11:33 a.m.
This reverts commit 6aa90452aa63cb1e1ffa84ff5f93f5873bf810a0.
Turns out that not linking against libvirt and gnulib is okay for
regular Linux (and FreeBSD) builds, but makes mingw very unhappy.
.../virnetserverclientmock_la-virnetserverclientmock.o:
In function `virNetSocketGetSELinuxContext':
.../virnetserverclientmock.c:61: undefined reference to `rpl_strdup'
.../libvirportallocatormock_la-virportallocatortest.o:
In function `init_syms':
.../virportallocatortest.c:61: undefined reference to `virFileClose'
---
Pushed as build breaker.
Bunny ears of shame will be deployed in the morning.
tests/Makefile.am | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/tests/Makefile.am b/tests/Makefile.am
index fb862cd..c5986f0 100644
--- a/tests/Makefile.am
+++ b/tests/Makefile.am
@@ -946,6 +946,7 @@ virnetserverclientmock_la_SOURCES = \
virnetserverclientmock.c
virnetserverclientmock_la_CFLAGS = $(AM_CFLAGS)
virnetserverclientmock_la_LDFLAGS = $(MOCKLIBS_LDFLAGS)
+virnetserverclientmock_la_LIBADD = $(GNULIB_LIBS)
if WITH_GNUTLS
virnettlscontexttest_SOURCES = \
@@ -1024,6 +1025,7 @@ libvirportallocatormock_la_SOURCES = \
virportallocatortest.c
libvirportallocatormock_la_CFLAGS = $(AM_CFLAGS) -DMOCK_HELPER=1
libvirportallocatormock_la_LDFLAGS = $(MOCKLIBS_LDFLAGS)
+libvirportallocatormock_la_LIBADD = ../src/libvirt.la
vircgrouptest_SOURCES = \
vircgrouptest.c testutils.h testutils.c
@@ -1049,6 +1051,8 @@ virpcitest_LDADD = $(LDADDS)
virpcimock_la_SOURCES = \
virpcimock.c
virpcimock_la_CFLAGS = $(AM_CFLAGS)
+virpcimock_la_LIBADD = $(GNULIB_LIBS) \
+ ../src/libvirt.la
virpcimock_la_LDFLAGS = $(MOCKLIBS_LDFLAGS)
nodeinfomock_la_SOURCES = \
@@ -1064,6 +1068,8 @@ virnetdevtest_LDADD = $(LDADDS)
virnetdevmock_la_SOURCES = \
virnetdevmock.c
virnetdevmock_la_CFLAGS = $(AM_CFLAGS) $(LIBNL_CFLAGS)
+virnetdevmock_la_LIBADD = $(GNULIB_LIBS) \
+ ../src/libvirt.la
virnetdevmock_la_LDFLAGS = $(MOCKLIBS_LDFLAGS)
virrotatingfiletest_SOURCES = \
--
2.5.0
12:09 p.m.
On Wed, Feb 10, 2016 at 06:33:21PM +0100, Andrea Bolognani wrote:
This reverts commit 6aa90452aa63cb1e1ffa84ff5f93f5873bf810a0.
Turns out that not linking against libvirt and gnulib is okay for
regular Linux (and FreeBSD) builds, but makes mingw very unhappy.
.../virnetserverclientmock_la-virnetserverclientmock.o:
In function `virNetSocketGetSELinuxContext':
.../virnetserverclientmock.c:61: undefined reference to `rpl_strdup'
.../libvirportallocatormock_la-virportallocatortest.o:
In function `init_syms':
.../virportallocatortest.c:61: undefined reference to `virFileClose'
---
Pushed as build breaker.
Bunny ears of shame will be deployed in the morning.
You should still be able to drop linkage of libvirt.la, just keep
the gnulib bits. The gnulib addition is harmless, because it is
a static library the linker will drop all functions which are not
actually used by the mock.o, so it will end up being a no-op on
Linux, and will only pull in rpl_strup on mingw32. We should never
link libvirt.la to mock libs though.
Regards,
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
2:54 a.m.
On Wed, 2016-02-10 at 18:09 +0000, Daniel P. Berrange wrote:
On Wed, Feb 10, 2016 at 06:33:21PM +0100, Andrea Bolognani wrote:
> This reverts commit 6aa90452aa63cb1e1ffa84ff5f93f5873bf810a0.
>
> Turns out that not linking against libvirt and gnulib is okay for
> regular Linux (and FreeBSD) builds, but makes mingw very unhappy.
>
> .../virnetserverclientmock_la-virnetserverclientmock.o:
> In function `virNetSocketGetSELinuxContext':
> .../virnetserverclientmock.c:61: undefined reference to `rpl_strdup'
> .../libvirportallocatormock_la-virportallocatortest.o:
> In function `init_syms':
> .../virportallocatortest.c:61: undefined reference to `virFileClose'
> ---
> Pushed as build breaker.
>
> Bunny ears of shame will be deployed in the morning.
You should still be able to drop linkage of libvirt.la, just keep
the gnulib bits. The gnulib addition is harmless, because it is
a static library the linker will drop all functions which are not
actually used by the mock.o, so it will end up being a no-op on
Linux, and will only pull in rpl_strup on mingw32. We should never
link libvirt.la to mock libs though.
Looks like it is complaining about virFileClose(), plus
virFilePrint() in the full output, as well. So maybe linking against
libvirt.la is needed after all?
If there's no harm in that, how about defining
MOCKLIBS_LDADD = \
$(GNULIB_LIBS) \
../src/libvirt.la
and using it for every mock library, like the existing
$(MOCKLIB_LDFLAGS)?
Cheers.
--
Andrea Bolognani
Software Engineer - Virtualization Team
3:49 a.m.
On Thu, Feb 11, 2016 at 09:54:45AM +0100, Andrea Bolognani wrote:
On Wed, 2016-02-10 at 18:09 +0000, Daniel P. Berrange wrote:
> On Wed, Feb 10, 2016 at 06:33:21PM +0100, Andrea Bolognani wrote:
> > This reverts commit 6aa90452aa63cb1e1ffa84ff5f93f5873bf810a0.
> >
> > Turns out that not linking against libvirt and gnulib is okay for
> > regular Linux (and FreeBSD) builds, but makes mingw very unhappy.
> >
> > .../virnetserverclientmock_la-virnetserverclientmock.o:
> > In function `virNetSocketGetSELinuxContext':
> > .../virnetserverclientmock.c:61: undefined reference to `rpl_strdup'
> > .../libvirportallocatormock_la-virportallocatortest.o:
> > In function `init_syms':
> > .../virportallocatortest.c:61: undefined reference to `virFileClose'
> > ---
> > Pushed as build breaker.
