $ret; diff --git a/build-aux/check-spacing.py b/build-aux/check-spacing.py new file mode 100755 index 0000000000..6b9f3ec1ba --- /dev/null +++ b/build-aux/check-spacing.py @@ -0,0 +1,229 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# check-spacing.pl: Report any usage of 'function (..args..)' +# Also check for other syntax issues, such as correct use of ';' +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_whitespace(filename): + errs = False + with open(filename, 'r') as fh: + quotedmetaprog = re.compile(r"""'[";,=]'""") + quotedstringprog = re.compile(r'''"(?:[^\\\"]|\\.)*"''') + commentstartprog = re.compile(r'''^(.*)/\*.*$''') + commentendprog = re.compile(r'''^.*\*/(.*)$''') + commentprog = re.compile(r'''^(.*)/\*.*\*/(.*)''') + funcprog = re.compile(r'''(\w+)\s\((?!\*)''') + keywordprog = re.compile( + r'''^.*\b(?:if|for|while|switch|return)\(.*$''') + functypedefprog = re.compile(r'''^.*\(\*\w+\)\s+\(.*$''') + whitespaceprog1 = re.compile(r'''^.*\s\).*$''') + whitespaceprog2 = re.compile(r'''^\s+\);?$''') + whitespaceprog3 = re.compile(r'''^.*\((?!$)\s.*''') + commasemiprog1 = re.compile(r'''.*\s[;,].*''') + commasemiprog2 = re.compile(r'''.*\S; ; .*''') + commasemiprog3 = re.compile(r'''^\s+;''') + semicolonprog = re.compile(r'''.*;[^ \\\n;)].*''') + commaprog = re.compile(r'''.*,[^ \\\n)}].*''') + assignprog1 = re.compile(r'''[^ ]\b[!<>&|\-+*\/%\^=]?=''') + assignprog2 = re.compile(r'''=[^= \\\n]''') + condstartprog = re.compile(r'''^\s*(if|while|for)\b.*\{$''') + statementprog = re.compile(r'''^[^;]*;[^;]*$''') + condendprog = re.compile(r'''^\s*}\s*$''') + + incomment = False + # Per-file variables for multiline Curly Bracket (cb_) check + cb_lineno = 0 + cb_code = "" + cb_scolon = False + + lineno = 0 + for line in fh: + lineno = lineno + 1 + data = line + # For temporary modifications + + # Kill any quoted , ; = or " + data = quotedmetaprog.sub("'X'", data) + + # Kill any quoted strings + data = quotedstringprog.sub('"XXX"', data) + + if data[0] == '#': + continue + + # Kill contents of multi-line comments + # and detect end of multi-line comments + if incomment: + if commentendprog.match(data): + data = commentendprog.sub('*/\2', data) + incomment = False + else: + data = "" + + # Kill single line comments, and detect + # start of multi-line comments + if commentprog.match(data): + data = commentprog.sub(r'''\1/* */\2''', data) + elif commentstartprog.match(data): + data = commentstartprog.sub(r'''\1/*''', data) + incomment = True + + # We need to match things like + # + # int foo (int bar, bool wizz); + # foo (bar, wizz); + # + # but not match things like: + # + # typedef int (*foo)(bar wizz) + # + # we can't do this (efficiently) without + # missing things like + # + # foo (*bar, wizz); + # + for match in funcprog.finditer(data): + kw = match.group(1) + + # Allow space after keywords only + if kw not in ["if", "for", "while", "switch", "return"]: + print("Whitespace after non-keyword:", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + break + + # Require whitespace immediately after keywords + if keywordprog.match(data): + print("No whitespace after keyword:", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # Forbid whitespace between )( of a function typedef + if functypedefprog.match(data): + print("Whitespace between ')' and '(':", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # Forbid whitespace following ( or prior to ) + # but allow whitespace before ) on a single line + # (optionally followed by a semicolon) + if ((whitespaceprog1.match(data) and + not whitespaceprog2.match(data)) + or whitespaceprog3.match(data)): + print("Whitespace after '(' or before ')':", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # Forbid whitespace before ";" or ",". Things like + # below are allowed: + # + # 1) The expression is empty for "for" loop. E.g. + # for (i = 0; ; i++) + # + # 2) An empty statement. E.g. + # while (write(statuswrite, &status, 1) == -1 && + # errno == EINTR) + # ; + # + if commasemiprog1.match(data) and not ( + commasemiprog2.match(data) or + commasemiprog3.match(data)): + print("Whitespace before semicolon or comma:", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # Require EOL, macro line continuation, or whitespace after ";". + # Allow "for (;;)" as an exception. + if semicolonprog.match(data): + print("Invalid character after semicolon:", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # Require EOL, space, or enum/struct end after comma. + if commaprog.match(data): + print("Invalid character after comma:", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # Require spaces around assignment '=', compounds and '==' + if assignprog1.match(data) or assignprog2.match(data): + print("Spacing around '=' or '==':", + file=sys.stderr) + print("%s:%d: %s" % (filename, lineno, line), + file=sys.stderr) + errs = True + + # One line conditional statements with one line bodies should + # not use curly brackets. + if condstartprog.match(data): + cb_lineno = lineno + cb_code = line + cb_scolon = False + + # We need to check for exactly one semicolon inside the body, + # because empty statements (e.g. with comment only) are + # allowed + if (cb_lineno == lineno - 1) and statementprog.match(data): + cb_code = cb_code + line + cb_scolon = True + + if (condendprog.match(data) and + (cb_lineno == lineno - 2) and + cb_scolon): + print("Curly brackets around single-line body:", + file=sys.stderr) + print("%s:%d:\n%s%s" % (filename, cb_lineno - lineno, + cb_code, line), + file=sys.stderr) + errs = True + + # There _should_ be no need to reset the values; but to + # keep my inner peace... + cb_lineno = 0 + cb_scolon = False + cb_code = "" + + return errs + + +ret = 0 +for filename in sys.argv[1:]: + if check_whitespace(filename): + ret = 1 + +sys.exit(ret) diff --git a/cfg.mk b/cfg.mk index bf4a8270d8..d52d42c06c 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1156,8 +1156,8 @@ prohibit-duplicate-header: $(PYTHON) $(top_srcdir)/build-aux/prohibit-duplicate-header.py spacing-check: - $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.c$$' | xargs \ - $(PERL) $(top_srcdir)/build-aux/check-spacing.pl || \ + $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.c$$' | $(RUNUTF8) xargs \ + $(PYTHON) $(top_srcdir)/build-aux/check-spacing.py || \ { echo '$(ME): incorrect formatting' 1>&2; exit 1; } mock-noinline: -- 2.21.0
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/check-spacing.pl | 198 -------------------------------- build-aux/check-spacing.py | 229 +++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 232 insertions(+), 201 deletions(-) delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py
diff --git a/build-aux/check-spacing.py b/build-aux/check-spacing.py new file mode 100755 index 0000000000..6b9f3ec1ba --- /dev/null +++ b/build-aux/check-spacing.py @@ -0,0 +1,229 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# check-spacing.pl: Report any usage of 'function (..args..)' +# Also check for other syntax issues, such as correct use of ';' +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_whitespace(filename): + errs = False + with open(filename, 'r') as fh: + quotedmetaprog = re.compile(r"""'[";,=]'""") + quotedstringprog = re.compile(r'''"(?:[^\\\"]|\\.)*"''') + commentstartprog = re.compile(r'''^(.*)/\*.*$''') + commentendprog = re.compile(r'''^.*\*/(.*)$''') + commentprog = re.compile(r'''^(.*)/\*.*\*/(.*)''') + funcprog = re.compile(r'''(\w+)\s\((?!\*)''') + keywordprog = re.compile( + r'''^.*\b(?:if|for|while|switch|return)\(.*$''') + functypedefprog = re.compile(r'''^.*\(\*\w+\)\s+\(.*$''') + whitespaceprog1 = re.compile(r'''^.*\s\).*$''') + whitespaceprog2 = re.compile(r'''^\s+\);?$''') + whitespaceprog3 = re.compile(r'''^.*\((?!$)\s.*''') + commasemiprog1 = re.compile(r'''.*\s[;,].*''') + commasemiprog2 = re.compile(r'''.*\S; ; .*''') + commasemiprog3 = re.compile(r'''^\s+;''') + semicolonprog = re.compile(r'''.*;[^ \\\n;)].*''') + commaprog = re.compile(r'''.*,[^ \\\n)}].*''') + assignprog1 = re.compile(r'''[^ ]\b[!<>&|\-+*\/%\^=]?=''') + assignprog2 = re.compile(r'''=[^= \\\n]''') + condstartprog = re.compile(r'''^\s*(if|while|for)\b.*\{$''') + statementprog = re.compile(r'''^[^;]*;[^;]*$''') + condendprog = re.compile(r'''^\s*}\s*$''') +
I think even with descriptive names for the regexes and the Python rewrite, this script is hard to read and comes too close to be a C parser. Also, the execution time goes from 1.5s to 6.5s, which is longer than all the other checks combined run on my 8-core machine. Can we get rid of it completely in favor of some proper formatting tool? I have played with clang-format to try to match our style, the main problems seem to be: * pre-processor directives are indented by the same offset as code (I see that cppi is specifically intended to add just one space) * function calls sometimes leave an empty opening parenthesis * always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places. On the plus side, we would be able to properly check for spacing after casts. Jano
+ incomment = False + # Per-file variables for multiline Curly Bracket (cb_) check + cb_lineno = 0 + cb_code = "" + cb_scolon = False + + lineno = 0 + for line in fh: + lineno = lineno + 1 + data = line + # For temporary modifications + + # Kill any quoted , ; = or " + data = quotedmetaprog.sub("'X'", data) + + # Kill any quoted strings + data = quotedstringprog.sub('"XXX"', data) + + if data[0] == '#': + continue
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/check-spacing.pl | 198 -------------------------------- build-aux/check-spacing.py | 229 +++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 232 insertions(+), 201 deletions(-) delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py
diff --git a/build-aux/check-spacing.py b/build-aux/check-spacing.py new file mode 100755 index 0000000000..6b9f3ec1ba --- /dev/null +++ b/build-aux/check-spacing.py @@ -0,0 +1,229 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# check-spacing.pl: Report any usage of 'function (..args..)' +# Also check for other syntax issues, such as correct use of ';' +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_whitespace(filename): + errs = False + with open(filename, 'r') as fh: + quotedmetaprog = re.compile(r"""'[";,=]'""") + quotedstringprog = re.compile(r'''"(?:[^\\\"]|\\.)*"''') + commentstartprog = re.compile(r'''^(.*)/\*.*$''') + commentendprog = re.compile(r'''^.*\*/(.*)$''') + commentprog = re.compile(r'''^(.*)/\*.*\*/(.*)''') + funcprog = re.compile(r'''(\w+)\s\((?!\*)''') + keywordprog = re.compile( + r'''^.*\b(?:if|for|while|switch|return)\(.*$''') + functypedefprog = re.compile(r'''^.*\(\*\w+\)\s+\(.*$''') + whitespaceprog1 = re.compile(r'''^.*\s\).*$''') + whitespaceprog2 = re.compile(r'''^\s+\);?$''') + whitespaceprog3 = re.compile(r'''^.*\((?!$)\s.*''') + commasemiprog1 = re.compile(r'''.*\s[;,].*''') + commasemiprog2 = re.compile(r'''.*\S; ; .*''') + commasemiprog3 = re.compile(r'''^\s+;''') + semicolonprog = re.compile(r'''.*;[^ \\\n;)].*''') + commaprog = re.compile(r'''.*,[^ \\\n)}].*''') + assignprog1 = re.compile(r'''[^ ]\b[!<>&|\-+*\/%\^=]?=''') + assignprog2 = re.compile(r'''=[^= \\\n]''') + condstartprog = re.compile(r'''^\s*(if|while|for)\b.*\{$''') + statementprog = re.compile(r'''^[^;]*;[^;]*$''') + condendprog = re.compile(r'''^\s*}\s*$''') +
I think even with descriptive names for the regexes and the Python rewrite, this script is hard to read and comes too close to be a C parser.
Yeah, trying to parse C code using regular expressions was not our most sensible decision.
Also, the execution time goes from 1.5s to 6.5s, which is longer than all the other checks combined run on my 8-core machine.
Thanks for pointing that out. I'll see if there's any low hanging fruit to optimize but I'm doubtful. I dont mind postponing this patch, as all patches in this series are independant of each other.
Can we get rid of it completely in favor of some proper formatting tool?
I think that would be a good idea because this script only handles a tiny subset of formatting rules we like.
I have played with clang-format to try to match our style, the main problems seem to be: * pre-processor directives are indented by the same offset as code (I see that cppi is specifically intended to add just one space)
That's an interesting approach. I wouldn't object to such indentation style myself.
* function calls sometimes leave an empty opening parenthesis
* always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places.
Yeah these last two points are the ones I struggled with too when I looked at clang-format 6 months back. The churn does worry me as it would make cherry-picking patches a big pain for downstreams. In the long term I think we'd win by having an explicit code formatting tool that everyone is expected to comply with, even if it isn't quite the same style that we currently use. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Fri, Sep 27, 2019 at 10:33:45 +0100, Daniel Berrange wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> ---
[...]
I have played with clang-format to try to match our style, the main problems seem to be: * pre-processor directives are indented by the same offset as code (I see that cppi is specifically intended to add just one space)
That's an interesting approach. I wouldn't object to such indentation style myself.
* function calls sometimes leave an empty opening parenthesis
* always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places.
Yeah these last two points are the ones I struggled with too when I looked at clang-format 6 months back.
The churn does worry me as it would make cherry-picking patches a big pain for downstreams.
In the long term I think we'd win by having an explicit code formatting tool that everyone is expected to comply with, even if it isn't quite the same style that we currently use.
I was also suggesting that we could use it at least in cases where the churn has already happened recently, e.g. when splitting parts of code into separate files. I have at least two series I've done recently and I have some new code which will be in separate files.
On Fri, 2019-09-27 at 10:33 +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
* function calls sometimes leave an empty opening parenthesis * always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places.
Yeah these last two points are the ones I struggled with too when I looked at clang-format 6 months back.
The churn does worry me as it would make cherry-picking patches a big pain for downstreams.
In the long term I think we'd win by having an explicit code formatting tool that everyone is expected to comply with, even if it isn't quite the same style that we currently use.
I have stated my support for this idea in the past, and it still applies. Automatic formatters sometimes produce output that's uglier than what a human would come up with, but remove a lot of fiddling and second-guessing, so overall the outcome is a positive one IMHO. Plus our coding style is quite weird in some areas, and the fact that something like clang-format, which IIRC is used for massive codebases with hundreds of contributors such as Chromium, can't be convinced to adapt to it, may be seen as a sign that we'd be better off adopting one of the more widespread styles instead. -- Andrea Bolognani / Red Hat / Virtualization
On Fri, Sep 27, 2019 at 10:33:45AM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/check-spacing.pl | 198 -------------------------------- build-aux/check-spacing.py | 229 +++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 232 insertions(+), 201 deletions(-) delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py
diff --git a/build-aux/check-spacing.py b/build-aux/check-spacing.py new file mode 100755 index 0000000000..6b9f3ec1ba --- /dev/null +++ b/build-aux/check-spacing.py @@ -0,0 +1,229 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# check-spacing.pl: Report any usage of 'function (..args..)' +# Also check for other syntax issues, such as correct use of ';' +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_whitespace(filename): + errs = False + with open(filename, 'r') as fh: + quotedmetaprog = re.compile(r"""'[";,=]'""") + quotedstringprog = re.compile(r'''"(?:[^\\\"]|\\.)*"''') + commentstartprog = re.compile(r'''^(.*)/\*.*$''') + commentendprog = re.compile(r'''^.*\*/(.*)$''') + commentprog = re.compile(r'''^(.*)/\*.*\*/(.*)''') + funcprog = re.compile(r'''(\w+)\s\((?!\*)''') + keywordprog = re.compile( + r'''^.*\b(?:if|for|while|switch|return)\(.*$''') + functypedefprog = re.compile(r'''^.*\(\*\w+\)\s+\(.*$''') + whitespaceprog1 = re.compile(r'''^.*\s\).*$''') + whitespaceprog2 = re.compile(r'''^\s+\);?$''') + whitespaceprog3 = re.compile(r'''^.*\((?!$)\s.*''') + commasemiprog1 = re.compile(r'''.*\s[;,].*''') + commasemiprog2 = re.compile(r'''.*\S; ; .*''') + commasemiprog3 = re.compile(r'''^\s+;''') + semicolonprog = re.compile(r'''.*;[^ \\\n;)].*''') + commaprog = re.compile(r'''.*,[^ \\\n)}].*''') + assignprog1 = re.compile(r'''[^ ]\b[!<>&|\-+*\/%\^=]?=''') + assignprog2 = re.compile(r'''=[^= \\\n]''') + condstartprog = re.compile(r'''^\s*(if|while|for)\b.*\{$''') + statementprog = re.compile(r'''^[^;]*;[^;]*$''') + condendprog = re.compile(r'''^\s*}\s*$''') +
I think even with descriptive names for the regexes and the Python rewrite, this script is hard to read and comes too close to be a C parser.
Yeah, trying to parse C code using regular expressions was not our most sensible decision.
Also, the execution time goes from 1.5s to 6.5s, which is longer than all the other checks combined run on my 8-core machine.
Thanks for pointing that out. I'll see if there's any low hanging fruit to optimize but I'm doubtful.
I dont mind postponing this patch, as all patches in this series are independant of each other.
Can we get rid of it completely in favor of some proper formatting tool?
I think that would be a good idea because this script only handles a tiny subset of formatting rules we like.
I have played with clang-format to try to match our style, the main problems seem to be: * pre-processor directives are indented by the same offset as code (I see that cppi is specifically intended to add just one space)
That's an interesting approach. I wouldn't object to such indentation style myself.
* function calls sometimes leave an empty opening parenthesis
* always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places.
Yeah these last two points are the ones I struggled with too when I looked at clang-format 6 months back.
Another tool is "uncrustify" which seems to have even more configurability than clang-format does. On my SSD it takes 30 seconds to run over all the .c file. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Fri, Sep 27, 2019 at 11:42:28AM +0100, Daniel P. Berrangé wrote:
On Fri, Sep 27, 2019 at 10:33:45AM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/check-spacing.pl | 198 -------------------------------- build-aux/check-spacing.py | 229 +++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 232 insertions(+), 201 deletions(-) delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py
diff --git a/build-aux/check-spacing.py b/build-aux/check-spacing.py new file mode 100755 index 0000000000..6b9f3ec1ba --- /dev/null +++ b/build-aux/check-spacing.py @@ -0,0 +1,229 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# check-spacing.pl: Report any usage of 'function (..args..)' +# Also check for other syntax issues, such as correct use of ';' +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_whitespace(filename): + errs = False + with open(filename, 'r') as fh: + quotedmetaprog = re.compile(r"""'[";,=]'""") + quotedstringprog = re.compile(r'''"(?:[^\\\"]|\\.)*"''') + commentstartprog = re.compile(r'''^(.*)/\*.*$''') + commentendprog = re.compile(r'''^.*\*/(.*)$''') + commentprog = re.compile(r'''^(.*)/\*.*\*/(.*)''') + funcprog = re.compile(r'''(\w+)\s\((?!\*)''') + keywordprog = re.compile( + r'''^.*\b(?:if|for|while|switch|return)\(.*$''') + functypedefprog = re.compile(r'''^.*\(\*\w+\)\s+\(.*$''') + whitespaceprog1 = re.compile(r'''^.*\s\).*$''') + whitespaceprog2 = re.compile(r'''^\s+\);?$''') + whitespaceprog3 = re.compile(r'''^.*\((?!$)\s.*''') + commasemiprog1 = re.compile(r'''.*\s[;,].*''') + commasemiprog2 = re.compile(r'''.*\S; ; .*''') + commasemiprog3 = re.compile(r'''^\s+;''') + semicolonprog = re.compile(r'''.*;[^ \\\n;)].*''') + commaprog = re.compile(r'''.*,[^ \\\n)}].*''') + assignprog1 = re.compile(r'''[^ ]\b[!<>&|\-+*\/%\^=]?=''') + assignprog2 = re.compile(r'''=[^= \\\n]''') + condstartprog = re.compile(r'''^\s*(if|while|for)\b.*\{$''') + statementprog = re.compile(r'''^[^;]*;[^;]*$''') + condendprog = re.compile(r'''^\s*}\s*$''') +
I think even with descriptive names for the regexes and the Python rewrite, this script is hard to read and comes too close to be a C parser.
Yeah, trying to parse C code using regular expressions was not our most sensible decision.
Also, the execution time goes from 1.5s to 6.5s, which is longer than all the other checks combined run on my 8-core machine.
Thanks for pointing that out. I'll see if there's any low hanging fruit to optimize but I'm doubtful.
I dont mind postponing this patch, as all patches in this series are independant of each other.
Can we get rid of it completely in favor of some proper formatting tool?
I think that would be a good idea because this script only handles a tiny subset of formatting rules we like.
I have played with clang-format to try to match our style, the main problems seem to be: * pre-processor directives are indented by the same offset as code (I see that cppi is specifically intended to add just one space)
That's an interesting approach. I wouldn't object to such indentation style myself.
* function calls sometimes leave an empty opening parenthesis
* always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places.
Yeah these last two points are the ones I struggled with too when I looked at clang-format 6 months back.
Another tool is "uncrustify" which seems to have even more configurability than clang-format does. On my SSD it takes 30 seconds to run over all the .c file.
I support both of the tools, as we can remove our own syntax-check to some extent and some remaining parts can be rewritten into python. This would help a lot with the Meson rewrite that I'm working on now. I also played with "uncrustify" and it looks reasonable and with some tweaking we could get close to our current coding style only with some tiny changes but clang-format works for me as well. Pavel
On Fri, Sep 27, 2019 at 11:42:28AM +0100, Daniel P. Berrangé wrote:
On Fri, Sep 27, 2019 at 10:33:45AM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/check-spacing.pl | 198 -------------------------------- build-aux/check-spacing.py | 229 +++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 232 insertions(+), 201 deletions(-) delete mode 100755 build-aux/check-spacing.pl create mode 100755 build-aux/check-spacing.py
diff --git a/build-aux/check-spacing.py b/build-aux/check-spacing.py new file mode 100755 index 0000000000..6b9f3ec1ba --- /dev/null +++ b/build-aux/check-spacing.py @@ -0,0 +1,229 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# check-spacing.pl: Report any usage of 'function (..args..)' +# Also check for other syntax issues, such as correct use of ';' +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + + +def check_whitespace(filename): + errs = False + with open(filename, 'r') as fh: + quotedmetaprog = re.compile(r"""'[";,=]'""") + quotedstringprog = re.compile(r'''"(?:[^\\\"]|\\.)*"''') + commentstartprog = re.compile(r'''^(.*)/\*.*$''') + commentendprog = re.compile(r'''^.*\*/(.*)$''') + commentprog = re.compile(r'''^(.*)/\*.*\*/(.*)''') + funcprog = re.compile(r'''(\w+)\s\((?!\*)''') + keywordprog = re.compile( + r'''^.*\b(?:if|for|while|switch|return)\(.*$''') + functypedefprog = re.compile(r'''^.*\(\*\w+\)\s+\(.*$''') + whitespaceprog1 = re.compile(r'''^.*\s\).*$''') + whitespaceprog2 = re.compile(r'''^\s+\);?$''') + whitespaceprog3 = re.compile(r'''^.*\((?!$)\s.*''') + commasemiprog1 = re.compile(r'''.*\s[;,].*''') + commasemiprog2 = re.compile(r'''.*\S; ; .*''') + commasemiprog3 = re.compile(r'''^\s+;''') + semicolonprog = re.compile(r'''.*;[^ \\\n;)].*''') + commaprog = re.compile(r'''.*,[^ \\\n)}].*''') + assignprog1 = re.compile(r'''[^ ]\b[!<>&|\-+*\/%\^=]?=''') + assignprog2 = re.compile(r'''=[^= \\\n]''') + condstartprog = re.compile(r'''^\s*(if|while|for)\b.*\{$''') + statementprog = re.compile(r'''^[^;]*;[^;]*$''') + condendprog = re.compile(r'''^\s*}\s*$''') +
I think even with descriptive names for the regexes and the Python rewrite, this script is hard to read and comes too close to be a C parser.
Yeah, trying to parse C code using regular expressions was not our most sensible decision.
Also, the execution time goes from 1.5s to 6.5s, which is longer than all the other checks combined run on my 8-core machine.
Thanks for pointing that out. I'll see if there's any low hanging fruit to optimize but I'm doubtful.
I dont mind postponing this patch, as all patches in this series are independant of each other.
Can we get rid of it completely in favor of some proper formatting tool?
I think that would be a good idea because this script only handles a tiny subset of formatting rules we like.
I have played with clang-format to try to match our style, the main problems seem to be: * pre-processor directives are indented by the same offset as code (I see that cppi is specifically intended to add just one space)
That's an interesting approach. I wouldn't object to such indentation style myself.
* function calls sometimes leave an empty opening parenthesis
* always rewrapping function arguments might create unnecessary churn * parameters wrapping might not be smart enough, e.g. we like to do virReportError(VIR_ERR_CODE, "%s", _("string")); in a lot of places.
Yeah these last two points are the ones I struggled with too when I looked at clang-format 6 months back.
Another tool is "uncrustify" which seems to have even more configurability than clang-format does. On my SSD it takes 30 seconds to run over all the .c file.
The attached lv.cfg file is an uncrustify config that is reasonably close to our current style. To see what it does, run this in your source tree: find src/ -name '*.c' | xargs uncrustify --replace -c uncrustify.cfg There's quite a few real mistakes it is correcting for us. Most of the big stuff is due to us having followed inconsistent rules in different parts of the source tree. No matter which tool we pick will suffer this as we have to pick one style switch/case indents are the big one - sometimes we line up case + switch, sometimes we don't. src/access/viraccessdriverpolkit.c | 16 +-- src/admin/admin_server_dispatch.c | 4 +- src/bhyve/bhyve_capabilities.c | 2 +- src/bhyve/bhyve_command.c | 66 ++++++------- src/bhyve/bhyve_driver.c | 14 +-- src/bhyve/bhyve_monitor.c | 4 +- src/bhyve/bhyve_parse_command.c | 96 +++++++++--------- src/bhyve/bhyve_process.c | 32 +++--- src/conf/capabilities.c | 2 +- src/conf/checkpoint_conf.c | 2 +- src/conf/cpu_conf.c | 8 +- src/conf/device_conf.c | 18 ++-- src/conf/domain_addr.c | 10 +- src/conf/domain_audit.c | 8 +- src/conf/domain_capabilities.c | 6 +- src/conf/domain_conf.c | 414 +++++++++++++++++++++++++++++++++++++++--------------------------------------- src/conf/domain_event.c | 434 +++++++++++++++++++++++++++++++++++++++++----------------------------------------- src/conf/interface_conf.c | 182 +++++++++++++++++------------------ src/conf/netdev_vlan_conf.c | 2 +- src/conf/network_conf.c | 60 ++++++------ src/conf/network_event.c | 20 ++-- src/conf/node_device_conf.c | 20 ++-- src/conf/node_device_event.c | 30 +++--- src/conf/node_device_util.c | 2 +- src/conf/numa_conf.c | 4 +- src/conf/nwfilter_conf.c | 684 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++----------------------------------------------------------------- src/conf/nwfilter_ipaddrmap.c | 2 +- src/conf/nwfilter_params.c | 12 +-- src/conf/secret_event.c | 30 +++--- src/conf/snapshot_conf.c | 8 +- src/conf/storage_adapter_conf.c | 6 +- src/conf/storage_conf.c | 76 ++++++--------- src/conf/storage_event.c | 32 +++--- src/conf/virchrdev.c | 32 +++--- src/conf/virinterfaceobj.c | 9 +- src/conf/virnetworkobj.c | 18 ++-- src/conf/virnetworkportdef.c | 12 +-- src/conf/virnodedeviceobj.c | 21 ++-- src/conf/virnwfilterobj.c | 2 +- src/conf/virsecretobj.c | 9 +- src/conf/virstorageobj.c | 24 +++-- src/cpu/cpu_ppc64.c | 8 +- src/cpu/cpu_s390.c | 8 +- src/cpu/cpu_x86.c | 76 +++++++-------- src/datatypes.c | 2 +- src/esx/esx_driver.c | 144 +++++++++++++-------------- src/esx/esx_interface_driver.c | 2 +- src/esx/esx_network_driver.c | 62 ++++++------ src/esx/esx_storage_backend_iscsi.c | 14 +-- src/esx/esx_storage_backend_vmfs.c | 30 +++--- src/esx/esx_storage_driver.c | 4 +- src/esx/esx_util.c | 6 +- src/esx/esx_vi.c | 332 +++++++++++++++++++++++++++++++-------------------------------- src/esx/esx_vi_methods.c | 58 +++++------ src/esx/esx_vi_types.c | 402 ++++++++++++++++++++++++++++++++++++++-------------------------------------- src/hyperv/hyperv_driver.c | 84 ++++++++-------- src/hyperv/hyperv_wmi.c | 374 +++++++++++++++++++++++++++++++++++----------------------------------- src/interface/interface_backend_netcf.c | 80 +++++++-------- src/interface/interface_backend_udev.c | 126 ++++++++++++------------ src/keycodemapdb | 0 src/libvirt-domain.c | 110 ++++++++++----------- src/libvirt-host.c | 4 +- src/libvirt-interface.c | 6 +- src/libvirt-lxc.c | 4 +- src/libvirt-network.c | 2 +- src/libvirt-nodedev.c | 4 +- src/libvirt-storage.c | 10 +- src/libvirt.c | 18 ++-- src/libxl/libxl_capabilities.c | 11 ++- src/libxl/libxl_conf.c | 404 ++++++++++++++++++++++++++++++++++++++-------------------------------------- src/libxl/libxl_domain.c | 40 ++++---- src/libxl/libxl_driver.c | 636 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++------------------------------------------------------------ src/libxl/libxl_migration.c | 30 +++--- src/libxl/xen_common.c | 38 ++++---- src/libxl/xen_xl.c | 184 +++++++++++++++++------------------ src/libxl/xen_xm.c | 54 +++++------ src/locking/domain_lock.c | 24 ++--- src/locking/lock_daemon.c | 126 ++++++++++++------------ src/locking/lock_manager.c | 4 +- src/logging/log_daemon.c | 122 +++++++++++------------ src/lxc/lxc_cgroup.c | 5 +- src/lxc/lxc_container.c | 14 +-- src/lxc/lxc_controller.c | 6 +- src/lxc/lxc_domain.c | 16 +-- src/lxc/lxc_driver.c | 94 +++++++++--------- src/lxc/lxc_fuse.c | 2 +- src/lxc/lxc_native.c | 18 ++-- src/lxc/lxc_process.c | 34 +++---- src/network/bridge_driver.c | 52 +++++----- src/node_device/node_device_hal.c | 6 +- src/node_device/node_device_udev.c | 6 +- src/nwfilter/nwfilter_dhcpsnoop.c | 104 ++++++++++---------- src/nwfilter/nwfilter_driver.c | 12 +-- src/nwfilter/nwfilter_ebiptables_driver.c | 160 +++++++++++++++--------------- src/nwfilter/nwfilter_gentech_driver.c | 2 +- src/nwfilter/nwfilter_learnipaddr.c | 42 ++++---- src/openvz/openvz_conf.c | 10 +- src/openvz/openvz_driver.c | 25 +++-- src/phyp/phyp_driver.c | 80 +++++++-------- src/qemu/qemu_agent.c | 32 +++--- src/qemu/qemu_alias.c | 4 +- src/qemu/qemu_block.c | 27 +++--- src/qemu/qemu_blockjob.c | 8 +- src/qemu/qemu_capabilities.c | 20 ++-- src/qemu/qemu_cgroup.c | 22 ++--- src/qemu/qemu_command.c | 574 ++++++++++++++++++++++++++++++++++++++++++++++++++++++------------------------------------------------------ src/qemu/qemu_conf.c | 54 +++++------ src/qemu/qemu_domain.c | 378 +++++++++++++++++++++++++++++++++++------------------------------------ src/qemu/qemu_domain_address.c | 36 +++---- src/qemu/qemu_driver.c | 526 +++++++++++++++++++++++++++++++++++++++++++++++++-------------------------------------------------- src/qemu/qemu_hostdev.c | 2 +- src/qemu/qemu_hotplug.c | 72 +++++++------- src/qemu/qemu_interface.c | 14 +-- src/qemu/qemu_migration.c | 104 ++++++++++---------- src/qemu/qemu_migration_cookie.c | 16 +-- src/qemu/qemu_migration_params.c | 4 +- src/qemu/qemu_monitor.c | 12 +-- src/qemu/qemu_monitor_json.c | 42 ++++---- src/qemu/qemu_monitor_text.c | 6 +- src/qemu/qemu_process.c | 94 +++++++++--------- src/qemu/qemu_security.c | 8 +- src/qemu/qemu_tpm.c | 34 +++---- src/remote/remote_daemon.c | 148 ++++++++++++++-------------- src/remote/remote_daemon_config.c | 6 +- src/remote/remote_daemon_dispatch.c | 22 ++--- src/remote/remote_driver.c | 28 +++--- src/rpc/virnetclient.c | 44 ++++----- src/rpc/virnetclientstream.c | 8 +- src/rpc/virnetlibsshsession.c | 4 +- src/rpc/virnetserverclient.c | 2 +- src/rpc/virnetserverprogram.c | 2 +- src/rpc/virnetserverservice.c | 2 +- src/rpc/virnetsocket.c | 32 +++--- src/rpc/virnetsshsession.c | 8 +- src/rpc/virnettlscontext.c | 4 +- src/secret/secret_driver.c | 2 +- src/secret/secret_util.c | 6 +- src/security/security_apparmor.c | 28 +++--- src/security/security_dac.c | 8 +- src/security/security_manager.c | 16 +-- src/security/security_selinux.c | 6 +- src/security/security_util.c | 4 +- src/security/virt-aa-helper.c | 204 +++++++++++++++++++-------------------- src/storage/storage_backend_disk.c | 92 +++++++++--------- src/storage/storage_backend_fs.c | 4 +- src/storage/storage_backend_gluster.c | 5 +- src/storage/storage_backend_logical.c | 28 +++--- src/storage/storage_backend_mpath.c | 4 +- src/storage/storage_backend_rbd.c | 42 ++++---- src/storage/storage_driver.c | 58 +++++------ src/storage/storage_util.c | 16 +-- src/test/test_driver.c | 456 +++++++++++++++++++++++++++++++++++++++++++------------------------------------------- src/util/viralloc.c | 4 +- src/util/virauthconfig.c | 4 +- src/util/virbuffer.c | 2 +- src/util/vircgroup.c | 22 ++--- src/util/vircommand.c | 6 +- src/util/virconf.c | 84 ++++++++-------- src/util/vircrypto.c | 4 +- src/util/virdbus.c | 4 +- src/util/virdnsmasq.c | 6 +- src/util/virerror.c | 357 +++++++++++++++++++++++++++++++++++++++++++------------------------ src/util/virevent.c | 12 +-- src/util/vireventpoll.c | 4 +- src/util/virfdstream.c | 8 +- src/util/virfile.c | 12 +-- src/util/virfirewall.c | 2 +- src/util/virfirewalld.c | 6 +- src/util/virhook.c | 28 +++--- src/util/virhostdev.c | 14 +-- src/util/virinitctl.c | 46 ++++----- src/util/virlease.c | 4 +- src/util/virlog.c | 38 ++++---- src/util/virmacaddr.c | 2 +- src/util/virnetdev.c | 140 +++++++++++++-------------- src/util/virnetdevbandwidth.c | 6 +- src/util/virnetdevbridge.c | 14 +-- src/util/virnetdevip.c | 8 +- src/util/virnetdevmacvlan.c | 14 +-- src/util/virnetdevvportprofile.c | 14 +-- src/util/virnetlink.c | 14 +-- src/util/virnodesuspend.c | 8 +- src/util/virnuma.c | 2 +- src/util/virpci.c | 20 ++-- src/util/virprocess.c | 22 ++--- src/util/virresctrl.c | 2 +- src/util/virrotatingfile.c | 2 +- src/util/virscsivhost.c | 2 +- src/util/virsocketaddr.c | 14 +-- src/util/virstorageencryption.c | 4 +- src/util/virstoragefile.c | 93 +++++++++--------- src/util/virsysinfo.c | 6 +- src/util/virsystemd.c | 12 +-- src/util/virtime.c | 10 +- src/util/virtpm.c | 8 +- src/util/virtypedparam.c | 18 ++-- src/util/viruri.c | 4 +- src/util/virusb.c | 8 +- src/util/virutil.c | 2 +- src/util/virvhba.c | 14 +-- src/util/virxml.c | 9 +- src/vbox/vbox_V5_0.c | 2 +- src/vbox/vbox_V5_1.c | 2 +- src/vbox/vbox_V5_2.c | 2 +- src/vbox/vbox_common.c | 250 +++++++++++++++++++++++------------------------ src/vbox/vbox_network.c | 2 +- src/vbox/vbox_snapshot_conf.c | 106 ++++++++++---------- src/vbox/vbox_storage.c | 2 +- src/vbox/vbox_tmpl.c | 86 ++++++++--------- src/vmware/vmware_conf.c | 68 ++++++------- src/vmware/vmware_driver.c | 8 +- src/vmx/vmx.c | 652 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------------------------------------------------------------- src/vz/vz_driver.c | 64 ++++++------ src/vz/vz_sdk.c | 94 +++++++++--------- src/vz/vz_utils.c | 8 +- 216 files changed, 6371 insertions(+), 6256 deletions(-) If anyone fancies taking this further feel free. I'm not going to work on uncrustify right now. I'll just drop this whitespace patch rewrite and focus on the other things as a priority. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Fri, Sep 27, 2019 at 12:24:05 +0100, Daniel Berrange wrote:
On Fri, Sep 27, 2019 at 11:42:28AM +0100, Daniel P. Berrangé wrote:
On Fri, Sep 27, 2019 at 10:33:45AM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> ---
[...]
The attached lv.cfg file is an uncrustify config that is reasonably close to our current style.
To see what it does, run this in your source tree:
find src/ -name '*.c' | xargs uncrustify --replace -c uncrustify.cfg
There's quite a few real mistakes it is correcting for us.
Most of the big stuff is due to us having followed inconsistent rules in different parts of the source tree. No matter which tool we pick will suffer this as we have to pick one style
switch/case indents are the big one - sometimes we line up case + switch, sometimes we don't.
[...]
If anyone fancies taking this further feel free. I'm not going to work on uncrustify right now. I'll just drop this whitespace patch rewrite and focus on the other things as a priority.
As I've mentioned in a different subthread: What do you think of applying this on some recently created files first? That way there is already a churn barrier originating from the refactor so the reformat itself will not make backports much more painful.
On Fri, Sep 27, 2019 at 01:32:25PM +0200, Peter Krempa wrote:
On Fri, Sep 27, 2019 at 12:24:05 +0100, Daniel Berrange wrote:
On Fri, Sep 27, 2019 at 11:42:28AM +0100, Daniel P. Berrangé wrote:
On Fri, Sep 27, 2019 at 10:33:45AM +0100, Daniel P. Berrangé wrote:
On Thu, Sep 26, 2019 at 06:08:14PM +0200, Ján Tomko wrote:
On Tue, Sep 24, 2019 at 03:58:46PM +0100, Daniel P. Berrangé wrote:
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-spacing.pl tool in Python.
This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> ---
[...]
The attached lv.cfg file is an uncrustify config that is reasonably close to our current style.
To see what it does, run this in your source tree:
find src/ -name '*.c' | xargs uncrustify --replace -c uncrustify.cfg
There's quite a few real mistakes it is correcting for us.
Most of the big stuff is due to us having followed inconsistent rules in different parts of the source tree. No matter which tool we pick will suffer this as we have to pick one style
switch/case indents are the big one - sometimes we line up case + switch, sometimes we don't.
[...]
If anyone fancies taking this further feel free. I'm not going to work on uncrustify right now. I'll just drop this whitespace patch rewrite and focus on the other things as a priority.
As I've mentioned in a different subthread: What do you think of applying this on some recently created files first? That way there is already a churn barrier originating from the refactor so the reformat itself will not make backports much more painful.
That's certainly reasonable. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
As part of an goal to eliminate Perl from libvirt build tools, rewrite the mock-noinline.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/mock-noinline.pl | 75 -------------------------------- build-aux/mock-noinline.py | 89 ++++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 92 insertions(+), 78 deletions(-) delete mode 100644 build-aux/mock-noinline.pl create mode 100644 build-aux/mock-noinline.py diff --git a/Makefile.am b/Makefile.am index 03bf1beb78..db77da890c 100644 --- a/Makefile.am +++ b/Makefile.am @@ -46,7 +46,7 @@ EXTRA_DIST = \ build-aux/gitlog-to-changelog \ build-aux/header-ifdef.pl \ build-aux/minimize-po.py \ - build-aux/mock-noinline.pl \ + build-aux/mock-noinline.py \ build-aux/prohibit-duplicate-header.py \ build-aux/useless-if-before-free \ build-aux/vc-list-files \ diff --git a/build-aux/mock-noinline.pl b/build-aux/mock-noinline.pl deleted file mode 100644 index 958e133885..0000000000 --- a/build-aux/mock-noinline.pl +++ /dev/null @@ -1,75 +0,0 @@ -#!/usr/bin/env perl - -my %noninlined; -my %mocked; - -# Functions in public header don't get the noinline annotation -# so whitelist them here -$noninlined{"virEventAddTimeout"} = 1; -# This one confuses the script as its defined in the mock file -# but is actually just a local helper -$noninlined{"virMockStatRedirect"} = 1; - -foreach my $arg (@ARGV) { - if ($arg =~ /\.h$/) { - #print "Scan header $arg\n"; - &scan_annotations($arg); - } elsif ($arg =~ /mock\.c$/) { - #print "Scan mock $arg\n"; - &scan_overrides($arg); - } -} - -my $warned = 0; -foreach my $func (keys %mocked) { - next if exists $noninlined{$func}; - - $warned++; - print STDERR "$func is mocked at $mocked{$func} but missing noinline annotation\n"; -} - -exit $warned ? 1 : 0; - - -sub scan_annotations { - my $file = shift; - - open FH, $file or die "cannot read $file: $!"; - - my $func; - while (<FH>) { - if (/^\s*(\w+)\(/ || /^(?:\w+\*?\s+)+(?:\*\s*)?(\w+)\(/) { - my $name = $1; - if ($name !~ /ATTRIBUTE/) { - $func = $name; - } - } elsif (/^\s*$/) { - $func = undef; - } - if (/ATTRIBUTE_NOINLINE/) { - if (defined $func) { - $noninlined{$func} = 1; - } - } - } - - close FH -} - -sub scan_overrides { - my $file = shift; - - open FH, $file or die "cannot read $file: $!"; - - my $func; - while (<FH>) { - if (/^(\w+)\(/ || /^\w+\s*(?:\*\s*)?(\w+)\(/) { - my $name = $1; - if ($name =~ /^vir/) { - $mocked{$name} = "$file:$."; - } - } - } - - close FH -} diff --git a/build-aux/mock-noinline.py b/build-aux/mock-noinline.py new file mode 100644 index 0000000000..53aeeb429c --- /dev/null +++ b/build-aux/mock-noinline.py @@ -0,0 +1,89 @@ +#!/usr/bin/env python +# +# Copyright (C) 2017-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import sys + +noninlined = {} +mocked = {} + +# Functions in public header don't get the noinline annotation +# so whitelist them here +noninlined["virEventAddTimeout"] = True +# This one confuses the script as its defined in the mock file +# but is actually just a local helper +noninlined["virMockStatRedirect"] = True + + +def scan_annotations(filename): + funcprog1 = re.compile(r'''^\s*(\w+)\(.*''') + funcprog2 = re.compile(r'''^(?:\w+\*?\s+)+(?:\*\s*)?(\w+)\(.*''') + with open(filename, "r") as fh: + func = None + for line in fh: + line = line.strip() + m = funcprog1.match(line) + if m is None: + m = funcprog2.match(line) + if m is not None: + name = m.group(1) + if name.find("ATTRIBUTE") == -1: + func = name + elif line == "": + func = None + + if line.find("ATTRIBUTE_NOINLINE") != -1: + if func is not None: + noninlined[func] = True + + +def scan_overrides(filename): + funcprog1 = re.compile(r'''^(\w+)\(.*''') + funcprog2 = re.compile(r'''^\w+\s*(?:\*\s*)?(\w+)\(.*''') + with open(filename, "r") as fh: + lineno = 0 + for line in fh: + lineno = lineno + 1 + + m = funcprog1.match(line) + if m is None: + m = funcprog2.match(line) + if m is not None: + name = m.group(1) + if name.startswith("vir"): + mocked[name] = "%s:%d" % (filename, lineno) + + +for filename in sys.argv[1:]: + if filename.endswith(".h"): + scan_annotations(filename) + elif filename.endswith("mock.c"): + scan_overrides(filename) + +warned = False +for func in mocked.keys(): + if func not in noninlined: + warned = True + print("%s is mocked at %s but missing noinline annotation" % + (func, mocked[func]), file=sys.stderr) + +if warned: + sys.exit(1) +sys.exit(0) diff --git a/cfg.mk b/cfg.mk index d52d42c06c..b73d3ae1bf 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1161,8 +1161,8 @@ spacing-check: { echo '$(ME): incorrect formatting' 1>&2; exit 1; } mock-noinline: - $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.[ch]$$' | xargs \ - $(PERL) $(top_srcdir)/build-aux/mock-noinline.pl + $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.[ch]$$' | $(RUNUTF8) xargs \ + $(PYTHON) $(top_srcdir)/build-aux/mock-noinline.py header-ifdef: $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.[h]$$' | xargs \ -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the header-ifdef.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- Makefile.am | 2 +- build-aux/header-ifdef.pl | 182 ------------------------------ build-aux/header-ifdef.py | 231 ++++++++++++++++++++++++++++++++++++++ cfg.mk | 4 +- 4 files changed, 234 insertions(+), 185 deletions(-) delete mode 100644 build-aux/header-ifdef.pl create mode 100644 build-aux/header-ifdef.py diff --git a/Makefile.am b/Makefile.am index db77da890c..4ba729d3f7 100644 --- a/Makefile.am +++ b/Makefile.am @@ -44,7 +44,7 @@ EXTRA_DIST = \ build-aux/augeas-gentest.py \ build-aux/check-spacing.py \ build-aux/gitlog-to-changelog \ - build-aux/header-ifdef.pl \ + build-aux/header-ifdef.py \ build-aux/minimize-po.py \ build-aux/mock-noinline.py \ build-aux/prohibit-duplicate-header.py \ diff --git a/build-aux/header-ifdef.pl b/build-aux/header-ifdef.pl deleted file mode 100644 index dba3dbcbdc..0000000000 --- a/build-aux/header-ifdef.pl +++ /dev/null @@ -1,182 +0,0 @@ -#!/usr/bin/perl -# -# Validate that header files follow a standard layout: -# -# /* -# ...copyright header... -# */ -# <one blank line> -# #pragma once -# ....content.... -# -#--- -# -# For any file ending priv.h, before the #pragma once -# We will have a further section -# -# #ifndef SYMBOL_ALLOW -# # error .... -# #endif /* SYMBOL_ALLOW */ -# <one blank line> -# -#--- -# -# For public headers (files in include/), use the standard header guard instead of #pragma once: -# #ifndef SYMBOL -# # define SYMBOL -# ....content.... -# #endif /* SYMBOL */ - -use strict; -use warnings; - -my $STATE_COPYRIGHT_COMMENT = 0; -my $STATE_COPYRIGHT_BLANK = 1; -my $STATE_PRIV_START = 2; -my $STATE_PRIV_ERROR = 3; -my $STATE_PRIV_END = 4; -my $STATE_PRIV_BLANK = 5; -my $STATE_GUARD_START = 6; -my $STATE_GUARD_DEFINE = 7; -my $STATE_GUARD_END = 8; -my $STATE_EOF = 9; -my $STATE_PRAGMA = 10; - -my $file = " "; -my $ret = 0; -my $ifdef = ""; -my $ifdefpriv = ""; -my $publicheader = 0; - -my $state = $STATE_EOF; -my $mistake = 0; - -sub mistake { - my $msg = shift; - warn $msg; - $mistake = 1; - $ret = 1; -} - -while (<>) { - if (not $file eq $ARGV) { - if ($state == $STATE_COPYRIGHT_COMMENT) { - &mistake("$file: missing copyright comment"); - } elsif ($state == $STATE_COPYRIGHT_BLANK) { - &mistake("$file: missing blank line after copyright header"); - } elsif ($state == $STATE_PRIV_START) { - &mistake("$file: missing '#ifndef $ifdefpriv'"); - } elsif ($state == $STATE_PRIV_ERROR) { - &mistake("$file: missing '# error ...priv allow...'"); - } elsif ($state == $STATE_PRIV_END) { - &mistake("$file: missing '#endif /* $ifdefpriv */'"); - } elsif ($state == $STATE_PRIV_BLANK) { - &mistake("$file: missing blank line after priv header check"); - } elsif ($state == $STATE_GUARD_START) { - if ($publicheader) { - &mistake("$file: missing '#ifndef $ifdef'"); - } else { - &mistake("$file: missing '#pragma once' header guard"); - } - } elsif ($state == $STATE_GUARD_DEFINE) { - &mistake("$file: missing '# define $ifdef'"); - } elsif ($state == $STATE_GUARD_END) { - &mistake("$file: missing '#endif /* $ifdef */'"); - } - - $ifdef = uc $ARGV; - $ifdef =~ s,.*/,,; - $ifdef =~ s,[^A-Z0-9],_,g; - $ifdef =~ s,__+,_,g; - unless ($ifdef =~ /^LIBVIRT_/ && $ARGV !~ /libvirt_internal.h/) { - $ifdef = "LIBVIRT_" . $ifdef; - } - $ifdefpriv = $ifdef . "_ALLOW"; - - $file = $ARGV; - $state = $STATE_COPYRIGHT_COMMENT; - $mistake = 0; - $publicheader = ($ARGV =~ /include\//); - } - - if ($mistake || - $ARGV =~ /config-post\.h$/ || - $ARGV =~ /vbox_(CAPI|XPCOM)/) { - $state = $STATE_EOF; - next; - } - - if ($state == $STATE_COPYRIGHT_COMMENT) { - if (m,\*/,) { - $state = $STATE_COPYRIGHT_BLANK; - } - } elsif ($state == $STATE_COPYRIGHT_BLANK) { - if (! /^$/) { - &mistake("$file: missing blank line after copyright header"); - } - if ($ARGV =~ /priv\.h$/) { - $state = $STATE_PRIV_START; - } else { - $state = $STATE_GUARD_START; - } - } elsif ($state == $STATE_PRIV_START) { - if (/^$/) { - &mistake("$file: too many blank lines after copyright header"); - } elsif (/#ifndef $ifdefpriv$/) { - $state = $STATE_PRIV_ERROR; - } else { - &mistake("$file: missing '#ifndef $ifdefpriv'"); - } - } elsif ($state == $STATE_PRIV_ERROR) { - if (/# error ".*"$/) { - $state = $STATE_PRIV_END; - } else { - &mistake("$file: missing '# error ...priv allow...'"); - } - } elsif ($state == $STATE_PRIV_END) { - if (m,#endif /\* $ifdefpriv \*/,) { - $state = $STATE_PRIV_BLANK; - } else { - &mistake("$file: missing '#endif /* $ifdefpriv */'"); - } - } elsif ($state == $STATE_PRIV_BLANK) { - if (! /^$/) { - &mistake("$file: missing blank line after priv guard"); - } - $state = $STATE_GUARD_START; - } elsif ($state == $STATE_GUARD_START) { - if (/^$/) { - &mistake("$file: too many blank lines after copyright header"); - } - if ($publicheader) { - if (/#ifndef $ifdef$/) { - $state = $STATE_GUARD_DEFINE; - } else { - &mistake("$file: missing '#ifndef $ifdef'"); - } - } else { - if (/#pragma once/) { - $state = $STATE_PRAGMA; - } else { - &mistake("$file: missing '#pragma once' header guard"); - } - } - } elsif ($state == $STATE_GUARD_DEFINE) { - if (/# define $ifdef$/) { - $state = $STATE_GUARD_END; - } else { - &mistake("$file: missing '# define $ifdef'"); - } - } elsif ($state == $STATE_GUARD_END) { - if (m,#endif /\* $ifdef \*/$,) { - $state = $STATE_EOF; - } - } elsif ($state == $STATE_PRAGMA) { - next; - } elsif ($state == $STATE_EOF) { - die "$file: unexpected content after '#endif /* $ifdef */'"; - } else { - die "$file: unexpected state $state"; - } -} -exit $ret; diff --git a/build-aux/header-ifdef.py b/build-aux/header-ifdef.py new file mode 100644 index 0000000000..a192e01bc4 --- /dev/null +++ b/build-aux/header-ifdef.py @@ -0,0 +1,231 @@ +#!/usr/bin/env python +# +# Copyright (C) 2018-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# Validate that header files follow a standard layout: +# +# /* +# ...copyright header... +# */ +# <one blank line> +# #pragma once +# ....content.... +# +# --- +# +# For any file ending priv.h, before the #pragma once +# We will have a further section +# +# #ifndef SYMBOL_ALLOW +# # error .... +# #endif /* SYMBOL_ALLOW */ +# <one blank line> +# +# --- +# +# For public headers (files in include/), use the standard +# header guard instead of #pragma once: +# #ifndef SYMBOL +# # define SYMBOL +# ....content.... +# #endif /* SYMBOL */ + +from __future__ import print_function + +import os.path +import re +import sys + +STATE_COPYRIGHT_COMMENT = 0 +STATE_COPYRIGHT_BLANK = 1 +STATE_PRIV_START = 2 +STATE_PRIV_ERROR = 3 +STATE_PRIV_END = 4 +STATE_PRIV_BLANK = 5 +STATE_GUARD_START = 6 +STATE_GUARD_DEFINE = 7 +STATE_GUARD_END = 8 +STATE_EOF = 9 +STATE_PRAGMA = 10 + + +def check_header(filename): + ifdef = "" + ifdefpriv = "" + + state = STATE_EOF + + ifdef = os.path.basename(filename).upper() + ifdef = re.sub(r"""[^A-Z0-9]""", "_", ifdef) + ifdef = re.sub(r"""__+""", "_", ifdef) + + if (not ifdef.startswith("LIBVIRT_") or + filename.find("libvirt_internal.h") != -1): + ifdef = "LIBVIRT_" + ifdef + + ifdefpriv = ifdef + "_ALLOW" + + state = STATE_COPYRIGHT_COMMENT + publicheader = False + if filename.find("include/") != -1: + publicheader = True + + with open(filename, "r") as fh: + for line in fh: + line = line.rstrip("\n") + if state == STATE_COPYRIGHT_COMMENT: + if line.find("*/") != -1: + state = STATE_COPYRIGHT_BLANK + elif state == STATE_COPYRIGHT_BLANK: + if line != "": + print("%s: missing blank line after copyright header" % + filename, file=sys.stderr) + return True + + if filename.endswith("priv.h"): + state = STATE_PRIV_START + else: + state = STATE_GUARD_START + elif state == STATE_PRIV_START: + if line == "": + print("%s: too many blank lines after copyright header" % + filename, file=sys.stderr) + return True + elif re.match(r"""#ifndef %s$""" % ifdefpriv, line): + state = STATE_PRIV_ERROR + else: + print("%s: missing '#ifndef %s'" % (filename, ifdefpriv), + file=sys.stderr) + return True + elif state == STATE_PRIV_ERROR: + if re.match(r"""# error ".*"$""", line): + state = STATE_PRIV_END + else: + print("%s: missing '# error ...priv allow...'" % + filename, file=sys.stderr) + return True + elif state == STATE_PRIV_END: + if re.match(r"""#endif /\* %s \*/""" % ifdefpriv, line): + state = STATE_PRIV_BLANK + else: + print("%s: missing '#endif /* %s */'" % + (filename, ifdefpriv), file=sys.stderr) + return True + elif state == STATE_PRIV_BLANK: + if line != "": + print("%s: missing blank line after priv guard" % + filename, file=sys.stderr) + return True + state = STATE_GUARD_START + elif state == STATE_GUARD_START: + if line == "": + print("%s: too many blank lines after copyright header" % + filename, file=sys.stderr) + return True + if publicheader: + if re.match(r"""#ifndef %s$""" % ifdef, line): + state = STATE_GUARD_DEFINE + else: + print("%s: missing '#ifndef %s'" % + (filename, ifdef), file=sys.stderr) + return True + else: + if re.match(r"""#pragma once""", line): + state = STATE_PRAGMA + else: + print("%s: missing '#pragma once' header guard" % + filename, file=sys.stderr) + return True + elif state == STATE_GUARD_DEFINE: + if re.match(r"""# define %s$""" % ifdef, line): + state = STATE_GUARD_END + else: + print("%s: missing '# define %s'" % + (filename, ifdef), file=sys.stderr) + return True + elif state == STATE_GUARD_END: + if re.match(r"""#endif /\* %s \*/$""" % ifdef, line): + state = STATE_EOF + elif state == STATE_PRAGMA: + next + elif state == STATE_EOF: + print("%s: unexpected content after '#endif /* %s */'" % + (filename, ifdef), file=sys.stderr) + return True + else: + print("%s: unexpected state $state" % + filename, file=sys.stderr) + return True + + if state == STATE_COPYRIGHT_COMMENT: + print("%s: missing copyright comment" % + filename, file=sys.stderr) + return True + elif state == STATE_COPYRIGHT_BLANK: + print("%s: missing blank line after copyright header" % + filename, file=sys.stderr) + return True + elif state == STATE_PRIV_START: + print("%s: missing '#ifndef %s'" % + (filename, ifdefpriv), file=sys.stderr) + return True + elif state == STATE_PRIV_ERROR: + print("%s: missing '# error ...priv allow...'" % + filename, file=sys.stderr) + return True + elif state == STATE_PRIV_END: + print("%s: missing '#endif /* %s */'" % + (filename, ifdefpriv), file=sys.stderr) + return True + elif state == STATE_PRIV_BLANK: + print("%s: missing blank line after priv header check" % + filename, file=sys.stderr) + return True + elif state == STATE_GUARD_START: + if publicheader: + print("%s: missing '#ifndef %s'" % + (filename, ifdef), file=sys.stderr) + return True + else: + print("%s: missing '#pragma once' header guard" % + filename, file=sys.stderr) + return True + elif state == STATE_GUARD_DEFINE: + print("%s: missing '# define %s'" % + (filename, ifdef), file=sys.stderr) + return True + elif state == STATE_GUARD_END: + print("%s: missing '#endif /* %s */'" % + (filename, ifdef), file=sys.stderr) + return True + + return False + + +ret = 0 + +for filename in sys.argv[1:]: + if filename.find("config-post.h") != -1: + continue + if filename.find("vbox_CAPI") != -1: + continue + if filename.find("vbox_XPCOM") != -1: + continue + if check_header(filename): + ret = 1 + +sys.exit(ret) diff --git a/cfg.mk b/cfg.mk index b73d3ae1bf..f0aed0365a 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1165,8 +1165,8 @@ mock-noinline: $(PYTHON) $(top_srcdir)/build-aux/mock-noinline.py header-ifdef: - $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.[h]$$' | xargs \ - $(PERL) $(top_srcdir)/build-aux/header-ifdef.pl + $(AM_V_GEN)$(VC_LIST) | $(GREP) '\.[h]$$' | $(RUNUTF8) xargs \ + $(PYTHON) $(top_srcdir)/build-aux/header-ifdef.py test-wrap-argv: $(AM_V_GEN)$(VC_LIST) | $(GREP) -E '\.(ldargs|args)' | xargs \ -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-aclperms.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 4 +-- src/check-aclperms.pl | 73 ---------------------------------------- src/check-aclperms.py | 78 +++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 80 insertions(+), 75 deletions(-) delete mode 100755 src/check-aclperms.pl create mode 100755 src/check-aclperms.py diff --git a/src/Makefile.am b/src/Makefile.am index 2956e4bf35..71858ea07f 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -358,11 +358,11 @@ check-aclrules: $(addprefix $(srcdir)/,$(filter-out /%,$(STATEFUL_DRIVER_SOURCE_FILES))) check-aclperms: - $(AM_V_GEN)$(PERL) $(srcdir)/check-aclperms.pl \ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-aclperms.py \ $(srcdir)/access/viraccessperm.h \ $(srcdir)/access/viraccessperm.c -EXTRA_DIST += check-driverimpls.pl check-aclrules.pl check-aclperms.pl +EXTRA_DIST += check-driverimpls.pl check-aclrules.pl check-aclperms.py check-local: check-protocol check-symfile check-symsorting \ check-drivername check-driverimpls check-aclrules \ diff --git a/src/check-aclperms.pl b/src/check-aclperms.pl deleted file mode 100755 index 55b6598313..0000000000 --- a/src/check-aclperms.pl +++ /dev/null @@ -1,73 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# This script just validates that the stringified version of -# a virAccessPerm enum matches the enum constant name. We do -# a lot of auto-generation of code, so when these don't match -# problems occur, preventing auth from succeeding at all. - -my $hdr = shift; -my $impl = shift; - -my %perms; - -my @perms; - -open HDR, $hdr or die "cannot read $hdr: $!"; - -while (<HDR>) { - if (/^\s+VIR_ACCESS_PERM_([_A-Z]+)(,?|\s|$)/) { - my $perm = $1; - - $perms{$perm} = 1 unless ($perm =~ /_LAST$/); - } -} - -close HDR; - - -open IMPL, $impl or die "cannot read $impl: $!"; - -my $group; -my $warned = 0; - -while (defined (my $line = <IMPL>)) { - if ($line =~ /VIR_ACCESS_PERM_([_A-Z]+)_LAST/) { - $group = $1; - } elsif ($line =~ /"[_a-z]+"/) { - my @bits = split /,/, $line; - foreach my $bit (@bits) { - if ($bit =~ /"([_a-z]+)"/) { - my $perm = uc($group . "_" . $1); - if (!exists $perms{$perm}) { - print STDERR "Unknown perm string $1 for group $group\n"; - $warned = 1; - } - delete $perms{$perm}; - } - } - } -} -close IMPL; - -foreach my $perm (keys %perms) { - print STDERR "Perm $perm had not string form\n"; - $warned = 1; -} - -exit $warned; diff --git a/src/check-aclperms.py b/src/check-aclperms.py new file mode 100755 index 0000000000..035c0f6999 --- /dev/null +++ b/src/check-aclperms.py @@ -0,0 +1,78 @@ +#!/usr/bin/env python +# +# Copyright (C) 2013-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# This script just validates that the stringified version of +# a virAccessPerm enum matches the enum constant name. We do +# a lot of auto-generation of code, so when these don't match +# problems occur, preventing auth from succeeding at all. + +from __future__ import print_function + +import re +import sys + +if len(sys.argv) != 3: + print("syntax: %s HEADER IMPL" % (sys.argv[0]), file=sys.stderr) + sys.exit(1) + +hdr = sys.argv[1] +impl = sys.argv[2] + +perms = {} + +with open(hdr) as fh: + symprog = re.compile(r"^\s+VIR_ACCESS_PERM_([_A-Z]+)(,?|\s|$).*") + for line in fh: + symmatch = symprog.match(line) + if symmatch is not None: + perm = symmatch.group(1) + + if not perm.endswith("_LAST"): + perms[perm] = 1 + +warned = False + +with open(impl) as fh: + group = None + symlastprog = re.compile(r".*VIR_ACCESS_PERM_([_A-Z]+)_LAST.*") + alnumprog = re.compile(r'''.*"([_a-z]+)".*''') + + for line in fh: + symlastmatch = symlastprog.match(line) + if symlastmatch is not None: + group = symlastmatch.group(1) + elif alnumprog.match(line) is not None: + bits = line.split(",") + for bit in bits: + m = alnumprog.match(bit) + if m is not None: + perm = (group + "_" + m.group(1)).upper() + if perm not in perms: + print("Unknown perm string %s for group %s" % + (m.group(1), group), file=sys.stderr) + warned = True + + del perms[perm] + +for perm in perms.keys(): + print("Perm %s had not string form" % perm, file=sys.stderr) + warned = True + +if warned: + sys.exit(1) +sys.exit(0) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-symsorting.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 6 +- src/check-symsorting.pl | 106 ----------------------------------- src/check-symsorting.py | 119 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 122 insertions(+), 109 deletions(-) delete mode 100755 src/check-symsorting.pl create mode 100755 src/check-symsorting.py diff --git a/src/Makefile.am b/src/Makefile.am index 71858ea07f..7c3e9ab981 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -284,12 +284,12 @@ check-symfile: check-admin-symfile: endif ! WITH_LINUX check-symsorting: - $(AM_V_GEN)$(PERL) $(srcdir)/check-symsorting.pl \ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-symsorting.py \ $(srcdir) $(SYM_FILES) check-admin-symsorting: - $(AM_V_GEN)$(PERL) $(srcdir)/check-symsorting.pl \ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-symsorting.py \ $(srcdir) $(ADMIN_SYM_FILES) -EXTRA_DIST += check-symfile.pl check-symsorting.pl +EXTRA_DIST += check-symfile.pl check-symsorting.py # Keep this list synced with RPC_PROBE_FILES PROTOCOL_STRUCTS = \ diff --git a/src/check-symsorting.pl b/src/check-symsorting.pl deleted file mode 100755 index 51e38bdedb..0000000000 --- a/src/check-symsorting.pl +++ /dev/null @@ -1,106 +0,0 @@ -#!/usr/bin/env perl - -# Copyright (C) 2012-2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. - -use strict; -use warnings; - -die "syntax: $0 SRCDIR SYMFILE..." unless int(@ARGV) >= 2; - -my $ret = 0; -my $srcdir = shift; -my $lastgroup = undef; -foreach my $symfile (@ARGV) { - open SYMFILE, $symfile or die "cannot read $symfile: $!"; - - my $line = 0; - my $groupfile = ""; - my @group; - - while (<SYMFILE>) { - chomp; - - if (/^#\s*((\w+\/)*(\w+\.h))\s*$/) { - $groupfile = $1; - } elsif (/^#/) { - # Ignore comments - } elsif (/^\s*$/) { - if (@group) { - &check_sorting(\@group, $symfile, $line, $groupfile); - } - @group = (); - $line = $.; - } else { - $_ =~ s/;//; - push @group, $_; - } - } - - close SYMFILE; - if (@group) { - &check_sorting(\@group, $symfile, $line, $groupfile); - } - $lastgroup = undef; -} - -sub check_sorting { - my $group = shift; - my $symfile = shift; - my $line = shift; - my $groupfile = shift; - - my @group = @{$group}; - my @sorted = sort { lc $a cmp lc $b } @group; - my $sorted = 1; - my $first; - my $last; - - # Check that groups are in order and groupfile exists - if (defined $lastgroup && lc $lastgroup ge lc $groupfile) { - print "Symbol block at $symfile:$line: block not sorted\n"; - print "Move $groupfile block before $lastgroup block\n"; - print "\n"; - $ret = 1; - } - if (! -e "$srcdir/$groupfile") { - print "Symbol block at $symfile:$line: $groupfile not found\n"; - print "\n"; - $ret = 1; - } - $lastgroup = $groupfile; - - # Check that symbols within a group are in order - for (my $i = 0 ; $i <= $#sorted ; $i++) { - if ($sorted[$i] ne $group[$i]) { - $first = $i unless defined $first; - $last = $i; - $sorted = 0; - } - } - if (!$sorted) { - @group = splice @group, $first, ($last-$first+1); - @sorted = splice @sorted, $first, ($last-$first+1); - print "Symbol block at $symfile:$line: symbols not sorted\n"; - print map { " " . $_ . "\n" } @group; - print "Correct ordering\n"; - print map { " " . $_ . "\n" } @sorted; - print "\n"; - $ret = 1; - } -} - -exit $ret; diff --git a/src/check-symsorting.py b/src/check-symsorting.py new file mode 100755 index 0000000000..a25175cac8 --- /dev/null +++ b/src/check-symsorting.py @@ -0,0 +1,119 @@ +#!/usr/bin/env python + +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import os.path +import re +import sys + +if len(sys.argv) < 3: + print("syntax: %s SRCDIR SYMFILE..." % sys.argv[0], file=sys.stderr) + sys.exit(1) + + +def check_sorting(group, symfile, line, groupfile, lastgroup): + sortedgroup = sorted(group, key=str.lower) + issorted = True + first = None + last = None + + err = False + # Check that groups are in order and groupfile exists + if lastgroup is not None and lastgroup.lower() > groupfile.lower(): + print("Symbol block at %s:%s: block not sorted" % + (symfile, line), file=sys.stderr) + print("Move %s block before %s block" % + (groupfile, lastgroup), file=sys.stderr) + print("", file=sys.stderr) + err = True + + if not os.path.exists(os.path.join(srcdir, groupfile)): + print("Symbol block at %s:%s: %s not found" % + (symfile, line, groupfile), file=sys.stderr) + print("", file=sys.stderr) + err = True + + # Check that symbols within a group are in order + for i in range(len(group)): + if sortedgroup[i] != group[i]: + if first is None: + first = i + + last = i + issorted = False + + if not issorted: + actual = group[first:(last-first+1)] + expect = sortedgroup[first:(last-first+1)] + print("Symbol block at %s:%s: symbols not sorted" % + (symfile, line), file=sys.stderr) + for g in actual: + print(" %s" % g, file=sys.stderr) + print("Correct ordering", file=sys.stderr) + for g in expect: + print(" %s" % g, file=sys.stderr) + print("", file=sys.stderr) + err = True + + return err + + +ret = 0 +srcdir = sys.argv[1] +lastgroup = None +for symfile in sys.argv[2:]: + with open(symfile, "r") as fh: + lineno = 0 + groupfile = "" + group = [] + thisline = 0 + + filenameprog = re.compile(r'''^#\s*((\w+\/)*(\w+\.h))\s*$''') + + for line in fh: + thisline = thisline + 1 + line = line.strip() + + filenamematch = filenameprog.match(line) + if filenamematch is not None: + groupfile = filenamematch.group(1) + elif line == "": + if len(group) > 0: + if check_sorting(group, symfile, lineno, + groupfile, lastgroup): + ret = 1 + + group = [] + lineno = thisline + lastgroup = groupfile + elif line[0] == '#': + # Ignore comments + pass + else: + line = line.strip(";") + group.append(line) + + if len(group) > 0: + if check_sorting(group, symfile, lineno, + groupfile, lastgroup): + ret = 1 + + lastgroup = None + +sys.exit(ret) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-symfile.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 10 +++--- src/check-symfile.pl | 70 -------------------------------------- src/check-symfile.py | 81 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 86 insertions(+), 75 deletions(-) delete mode 100755 src/check-symfile.pl create mode 100755 src/check-symfile.py diff --git a/src/Makefile.am b/src/Makefile.am index 7c3e9ab981..d10bfdf448 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -274,11 +274,11 @@ PDWTAGS = \ # rule for libvirt.la. However, checking symbols relies on Linux ELF layout if WITH_LINUX check-symfile: libvirt.syms libvirt.la - $(AM_V_GEN)$(PERL) $(srcdir)/check-symfile.pl libvirt.syms \ - .libs/libvirt.so + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-symfile.py \ + libvirt.syms .libs/libvirt.so check-admin-symfile: libvirt_admin.syms libvirt-admin.la - $(AM_V_GEN)$(PERL) $(srcdir)/check-symfile.pl libvirt_admin.syms \ - .libs/libvirt-admin.so + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-symfile.py \ + libvirt_admin.syms .libs/libvirt-admin.so else ! WITH_LINUX check-symfile: check-admin-symfile: @@ -289,7 +289,7 @@ check-symsorting: check-admin-symsorting: $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-symsorting.py \ $(srcdir) $(ADMIN_SYM_FILES) -EXTRA_DIST += check-symfile.pl check-symsorting.py +EXTRA_DIST += check-symfile.py check-symsorting.py # Keep this list synced with RPC_PROBE_FILES PROTOCOL_STRUCTS = \ diff --git a/src/check-symfile.pl b/src/check-symfile.pl deleted file mode 100755 index 4f88300864..0000000000 --- a/src/check-symfile.pl +++ /dev/null @@ -1,70 +0,0 @@ -#!/usr/bin/env perl - -# Copyright (C) 2012-2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. - -die "syntax: $0 SYMFILE ELFLIB(S)" unless int(@ARGV) >= 2; - -my $symfile = shift @ARGV; -my @elflibs = @ARGV; - -my %wantsyms; -my %gotsyms; - -my $ret = 0; - -open SYMFILE, $symfile or die "cannot read $symfile: $!"; - -while (<SYMFILE>) { - next if /{/; - next if /}/; - next if /global:/; - next if /local:/; - next if /^\s*$/; - next if /^\s*#/; - next if /\*/; - - die "malformed line $_" unless /^\s*(\S+);$/; - - if (exists $wantsyms{$1}) { - print STDERR "Symbol $1 is listed twice\n"; - $ret = 1; - } else { - $wantsyms{$1} = 1; - } -} -close SYMFILE; - -foreach my $elflib (@elflibs) { - open NM, "-|", "nm", $elflib or die "cannot run 'nm $elflib': $!"; - - while (<NM>) { - next unless /^\S+\s(?:[TBD])\s(\S+)\s*$/; - - $gotsyms{$1} = 1; - } - - close NM; -} - -foreach my $sym (keys(%wantsyms)) { - next if exists $gotsyms{$sym}; - - print STDERR "Expected symbol $sym is not in ELF library\n"; - $ret = 1; -} - -exit($ret); diff --git a/src/check-symfile.py b/src/check-symfile.py new file mode 100755 index 0000000000..0bb1c1e00d --- /dev/null +++ b/src/check-symfile.py @@ -0,0 +1,81 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import re +import subprocess +import sys + +if len(sys.argv) < 3: + print("syntax: %s SYMFILE ELFLIB(S)" % sys.argv[0], file=sys.stderr) + +symfile = sys.argv[1] +elflibs = sys.argv[2:] + +wantsyms = {} +gotsyms = {} + +ret = 0 + +with open(symfile, "r") as fh: + for line in fh: + line = line.strip() + if line.find("{") != -1: + continue + if line.find("}") != -1: + continue + if line in ["global:", "local:"]: + continue + if line == "": + continue + if line[0] == '#': + continue + if line.find("*") != -1: + continue + + line = line.strip(";") + + if line in wantsyms: + print("Symbol $1 is listed twice", file=sys.stderr) + ret = 1 + else: + wantsyms[line] = True + +for elflib in elflibs: + nm = subprocess.Popen(["nm", elflib], shell=False, + stdout=subprocess.PIPE).stdout + + symprog = re.compile(r'''^\S+\s(?:[TBD])\s(\S+)\s*$''') + for line in nm: + line = line.decode("utf-8") + symmatch = symprog.match(line) + if symmatch is None: + continue + + gotsyms[symmatch.group(1)] = True + + +for sym in wantsyms.keys(): + if sym in gotsyms: + continue + + print("Expected symbol '%s' is not in ELF library" % sym, file=sys.stderr) + ret = 1 + +sys.exit(ret) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the dtrace2systemtap.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 9 ++- src/dtrace2systemtap.pl | 130 ------------------------------------ src/dtrace2systemtap.py | 143 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 147 insertions(+), 135 deletions(-) delete mode 100755 src/dtrace2systemtap.pl create mode 100755 src/dtrace2systemtap.py diff --git a/src/Makefile.am b/src/Makefile.am index d10bfdf448..5a1df7a93c 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -604,7 +604,6 @@ nodist_libvirt_la_SOURCES = libvirt_probes.h if WITH_REMOTE nodist_libvirt_driver_remote_la_SOURCES = libvirt_probes.h endif WITH_REMOTE -DTRACE2SYSTEMTAP_FLAGS = --with-modules BUILT_SOURCES += libvirt_probes.h libvirt_probes.stp libvirt_functions.stp @@ -639,10 +638,10 @@ RPC_PROBE_FILES += $(srcdir)/rpc/virnetprotocol.x \ libvirt_functions.stp: $(RPC_PROBE_FILES) $(srcdir)/rpc/gensystemtap.pl $(AM_V_GEN)$(PERL) -w $(srcdir)/rpc/gensystemtap.pl $(RPC_PROBE_FILES) > $@ -%_probes.stp: %_probes.d $(srcdir)/dtrace2systemtap.pl \ +%_probes.stp: %_probes.d $(srcdir)/dtrace2systemtap.py \ $(top_builddir)/config.status - $(AM_V_GEN)$(PERL) -w $(srcdir)/dtrace2systemtap.pl \ - $(DTRACE2SYSTEMTAP_FLAGS) $(bindir) $(sbindir) $(libdir) $< > $@ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/dtrace2systemtap.py \ + $(bindir) $(sbindir) $(libdir) $< > $@ CLEANFILES += libvirt_probes.h libvirt_probes.o libvirt_probes.lo \ libvirt_functions.stp libvirt_probes.stp @@ -763,7 +762,7 @@ endif LIBVIRT_INIT_SCRIPT_SYSTEMD endif WITH_LIBVIRTD -EXTRA_DIST += dtrace2systemtap.pl +EXTRA_DIST += dtrace2systemtap.py if WITH_LIBVIRTD diff --git a/src/dtrace2systemtap.pl b/src/dtrace2systemtap.pl deleted file mode 100755 index c5fce248b4..0000000000 --- a/src/dtrace2systemtap.pl +++ /dev/null @@ -1,130 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2011-2012 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# -# Generate a set of systemtap probe definitions corresponding to -# DTrace probe markers in libvirt.so -# -# perl dtrace2systemtap.pl probes.d > libvirt_probes.stp -# - -use strict; -use warnings; - -my $file; -my @files; -my %files; - -my $with_modules = 0; -if ($ARGV[0] eq "--with-modules") { - # set if we want to honor the "module" setting in the .d file - $with_modules = 1; - shift @ARGV; -} - -my $bindir = shift @ARGV; -my $sbindir = shift @ARGV; -my $libdir = shift @ARGV; - -my $probe; -my $args; - -# Read the DTraceprobes definition -while (<>) { - next if m,^\s*$,; - - next if /^\s*provider\s+\w+\s*{\s*$/; - next if /^\s*};\s*$/; - - if (m,^\s*\#,) { - if (m,^\s*\#\s*file:\s*(\S+)\s*$,) { - $file = $1; - push @files, $file; - $files{$file} = { prefix => undef, probes => [] }; - } elsif (m,^\s*\#\s*prefix:\s*(\S+)\s*$,) { - $files{$file}->{prefix} = $1; - } elsif (m,^\s*\#\s*binary:\s*(\S+)\s*$,) { - $files{$file}->{binary} = $1; - } elsif (m,^\s*\#\s*module:\s*(\S+)\s*$,) { - $files{$file}->{module} = $1; - } else { - # ignore unknown comments - } - } else { - if (m,\s*probe\s+([a-zA-Z0-9_]+)\((.*?)(\);)?$,) { - $probe = $1; - $args = $2; - if ($3) { - push @{$files{$file}->{probes}}, [$probe, $args]; - $probe = $args = undef; - } - } elsif ($probe) { - if (m,^(.*?)(\);)?$,) { - $args .= $1; - if ($2) { - push @{$files{$file}->{probes}}, [$probe, $args]; - $probe = $args = undef; - } - } else { - die "unexpected data $_ on line $."; - } - } else { - die "unexpected data $_ on line $."; - } - } -} - -# Write out the SystemTap probes -foreach my $file (@files) { - my $prefix = $files{$file}->{prefix}; - my @probes = @{$files{$file}->{probes}}; - - print "# $file\n\n"; - foreach my $probe (@probes) { - my $name = $probe->[0]; - my $args = $probe->[1]; - - my $pname = $name; - $pname =~ s/${prefix}_/libvirt.$prefix./; - - my $binary = "$libdir/libvirt.so"; - if (exists $files{$file}->{binary}) { - $binary = $sbindir . "/" . $files{$file}->{binary}; - } - if ($with_modules && exists $files{$file}->{module}) { - $binary = $libdir . "/" . $files{$file}->{module}; - } - - print "probe $pname = process(\"$binary\").mark(\"$name\") {\n"; - - my @args = split /,/, $args; - for (my $i = 0 ; $i <= $#args ; $i++) { - my $arg = $args[$i]; - my $isstr = $arg =~ /char\s+\*/; - $arg =~ s/^.*\s\*?(\S+)$/$1/; - - if ($isstr) { - print " $arg = user_string(\$arg", $i + 1, ");\n"; - } else { - print " $arg = \$arg", $i + 1, ";\n"; - } - } - print "}\n\n"; - } - print "\n"; -} diff --git a/src/dtrace2systemtap.py b/src/dtrace2systemtap.py new file mode 100755 index 0000000000..8ee5632d6b --- /dev/null +++ b/src/dtrace2systemtap.py @@ -0,0 +1,143 @@ +#!/usr/bin/env python +# +# Copyright (C) 2011-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# +# Generate a set of systemtap probe definitions corresponding to +# DTrace probe markers in libvirt.so +# +# python dtrace2systemtap.py probes.d > libvirt_probes.stp +# + +from __future__ import print_function + +import re +import sys + +file = None +filelist = [] +files = {} + +bindir = sys.argv[1] +sbindir = sys.argv[2] +libdir = sys.argv[3] +dtrace = sys.argv[4] + +probe = None +args = None + +# Read the DTraceprobes definition + +with open(dtrace, "r") as fh: + lineno = 0 + for line in fh: + lineno = lineno + 1 + line = line.strip() + if line == "": + continue + if line.find("provider ") != -1 and line.find("{") != -1: + continue + if line.find("};") != -1: + continue + + if line.startswith("#"): + m = re.match(r'''^\#\s*file:\s*(\S+)$''', line) + if m is not None: + file = m.group(1) + filelist.append(file) + files[file] = {"prefix": None, "probes": []} + continue + + m = re.match(r'''^\#\s*prefix:\s*(\S+)$''', line) + if m is not None: + files[file]["prefix"] = m.group(1) + continue + + m = re.match(r'''^\#\s*binary:\s*(\S+)$''', line) + if m is not None: + files[file]["binary"] = m.group(1) + continue + + m = re.match(r'''^\#\s*module:\s*(\S+)$''', line) + if m is not None: + files[file]["module"] = m.group(1) + + # ignore unknown comments + else: + m = re.match(r'''probe\s+([a-zA-Z0-9_]+)\((.*?)(\);)?$''', line) + if m is not None: + probe = m.group(1) + args = m.group(2) + if m.group(3) is not None: + files[file]["probes"].append([probe, args]) + probe = None + args = None + elif probe is not None: + m = re.match(r'''^(.*?)(\);)?$''', line) + if m is not None: + args = args + m.group(1) + if m.group(2) is not None: + files[file]["probes"].append([probe, args]) + probe = None + args = None + else: + raise Exception("unexpected data %s on line %d" % + (line, lineno)) + else: + raise Exception("unexpected data %s on line %d" % + (line, lineno)) + +# Write out the SystemTap probes +for file in filelist: + prefix = files[file]["prefix"] + probes = files[file]["probes"] + + print("# %s\n" % file) + for probe in probes: + name = probe[0] + args = probe[1] + + pname = name.replace(prefix + "_", "libvirt." + prefix + ".") + + binary = libdir + "/libvirt.so" + if "binary" in files[file]: + binary = sbindir + "/" + files[file]["binary"] + if "module" in files[file]: + binary = libdir + "/" + files[file]["module"] + + print("probe %s = process(\"%s\").mark(\"%s\") {" % + (pname, binary, name)) + + argbits = args.split(",") + for idx in range(len(argbits)): + arg = argbits[idx] + isstr = False + if arg.find("char *") != -1: + isstr = True + + m = re.match(r'''^.*\s\*?(\S+)$''', arg) + if m is not None: + arg = m.group(1) + else: + raise Exception("Malformed arg %s" % arg) + + if isstr: + print(" %s = user_string($arg%d);" % (arg, idx + 1)) + else: + print(" %s = $arg%d;" % (arg, idx + 1)) + print("}\n") + print("") -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the gensystemtap.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- cfg.mk | 4 +- src/Makefile.am | 5 +- src/rpc/Makefile.inc.am | 2 +- src/rpc/gensystemtap.pl | 193 ---------------------------------------- src/rpc/gensystemtap.py | 184 ++++++++++++++++++++++++++++++++++++++ 5 files changed, 189 insertions(+), 199 deletions(-) delete mode 100755 src/rpc/gensystemtap.pl create mode 100755 src/rpc/gensystemtap.py diff --git a/cfg.mk b/cfg.mk index f0aed0365a..18bab6e9a4 100644 --- a/cfg.mk +++ b/cfg.mk @@ -407,6 +407,7 @@ sc_prohibit_risky_id_promotion: # since gnulib has more guarantees for snprintf portability sc_prohibit_sprintf: @prohibit='\<[s]printf\>' \ + in_vc_files='\.[ch]$$' \ halt='use snprintf, not sprintf' \ $(_sc_search_regexp) @@ -1272,9 +1273,6 @@ exclude_file_name_regexp--sc_prohibit_readlink = \ exclude_file_name_regexp--sc_prohibit_setuid = ^src/util/virutil\.c|tools/virt-login-shell\.c$$ -exclude_file_name_regexp--sc_prohibit_sprintf = \ - ^(cfg\.mk|docs/hacking\.html\.in|.*\.stp|.*\.pl)$$ - exclude_file_name_regexp--sc_prohibit_strncpy = ^src/util/virstring\.c$$ exclude_file_name_regexp--sc_prohibit_strtol = ^examples/.*$$ diff --git a/src/Makefile.am b/src/Makefile.am index 5a1df7a93c..5226871a73 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -635,8 +635,9 @@ RPC_PROBE_FILES += $(srcdir)/rpc/virnetprotocol.x \ $(srcdir)/remote/qemu_protocol.x \ $(srcdir)/admin/admin_protocol.x -libvirt_functions.stp: $(RPC_PROBE_FILES) $(srcdir)/rpc/gensystemtap.pl - $(AM_V_GEN)$(PERL) -w $(srcdir)/rpc/gensystemtap.pl $(RPC_PROBE_FILES) > $@ +libvirt_functions.stp: $(RPC_PROBE_FILES) $(srcdir)/rpc/gensystemtap.py + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/rpc/gensystemtap.py \ + $(RPC_PROBE_FILES) > $@ %_probes.stp: %_probes.d $(srcdir)/dtrace2systemtap.py \ $(top_builddir)/config.status diff --git a/src/rpc/Makefile.inc.am b/src/rpc/Makefile.inc.am index b8ca53c69a..b43ee4ab7a 100644 --- a/src/rpc/Makefile.inc.am +++ b/src/rpc/Makefile.inc.am @@ -3,7 +3,7 @@ EXTRA_DIST += \ rpc/gendispatch.pl \ rpc/genprotocol.pl \ - rpc/gensystemtap.pl \ + rpc/gensystemtap.py \ rpc/virnetprotocol.x \ rpc/virkeepaliveprotocol.x \ $(NULL) diff --git a/src/rpc/gensystemtap.pl b/src/rpc/gensystemtap.pl deleted file mode 100755 index 6693d4d6f5..0000000000 --- a/src/rpc/gensystemtap.pl +++ /dev/null @@ -1,193 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2011-2012 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# Generate a set of systemtap functions for translating various -# RPC enum values into strings -# -# perl gensystemtap.pl */*.x > libvirt_functions.stp -# - -use strict; - -my %funcs; - -my %type; -my %status; -my %auth; - -my $instatus = 0; -my $intype = 0; -my $inauth = 0; -while (<>) { - if (/enum\s+virNetMessageType/) { - $intype = 1; - } elsif (/enum\s+virNetMessageStatus/) { - $instatus = 1; - } elsif (/enum remote_auth_type/) { - $inauth = 1; - } elsif (/}/) { - $instatus = $intype = $inauth = 0; - } elsif ($instatus) { - if (/^\s+VIR_NET_(\w+)\s*=\s*(\d+),?$/) { - $status{$2} = lc $1; - } - } elsif ($intype) { - if (/^\s+VIR_NET_(\w+)\s*=\s*(\d+),?$/) { - $type{$2} = lc $1; - } - } elsif ($inauth) { - if (/^\s+REMOTE_AUTH_(\w+)\s*=\s*(\d+),?$/) { - $auth{$2} = lc $1; - } - } else { - if (/(?:VIR_)?(\w+?)(?:_PROTOCOL)?_PROGRAM\s*=\s*0x([a-fA-F0-9]+)\s*;/) { - $funcs{lc $1} = { id => hex($2), version => undef, progs => [] }; - } elsif (/(?:VIR_)?(\w+?)(?:_PROTOCOL)?_(?:PROGRAM|PROTOCOL)_VERSION\s*=\s*(\d+)\s*;/) { - $funcs{lc $1}->{version} = $2; - } elsif (/(?:VIR_)?(\w+?)(?:_PROTOCOL)?_PROC_(.*?)\s+=\s+(\d+)/) { - $funcs{lc $1}->{progs}->[$3] = lc $2; - } - } -} - -print <<EOF; -function libvirt_rpc_auth_name(type, verbose) -{ -EOF -my $first = 1; -foreach my $type (sort(keys %auth)) { - my $cond = $first ? "if" : "} else if"; - $first = 0; - print " $cond (type == ", $type, ") {\n"; - print " typestr = \"", $auth{$type}, "\"\n"; -} -print <<EOF; - } else { - typestr = "unknown"; - verbose = 1; - } - if (verbose) { - typestr = typestr . sprintf(":%d", type) - } - return typestr; -} -EOF - -print <<EOF; -function libvirt_rpc_type_name(type, verbose) -{ -EOF -$first = 1; -foreach my $type (sort(keys %type)) { - my $cond = $first ? "if" : "} else if"; - $first = 0; - print " $cond (type == ", $type, ") {\n"; - print " typestr = \"", $type{$type}, "\"\n"; -} -print <<EOF; - } else { - typestr = "unknown"; - verbose = 1; - } - if (verbose) { - typestr = typestr . sprintf(":%d", type) - } - return typestr; -} -EOF - -print <<EOF; -function libvirt_rpc_status_name(status, verbose) -{ -EOF -$first = 1; -foreach my $status (sort(keys %status)) { - my $cond = $first ? "if" : "} else if"; - $first = 0; - print " $cond (status == ", $status, ") {\n"; - print " statusstr = \"", $status{$status}, "\"\n"; -} -print <<EOF; - } else { - statusstr = "unknown"; - verbose = 1; - } - if (verbose) { - statusstr = statusstr . sprintf(":%d", status) - } - return statusstr; -} -EOF - -print <<EOF; -function libvirt_rpc_program_name(program, verbose) -{ -EOF -$first = 1; -foreach my $prog (sort(keys %funcs)) { - my $cond = $first ? "if" : "} else if"; - $first = 0; - print " $cond (program == ", $funcs{$prog}->{id}, ") {\n"; - print " programstr = \"", $prog, "\"\n"; -} -print <<EOF; - } else { - programstr = "unknown"; - verbose = 1; - } - if (verbose) { - programstr = programstr . sprintf(":%d", program) - } - return programstr; -} -EOF - - -print <<EOF; -function libvirt_rpc_procedure_name(program, version, proc, verbose) -{ -EOF -$first = 1; -foreach my $prog (sort(keys %funcs)) { - my $cond = $first ? "if" : "} else if"; - $first = 0; - print " $cond (program == ", $funcs{$prog}->{id}, " && version == ", $funcs{$prog}->{version}, ") {\n"; - - my $pfirst = 1; - for (my $id = 1 ; $id <= $#{$funcs{$prog}->{progs}} ; $id++) { - my $cond = $pfirst ? "if" : "} else if"; - $pfirst = 0; - print " $cond (proc == $id) {\n"; - print " procstr = \"", $funcs{$prog}->{progs}->[$id], "\";\n"; - } - print " } else {\n"; - print " procstr = \"unknown\";\n"; - print " verbose = 1;\n"; - print " }\n"; -} -print <<EOF; - } else { - procstr = "unknown"; - verbose = 1; - } - if (verbose) { - procstr = procstr . sprintf(":%d", proc) - } - return procstr; -} -EOF diff --git a/src/rpc/gensystemtap.py b/src/rpc/gensystemtap.py new file mode 100755 index 0000000000..3bd06927af --- /dev/null +++ b/src/rpc/gensystemtap.py @@ -0,0 +1,184 @@ +#!/usr/bin/env python +# +# Copyright (C) 2011-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# Generate a set of systemtap functions for translating various +# RPC enum values into strings +# +# python gensystemtap.py */*.x > libvirt_functions.stp +# + +from __future__ import print_function + +import re +import sys + +funcs = {} + +types = {} +status = {} +auth = {} + + +def load_file(fh): + instatus = False + intype = False + inauth = False + + for line in fh: + if re.match(r'''\s*enum\s+virNetMessageType\s*''', line): + intype = True + elif re.match(r'''\s*enum\s+virNetMessageStatus\s*''', line): + instatus = True + elif re.match(r'''\s*enum remote_auth_type.*''', line): + inauth = True + elif line.find("}") != -1: + intype = False + instatus = False + inauth = False + elif instatus: + m = re.match(r'''^\s+VIR_NET_(\w+)\s*=\s*(\d+),?$''', line) + if m is not None: + status[m.group(2)] = m.group(1).lower() + elif intype: + m = re.match(r'''^\s+VIR_NET_(\w+)\s*=\s*(\d+),?$''', line) + if m is not None: + types[m.group(2)] = m.group(1).lower() + elif inauth: + m = re.match(r'''^\s+REMOTE_AUTH_(\w+)\s*=\s*(\d+),?$''', line) + if m is not None: + auth[m.group(2)] = m.group(1).lower() + else: + m = re.match(r'''^.*?(?:VIR_)?(\w+?)(?:_PROTOCOL)?''' + + r'''_PROGRAM\s*=\s*0x([a-fA-F0-9]+)\s*;''', line) + if m is not None: + funcs[m.group(1).lower()] = { + "id": int(m.group(2), 16), + "version": None, + "progs": [] + } + continue + + m = re.match(r'''^.*?(?:VIR_)?(\w+?)(?:_PROTOCOL)?_''' + + r'''(?:PROGRAM|PROTOCOL)_VERSION\s*=\s*(\d+)\s*;''', + line) + if m is not None: + funcs[m.group(1).lower()]["version"] = m.group(2) + continue + + m = re.match(r'''^.*?(?:VIR_)?(\w+?)(?:_PROTOCOL)?''' + + r'''_PROC_(.*?)\s+=\s+(\d+)''', line) + if m is not None: + funcs[m.group(1).lower()]["progs"].insert( + int(m.group(3)), m.group(2).lower()) + + +for file in sys.argv[1:]: + with open(file, "r") as fh: + load_file(fh) + + +def genfunc(name, varname, types): + print("function %s(%s, verbose)" % (name, varname)) + print("{") + + first = True + for typename in sorted(types.keys()): + cond = "} else if" + if first: + cond = "if" + first = False + + print(" %s (%s == %s) {" % (cond, varname, typename)) + print(" %sstr = \"%s\"" % (varname, types[typename])) + + print(" } else {") + print(" %sstr = \"unknown\";" % varname) + print(" verbose = 1;") + print(" }") + print(" if (verbose) {") + print(" %sstr = %sstr . sprintf(\":%%d\", %s)" % + (varname, varname, varname)) + print(" }") + print(" return %sstr;" % varname) + print("}") + + +genfunc("libvirt_rpc_auth_name", "type", auth) +genfunc("libvirt_rpc_type_name", "type", types) +genfunc("libvirt_rpc_status_name", "status", status) + +print("function libvirt_rpc_program_name(program, verbose)") +print("{") + +first = True +for funcname in sorted(funcs.keys()): + cond = "} else if" + if first: + cond = "if" + first = False + + print(" %s (program == %s) {" % (cond, funcs[funcname]["id"])) + print(" programstr = \"%s\"" % funcname) + +print(" } else {") +print(" programstr = \"unknown\";") +print(" verbose = 1;") +print(" }") +print(" if (verbose) {") +print(" programstr = programstr . sprintf(\":%d\", program)") +print(" }") +print(" return programstr;") +print("}") + +print("function libvirt_rpc_procedure_name(program, version, proc, verbose)") +print("{") + +first = True +for prog in sorted(funcs.keys()): + cond = "} else if" + if first: + cond = "if" + first = False + + print(" %s (program == %s && version == %s) {" % + (cond, funcs[prog]["id"], funcs[prog]["version"])) + + pfirst = True + for id in range(len(funcs[prog]["progs"])): + pcond = "} else if" + if pfirst: + pcond = "if" + pfirst = False + + print(" %s (proc == %s) {" % (pcond, id + 1)) + print(" procstr = \"%s\";" % funcs[prog]["progs"][id]) + + print(" } else {") + print(" procstr = \"unknown\";") + print(" verbose = 1;") + print(" }") + +print(" } else {") +print(" procstr = \"unknown\";") +print(" verbose = 1;") +print(" }") +print(" if (verbose) {") +print(" procstr = procstr . sprintf(\":%d\", proc)") +print(" }") +print(" return procstr;") +print("}") -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-drivername.pl tool in Python. This was mostly a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. In testing though it was discovered the existing code was broken since it hadn't been updated after driver.h was split into many files. Since the old code is being thrown away, the fix was done as part of the rewrite rather than split into a separate commit. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 18 +++++-- src/check-drivername.pl | 83 ----------------------------- src/check-drivername.py | 114 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 129 insertions(+), 86 deletions(-) delete mode 100755 src/check-drivername.pl create mode 100644 src/check-drivername.py diff --git a/src/Makefile.am b/src/Makefile.am index 5226871a73..0d387a1335 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -335,15 +335,27 @@ check-protocol: endif !WITH_REMOTE EXTRA_DIST += $(PROTOCOL_STRUCTS) +DRIVERS = \ + $(srcdir)/driver-hypervisor.h \ + $(srcdir)/driver-interface.h \ + $(srcdir)/driver-network.h \ + $(srcdir)/driver-nodedev.h \ + $(srcdir)/driver-nwfilter.h \ + $(srcdir)/driver-secret.h \ + $(srcdir)/driver-state.h \ + $(srcdir)/driver-storage.h \ + $(srcdir)/driver-stream.h \ + $(NULL) + check-drivername: - $(AM_V_GEN)$(PERL) $(srcdir)/check-drivername.pl \ - $(srcdir)/driver.h \ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-drivername.py \ + $(DRIVERS) \ $(srcdir)/libvirt_public.syms \ $(srcdir)/libvirt_admin_public.syms \ $(srcdir)/libvirt_qemu.syms \ $(srcdir)/libvirt_lxc.syms -EXTRA_DIST += check-drivername.pl +EXTRA_DIST += check-drivername.py check-driverimpls: $(AM_V_GEN)$(PERL) $(srcdir)/check-driverimpls.pl \ diff --git a/src/check-drivername.pl b/src/check-drivername.pl deleted file mode 100755 index 3a62193e33..0000000000 --- a/src/check-drivername.pl +++ /dev/null @@ -1,83 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# - -use strict; -use warnings; - -my $drvfile = shift; -my @symfiles = @ARGV; - -my %symbols; - -foreach my $symfile (@symfiles) { - open SYMFILE, "<", $symfile - or die "cannot read $symfile: $!"; - while (<SYMFILE>) { - if (/^\s*(vir\w+)\s*;\s*$/) { - $symbols{$1} = 1; - } - } - - close SYMFILE; -} - -open DRVFILE, "<", $drvfile - or die "cannot read $drvfile: $!"; - -my $status = 0; - -while (<DRVFILE>) { - next if /virDrvConnectSupportsFeature/; - if (/\*(virDrv\w+)\s*\)/) { - - my $drv = $1; - - next if $drv =~ /virDrvState/; - next if $drv =~ /virDrvDomainMigrate(Prepare|Perform|Confirm|Begin|Finish)/; - - my $sym = $drv; - $sym =~ s/virDrv/vir/; - - unless (exists $symbols{$sym}) { - print "Driver method name $drv doesn't match public API name\n"; - $status = 1; - } - } elsif (/^\*(vir\w+)\s*\)/) { - my $name = $1; - print "Bogus name $1\n"; - $status = 1; - } elsif (/^\s*(virDrv\w+)\s+(\w+);\s*/) { - my $drv = $1; - my $field = $2; - - my $tmp = $drv; - $tmp =~ s/virDrv//; - $tmp =~ s/^NWFilter/nwfilter/; - $tmp =~ s/^(\w)/lc $1/e; - - unless ($tmp eq $field) { - print "Driver struct field $field should be named $tmp\n"; - $status = 1; - } - } -} - -close DRVFILE; - -exit $status; diff --git a/src/check-drivername.py b/src/check-drivername.py new file mode 100644 index 0000000000..2f024f5dee --- /dev/null +++ b/src/check-drivername.py @@ -0,0 +1,114 @@ +#!/usr/bin/env python +# +# Copyright (C) 2013-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# + +from __future__ import print_function + +import re +import sys + +drvfiles = [] +symfiles = [] +for arg in sys.argv: + if arg.endswith(".h"): + drvfiles.append(arg) + else: + symfiles.append(arg) + +symbols = {} + +for symfile in symfiles: + with open(symfile, "r") as fh: + for line in fh: + m = re.match(r'''^\s*(vir\w+)\s*;\s*$''', line) + if m is not None: + symbols[m.group(1)] = True + +status = 0 +for drvfile in drvfiles: + with open(drvfile, "r") as fh: + for line in fh: + if line.find("virDrvConnectSupportsFeature") != -1: + continue + + m = re.match(r'''.*\*(virDrv\w+)\s*\).*''', line) + if m is not None: + drv = m.group(1) + + skip = [ + "virDrvStateInitialize", + "virDrvStateCleanup", + "virDrvStateReload", + "virDrvStateStop", + "virDrvConnectURIProbe", + "virDrvDomainMigratePrepare", + "virDrvDomainMigratePrepare2", + "virDrvDomainMigratePrepare3", + "virDrvDomainMigratePrepare3Params", + "virDrvDomainMigratePrepareTunnel", + "virDrvDomainMigratePrepareTunnelParams", + "virDrvDomainMigratePrepareTunnel3", + "virDrvDomainMigratePrepareTunnel3Params", + "virDrvDomainMigratePerform", + "virDrvDomainMigratePerform3", + "virDrvDomainMigratePerform3Params", + "virDrvDomainMigrateConfirm", + "virDrvDomainMigrateConfirm3", + "virDrvDomainMigrateConfirm3Params", + "virDrvDomainMigrateBegin", + "virDrvDomainMigrateBegin3", + "virDrvDomainMigrateBegin3Params", + "virDrvDomainMigrateFinish", + "virDrvDomainMigrateFinish2", + "virDrvDomainMigrateFinish3", + "virDrvDomainMigrateFinish3Params", + "virDrvStreamInData", + ] + if drv in skip: + continue + + sym = drv.replace("virDrv", "vir") + + if sym not in symbols: + print("Driver method name %s doesn't match public API" % + drv) + continue + + m = re.match(r'''^\*(vir\w+)\s*\)''', line) + if m is not None: + name = m.group(1) + print("Bogus name %s" % name) + status = 1 + continue + + m = re.match(r'''^\s*(virDrv\w+)\s+(\w+);\s*''', line) + if m is not None: + drv = m.group(1) + field = m.group(2) + + tmp = drv.replace("virDrv", "") + if tmp.startswith("NWFilter"): + tmp = "nwfilter" + tmp[8:] + tmp = tmp[0:1].lower() + tmp[1:] + + if tmp != field: + print("Driver struct field %s should be named %s" % + (field, tmp)) + status = 1 + +sys.exit(status) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-driverimpls.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 4 +- src/check-driverimpls.pl | 80 ------------------------------ src/check-driverimpls.py | 102 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 104 insertions(+), 82 deletions(-) delete mode 100755 src/check-driverimpls.pl create mode 100755 src/check-driverimpls.py diff --git a/src/Makefile.am b/src/Makefile.am index 0d387a1335..1066f7f7cc 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -358,7 +358,7 @@ check-drivername: EXTRA_DIST += check-drivername.py check-driverimpls: - $(AM_V_GEN)$(PERL) $(srcdir)/check-driverimpls.pl \ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-driverimpls.py \ $(filter /%,$(DRIVER_SOURCE_FILES)) \ $(filter $(srcdir)/%,$(DRIVER_SOURCE_FILES)) \ $(addprefix $(srcdir)/,$(filter-out $(srcdir)/%, \ @@ -374,7 +374,7 @@ check-aclperms: $(srcdir)/access/viraccessperm.h \ $(srcdir)/access/viraccessperm.c -EXTRA_DIST += check-driverimpls.pl check-aclrules.pl check-aclperms.py +EXTRA_DIST += check-driverimpls.py check-aclrules.pl check-aclperms.py check-local: check-protocol check-symfile check-symsorting \ check-drivername check-driverimpls check-aclrules \ diff --git a/src/check-driverimpls.pl b/src/check-driverimpls.pl deleted file mode 100755 index 3c0d54724c..0000000000 --- a/src/check-driverimpls.pl +++ /dev/null @@ -1,80 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# - -use strict; -use warnings; - -my $intable = 0; -my $table; -my $mainprefix; - -my $status = 0; -while (<>) { - if ($intable) { - if (/}/) { - $intable = 0; - $table = undef; - $mainprefix = undef; - } elsif (/\.(\w+)\s*=\s*(\w+),?/) { - my $api = $1; - my $impl = $2; - - next if $api eq "no"; - next if $api eq "name"; - next if $impl eq "NULL"; - - my $suffix = $impl; - my $prefix = $impl; - $prefix =~ s/^([a-z]+)(.*?)$/$1/; - - if (defined $mainprefix) { - if ($mainprefix ne $prefix) { - print "$ARGV:$. Bad prefix '$prefix' for API '$api', expecting '$mainprefix'\n"; - $status = 1; - } - } else { - $mainprefix = $prefix; - } - - if ($api !~ /^$mainprefix/) { - $suffix =~ s/^[a-z]+//; - $suffix =~ s/^([A-Z]+)/lc $1/e; - } - - if ($api ne $suffix) { - my $want = $api; - $want =~ s/^nwf/NWF/; - if ($api !~ /^$mainprefix/) { - $want =~ s/^([a-z])/uc $1/e; - $want = $mainprefix . $want; - } - print "$ARGV:$. Bad impl name '$impl' for API '$api', expecting '$want'\n"; - $status = 1; - } - } - } elsif (/^(?:static\s+)?(vir(?:\w+)?Driver)\s+(?!.*;)/) { - next if $1 eq "virNWFilterCallbackDriver" || - $1 eq "virNWFilterTechDriver" || - $1 eq "virConnectDriver"; - $intable = 1; - $table = $1; - } -} - -exit $status; diff --git a/src/check-driverimpls.py b/src/check-driverimpls.py new file mode 100755 index 0000000000..30c92698b3 --- /dev/null +++ b/src/check-driverimpls.py @@ -0,0 +1,102 @@ +#!/usr/bin/env python +# +# Copyright (C) 2013-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# + +from __future__ import print_function + +import re +import sys + + +def checkdriverimpls(filename): + intable = False + mainprefix = None + + errs = False + with open(filename, "r") as fh: + lineno = 0 + for line in fh: + lineno = lineno + 1 + if intable: + if line.find("}") != -1: + intable = False + mainprefix = None + continue + + m = re.match(r'''\s*\.(\w+)\s*=\s*(\w+),?\s*/''', line) + if m is not None: + api = m.group(1) + impl = m.group(2) + + if api in ["no", "name"]: + continue + if impl in ["NULL"]: + continue + + suffix = impl + prefix = re.sub(r'''^([a-z]+)(.*?)$''', r'''\1''', impl) + + if mainprefix is not None: + if mainprefix != prefix: + print("%s:%d Bad prefix '%s' for API '%s', " + + "expecting '%s'" % + (filename, lineno, prefix, api, mainprefix), + file=sys.stderr) + errs = True + else: + mainprefix = prefix + + if not api.startswith(mainprefix): + suffix = re.sub(r'''^[a-z]+''', "", suffix) + suffix = re.sub(r'''^([A-Z]+)''', + lambda m: m.group(1).lower(), suffix) + + if api != suffix: + want = api + if want.startswith("nwf"): + want = "NWF" + want[3:] + + if not api.startswith(mainprefix): + want = re.sub(r'''^([a-z])''', + lambda m: m.group(1).upper(), want) + want = mainprefix + want + + print("%s:%d Bad impl name '%s' for API " + + "'%s', expecting '%s'" % + (filename, lineno, impl, api, want), + file=sys.stderr) + errs = True + else: + m = re.match(r'''^(?:static\s+)?(vir(?:\w+)?Driver)''' + + r'''\s+(?!.*;)''', line) + if m is not None: + drv = m.group(1) + if drv in ["virNWFilterCallbackDriver", + "virNWFilterTechDriver", + "virConnectDriver"]: + continue + intable = True + + return errs + + +status = 0 +for filename in sys.argv[1:]: + if checkdriverimpls(filename): + status = 1 +sys.exit(status) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-aclrules.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/Makefile.am | 4 +- src/check-aclrules.pl | 252 ---------------------------------------- src/check-aclrules.py | 263 ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 265 insertions(+), 254 deletions(-) delete mode 100755 src/check-aclrules.pl create mode 100755 src/check-aclrules.py diff --git a/src/Makefile.am b/src/Makefile.am index 1066f7f7cc..ed1e1ca32b 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -365,7 +365,7 @@ check-driverimpls: $(filter-out /%,$(DRIVER_SOURCE_FILES)))) check-aclrules: - $(AM_V_GEN)$(PERL) $(srcdir)/check-aclrules.pl \ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-aclrules.py \ $(REMOTE_PROTOCOL) \ $(addprefix $(srcdir)/,$(filter-out /%,$(STATEFUL_DRIVER_SOURCE_FILES))) @@ -374,7 +374,7 @@ check-aclperms: $(srcdir)/access/viraccessperm.h \ $(srcdir)/access/viraccessperm.c -EXTRA_DIST += check-driverimpls.py check-aclrules.pl check-aclperms.py +EXTRA_DIST += check-driverimpls.py check-aclrules.py check-aclperms.py check-local: check-protocol check-symfile check-symsorting \ check-drivername check-driverimpls check-aclrules \ diff --git a/src/check-aclrules.pl b/src/check-aclrules.pl deleted file mode 100755 index 0d4cac17ca..0000000000 --- a/src/check-aclrules.pl +++ /dev/null @@ -1,252 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2013-2014 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# This script validates that the driver implementation of any -# public APIs contain ACL checks. -# -# As the script reads each source file, it attempts to identify -# top level function names. -# -# When reading the body of the functions, it looks for anything -# that looks like an API called named XXXEnsureACL. It will -# validate that the XXX prefix matches the name of the function -# it occurs in. -# -# When it later finds the virDriverPtr table, for each entry -# point listed, it will validate if there was a previously -# detected EnsureACL call recorded. -# -use strict; -use warnings; - -my $status = 0; - -my $brace = 0; -my $maybefunc; -my $intable = 0; -my $table; - -my %acls; -my %aclfilters; - -my %whitelist = ( - "connectClose" => 1, - "connectIsEncrypted" => 1, - "connectIsSecure" => 1, - "connectIsAlive" => 1, - "networkOpen" => 1, - "networkClose" => 1, - "nwfilterOpen" => 1, - "nwfilterClose" => 1, - "secretOpen" => 1, - "secretClose" => 1, - "storageOpen" => 1, - "storageClose" => 1, - "interfaceOpen" => 1, - "interfaceClose" => 1, - "connectURIProbe" => 1, - "localOnly" => 1, - "domainQemuAttach" => 1, - ); - -# XXX this vzDomainMigrateConfirm3Params looks -# bogus - determine why it doesn't have a valid -# ACL check. -my %implwhitelist = ( - "vzDomainMigrateConfirm3Params" => 1, - ); - -my $lastfile; - -sub fixup_name { - my $name = shift; - - $name =~ s/Nwfilter/NWFilter/; - $name =~ s/Xml$/XML/; - $name =~ s/Uri$/URI/; - $name =~ s/Uuid$/UUID/; - $name =~ s/Id$/ID/; - $name =~ s/Mac$/MAC/; - $name =~ s/Cpu$/CPU/; - $name =~ s/Os$/OS/; - $name =~ s/Nmi$/NMI/; - $name =~ s/Pm/PM/; - $name =~ s/Fstrim$/FSTrim/; - $name =~ s/Scsi/SCSI/; - $name =~ s/Wwn$/WWN/; - - return $name; -} - -sub name_to_ProcName { - my $name = shift; - - my @elems; - if ($name =~ /_/ || (lc $name) eq "open" || (lc $name) eq "close") { - @elems = split /_/, $name; - @elems = map lc, @elems; - @elems = map ucfirst, @elems; - } else { - @elems = $name; - } - @elems = map { fixup_name($_) } @elems; - my $procname = join "", @elems; - - $procname =~ s/^([A-Z])/lc $1/e; - - return $procname; -} - - -my $proto = shift @ARGV; - -open PROTO, "<$proto" or die "cannot read $proto"; - -my %filtered; -my $incomment = 0; -my $filtered = 0; -while (<PROTO>) { - if (m,/\*\*,) { - $incomment = 1; - $filtered = 0; - } elsif ($incomment) { - if (m,\*\s\@aclfilter,) { - $filtered = 1; - } elsif ($filtered && - m,REMOTE_PROC_(.*)\s+=\s*\d+,) { - my $api = name_to_ProcName($1); - # Event filtering is handled in daemon/remote.c instead of drivers - if (! m,_EVENT_REGISTER,) { - $filtered{$api} = 1; - } - $incomment = 0; - } - } -} - -close PROTO; - -while (<>) { - if (!defined $lastfile || - $lastfile ne $ARGV) { - %acls = (); - $brace = 0; - $maybefunc = undef; - $lastfile = $ARGV; - } - if ($brace == 0) { - # Looks for anything which appears to be a function - # body name. Doesn't matter if we pick up bogus stuff - # here, as long as we don't miss valid stuff - if (m,\b(\w+)\(,) { - $maybefunc = $1; - } - } elsif ($brace > 0) { - if (m,(\w+)EnsureACL,) { - # Record the fact that maybefunc contains an - # ACL call, and make sure it is the right call! - my $func = $1; - $func =~ s/^vir//; - if (!defined $maybefunc) { - print "$ARGV:$. Unexpected check '$func' outside function\n"; - $status = 1; - } else { - unless ($maybefunc =~ /$func$/i) { - print "$ARGV:$. Mismatch check 'vir${func}EnsureACL' for function '$maybefunc'\n"; - $status = 1; - } - } - $acls{$maybefunc} = 1; - } elsif (m,(\w+)CheckACL,) { - # Record the fact that maybefunc contains an - # ACL filter call, and make sure it is the right call! - my $func = $1; - $func =~ s/^vir//; - if (!defined $maybefunc) { - print "$ARGV:$. Unexpected check '$func' outside function\n"; - $status = 1; - } else { - unless ($maybefunc =~ /$func$/i) { - print "$ARGV:$. Mismatch check 'vir${func}CheckACL' for function '$maybefunc'\n"; - $status = 1; - } - } - $aclfilters{$maybefunc} = 1; - } elsif (m,\b(\w+)\(,) { - # Handles case where we replaced an API with a new - # one which adds new parameters, and we're left with - # a simple stub calling the new API. - my $callfunc = $1; - if (exists $acls{$callfunc}) { - $acls{$maybefunc} = 1; - } - if (exists $aclfilters{$callfunc}) { - $aclfilters{$maybefunc} = 1; - } - } - } - - # Pass the vir*DriverPtr tables and make sure that - # every func listed there, has an impl which calls - # an ACL function - if ($intable) { - if (/\}/) { - $intable = 0; - $table = undef; - } elsif (/\.(\w+)\s*=\s*(\w+),?/) { - my $api = $1; - my $impl = $2; - - next if $impl eq "NULL"; - - if ($api ne "no" && - $api ne "name" && - $table ne "virStateDriver" && - !exists $acls{$impl} && - !exists $whitelist{$api} && - !exists $implwhitelist{$impl}) { - print "$ARGV:$. Missing ACL check in function '$impl' for '$api'\n"; - $status = 1; - } - - if (exists $filtered{$api} && - !exists $aclfilters{$impl}) { - print "$ARGV:$. Missing ACL filter in function '$impl' for '$api'\n"; - $status = 1; - } - } - } elsif (/^(?:static\s+)?(vir(?:\w+)?Driver)\s+/) { - if ($1 ne "virNWFilterCallbackDriver" && - $1 ne "virNWFilterTechDriver" && - $1 ne "virDomainConfNWFilterDriver") { - $intable = 1; - $table = $1; - } - } - - - my $count; - $count = s/{//g; - $brace += $count; - $count = s/}//g; - $brace -= $count; -} continue { - close ARGV if eof; -} - -exit $status; diff --git a/src/check-aclrules.py b/src/check-aclrules.py new file mode 100755 index 0000000000..a7ccc794be --- /dev/null +++ b/src/check-aclrules.py @@ -0,0 +1,263 @@ +#!/usr/bin/env python +# +# Copyright (C) 2013-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# This script validates that the driver implementation of any +# public APIs contain ACL checks. +# +# As the script reads each source file, it attempts to identify +# top level function names. +# +# When reading the body of the functions, it looks for anything +# that looks like an API called named XXXEnsureACL. It will +# validate that the XXX prefix matches the name of the function +# it occurs in. +# +# When it later finds the virDriverPtr table, for each entry +# point listed, it will validate if there was a previously +# detected EnsureACL call recorded. +# + +from __future__ import print_function + +import re +import sys + +whitelist = { + "connectClose": True, + "connectIsEncrypted": True, + "connectIsSecure": True, + "connectIsAlive": True, + "networkOpen": True, + "networkClose": True, + "nwfilterOpen": True, + "nwfilterClose": True, + "secretOpen": True, + "secretClose": True, + "storageOpen": True, + "storageClose": True, + "interfaceOpen": True, + "interfaceClose": True, + "connectURIProbe": True, + "localOnly": True, + "domainQemuAttach": True, +} + +# XXX this vzDomainMigrateConfirm3Params looks +# bogus - determine why it doesn't have a valid +# ACL check. +implwhitelist = { + "vzDomainMigrateConfirm3Params": True, +} + +lastfile = None + + +def fixup_name(name): + name.replace("Nwfilter", "NWFilter") + name.replace("Pm", "PM") + name.replace("Scsi", "SCSI") + if name.endswith("Xml"): + name = name[:-3] + "XML" + elif name.endswith("Uri"): + name = name[:-3] + "URI" + elif name.endswith("Uuid"): + name = name[:-4] + "UUID" + elif name.endswith("Id"): + name = name[:-2] + "ID" + elif name.endswith("Mac"): + name = name[:-3] + "MAC" + elif name.endswith("Cpu"): + name = name[:-3] + "MAC" + elif name.endswith("Os"): + name = name[:-2] + "OS" + elif name.endswith("Nmi"): + name = name[:-3] + "NMI" + elif name.endswith("Fstrim"): + name = name[:-6] + "FSTrim" + elif name.endswith("Wwn"): + name = name[:-3] + "WWN" + + return name + + +def name_to_ProcName(name): + elems = [] + if name.find("_") != -1 or name.lower() in ["open", "close"]: + elems = [n.lower().capitalize() for n in name.split("_")] + else: + elems = [name] + + elems = [fixup_name(n) for n in elems] + procname = "".join(elems) + + return procname[0:1].lower() + procname[1:] + + +proto = sys.argv[1] + +filteredmap = {} +with open(proto, "r") as fh: + incomment = False + filtered = False + + for line in fh: + if line.find("/**") != -1: + incomment = True + filtered = False + elif incomment: + if line.find("* @aclfilter") != -1: + filtered = True + elif filtered: + m = re.match(r'''.*REMOTE_PROC_(.*)\s+=\s*\d+.*''', line) + if m is not None: + api = name_to_ProcName(m.group(1)) + # Event filtering is handled in daemon/remote.c + # instead of drivers + if line.find("_EVENT_REGISTER") == -1: + filteredmap[api] = True + incomment = False + + +def process_file(filename): + brace = 0 + maybefunc = None + intable = False + table = None + + acls = {} + aclfilters = {} + errs = False + with open(filename, "r") as fh: + lineno = 0 + for line in fh: + lineno = lineno + 1 + if brace == 0: + # Looks for anything which appears to be a function + # body name. Doesn't matter if we pick up bogus stuff + # here, as long as we don't miss valid stuff + m = re.match(r'''.*\b(\w+)\(.*''', line) + if m is not None: + maybefunc = m.group(1) + elif brace > 0: + ensureacl = re.match(r'''.*?(\w+)EnsureACL.*''', line) + checkacl = re.match(r'''.*?(\w+)CheckACL.*''', line) + stub = re.match(r'''.*?\b(\w+)\(.*''', line) + if ensureacl is not None: + # Record the fact that maybefunc contains an + # ACL call, and make sure it is the right call! + func = ensureacl.group(1) + if func.startswith("vir"): + func = func[3:] + + if maybefunc is None: + print("%s:%d Unexpected check '%s' outside function" % + (filename, lineno, func), file=sys.stderr) + errs = True + else: + if not maybefunc.lower().endswith(func.lower()): + print("%s:%d Mismatch check 'vir%sEnsureACL'" + + "for function '%s'" % + (filename, lineno, func, maybefunc), + file=sys.stderr) + errs = True + acls[maybefunc] = True + elif checkacl: + # Record the fact that maybefunc contains an + # ACL filter call, and make sure it is the right call! + func = checkacl.group(1) + if func.startswith("vir"): + func = func[3:] + + if maybefunc is None: + print("%s:%d Unexpected check '%s' outside function" % + (filename, lineno, func), file=sys.stderr) + errs = True + else: + if not maybefunc.lower().endswith(func.lower()): + print("%s:%d Mismatch check 'vir%sEnsureACL' " + + "for function '%s'" % + (filename, lineno, func, maybefunc), + file=sys.stderr) + errs = True + aclfilters[maybefunc] = True + elif stub: + # Handles case where we replaced an API with a new + # one which adds new parameters, and we're left with + # a simple stub calling the new API. + callfunc = stub.group(1) + if callfunc in acls: + acls[maybefunc] = True + + if callfunc in aclfilters: + aclfilters[maybefunc] = True + + # Pass the vir*DriverPtr tables and make sure that + # every func listed there, has an impl which calls + # an ACL function + if intable: + assign = re.match(r'''.*?\.(\w+)\s*=\s*(\w+),?.*''', line) + if line.find("}") != -1: + intable = False + table = None + elif assign is not None: + api = assign.group(1) + impl = assign.group(2) + + if (impl != "NULL" and + api not in ["no", "name"] and + table != "virStateDriver"): + if (impl not in acls and + api not in whitelist and + impl not in implwhitelist): + print("%s:%d Missing ACL check in " + + "function '%s' for '%s'" % + (filename, lineno, impl, api), + file=sys.stderr) + errs = True + + if api in filteredmap and impl not in aclfilters: + print("%s:%d Missing ACL filter in " + + "function '%s' for '%s'" % + (filename, lineno, impl, api), + file=sys.stderr) + errs = True + else: + m = re.match(r'''^(?:static\s+)?(vir(?:\w+)?Driver)\s+''', + line) + if m is not None: + name = m.group(1) + if name not in ["virNWFilterCallbackDriver", + "virNWFilterTechDriver", + "virDomainConfNWFilterDriver"]: + intable = True + table = name + + if line.find("{") != -1: + brace = brace + 1 + if line.find("}") != -1: + brace = brace - 1 + + return errs + + +status = 0 +for filename in sys.argv[2:]: + if process_file(filename): + status = 1 + +sys.exit(status) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the genpolkit.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- src/access/Makefile.inc.am | 7 ++- src/access/genpolkit.pl | 119 ------------------------------------ src/access/genpolkit.py | 122 +++++++++++++++++++++++++++++++++++++ 3 files changed, 126 insertions(+), 122 deletions(-) delete mode 100755 src/access/genpolkit.pl create mode 100755 src/access/genpolkit.py diff --git a/src/access/Makefile.inc.am b/src/access/Makefile.inc.am index 4dc742f4e5..1550cce720 100644 --- a/src/access/Makefile.inc.am +++ b/src/access/Makefile.inc.am @@ -44,7 +44,7 @@ GENERATED_SYM_FILES += $(ACCESS_DRIVER_SYM_FILES) EXTRA_DIST += \ $(ACCESS_DRIVER_POLKIT_POLICY) \ - access/genpolkit.pl \ + access/genpolkit.py \ $(NULL) @@ -63,8 +63,9 @@ libvirt_driver_access_la_LIBADD = $(ACCESS_DRIVER_POLKIT_POLICY): $(srcdir)/access/viraccessperm.h \ - $(srcdir)/access/genpolkit.pl Makefile.am - $(AM_V_GEN)$(PERL) $(srcdir)/access/genpolkit.pl < $< > $@ || rm -f $@ + $(srcdir)/access/genpolkit.py Makefile.am + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) \ + $(srcdir)/access/genpolkit.py $< > $@ || rm -f $@ if WITH_POLKIT libvirt_driver_access_la_SOURCES += $(ACCESS_DRIVER_POLKIT_SOURCES) diff --git a/src/access/genpolkit.pl b/src/access/genpolkit.pl deleted file mode 100755 index f8f20caf65..0000000000 --- a/src/access/genpolkit.pl +++ /dev/null @@ -1,119 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2012-2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# - -use strict; -use warnings; - -my @objects = ( - "CONNECT", "DOMAIN", "INTERFACE", "NETWORK_PORT", - "NETWORK","NODE_DEVICE", "NWFILTER_BINDING", "NWFILTER", - "SECRET", "STORAGE_POOL", "STORAGE_VOL", - ); - -my $objects = join ("|", @objects); - -# Data we're going to be generating looks like this -# -# <policyconfig> -# <action id="org.libvirt.unix.monitor"> -# <description>Monitor local virtualized systems</description> -# <message>System policy prevents monitoring of local virtualized systems</message> -# <defaults> -# <allow_any>yes</allow_any> -# <allow_inactive>yes</allow_inactive> -# <allow_active>yes</allow_active> -# </defaults> -# </action> -# ...more <action> rules... -# </policyconfig> - -my %opts; -my $in_opts = 0; - -my %perms; - -while (<>) { - if ($in_opts) { - if (m,\*/,) { - $in_opts = 0; - } elsif (/\*\s*\@(\w+):\s*(.*?)\s*$/) { - $opts{$1} = $2; - } - } elsif (m,/\*\*,) { - $in_opts = 1; - } elsif (/VIR_ACCESS_PERM_($objects)_((?:\w|_)+),/) { - my $object = lc $1; - my $perm = lc $2; - next if $perm eq "last"; - - $object =~ s/_/-/g; - $perm =~ s/_/-/g; - - $perms{$object} = {} unless exists $perms{$object}; - $perms{$object}->{$perm} = { - desc => $opts{desc}, - message => $opts{message}, - anonymous => $opts{anonymous} - }; - %opts = (); - } -} - -print <<EOF; -<?xml version="1.0" encoding="UTF-8"?> -<!DOCTYPE policyconfig PUBLIC "-//freedesktop//DTD polkit Policy Configuration 1.0//EN" - "http://www.freedesktop.org/software/polkit/policyconfig-1.dtd"> -<policyconfig> - <vendor>Libvirt Project</vendor> - <vendor_url>https://libvirt.org</vendor_url> -EOF - -foreach my $object (sort { $a cmp $b } keys %perms) { - foreach my $perm (sort { $a cmp $b } keys %{$perms{$object}}) { - my $description = $perms{$object}->{$perm}->{desc}; - my $message = $perms{$object}->{$perm}->{message}; - my $anonymous = $perms{$object}->{$perm}->{anonymous}; - - die "missing description for $object.$perm" unless - defined $description; - die "missing message for $object.$perm" unless - defined $message; - - my $allow_any = $anonymous ? "yes" : "no"; - my $allow_inactive = $allow_any; - my $allow_active = $allow_any; - - print <<EOF; - <action id="org.libvirt.api.$object.$perm"> - <description>$description</description> - <message>$message</message> - <defaults> - <allow_any>$allow_any</allow_any> - <allow_inactive>$allow_inactive</allow_inactive> - <allow_active>$allow_active</allow_active> - </defaults> - </action> -EOF - - } -} - -print <<EOF; -</policyconfig> -EOF diff --git a/src/access/genpolkit.py b/src/access/genpolkit.py new file mode 100755 index 0000000000..04628d2d6a --- /dev/null +++ b/src/access/genpolkit.py @@ -0,0 +1,122 @@ +#!/usr/bin/env python +# +# Copyright (C) 2012-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# + +from __future__ import print_function + +import re +import sys + +objects = [ + "CONNECT", "DOMAIN", "INTERFACE", "NETWORK_PORT", + "NETWORK", "NODE_DEVICE", "NWFILTER_BINDING", + "NWFILTER", "SECRET", "STORAGE_POOL", "STORAGE_VOL", +] + +objectstr = "|".join(objects) + +# Data we're going to be generating looks like this +# +# <policyconfig> +# <action id="org.libvirt.unix.monitor"> +# <description>Monitor local virtualized systems</description> +# <message>System policy prevents monitoring of +# local virtualized systems</message> +# <defaults> +# <allow_any>yes</allow_any> +# <allow_inactive>yes</allow_inactive> +# <allow_active>yes</allow_active> +# </defaults> +# </action> +# ...more <action> rules... +# </policyconfig> + +opts = {} +in_opts = False + +perms = {} + +aclfile = sys.argv[1] +with open(aclfile, "r") as fh: + for line in fh: + if in_opts: + if line.find("*/") != -1: + in_opts = False + else: + m = re.match(r'''.*\*\s*\@(\w+):\s*(.*?)\s*$''', line) + if m is not None: + opts[m.group(1)] = m.group(2) + elif line.find("**") != -1: + in_opts = True + else: + m = re.match(r'''.*VIR_ACCESS_PERM_(%s)_((?:\w|_)+),''' % + objectstr, line) + if m is not None: + obj = m.group(1).lower() + perm = m.group(2).lower() + if perm == "last": + continue + + obj = obj.replace("_", "-") + perm = perm.replace("_", "-") + + if obj not in perms: + perms[obj] = {} + perms[obj][perm] = { + "desc": opts.get("desc", None), + "message": opts.get("message", None), + "anonymous": opts.get("anonymous", None), + } + opts = {} + +print('<?xml version="1.0" encoding="UTF-8"?>') +print('<!DOCTYPE policyconfig PUBLIC ' + + '"-//freedesktop//DTD polkit Policy Configuration 1.0//EN"') +print(' "http://www.freedesktop.org/software/polkit/policyconfig-1.dtd">') +print('<policyconfig>') +print(' <vendor>Libvirt Project</vendor>') +print(' <vendor_url>https://libvirt.org</vendor_url>') + +for obj in sorted(perms.keys()): + for perm in sorted(perms[obj].keys()): + description = perms[obj][perm]["desc"] + message = perms[obj][perm]["message"] + anonymous = perms[obj][perm]["anonymous"] + + if description is None: + raise Exception("missing description for %s.%s" % (obj, perm)) + if message is None: + raise Exception("missing message for %s.%s" % (obj, perm)) + + allow_any = "no" + if anonymous: + allow_any = "yes" + allow_inactive = allow_any + allow_active = allow_any + + print(' <action id="org.libvirt.api.%s.%s">' % (obj, perm)) + print(' <description>%s</description>' % description) + print(' <message>%s</message>' % message) + print(' <defaults>') + print(' <allow_any>%s</allow_any>' % allow_any) + print(' <allow_inactive>%s</allow_inactive>' % allow_inactive) + print(' <allow_active>%s</allow_active>' % allow_active) + print(' </defaults>') + print(' </action>') + +print('</policyconfig>') -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the pdwtags processing script in Python. The original inline shell and perl code was completely unintelligible. The new python code is a manual conversion that attempts todo basically the same thing. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- cfg.mk | 3 +- src/Makefile.am | 99 ++++--------------------- src/check-remote-protocol.py | 137 +++++++++++++++++++++++++++++++++++ 3 files changed, 154 insertions(+), 85 deletions(-) create mode 100644 src/check-remote-protocol.py diff --git a/cfg.mk b/cfg.mk index 18bab6e9a4..66df1448b4 100644 --- a/cfg.mk +++ b/cfg.mk @@ -353,6 +353,7 @@ sc_prohibit_mkstemp: # access with F_OK or R_OK is okay, though. sc_prohibit_access_xok: @prohibit='access(at)? *\(.*X_OK' \ + in_vc_files='\.[ch]$$' \ halt='use virFileIsExecutable instead of access(,X_OK)' \ $(_sc_search_regexp) @@ -1227,7 +1228,7 @@ exclude_file_name_regexp--sc_prohibit_PATH_MAX = \ ^cfg\.mk$$ exclude_file_name_regexp--sc_prohibit_access_xok = \ - ^(cfg\.mk|src/util/virutil\.c)$$ + ^(src/util/virutil\.c)$$ exclude_file_name_regexp--sc_prohibit_asprintf = \ ^(cfg\.mk|bootstrap.conf$$|examples/|src/util/virstring\.[ch]$$|tests/vircgroupmock\.c|tools/virt-login-shell\.c|tools/nss/libvirt_nss\.c$$) diff --git a/src/Makefile.am b/src/Makefile.am index ed1e1ca32b..05ad832668 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -192,84 +192,6 @@ DRIVER_SOURCES += \ -# Ensure that we don't change the struct or member names or member ordering -# in remote_protocol.x The embedded perl below needs a few comments, and -# presumes you know what pdwtags output looks like: -# * use -0777 -n to slurp the entire file into $_. -# * the "split" splits on the /* DD */ comments, so that $p iterates -# through the struct definitions. -# * process only "struct remote_..." entries -# * remove comments and preceding TAB throughout -# * remove empty lines throughout -# * remove white space at end of buffer - -# With pdwtags 1.8, --verbose output includes separators like these: -# /* 93 */ -# /* <0> (null):0 */ -# with the second line omitted for intrinsic types. -# Whereas with pdwtags 1.3, they look like this: -# /* <2d2> /usr/include/libio.h:180 */ -# The alternation of the following regexps matches both cases. -r1 = /\* \d+ \*/ -r2 = /\* <[[:xdigit:]]+> \S+:\d+ \*/ -libs_prefix = remote_|qemu_|lxc_|admin_ -other_prefix = keepalive|vir(Net|LockSpace|LXCMonitor) -struct_prefix = ($(libs_prefix)|$(other_prefix)) - -# Depending on configure options, libtool creates one or both of -# remote/{,.libs/}libvirt_driver_remote_la-remote_protocol.o. We want -# the newest of the two, in case configure options changed and a stale -# file is left around from an earlier build. -# The pdwtags output is completely different when building with clang -# which causes the comparison against expected output to fail, so skip -# if using clang as CC. -PDWTAGS = \ - $(AM_V_GEN)if $(CC) -v 2>&1 | grep -q clang; then \ - echo 'WARNING: skipping pdwtags test with Clang' >&2; \ - exit 0; \ - fi; \ - if (pdwtags --help) > /dev/null 2>&1; then \ - o=`ls -t $(<:.lo=.$(OBJEXT)) \ - $(subst /,/.libs/,$(<:.lo=.$(OBJEXT))) \ - 2>/dev/null | sed -n 1p`; \ - test -f "$$o" || { echo ".o for $< not found" >&2; exit 1; }; \ - pdwtags --verbose $$o > $(@F)-t1 2> $(@F)-t2; \ - if test ! -s $(@F)-t1 && test -s $(@F)-t2; then \ - rm -rf $(@F)-t?; \ - echo 'WARNING: pdwtags appears broken; skipping the $@ test' >&2;\ - else \ - $(PERL) -0777 -n \ - -e 'foreach my $$p (split m!\n*(?:$(r1)|$(r2))\n!) {' \ - -e ' if ($$p =~ /^(struct|enum) $(struct_prefix)/ ||' \ - -e ' $$p =~ /^enum \{/) {' \ - -e ' $$p =~ s!\t*/\*.*?\*/!!sg;' \ - -e ' $$p =~ s!\s+\n!\n!sg;' \ - -e ' $$p =~ s!\s+$$!!;' \ - -e ' $$p =~ s!\t! !g;' \ - -e ' print "$$p\n";' \ - -e ' $$n++;' \ - -e ' }' \ - -e '}' \ - -e 'BEGIN {' \ - -e ' print "/* -*- c -*- */\n";' \ - -e '}' \ - -e 'END {' \ - -e ' if ($$n < 1) {' \ - -e ' warn "WARNING: your pdwtags program is too old\n";' \ - -e ' warn "WARNING: skipping the $@ test\n";' \ - -e ' warn "WARNING: install dwarves-1.3 or newer\n";' \ - -e ' exit 8;' \ - -e ' }' \ - -e '}' \ - < $(@F)-t1 > $(@F)-t3; \ - case $$? in 8) rm -f $(@F)-t?; exit 0;; 0) ;; *) exit 1;; esac;\ - diff -u $(@)s $(@F)-t3; st=$$?; rm -f $(@F)-t?; exit $$st; \ - fi; \ - else \ - echo 'WARNING: you lack pdwtags; skipping the $@ test' >&2; \ - echo 'WARNING: install the dwarves package to get pdwtags' >&2; \ - fi - # .libs/libvirt.so is built by libtool as a side-effect of the Makefile # rule for libvirt.la. However, checking symbols relies on Linux ELF layout if WITH_LINUX @@ -306,34 +228,43 @@ PROTOCOL_STRUCTS = \ if WITH_REMOTE check-protocol: $(PROTOCOL_STRUCTS) $(PROTOCOL_STRUCTS:structs=struct) +# Ensure that we don't change the struct or member names or member ordering +# in remote_protocol.x The process-pdwtags.py post-processes output to +# extract the bits we want. + # The .o file that pdwtags parses is created as a side effect of running # libtool; but from make's perspective we depend on the .lo file. $(srcdir)/remote_protocol-struct \ $(srcdir)/qemu_protocol-struct \ $(srcdir)/lxc_protocol-struct: \ $(srcdir)/%-struct: remote/libvirt_driver_remote_la-%.lo - $(PDWTAGS) + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-remote-protocol.py \ + "$(CC)" "$(OBJEXT)" $< $(@)s $(srcdir)/virnetprotocol-struct $(srcdir)/virkeepaliveprotocol-struct: \ $(srcdir)/%-struct: rpc/libvirt_net_rpc_la-%.lo - $(PDWTAGS) + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-remote-protocol.py \ + "$(CC)" "$(OBJEXT)" $< $(@)s if WITH_LXC $(srcdir)/lxc_monitor_protocol-struct: \ $(srcdir)/%-struct: lxc/libvirt_driver_lxc_impl_la-%.lo - $(PDWTAGS) + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-remote-protocol.py \ + "$(CC)" "$(OBJEXT)" $< $(@)s endif WITH_LXC $(srcdir)/lock_protocol-struct: \ $(srcdir)/%-struct: locking/lockd_la-%.lo - $(PDWTAGS) + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-remote-protocol.py \ + "$(CC)" "$(OBJEXT)" $< $(@)s $(srcdir)/admin_protocol-struct: \ $(srcdir)/%-struct: admin/libvirt_admin_la-%.lo - $(PDWTAGS) + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/check-remote-protocol.py \ + "$(CC)" "$(OBJEXT)" $< $(@)s else !WITH_REMOTE # The $(PROTOCOL_STRUCTS) files must live in git, because they cannot be # re-generated when configured --without-remote. check-protocol: endif !WITH_REMOTE -EXTRA_DIST += $(PROTOCOL_STRUCTS) +EXTRA_DIST += $(PROTOCOL_STRUCTS) check-remote-protocol.py DRIVERS = \ $(srcdir)/driver-hypervisor.h \ diff --git a/src/check-remote-protocol.py b/src/check-remote-protocol.py new file mode 100644 index 0000000000..3b7b915e88 --- /dev/null +++ b/src/check-remote-protocol.py @@ -0,0 +1,137 @@ +#!/usr/bin/env python +# +# Copyright (C) 2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# This uses pdwtags to check remote protocol defs +# +# * the "split" splits on the /* DD */ comments, so that $p iterates +# through the struct definitions. +# * process only "struct remote_..." entries +# * remove comments and preceding TAB throughout +# * remove empty lines throughout +# * remove white space at end of buffer + +from __future__ import print_function + +import os +import os.path +import re +import subprocess +import sys + +cc = sys.argv[1] +objext = sys.argv[2] +proto_lo = sys.argv[3] +expected = sys.argv[4] + +proto_lo = proto_lo.replace("/", "/.libs/") + +ccargv = cc.split(" ") +ccargv.append("-v") +ccproc = subprocess.Popen(ccargv, stdout=subprocess.PIPE, + stderr=subprocess.STDOUT) +out, err = ccproc.communicate() +out = out.decode("utf-8") +if out.find("clang") != -1: + print("WARNING: skipping pdwtags test with Clang", file=sys.stderr) + sys.exit(0) + + +def which(program): + def is_exe(fpath): + return (os.path.isfile(fpath) and + os.access(fpath, os.X_OK)) + + fpath, fname = os.path.split(program) + if fpath: + if is_exe(program): + return program + else: + for path in os.environ["PATH"].split(os.pathsep): + exe_file = os.path.join(path, program) + if is_exe(exe_file): + return exe_file + + return None + + +pdwtags = which("pdwtags") +if pdwtags is None: + print("WARNING: you lack pdwtags; skipping the protocol test", + file=sys.stderr) + print("WARNING: install the dwarves package to get pdwtags", + file=sys.stderr) + sys.exit(0) + +proto_o = proto_lo.replace(".lo", ".o") + +if not os.path.exists(proto_o): + raise Exception("Missing %s", proto_o) + +pdwtagsproc = subprocess.Popen(["pdwtags", "--verbose", proto_o], + stdout=subprocess.PIPE, stderr=subprocess.PIPE) +out, err = pdwtagsproc.communicate() +out = out.decode("utf-8") +err = err.decode("utf-8") + +if out == "" and err != "": + print("WARNING: no output, pdwtags appears broken:", file=sys.stderr) + for l in err.strip().split("\n"): + print("WARNING: %s" % l, file=sys.stderr) + print("WARNING: skipping the remote protocol test", file=sys.stderr) + sys.exit(0) + +# With pdwtags 1.8, --verbose output includes separators like these: +# /* 93 */ +# /* <0> (null):0 */ +# with the second line omitted for intrinsic types. +# Whereas with pdwtags 1.3, they look like this: +# /* <2d2> /usr/include/libio.h:180 */ +# The alternation of the following regexps matches both cases. +r1 = r'''/\* \d+ \*/''' +r2 = r'''/\* <[0-9a-fA-F]+> \S+:\d+ \*/''' + +libs_prefix = "remote_|qemu_|lxc_|admin_" +other_prefix = "keepalive|vir(Net|LockSpace|LXCMonitor)" +struct_prefix = "(" + libs_prefix + "|" + other_prefix + ")" + +n = 0 +bits = re.split(r'''\n*(?:%s|%s)\n''' % (r1, r2), out) +actual = ["/* -*- c -*- */"] + +for bit in bits: + if (re.match(r'''^(struct|enum)\s+''' + struct_prefix, bit) or + bit.startswith("enum {")): + bit = re.sub(r'''\t*/\*.*?\*/''', "", bit) + bit = re.sub(r'''\s+\n''', '''\n''', bit) + bit = re.sub(r'''\s+$''', "", bit) + bit = re.sub(r'''\t''', " ", bit) + actual.append(bit) + n = n + 1 + +if n < 1: + print("WARNING: No structs/enums matched. Your", file=sys.stderr) + print("WARNING: pdwtags program is probably too old", file=sys.stderr) + print("WARNING: skipping the remote protocol test", file=sys.stderr) + print("WARNING: install dwarves-1.3 or newer", file=sys.stderr) + sys.exit(8) + +diff = subprocess.Popen(["diff", "-u", expected, "-"], stdin=subprocess.PIPE) +actualstr = "\n".join(actual) + "\n" +diff.communicate(input=actualstr.encode("utf-8")) + +sys.exit(diff.returncode) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the test-wrap-argv.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- cfg.mk | 4 +- tests/test-wrap-argv.pl | 174 ---------------------------------------- tests/test-wrap-argv.py | 170 +++++++++++++++++++++++++++++++++++++++ tests/testutils.c | 16 ++-- 4 files changed, 180 insertions(+), 184 deletions(-) delete mode 100755 tests/test-wrap-argv.pl create mode 100755 tests/test-wrap-argv.py diff --git a/cfg.mk b/cfg.mk index 66df1448b4..d5d93e2c1f 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1171,8 +1171,8 @@ header-ifdef: $(PYTHON) $(top_srcdir)/build-aux/header-ifdef.py test-wrap-argv: - $(AM_V_GEN)$(VC_LIST) | $(GREP) -E '\.(ldargs|args)' | xargs \ - $(PERL) $(top_srcdir)/tests/test-wrap-argv.pl --check + $(AM_V_GEN)$(VC_LIST) | $(GREP) -E '\.(ldargs|args)' | $(RUNUTF8) xargs \ + $(PYTHON) $(top_srcdir)/tests/test-wrap-argv.py --check group-qemu-caps: $(AM_V_GEN)$(PERL) $(top_srcdir)/tests/group-qemu-caps.pl --check $(top_srcdir)/ diff --git a/tests/test-wrap-argv.pl b/tests/test-wrap-argv.pl deleted file mode 100755 index 7867e9d719..0000000000 --- a/tests/test-wrap-argv.pl +++ /dev/null @@ -1,174 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2015 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# This script is intended to be passed a list of .args files, used -# to store command line ARGV for the test suites. It will reformat -# them such that there is at most one '-param value' on each line -# of the file. Parameter values that are longer than 80 chars will -# also be split. -# -# If --in-place is supplied as the first parameter of this script, -# the files will be changed in place. -# If --check is the first parameter, the script will return -# a non-zero value if a file is not wrapped correctly. -# Otherwise the rewrapped files are printed to the standard output. - -$in_place = 0; -$check = 0; - -if (@ARGV[0] eq "--in-place" or @ARGV[0] eq "-i") { - $in_place = 1; - shift @ARGV; -} elsif (@ARGV[0] eq "--check") { - $check = 1; - shift @ARGV; -} - -$ret = 0; -foreach my $file (@ARGV) { - if (&rewrap($file) < 0) { - $ret = 1; - } -} - -exit $ret; - -sub rewrap { - my $file = shift; - - # Read the original file - open FILE, "<", $file or die "cannot read $file: $!"; - my @orig_lines = <FILE>; - close FILE; - my @lines = @orig_lines; - foreach (@lines) { - # If there is a trailing '\' then kill the new line - if (/\\$/) { - chomp; - $_ =~ s/\\$//; - } - } - - # Skip empty files - return unless @lines; - - # Kill the last new line in the file - chomp @lines[$#lines]; - - # Reconstruct the master data by joining all lines - # and then split again based on the real desired - # newlines - @lines = split /\n/, join('', @lines); - - # Now each @lines represents a single command, we - # can process them - @lines = map { &rewrap_line($_) } @lines; - - if ($in_place) { - open FILE, ">", $file or die "cannot write $file: $!"; - foreach my $line (@lines) { - print FILE $line; - } - close FILE; - } elsif ($check) { - my $nl = join('', @lines); - my $ol = join('', @orig_lines); - unless ($nl eq $ol) { - open DIFF, "| diff -u $file -" or die "cannot run diff: $!"; - print DIFF $nl; - close DIFF; - - print STDERR "Incorrect line wrapping in $file\n"; - print STDERR "Use test-wrap-argv.pl to wrap test data files\n"; - return -1; - } - } else { - foreach my $line (@lines) { - print $line; - } - } - return 0; -} - -sub rewrap_line { - my $line = shift; - my @bits = split / /, join('', $line); - - # @bits contains env vars, then the command line - # and then the arguments - my @env; - my $cmd; - my @args; - - if ($bits[0] !~ /=/) { - $cmd = shift @bits; - } - - foreach my $bit (@bits) { - # If no command is defined yet, we must still - # have env vars - if (!defined $cmd) { - # Look for leading / to indicate command name - if ($bit =~ m,^/,) { - $cmd = $bit; - } else { - push @env, $bit; - } - } else { - # If there's a leading '-' then this is a new - # parameter, otherwise its a value for the prev - # parameter. - if ($bit =~ m,^-,) { - push @args, $bit; - } else { - $args[$#args] .= " " . $bit; - } - } - } - - # We might have to split line argument values... - @args = map { &rewrap_arg($_) } @args; - # Print env + command first - return join(" \\\n", @env, $cmd, @args), "\n"; -} - -sub rewrap_arg { - my $arg = shift; - my @ret; - my $max_len = 78; - - while (length($arg) > $max_len) { - my $split = rindex $arg, ",", $max_len; - if ($split == -1) { - $split = rindex $arg, ":", $max_len; - } - if ($split == -1) { - $split = rindex $arg, " ", $max_len; - } - if ($split == -1) { - warn "cannot find nice place to split '$arg' below 80 chars\n"; - $split = $max_len - 1; - } - $split++; - - push @ret, substr $arg, 0, $split; - $arg = substr $arg, $split; - } - push @ret, $arg; - return join("\\\n", @ret); -} diff --git a/tests/test-wrap-argv.py b/tests/test-wrap-argv.py new file mode 100755 index 0000000000..dd9e7de824 --- /dev/null +++ b/tests/test-wrap-argv.py @@ -0,0 +1,170 @@ +#!/usr/bin/env python +# +# Copyright (C) 2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# This script is intended to be passed a list of .args files, used +# to store command line ARGV for the test suites. It will reformat +# them such that there is at most one '-param value' on each line +# of the file. Parameter values that are longer than 80 chars will +# also be split. +# +# If --in-place is supplied as the first parameter of this script, +# the files will be changed in place. +# If --check is the first parameter, the script will return +# a non-zero value if a file is not wrapped correctly. +# Otherwise the rewrapped files are printed to the standard output. + +from __future__ import print_function + +import argparse +import subprocess +import sys + + +def rewrap_line(line): + bits = line.split(" ") + + # bits contains env vars, then the command line + # and then the arguments + env = [] + cmd = None + args = [] + + if bits[0].find("=") == -1: + cmd = bits[0] + bits = bits[1:] + + for bit in bits: + # If no command is defined yet, we must still + # have env vars + if cmd is None: + # Look for leading / to indicate command name + if bit.startswith("/"): + cmd = bit + else: + env.append(bit) + else: + # If there's a leading '-' then this is a new + # parameter, otherwise its a value for the prev + # parameter. + if bit.startswith("-"): + args.append(bit) + else: + args[-1] = args[-1] + " " + bit + + # We might have to split line argument values... + args = [rewrap_arg(arg) for arg in args] + + # Print env + command first + return " \\\n".join(env + [cmd] + args) + "\n" + + +def rewrap_arg(arg): + ret = [] + max_len = 78 + + while len(arg) > max_len: + split = arg.rfind(",", 0, max_len + 1) + if split == -1: + split = arg.rfind(":", 0, max_len + 1) + if split == -1: + split = arg.rfind(" ", 0, max_len + 1) + if split == -1: + print("cannot find nice place to split '%s' below 80 chars" % + arg, file=sys.stderr) + split = max_len - 1 + + split = split + 1 + + ret.append(arg[0:split]) + arg = arg[split:] + + ret.append(arg) + return "\\\n".join(ret) + + +def rewrap(filename, in_place, check): + # Read the original file + with open(filename, 'r') as fh: + orig_lines = [] + for line in fh: + orig_lines.append(line) + + if len(orig_lines) == 0: + return + + lines = [] + for line in orig_lines: + if line.endswith("\\\n"): + line = line[:-2] + lines.append(line) + + # Kill the last new line in the file + lines[-1] = lines[-1].rstrip("\n") + + # Reconstruct the master data by joining all lines + # and then split again based on the real desired + # newlines + lines = "".join(lines).split("\n") + + # Now each 'lines' entry represents a single command, we + # can process them + new_lines = [] + for line in lines: + new_lines.append(rewrap_line(line)) + + if in_place: + with open(filename, "w") as fh: + for line in new_lines: + print(line, file=fh) + elif check: + orig = "".join(orig_lines) + new = "".join(new_lines) + if new != orig: + diff = subprocess.Popen(["diff", "-u", filename, "-"], + stdin=subprocess.PIPE) + diff.communicate(input=new.encode('utf-8')) + + print("Incorrect line wrapping in $file", + file=sys.stderr) + print("Use test-wrap-argv.py to wrap test data files", + file=sys.stderr) + return False + else: + for line in new_lines: + print(line) + + return True + + +parser = argparse.ArgumentParser(description='Test arg line wrapper') +parser.add_argument('--in-place', '-i', action="store_true", + help='modify files in-place') +parser.add_argument('--check', action="store_true", + help='check existing files only') +parser.add_argument('files', nargs="+", + help="filenames to check") +args = parser.parse_args() + +errs = False +for filename in args.files: + if not rewrap(filename, args.in_place, args.check): + errs = True + +if errs: + sys.exit(1) +sys.exit(0) diff --git a/tests/testutils.c b/tests/testutils.c index 1b663f9d5d..186ddeca8e 100644 --- a/tests/testutils.c +++ b/tests/testutils.c @@ -59,7 +59,7 @@ static size_t testCounter; static virBitmapPtr testBitmap; char *progname; -static char *perl; +static char *python; static int virTestUseTerminalColors(void) { @@ -399,15 +399,15 @@ virTestRewrapFile(const char *filename) virStringHasSuffix(filename, ".ldargs"))) return 0; - if (!perl) { - fprintf(stderr, "cannot rewrap %s: unable to find perl in path", filename); + if (!python) { + fprintf(stderr, "cannot rewrap %s: unable to find python in path", filename); return -1; } - if (virAsprintf(&script, "%s/test-wrap-argv.pl", abs_srcdir) < 0) + if (virAsprintf(&script, "%s/test-wrap-argv.py", abs_srcdir) < 0) goto cleanup; - cmd = virCommandNewArgList(perl, script, "--in-place", filename, NULL); + cmd = virCommandNewArgList(python, script, "--in-place", filename, NULL); if (virCommandRun(cmd, NULL) < 0) goto cleanup; @@ -900,8 +900,8 @@ int virTestMain(int argc, } } - /* Find perl early because some tests override PATH */ - perl = virFindFileInPath("perl"); + /* Find python early because some tests override PATH */ + python = virFindFileInPath("python"); ret = (func)(); @@ -912,7 +912,7 @@ int virTestMain(int argc, fprintf(stderr, " %-3zu %s\n", testCounter, ret == 0 ? "OK" : "FAIL"); } virLogReset(); - VIR_FREE(perl); + VIR_FREE(python); return ret; } -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the group-qemu-caps.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- cfg.mk | 3 +- tests/group-qemu-caps.pl | 124 --------------------------------------- tests/group-qemu-caps.py | 123 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 125 insertions(+), 125 deletions(-) delete mode 100755 tests/group-qemu-caps.pl create mode 100755 tests/group-qemu-caps.py diff --git a/cfg.mk b/cfg.mk index d5d93e2c1f..25785423a2 100644 --- a/cfg.mk +++ b/cfg.mk @@ -1175,7 +1175,8 @@ test-wrap-argv: $(PYTHON) $(top_srcdir)/tests/test-wrap-argv.py --check group-qemu-caps: - $(AM_V_GEN)$(PERL) $(top_srcdir)/tests/group-qemu-caps.pl --check $(top_srcdir)/ + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(top_srcdir)/tests/group-qemu-caps.py \ + --check --prefix $(top_srcdir)/ # sc_po_check can fail if generated files are not built first sc_po_check: \ diff --git a/tests/group-qemu-caps.pl b/tests/group-qemu-caps.pl deleted file mode 100755 index 829e63a562..0000000000 --- a/tests/group-qemu-caps.pl +++ /dev/null @@ -1,124 +0,0 @@ -#!/usr/bin/env perl -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# -# Regroup array values into smaller groups separated by numbered comments. -# -# If --check is the first parameter, the script will return -# a non-zero value if a file is not grouped correctly. -# Otherwise the files are regrouped in place. - -use strict; -use warnings; - -my $check = 0; - -if (defined $ARGV[0] && $ARGV[0] eq "--check") { - $check = 1; - shift @ARGV; -} - -my $prefix = ''; -if (defined $ARGV[0]) { - $prefix = $ARGV[0]; - shift @ARGV; -} - -my $ret = 0; -if (®roup_caps($prefix . 'src/qemu/qemu_capabilities.c', - 'virQEMUCaps grouping marker', - '\);', - 0, - " ") < 0) { - $ret = 1; -} -if (®roup_caps($prefix . 'src/qemu/qemu_capabilities.h', - 'virQEMUCapsFlags grouping marker', - 'QEMU_CAPS_LAST \/\* this must', - 1, - " ") < 0) { - $ret = 1; -} - -exit $ret; - -sub regroup_caps { - my $filename = shift; - my $start_regex = shift; - my $end_regex = shift; - my $trailing_newline = shift; - my $counter_prefix = shift; - my $step = 5; - - open FILE, '<', $filename or die "cannot open $filename: $!"; - my @original = <FILE>; - close FILE; - - my @fixed; - my $game_on = 0; - my $counter = 0; - foreach (@original) { - if ($game_on) { - next if ($_ =~ '/\* [0-9]+ \*/'); - next if (/^\s+$/); - if ($counter % $step == 0) { - if ($counter != 0) { - push @fixed, "\n"; - } - push @fixed, "$counter_prefix/* $counter */\n"; - } - if (!($_ =~ '/\*' && !($_ =~ '\*/'))) { - # count two-line comments as one line - $counter++; - } - } - if (/$start_regex/) { - $game_on = 1; - } elsif ($game_on && $_ =~ /$end_regex/) { - if (($counter -1) % $step == 0) { - pop @fixed; # /* $counter */ - if ($counter != 1) { - pop @fixed; # \n - } - } - if ($trailing_newline) { - push @fixed, "\n"; - } - $game_on = 0; - } - push @fixed, $_; - } - - if ($check) { - my $nl = join('', @fixed); - my $ol = join('', @original); - unless ($nl eq $ol) { - open DIFF, "| diff -u $filename -" or die "cannot run diff: $!"; - print DIFF $nl; - close DIFF; - - print STDERR "Incorrect array grouping in $filename\n"; - print STDERR "Use group-qemu-caps.pl to group long array members\n"; - return -1; - } - } else { - open FILE, '>', $filename or die "cannot open $filename: $!"; - foreach my $line (@fixed) { - print FILE $line; - } - close FILE; - } -} diff --git a/tests/group-qemu-caps.py b/tests/group-qemu-caps.py new file mode 100755 index 0000000000..f73a94b998 --- /dev/null +++ b/tests/group-qemu-caps.py @@ -0,0 +1,123 @@ +#!/usr/bin/env python +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# +# Regroup array values into smaller groups separated by numbered comments. +# +# If --check is the first parameter, the script will return +# a non-zero value if a file is not grouped correctly. +# Otherwise the files are regrouped in place. + +from __future__ import print_function + +import argparse +import re +import subprocess +import sys + + +def regroup_caps(check, filename, start_regex, end_regex, + trailing_newline, counter_prefix): + step = 5 + + original = [] + with open(filename, "r") as fh: + for line in fh: + original.append(line) + + fixed = [] + game_on = False + counter = 0 + for line in original: + line = line.rstrip("\n") + if game_on: + if re.match(r'''.*/\* [0-9]+ \*/.*''', line): + continue + if re.match(r'''^\s*$''', line): + continue + if counter % step == 0: + if counter != 0: + fixed.append("\n") + fixed.append("%s/* %d */\n" % (counter_prefix, counter)) + + if not (line.find("/*") != -1 and line.find("*/") == -1): + # count two-line comments as one line + counter = counter + 1 + + if re.match(start_regex, line): + game_on = True + elif game_on and re.match(end_regex, line): + if (counter - 1) % step == 0: + fixed = fixed[:-1] # /* $counter */ + if counter != 1: + fixed = fixed[:-1] # \n + + if trailing_newline: + fixed.append("\n") + + game_on = False + + fixed.append(line + "\n") + + if check: + orig = "".join(original) + new = "".join(fixed) + if new != orig: + diff = subprocess.Popen(["diff", "-u", filename, "-"], + stdin=subprocess.PIPE) + diff.communicate(input=new.encode('utf-8')) + + print("Incorrect line wrapping in $file", + file=sys.stderr) + print("Use test-wrap-argv.py to wrap test data files", + file=sys.stderr) + return False + else: + with open(filename, "w") as fh: + for line in fixed: + print(line, file=fh, end='') + + return True + + +parser = argparse.ArgumentParser(description='Test arg line wrapper') +parser.add_argument('--check', action="store_true", + help='check existing files only') +parser.add_argument('--prefix', default='', + help='source code tree prefix') +args = parser.parse_args() + +errs = False + +if not regroup_caps(args.check, + args.prefix + 'src/qemu/qemu_capabilities.c', + r'.*virQEMUCaps grouping marker.*', + r'.*\);.*', + 0, + " "): + errs = True + +if not regroup_caps(args.check, + args.prefix + 'src/qemu/qemu_capabilities.h', + r'.*virQEMUCapsFlags grouping marker.*', + r'.*QEMU_CAPS_LAST \/\* this must.*', + 1, + " "): + errs = True + +if errs: + sys.exit(1) +sys.exit(0) -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the check-file-access.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- tests/Makefile.am | 4 +- tests/check-file-access.pl | 126 -------------------------------- tests/check-file-access.py | 123 +++++++++++++++++++++++++++++++ tests/file_access_whitelist.txt | 2 +- 4 files changed, 126 insertions(+), 129 deletions(-) delete mode 100755 tests/check-file-access.pl create mode 100755 tests/check-file-access.py diff --git a/tests/Makefile.am b/tests/Makefile.am index e1bcd10c7c..aba2310e21 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -450,14 +450,14 @@ EXTRA_DIST += $(test_scripts) if WITH_LINUX check-access: file-access-clean VIR_TEST_FILE_ACCESS=1 $(MAKE) $(AM_MAKEFLAGS) check - $(PERL) check-file-access.pl | sort -u + $(RUNUTF8) $(PYTHON) check-file-access.py | sort -u file-access-clean: > test_file_access.txt endif WITH_LINUX EXTRA_DIST += \ - check-file-access.pl \ + check-file-access.py \ file_access_whitelist.txt if WITH_TESTS diff --git a/tests/check-file-access.pl b/tests/check-file-access.pl deleted file mode 100755 index ea0b7a18a2..0000000000 --- a/tests/check-file-access.pl +++ /dev/null @@ -1,126 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2016 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# -# This script is supposed to check test_file_access.txt file and -# warn about file accesses outside our working tree. -# -# - -use strict; -use warnings; - -my $access_file = "test_file_access.txt"; -my $whitelist_file = "file_access_whitelist.txt"; - -my @known_actions = ("open", "fopen", "access", "stat", "lstat", "connect"); - -my @files; -my @whitelist; - -open FILE, "<", $access_file or die "Unable to open $access_file: $!"; -while (<FILE>) { - chomp; - if (/^(\S*):\s*(\S*):\s*(\S*)(\s*:\s*(.*))?$/) { - my %rec; - ${rec}{path} = $1; - ${rec}{action} = $2; - ${rec}{progname} = $3; - if (defined $5) { - ${rec}{testname} = $5; - } - push (@files, \%rec); - } else { - die "Malformed line $_"; - } -} -close FILE; - -open FILE, "<", $whitelist_file or die "Unable to open $whitelist_file: $!"; -while (<FILE>) { - chomp; - if (/^\s*#.*$/) { - # comment - } elsif (/^(\S*):\s*(\S*)(:\s*(\S*)(\s*:\s*(.*))?)?$/ and - grep /^$2$/, @known_actions) { - # $path: $action: $progname: $testname - my %rec; - ${rec}{path} = $1; - ${rec}{action} = $3; - if (defined $4) { - ${rec}{progname} = $4; - } - if (defined $6) { - ${rec}{testname} = $6; - } - push (@whitelist, \%rec); - } elsif (/^(\S*)(:\s*(\S*)(\s*:\s*(.*))?)?$/) { - # $path: $progname: $testname - my %rec; - ${rec}{path} = $1; - if (defined $3) { - ${rec}{progname} = $3; - } - if (defined $5) { - ${rec}{testname} = $5; - } - push (@whitelist, \%rec); - } else { - die "Malformed line $_"; - } -} -close FILE; - -# Now we should check if %traces is included in $whitelist. For -# now checking just keys is sufficient -my $error = 0; -for my $file (@files) { - my $match = 0; - - for my $rule (@whitelist) { - if (not %${file}{path} =~ m/^$rule->{path}$/) { - next; - } - - if (defined %${rule}{action} and - not %${file}{action} =~ m/^$rule->{action}$/) { - next; - } - - if (defined %${rule}{progname} and - not %${file}{progname} =~ m/^$rule->{progname}$/) { - next; - } - - if (defined %${rule}{testname} and - defined %${file}{testname} and - not %${file}{testname} =~ m/^$rule->{testname}$/) { - next; - } - - $match = 1; - } - - if (not $match) { - $error = 1; - print "$file->{path}: $file->{action}: $file->{progname}"; - print ": $file->{testname}" if defined %${file}{testname}; - print "\n"; - } -} - -exit $error; diff --git a/tests/check-file-access.py b/tests/check-file-access.py new file mode 100755 index 0000000000..d0e3ef68b4 --- /dev/null +++ b/tests/check-file-access.py @@ -0,0 +1,123 @@ +#!/usr/bin/env python +# +# Copyright (C) 2016-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# +# This script is supposed to check test_file_access.txt file and +# warn about file accesses outside our working tree. +# +# + +from __future__ import print_function + +import re +import sys + +access_file = "test_file_access.txt" +whitelist_file = "file_access_whitelist.txt" + +known_actions = ["open", "fopen", "access", "stat", "lstat", "connect"] + +files = [] +whitelist = [] + +with open(access_file, "r") as fh: + for line in fh: + line = line.rstrip("\n") + + m = re.match(r'''^(\S*):\s*(\S*):\s*(\S*)(\s*:\s*(.*))?$''', line) + if m is not None: + rec = { + "path": m.group(1), + "action": m.group(2), + "progname": m.group(3), + "testname": m.group(5), + } + files.append(rec) + else: + raise Exception("Malformed line %s" % line) + +with open(whitelist_file, "r") as fh: + for line in fh: + line = line.rstrip("\n") + + if re.match(r'''^\s*#.*$''', line): + continue # comment + if line == "": + continue + + m = re.match(r'''^(\S*):\s*(\S*)(:\s*(\S*)(\s*:\s*(.*))?)?$''', line) + if m is not None and m.group(2) in known_actions: + # $path: $action: $progname: $testname + rec = { + "path": m.group(1), + "action": m.group(3), + "progname": m.group(4), + "testname": m.group(6), + } + whitelist.append(rec) + else: + m = re.match(r'''^(\S*)(:\s*(\S*)(\s*:\s*(.*))?)?$''', line) + if m is not None: + # $path: $progname: $testname + rec = { + "path": m.group(1), + "action": None, + "progname": m.group(3), + "testname": m.group(5), + } + whitelist.append(rec) + else: + raise Exception("Malformed line %s" % line) + + +# Now we should check if %traces is included in $whitelist. For +# now checking just keys is sufficient +err = False +for file in files: + match = False + + for rule in whitelist: + if not re.match("^" + rule["path"], file["path"]): + continue + + if (rule["action"] is not None and + not re.match("^" + rule["action"], file["action"])): + continue + + if (rule["progname"] is not None and + not re.match("^" + rule["progname"], file["progname"])): + continue + + if (rule["testname"] is not None and + file["testname"] is not None and + not re.match("^" + rule["testname"], file["testname"])): + continue + + match = True + + if not match: + err = True + print("%s: %s: %s" % + (file["path"], file["action"], file["progname"]), + file=sys.stderr, end="") + if file["testname"] is not None: + print(": %s" % file["testname"], file=sys.stderr, end="") + print("", file=sys.stderr) + +if err: + sys.exit(1) +sys.exit(0) diff --git a/tests/file_access_whitelist.txt b/tests/file_access_whitelist.txt index 3fb318cbab..5ec7ee63bb 100644 --- a/tests/file_access_whitelist.txt +++ b/tests/file_access_whitelist.txt @@ -5,7 +5,7 @@ # $path: $progname: $testname # $path: $action: $progname: $testname # -# All these variables are evaluated as perl RE. So to allow +# All these variables are evaluated as python RE. So to allow # /dev/sda and /dev/sdb, you can just '/dev/sd[a-b]', or to allow # /proc/$pid/status you can '/proc/\d+/status' and so on. # Moreover, $action, $progname and $testname can be empty, in which -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the hvsupport.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. The new impl generates byte-for-byte identical output to the old impl. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- docs/Makefile.am | 8 +- docs/hvsupport.pl | 458 ---------------------------------------- docs/hvsupport.py | 516 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 520 insertions(+), 462 deletions(-) delete mode 100755 docs/hvsupport.pl create mode 100755 docs/hvsupport.py diff --git a/docs/Makefile.am b/docs/Makefile.am index 8b8074c3c0..b9d69db8c8 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -172,7 +172,7 @@ EXTRA_DIST= \ $(internals_html_in) $(internals_html) $(fonts) \ $(kbase_html_in) $(kbase_html) \ aclperms.htmlinc \ - hvsupport.pl \ + hvsupport.py \ $(schema_DATA) acl_generated = aclperms.htmlinc @@ -207,12 +207,12 @@ web: $(dot_html) $(internals_html) $(kbase_html) \ hvsupport.html: $(srcdir)/hvsupport.html.in -$(srcdir)/hvsupport.html.in: $(srcdir)/hvsupport.pl $(api_DATA) \ +$(srcdir)/hvsupport.html.in: $(srcdir)/hvsupport.py $(api_DATA) \ $(top_srcdir)/src/libvirt_public.syms \ $(top_srcdir)/src/libvirt_qemu.syms $(top_srcdir)/src/libvirt_lxc.syms \ $(top_srcdir)/src/driver.h - $(AM_V_GEN)$(PERL) $(srcdir)/hvsupport.pl $(top_srcdir)/src > $@ \ - || { rm $@ && exit 1; } + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/hvsupport.py \ + $(top_srcdir)/src > $@ || { rm $@ && exit 1; } news.html.in: \ $(srcdir)/news.xml \ diff --git a/docs/hvsupport.pl b/docs/hvsupport.pl deleted file mode 100755 index 494b8a27ec..0000000000 --- a/docs/hvsupport.pl +++ /dev/null @@ -1,458 +0,0 @@ -#!/usr/bin/env perl - -use strict; -use warnings; - -use File::Find; - -die "syntax: $0 SRCDIR\n" unless int(@ARGV) == 1; - -my $srcdir = shift @ARGV; - -my $symslibvirt = "$srcdir/libvirt_public.syms"; -my $symsqemu = "$srcdir/libvirt_qemu.syms"; -my $symslxc = "$srcdir/libvirt_lxc.syms"; -my @drivertable = ( - "$srcdir/driver-hypervisor.h", - "$srcdir/driver-interface.h", - "$srcdir/driver-network.h", - "$srcdir/driver-nodedev.h", - "$srcdir/driver-nwfilter.h", - "$srcdir/driver-secret.h", - "$srcdir/driver-state.h", - "$srcdir/driver-storage.h", - "$srcdir/driver-stream.h", - ); - -my %groupheaders = ( - "virHypervisorDriver" => "Hypervisor APIs", - "virNetworkDriver" => "Virtual Network APIs", - "virInterfaceDriver" => "Host Interface APIs", - "virNodeDeviceDriver" => "Host Device APIs", - "virStorageDriver" => "Storage Pool APIs", - "virSecretDriver" => "Secret APIs", - "virNWFilterDriver" => "Network Filter APIs", - ); - - -my @srcs; -find({ - wanted => sub { - if (m!$srcdir/.*/\w+_(driver|common|tmpl|monitor|hal|udev)\.c$!) { - push @srcs, $_ if $_ !~ /vbox_driver\.c/; - } - }, no_chdir => 1}, $srcdir); - -# Map API functions to the header and documentation files they're in -# so that we can generate proper hyperlinks to their documentation. -# -# The function names are grep'd from the XML output of apibuild.py. -sub getAPIFilenames { - my $filename = shift; - - my %files; - my $line; - - open FILE, "<", $filename or die "cannot read $filename: $!"; - - while (defined($line = <FILE>)) { - if ($line =~ /function name='([^']+)' file='([^']+)'/) { - $files{$1} = $2; - } - } - - close FILE; - - if (keys %files == 0) { - die "No functions found in $filename. Has the apibuild.py output changed?"; - } - return \%files; -} - -sub parseSymsFile { - my $apisref = shift; - my $prefix = shift; - my $filename = shift; - my $xmlfilename = shift; - - my $line; - my $vers; - my $prevvers; - - my $filenames = getAPIFilenames($xmlfilename); - - open FILE, "<$filename" - or die "cannot read $filename: $!"; - - while (defined($line = <FILE>)) { - chomp $line; - next if $line =~ /^\s*#/; - next if $line =~ /^\s*$/; - next if $line =~ /^\s*(global|local):/; - if ($line =~ /^\s*${prefix}_(\d+\.\d+\.\d+)\s*{\s*$/) { - if (defined $vers) { - die "malformed syms file"; - } - $vers = $1; - } elsif ($line =~ /\s*}\s*;\s*$/) { - if (defined $prevvers) { - die "malformed syms file"; - } - $prevvers = $vers; - $vers = undef; - } elsif ($line =~ /\s*}\s*${prefix}_(\d+\.\d+\.\d+)\s*;\s*$/) { - if ($1 ne $prevvers) { - die "malformed syms file $1 != $vers"; - } - $prevvers = $vers; - $vers = undef; - } elsif ($line =~ /\s*(\w+)\s*;\s*$/) { - $$apisref{$1} = {}; - $$apisref{$1}->{vers} = $vers; - $$apisref{$1}->{file} = $$filenames{$1}; - } else { - die "unexpected data $line\n"; - } - } - - close FILE; -} - -my %apis; -# Get the list of all public APIs and their corresponding version -parseSymsFile(\%apis, "LIBVIRT", $symslibvirt, "$srcdir/../docs/libvirt-api.xml"); - -# And the same for the QEMU specific APIs -parseSymsFile(\%apis, "LIBVIRT_QEMU", $symsqemu, "$srcdir/../docs/libvirt-qemu-api.xml"); - -# And the same for the LXC specific APIs -parseSymsFile(\%apis, "LIBVIRT_LXC", $symslxc, "$srcdir/../docs/libvirt-lxc-api.xml"); - - -# Some special things which aren't public APIs, -# but we want to report -$apis{virConnectSupportsFeature}->{vers} = "0.3.2"; -$apis{virDomainMigratePrepare}->{vers} = "0.3.2"; -$apis{virDomainMigratePerform}->{vers} = "0.3.2"; -$apis{virDomainMigrateFinish}->{vers} = "0.3.2"; -$apis{virDomainMigratePrepare2}->{vers} = "0.5.0"; -$apis{virDomainMigrateFinish2}->{vers} = "0.5.0"; -$apis{virDomainMigratePrepareTunnel}->{vers} = "0.7.2"; - -$apis{virDomainMigrateBegin3}->{vers} = "0.9.2"; -$apis{virDomainMigratePrepare3}->{vers} = "0.9.2"; -$apis{virDomainMigratePrepareTunnel3}->{vers} = "0.9.2"; -$apis{virDomainMigratePerform3}->{vers} = "0.9.2"; -$apis{virDomainMigrateFinish3}->{vers} = "0.9.2"; -$apis{virDomainMigrateConfirm3}->{vers} = "0.9.2"; - -$apis{virDomainMigrateBegin3Params}->{vers} = "1.1.0"; -$apis{virDomainMigratePrepare3Params}->{vers} = "1.1.0"; -$apis{virDomainMigratePrepareTunnel3Params}->{vers} = "1.1.0"; -$apis{virDomainMigratePerform3Params}->{vers} = "1.1.0"; -$apis{virDomainMigrateFinish3Params}->{vers} = "1.1.0"; -$apis{virDomainMigrateConfirm3Params}->{vers} = "1.1.0"; - - - -# Now we want to get the mapping between public APIs -# and driver struct fields. This lets us later match -# update the driver impls with the public APis. - -my $line; - -# Group name -> hash of APIs { fields -> api name } -my %groups; -my $ingrp; -foreach my $drivertable (@drivertable) { - open FILE, "<$drivertable" - or die "cannot read $drivertable: $!"; - - while (defined($line = <FILE>)) { - if ($line =~ /struct _(vir\w*Driver)/) { - my $grp = $1; - if ($grp ne "virStateDriver" && - $grp ne "virStreamDriver") { - $ingrp = $grp; - $groups{$ingrp} = { apis => {}, drivers => {} }; - } - } elsif ($ingrp) { - if ($line =~ /^\s*vir(?:Drv)(\w+)\s+(\w+);\s*$/) { - my $field = $2; - my $name = $1; - - my $api; - if (exists $apis{"vir$name"}) { - $api = "vir$name"; - } elsif ($name =~ /\w+(Open|Close|URIProbe)/) { - next; - } else { - die "driver $name does not have a public API"; - } - $groups{$ingrp}->{apis}->{$field} = $api; - } elsif ($line =~ /};/) { - $ingrp = undef; - } - } - } - - close FILE; -} - - -# Finally, we read all the primary driver files and extract -# the driver API tables from each one. - -foreach my $src (@srcs) { - open FILE, "<$src" or - die "cannot read $src: $!"; - - my $groups_regex = join("|", keys %groups); - $ingrp = undef; - my $impl; - while (defined($line = <FILE>)) { - if (!$ingrp) { - # skip non-matching lines early to save time - next if not $line =~ /$groups_regex/; - - if ($line =~ /^\s*(?:static\s+)?($groups_regex)\s+(\w+)\s*=\s*{/ || - $line =~ /^\s*(?:static\s+)?($groups_regex)\s+NAME\(\w+\)\s*=\s*{/) { - $ingrp = $1; - $impl = $src; - - if ($impl =~ m,.*/node_device_(\w+)\.c,) { - $impl = $1; - } else { - $impl =~ s,.*/(\w+?)_((\w+)_)?(\w+)\.c,$1,; - } - - if ($groups{$ingrp}->{drivers}->{$impl}) { - die "Group $ingrp already contains $impl"; - } - - $groups{$ingrp}->{drivers}->{$impl} = {}; - } - - } else { - if ($line =~ m!\s*\.(\w+)\s*=\s*(\w+)\s*,?\s*(?:/\*\s*(\d+\.\d+\.\d+)\s*(?:-\s*(\d+\.\d+\.\d+))?\s*\*/\s*)?$!) { - my $api = $1; - my $meth = $2; - my $vers = $3; - my $deleted = $4; - - next if $api eq "no" || $api eq "name"; - - if ($meth eq "NULL" && !defined $deleted) { - die "Method impl for $api is NULL, but no deleted version is provided"; - } - if ($meth ne "NULL" && defined $deleted) { - die "Method impl for $api is non-NULL, but deleted version is provided"; - } - - die "Method $meth in $src is missing version" unless defined $vers || $api eq "connectURIProbe"; - - if (!exists($groups{$ingrp}->{apis}->{$api})) { - next if $api =~ /\w(Open|Close|URIProbe)/; - - die "Found unexpected method $api in $ingrp\n"; - } - - $groups{$ingrp}->{drivers}->{$impl}->{$api} = {}; - $groups{$ingrp}->{drivers}->{$impl}->{$api}->{vers} = $vers; - $groups{$ingrp}->{drivers}->{$impl}->{$api}->{deleted} = $deleted; - if ($api eq "domainMigratePrepare" || - $api eq "domainMigratePrepare2" || - $api eq "domainMigratePrepare3") { - if (!$groups{$ingrp}->{drivers}->{$impl}->{"domainMigrate"}) { - $groups{$ingrp}->{drivers}->{$impl}->{"domainMigrate"} = {}; - $groups{$ingrp}->{drivers}->{$impl}->{"domainMigrate"}->{vers} = $vers; - } - } - - } elsif ($line =~ /}/) { - $ingrp = undef; - } - } - } - - close FILE; -} - - -# The '.open' driver method is used for 3 public APIs, so we -# have a bit of manual fixup todo with the per-driver versioning -# and support matrix - -$groups{virHypervisorDriver}->{apis}->{"openAuth"} = "virConnectOpenAuth"; -$groups{virHypervisorDriver}->{apis}->{"openReadOnly"} = "virConnectOpenReadOnly"; -$groups{virHypervisorDriver}->{apis}->{"domainMigrate"} = "virDomainMigrate"; - -my $openAuthVers = (0 * 1000 * 1000) + (4 * 1000) + 0; - -foreach my $drv (keys %{$groups{"virHypervisorDriver"}->{drivers}}) { - my $openVersStr = $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"connectOpen"}->{vers}; - my $openVers; - if ($openVersStr =~ /(\d+)\.(\d+)\.(\d+)/) { - $openVers = ($1 * 1000 * 1000) + ($2 * 1000) + $3; - } - - # virConnectOpenReadOnly always matches virConnectOpen version - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"connectOpenReadOnly"} = - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"connectOpen"}; - - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"connectOpenAuth"} = {}; - - # virConnectOpenAuth is always 0.4.0 if the driver existed - # before this time, otherwise it matches the version of - # the driver's virConnectOpen entry - if ($openVersStr eq "Y" || - $openVers >= $openAuthVers) { - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"connectOpenAuth"}->{vers} = $openVersStr; - } else { - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"connectOpenAuth"}->{vers} = "0.4.0"; - } -} - - -# Another special case for the virDomainCreateLinux which was replaced -# with virDomainCreateXML -$groups{virHypervisorDriver}->{apis}->{"domainCreateLinux"} = "virDomainCreateLinux"; - -my $createAPIVers = (0 * 1000 * 1000) + (0 * 1000) + 3; - -foreach my $drv (keys %{$groups{"virHypervisorDriver"}->{drivers}}) { - my $createVersStr = $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"domainCreateXML"}->{vers}; - next unless defined $createVersStr; - my $createVers; - if ($createVersStr =~ /(\d+)\.(\d+)\.(\d+)/) { - $createVers = ($1 * 1000 * 1000) + ($2 * 1000) + $3; - } - - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"domainCreateLinux"} = {}; - - # virCreateLinux is always 0.0.3 if the driver existed - # before this time, otherwise it matches the version of - # the driver's virCreateXML entry - if ($createVersStr eq "Y" || - $createVers >= $createAPIVers) { - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"domainCreateLinux"}->{vers} = $createVersStr; - } else { - $groups{"virHypervisorDriver"}->{drivers}->{$drv}->{"domainCreateLinux"}->{vers} = "0.0.3"; - } -} - - -# Finally we generate the HTML file with the tables - -print <<EOF; -<?xml version="1.0" encoding="UTF-8"?> -<!DOCTYPE html> -<html xmlns="http://www.w3.org/1999/xhtml"> -<body class="hvsupport"> -<h1>libvirt API support matrix</h1> - -<ul id="toc"></ul> - -<p> -This page documents which <a href="html/">libvirt calls</a> work on -which libvirt drivers / hypervisors, and which version the API appeared -in. If a hypervisor driver later dropped support for the API, the version -when it was removed is also mentioned (highlighted in -<span class="removedhv">dark red</span>). -</p> - -EOF - - foreach my $grp (sort { $a cmp $b } keys %groups) { - print "<h2><a id=\"$grp\">", $groupheaders{$grp}, "</a></h2>\n"; - print <<EOF; -<table class="top_table"> -<thead> -<tr> -<th>API</th> -<th>Version</th> -EOF - - foreach my $drv (sort { $a cmp $b } keys %{$groups{$grp}->{drivers}}) { - print " <th>$drv</th>\n"; - } - - print <<EOF; -</tr> -</thead> -<tbody> -EOF - - my $row = 0; - foreach my $field (sort { - $groups{$grp}->{apis}->{$a} - cmp - $groups{$grp}->{apis}->{$b} - } keys %{$groups{$grp}->{apis}}) { - my $api = $groups{$grp}->{apis}->{$field}; - my $vers = $apis{$api}->{vers}; - my $htmlgrp = $apis{$api}->{file}; - print <<EOF; -<tr> -<td> -EOF - - if (defined $htmlgrp) { - print <<EOF; -<a href=\"html/libvirt-$htmlgrp.html#$api\">$api</a> -EOF - - } else { - print $api; - } - print <<EOF; -</td> -<td>$vers</td> -EOF - - foreach my $drv (sort {$a cmp $b } keys %{$groups{$grp}->{drivers}}) { - print "<td>"; - if (exists $groups{$grp}->{drivers}->{$drv}->{$field}) { - if ($groups{$grp}->{drivers}->{$drv}->{$field}->{vers}) { - print $groups{$grp}->{drivers}->{$drv}->{$field}->{vers}; - } - if ($groups{$grp}->{drivers}->{$drv}->{$field}->{deleted}) { - print " - <span class=\"removedhv\">", $groups{$grp}->{drivers}->{$drv}->{$field}->{deleted}, "</span>"; - } - } - print "</td>\n"; - } - - print <<EOF; -</tr> -EOF - - $row++; - if (($row % 15) == 0) { - print <<EOF; -<tr> -<th>API</th> -<th>Version</th> -EOF - - foreach my $drv (sort { $a cmp $b } keys %{$groups{$grp}->{drivers}}) { - print " <th>$drv</th>\n"; - } - - print <<EOF; -</tr> -EOF - } - - } - - print <<EOF; -</tbody> -</table> -EOF -} - -print <<EOF; -</body> -</html> -EOF diff --git a/docs/hvsupport.py b/docs/hvsupport.py new file mode 100755 index 0000000000..cc05cce565 --- /dev/null +++ b/docs/hvsupport.py @@ -0,0 +1,516 @@ +#!/usr/bin/env python +# +# Copyright (C) 2011-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. + +from __future__ import print_function + +import sys +import os.path +import re + +if len(sys.argv) != 2: + print("syntax: %s SRCDIR\n" % sys.argv[0], file=sys.stderr) + +srcdir = sys.argv[1] + +symslibvirt = os.path.join(srcdir, "libvirt_public.syms") +symsqemu = os.path.join(srcdir, "libvirt_qemu.syms") +symslxc = os.path.join(srcdir, "libvirt_lxc.syms") +drivertablefiles = [ + os.path.join(srcdir, "driver-hypervisor.h"), + os.path.join(srcdir, "driver-interface.h"), + os.path.join(srcdir, "driver-network.h"), + os.path.join(srcdir, "driver-nodedev.h"), + os.path.join(srcdir, "driver-nwfilter.h"), + os.path.join(srcdir, "driver-secret.h"), + os.path.join(srcdir, "driver-state.h"), + os.path.join(srcdir, "driver-storage.h"), + os.path.join(srcdir, "driver-stream.h"), +] + +groupheaders = { + "virHypervisorDriver": "Hypervisor APIs", + "virNetworkDriver": "Virtual Network APIs", + "virInterfaceDriver": "Host Interface APIs", + "virNodeDeviceDriver": "Host Device APIs", + "virStorageDriver": "Storage Pool APIs", + "virSecretDriver": "Secret APIs", + "virNWFilterDriver": "Network Filter APIs", +} + + +srcs = [] +for root, dirs, files in os.walk(srcdir): + for file in files: + if ((file.endswith("driver.c") and + not file.endswith("vbox_driver.c")) or + file.endswith("common.c") or + file.endswith("tmpl.c") or + file.endswith("monitor.c") or + file.endswith("hal.c") or + file.endswith("udev.c")): + srcs.append(os.path.join(root, file)) + + +# Map API functions to the header and documentation files they're in +# so that we can generate proper hyperlinks to their documentation. +# +# The function names are grep'd from the XML output of apibuild.py. +def getAPIFilenames(filename): + files = {} + + with open(filename) as fh: + prog = re.compile(r"\s*<function name='([^']+)' file='([^']+)'.*") + for line in fh: + res = prog.match(line) + if res is not None: + files[res.group(1)] = res.group(2) + + if len(files) == 0: + raise Exception(("No functions found in %s. " + + "Has the apibuild.py output changed?") % + filename) + + return files + + +def parseSymsFile(apisref, prefix, filename, xmlfilename): + vers = None + prevvers = None + + filenames = getAPIFilenames(xmlfilename) + + with open(filename) as fh: + groupstartprog = re.compile(r"^\s*%s_(\d+\.\d+\.\d+)\s*{\s*$" % + prefix) + groupendprog1 = re.compile(r"^\s*}\s*;\s*$") + groupendprog2 = re.compile(r"^\s*}\s*%s_(\d+\.\d+\.\d+)\s*;\s*$" % + prefix) + symbolprog = re.compile(r"^\s*(\w+)\s*;\s*$") + for line in fh: + line = line.strip() + + if line == "": + continue + if line[0] == '#': + continue + if line.startswith("global:"): + continue + if line.startswith("local:"): + continue + + groupstartmatch = groupstartprog.match(line) + groupendmatch1 = groupendprog1.match(line) + groupendmatch2 = groupendprog2.match(line) + symbolmatch = symbolprog.match(line) + if groupstartmatch is not None: + if vers is not None: + raise Exception("malformed syms file when starting group") + + vers = groupstartmatch.group(1) + elif groupendmatch1 is not None: + if prevvers is not None: + raise Exception("malformed syms file when ending group") + + prevvers = vers + vers = None + elif groupendmatch2 is not None: + if groupendmatch2.group(1) != prevvers: + raise Exception("malformed syms file %s != %s " + + "when ending group" % + (groupendmatch2.group(1), prevvers)) + + prevvers = vers + vers = None + elif symbolmatch is not None: + name = symbolmatch.group(1) + apisref[name] = { + "vers": vers, + "file": filenames.get(name), + } + else: + raise Exception("unexpected data %s" % line) + + +apis = {} +# Get the list of all public APIs and their corresponding version +parseSymsFile(apis, "LIBVIRT", symslibvirt, + os.path.join(srcdir, "../docs/libvirt-api.xml")) + +# And the same for the QEMU specific APIs +parseSymsFile(apis, "LIBVIRT_QEMU", symsqemu, + os.path.join(srcdir, "../docs/libvirt-qemu-api.xml")) + +# And the same for the LXC specific APIs +parseSymsFile(apis, "LIBVIRT_LXC", symslxc, + os.path.join(srcdir, "../docs/libvirt-lxc-api.xml")) + + +# Some special things which aren't public APIs, +# but we want to report +apis["virConnectSupportsFeature"] = { + "vers": "0.3.2" +} +apis["virDomainMigratePrepare"] = { + "vers": "0.3.2" +} +apis["virDomainMigratePerform"] = { + "vers": "0.3.2" +} +apis["virDomainMigrateFinish"] = { + "vers": "0.3.2" +} +apis["virDomainMigratePrepare2"] = { + "vers": "0.5.0" +} +apis["virDomainMigrateFinish2"] = { + "vers": "0.5.0" +} +apis["virDomainMigratePrepareTunnel"] = { + "vers": "0.7.2" +} + +apis["virDomainMigrateBegin3"] = { + "vers": "0.9.2" +} +apis["virDomainMigratePrepare3"] = { + "vers": "0.9.2" +} +apis["virDomainMigratePrepareTunnel3"] = { + "vers": "0.9.2" +} +apis["virDomainMigratePerform3"] = { + "vers": "0.9.2" +} +apis["virDomainMigrateFinish3"] = { + "vers": "0.9.2" +} +apis["virDomainMigrateConfirm3"] = { + "vers": "0.9.2" +} + +apis["virDomainMigrateBegin3Params"] = { + "vers": "1.1.0" +} +apis["virDomainMigratePrepare3Params"] = { + "vers": "1.1.0" +} +apis["virDomainMigratePrepareTunnel3Params"] = { + "vers": "1.1.0" +} +apis["virDomainMigratePerform3Params"] = { + "vers": "1.1.0" +} +apis["virDomainMigrateFinish3Params"] = { + "vers": "1.1.0" +} +apis["virDomainMigrateConfirm3Params"] = { + "vers": "1.1.0" +} + + +# Now we want to get the mapping between public APIs +# and driver struct fields. This lets us later match +# update the driver impls with the public APis. + +# Group name -> hash of APIs { fields -> api name } +groups = {} +ingrp = None +for drivertablefile in drivertablefiles: + with open(drivertablefile) as fh: + starttableprog = re.compile(r"struct _(vir\w*Driver)") + endtableprog = re.compile(r"};") + callbackprog = re.compile(r"^\s*vir(?:Drv)(\w+)\s+(\w+);\s*$") + ignoreapiprog = re.compile(r"\w+(Open|Close|URIProbe)") + for line in fh: + starttablematch = starttableprog.match(line) + if starttablematch is not None: + grp = starttablematch.group(1) + if grp != "virStateDriver" and grp != "virStreamDriver": + ingrp = grp + groups[ingrp] = { + "apis": {}, + "drivers": {} + } + elif ingrp is not None: + callbackmatch = callbackprog.match(line) + if callbackmatch is not None: + name = callbackmatch.group(1) + field = callbackmatch.group(2) + + api = "vir" + name + if api in apis: + groups[ingrp]["apis"][field] = api + elif ignoreapiprog.match(api) is not None: + continue + else: + raise Exception(("driver %s does not have " + + "a public API") % name) + elif endtableprog.match(line): + ingrp = None + + +# Finally, we read all the primary driver files and extract +# the driver API tables from each one. + +for src in srcs: + with open(src) as fh: + groupsre = "|".join(groups.keys()) + groupsprog2 = re.compile(r"^\s*(static\s+)?(" + + groupsre + + r")\s+(\w+)\s*=\s*{") + groupsprog3 = re.compile(r"^\s*(static\s+)?(" + + groupsre + + r")\s+NAME\(\w+\)\s*=\s*{") + nodedevimplprog = re.compile(r".*/node_device_(\w+)\.c") + miscimplprog = re.compile(r".*/(\w+?)_((\w+)_)?(\w+)\.c") + callbackprog = re.compile(r"\s*\.(\w+)\s*=\s*(\w+)\s*,?\s*" + + r"(?:/\*\s*(\d+\.\d+\.\d+)\s*" + + r"(?:-\s*(\d+\.\d+\.\d+))?\s*\*/\s*)?$") + skipapiprog = re.compile(r"\w+(Open|Close|URIProbe)") + + ingrp = None + impl = None + for line in fh: + if ingrp is None: + m = groupsprog2.match(line) + if m is None: + m = groupsprog3.match(line) + if m is not None: + ingrp = m.group(2) + impl = src + + implmatch = nodedevimplprog.match(impl) + if implmatch is None: + implmatch = miscimplprog.match(impl) + if implmatch is None: + raise Exception("Unexpected impl format '%s'" % impl) + impl = implmatch.group(1) + + if impl in groups[ingrp]["drivers"]: + raise Exception( + "Group %s already contains %s" % (ingrp, impl)) + + groups[ingrp]["drivers"][impl] = {} + else: + callbackmatch = callbackprog.match(line) + if callbackmatch is not None: + api = callbackmatch.group(1) + meth = callbackmatch.group(2) + vers = callbackmatch.group(3) + deleted = callbackmatch.group(4) + + if api == "no" or api == "name": + continue + + if meth == "NULL" and deleted is None: + raise Exception( + ("Method impl for %s is NULL, but " + + "no deleted version is provided") % api) + + if meth != "NULL" and deleted is not None: + raise Exception( + ("Method impl for %s is non-NULL, but " + + "deleted version is provided") % api) + + if vers is None and api != "connectURIProbe": + raise Exception( + "Method %s in %s is missing version" % + (meth, src)) + + if api not in groups[ingrp]["apis"]: + if skipapiprog.match(api): + continue + + raise Exception("Found unexpected method " + + "%s in %s" % (api, ingrp)) + + groups[ingrp]["drivers"][impl][api] = { + "vers": vers, + "deleted": deleted, + } + + if (api == "domainMigratePrepare" or + api == "domainMigratePrepare2" or + api == "domainMigratePrepare3"): + if ("domainMigrate" not in + groups[ingrp]["drivers"][impl]): + groups[ingrp]["drivers"][impl]["domainMigrate"] = { + "vers": vers, + } + elif line.find("}") != -1: + ingrp = None + + +# The '.open' driver method is used for 3 public APIs, so we +# have a bit of manual fixup todo with the per-driver versioning +# and support matrix + +groups["virHypervisorDriver"]["apis"]["openAuth"] = \ + "virConnectOpenAuth" +groups["virHypervisorDriver"]["apis"]["openReadOnly"] = \ + "virConnectOpenReadOnly" +groups["virHypervisorDriver"]["apis"]["domainMigrate"] = \ + "virDomainMigrate" + +openAuthVers = (0 * 1000 * 1000) + (4 * 1000) + 0 + +drivers = groups["virHypervisorDriver"]["drivers"] +for drv in drivers.keys(): + openVersStr = drivers[drv]["connectOpen"]["vers"] + openVers = 0 + if openVersStr != "Y": + openVersBits = openVersStr.split(".") + if len(openVersBits) != 3: + raise Exception("Expected 3 digit version for %s" % openVersStr) + openVers = ((int(openVersBits[0]) * 1000 * 1000) + + (int(openVersBits[1]) * 1000) + + int(openVersBits[2])) + + # virConnectOpenReadOnly always matches virConnectOpen version + drivers[drv]["connectOpenReadOnly"] = \ + drivers[drv]["connectOpen"] + + # virConnectOpenAuth is always 0.4.0 if the driver existed + # before this time, otherwise it matches the version of + # the driver's virConnectOpen entry + if openVersStr == "Y" or openVers >= openAuthVers: + vers = openVersStr + else: + vers = "0.4.0" + drivers[drv]["connectOpenAuth"] = { + "vers": vers, + } + + +# Another special case for the virDomainCreateLinux which was replaced +# with virDomainCreateXML +groups["virHypervisorDriver"]["apis"]["domainCreateLinux"] = \ + "virDomainCreateLinux" + +createAPIVers = (0 * 1000 * 1000) + (0 * 1000) + 3 + +for drv in drivers.keys(): + if "domainCreateXML" not in drivers[drv]: + continue + createVersStr = drivers[drv]["domainCreateXML"]["vers"] + createVers = 0 + if createVersStr != "Y": + createVersBits = createVersStr.split(".") + if len(createVersBits) != 3: + raise Exception("Expected 3 digit version for %s" % createVersStr) + createVers = ((int(createVersBits[0]) * 1000 * 1000) + + (int(createVersBits[1]) * 1000) + + int(createVersBits[2])) + + # virCreateLinux is always 0.0.3 if the driver existed + # before this time, otherwise it matches the version of + # the driver's virCreateXML entry + if createVersStr == "Y" or createVers >= createAPIVers: + vers = createVersStr + else: + vers = "0.0.3" + + drivers[drv]["domainCreateLinux"] = { + "vers": vers, + } + + +# Finally we generate the HTML file with the tables + +print('''<?xml version="1.0" encoding="UTF-8"?> +<!DOCTYPE html> +<html xmlns="http://www.w3.org/1999/xhtml"> +<body class="hvsupport"> +<h1>libvirt API support matrix</h1> + +<ul id="toc"></ul> + +<p> +This page documents which <a href="html/">libvirt calls</a> work on +which libvirt drivers / hypervisors, and which version the API appeared +in. If a hypervisor driver later dropped support for the API, the version +when it was removed is also mentioned (highlighted in +<span class="removedhv">dark red</span>). +</p> +''') + +for grp in sorted(groups.keys()): + print("<h2><a id=\"%s\">%s</a></h2>" % (grp, groupheaders[grp])) + print('''<table class="top_table"> +<thead> +<tr> +<th>API</th> +<th>Version</th>''') + + for drv in sorted(groups[grp]["drivers"].keys()): + print(" <th>%s</th>" % drv) + + print('''</tr> +</thead> +<tbody>''') + + row = 0 + + def sortkey(field): + return groups[grp]["apis"][field] + + for field in sorted(groups[grp]["apis"].keys(), key=sortkey): + api = groups[grp]["apis"][field] + vers = apis[api]["vers"] + htmlgrp = apis[api].get("file") + print("<tr>") + + if htmlgrp is not None: + print(('''<td>\n<a href=\"html/libvirt-%s.html#%s\">''' + + '''%s</a>\n</td>''') % + (htmlgrp, api, api)) + else: + print("<td>\n%s</td>" % api) + + print("<td>%s</td>" % vers) + + for drv in sorted(groups[grp]["drivers"].keys()): + info = "" + if field in groups[grp]["drivers"][drv]: + vers = groups[grp]["drivers"][drv][field]["vers"] + if vers is not None: + info = info + vers + + deleted = groups[grp]["drivers"][drv][field].get("deleted") + if deleted is not None: + info = info + (''' - <span class="removedhv">''' + + '''%s</span>''' % deleted) + + print("<td>%s</td>" % info) + + print("</tr>") + + row = row + 1 + if (row % 15) == 0: + print('''<tr> +<th>API</th> +<th>Version</th>''') + + for drv in sorted(groups[grp]["drivers"].keys()): + print(" <th>%s</th>" % drv) + + print("</tr>") + + print("</tbody>\n</table>") + +print("</body>\n</html>") -- 2.21.0
As part of an goal to eliminate Perl from libvirt build tools, rewrite the genaclperms.pl tool in Python. This was a straight conversion, manually going line-by-line to change the syntax from Perl to Python. Thus the overall structure of the file and approach is the same. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com> --- docs/Makefile.am | 6 +-- docs/genaclperms.pl | 125 -------------------------------------------- docs/genaclperms.py | 123 +++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 126 insertions(+), 128 deletions(-) delete mode 100755 docs/genaclperms.pl create mode 100755 docs/genaclperms.py diff --git a/docs/Makefile.am b/docs/Makefile.am index b9d69db8c8..030f53db9a 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -163,7 +163,7 @@ schemadir = $(pkgdatadir)/schemas schema_DATA = $(wildcard $(srcdir)/schemas/*.rng) EXTRA_DIST= \ - apibuild.py genaclperms.pl \ + apibuild.py genaclperms.py \ site.xsl subsite.xsl newapi.xsl page.xsl \ wrapstring.xsl \ $(dot_html) $(dot_html_in) $(gif) $(apihtml) $(apipng) \ @@ -178,8 +178,8 @@ EXTRA_DIST= \ acl_generated = aclperms.htmlinc $(srcdir)/aclperms.htmlinc: $(top_srcdir)/src/access/viraccessperm.h \ - $(srcdir)/genaclperms.pl Makefile.am - $(AM_V_GEN)$(PERL) $(srcdir)/genaclperms.pl $< > $@ + $(srcdir)/genaclperms.py Makefile.am + $(AM_V_GEN)$(RUNUTF8) $(PYTHON) $(srcdir)/genaclperms.py $< > $@ MAINTAINERCLEANFILES = \ $(addprefix $(srcdir)/,$(dot_html)) \ diff --git a/docs/genaclperms.pl b/docs/genaclperms.pl deleted file mode 100755 index e20b4c11c3..0000000000 --- a/docs/genaclperms.pl +++ /dev/null @@ -1,125 +0,0 @@ -#!/usr/bin/env perl -# -# Copyright (C) 2013 Red Hat, Inc. -# -# This library is free software; you can redistribute it and/or -# modify it under the terms of the GNU Lesser General Public -# License as published by the Free Software Foundation; either -# version 2.1 of the License, or (at your option) any later version. -# -# This library is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU -# Lesser General Public License for more details. -# -# You should have received a copy of the GNU Lesser General Public -# License along with this library. If not, see -# <http://www.gnu.org/licenses/>. -# - -use strict; -use warnings; - -my @objects = ( - "CONNECT", "DOMAIN", "INTERFACE", - "NETWORK_PORT", "NETWORK", "NODE_DEVICE", - "NWFILTER_BINDING", "NWFILTER", - "SECRET", "STORAGE_POOL", "STORAGE_VOL", - ); - -my %class; - -foreach my $object (@objects) { - my $class = lc $object; - - $class =~ s/(^\w|_\w)/uc $1/eg; - $class =~ s/_//g; - $class =~ s/Nwfilter/NWFilter/; - $class = "vir" . $class . "Ptr"; - - $class{$object} = $class; -} - -my $objects = join ("|", @objects); - -my %opts; -my $in_opts = 0; - -my %perms; - -while (<>) { - if ($in_opts) { - if (m,\*/,) { - $in_opts = 0; - } elsif (/\*\s*\@(\w+):\s*(.*?)\s*$/) { - $opts{$1} = $2; - } - } elsif (m,/\*\*,) { - $in_opts = 1; - } elsif (/VIR_ACCESS_PERM_($objects)_((?:\w|_)+),/) { - my $object = $1; - my $perm = lc $2; - next if $perm eq "last"; - - $perm =~ s/_/-/g; - - $perms{$object} = {} unless exists $perms{$object}; - $perms{$object}->{$perm} = { - desc => $opts{desc}, - message => $opts{message}, - anonymous => $opts{anonymous} - }; - %opts = (); - } -} - -print <<EOF; -<?xml version="1.0" encoding="UTF-8"?> -<!DOCTYPE html> -<html xmlns="http://www.w3.org/1999/xhtml"> - <body> -EOF - -foreach my $object (sort { $a cmp $b } keys %perms) { - my $class = $class{$object}; - my $olink = lc "object_" . $object; - print <<EOF; -<h3><a id="$olink">$class</a></h3> -<table class="acl"> - <thead> - <tr> - <th>Permission</th> - <th>Description</th> - </tr> - </thead> - <tbody> -EOF - - foreach my $perm (sort { $a cmp $b } keys %{$perms{$object}}) { - my $description = $perms{$object}->{$perm}->{desc}; - - die "missing description for $object.$perm" unless - defined $description; - - my $plink = lc "perm_" . $object . "_" . $perm; - $plink =~ s/-/_/g; - - print <<EOF; - <tr> - <td><a id="$plink">$perm</a></td> - <td>$description</td> - </tr> -EOF - - } - - print <<EOF; - </tbody> -</table> -EOF -} - -print <<EOF; - </body> -</html> -EOF diff --git a/docs/genaclperms.py b/docs/genaclperms.py new file mode 100755 index 0000000000..1d3188062f --- /dev/null +++ b/docs/genaclperms.py @@ -0,0 +1,123 @@ +#!/usr/bin/env python +# +# Copyright (C) 2013-2019 Red Hat, Inc. +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library. If not, see +# <http://www.gnu.org/licenses/>. +# + +from __future__ import print_function + +import re +import sys + +objects = [ + "CONNECT", "DOMAIN", "INTERFACE", + "NETWORK_PORT", "NETWORK", "NODE_DEVICE", + "NWFILTER_BINDING", "NWFILTER", + "SECRET", "STORAGE_POOL", "STORAGE_VOL", +] + + +classes = {} + +for obj in objects: + klass = obj.lower() + + klass = re.sub(r'''(^\w|_\w)''', lambda a: a.group(1).upper(), klass) + klass = klass.replace("_", "") + klass = klass.replace("Nwfilter", "NWFilter") + klass = "vir" + klass + "Ptr" + + classes[obj] = klass + + +objectstr = "|".join(objects) + +opts = {} +in_opts = {} + +perms = {} + +aclfile = sys.argv[1] +with open(aclfile, "r") as fh: + for line in fh: + if in_opts: + if line.find("*/") != -1: + in_opts = False + else: + m = re.match(r'''.*\*\s*\@(\w+):\s*(.*?)\s*$''', line) + if m is not None: + opts[m.group(1)] = m.group(2) + elif line.find("**") != -1: + in_opts = True + else: + m = re.match(r'''.*VIR_ACCESS_PERM_(%s)_((?:\w|_)+),''' % + objectstr, line) + if m is not None: + obj = m.group(1) + perm = m.group(2).lower() + if perm == "last": + continue + + perm = perm.replace("_", "-") + + if obj not in perms: + perms[obj] = {} + perms[obj][perm] = { + "desc": opts.get("desc", None), + "message": opts.get("message", None), + "anonymous": opts.get("anonymous", None), + } + opts = {} + +print('<?xml version="1.0" encoding="UTF-8"?>') +print('<!DOCTYPE html>') +print('<html xmlns="http://www.w3.org/1999/xhtml">') +print(' <body>') + +for obj in sorted(perms.keys()): + klass = classes[obj] + + olink = "object_" + obj.lower() + + print(' <h3><a id="%s">%s</a></h3>' % (olink, klass)) + print(' <table class="acl">') + print(' <thead>') + print(' <tr>') + print(' <th>Permission</th>') + print(' <th>Description</th>') + print(' </tr>') + print(' </thead>') + print(' <tbody>') + + for perm in sorted(perms[obj].keys()): + description = perms[obj][perm]["desc"] + + if description is None: + raise Exception("missing description for %s.%s" % (obj, perm)) + + plink = "perm_" + obj.lower() + "_" + perm.lower() + plink = plink.replace("-", "_") + + print(' <tr>') + print(' <td><a id="%s">%s</a></td>' % (plink, perm)) + print(' <td>%s</td>' % description) + print(' </tr>') + + print(' </tbody>') + print(' </table>') + +print(' </body>') +print('</html>') -- 2.21.0
participants (6)
-
Andrea Bolognani -
Daniel P. Berrangé -
Erik Skultety -
Ján Tomko -
Pavel Hrdina -
Peter Krempa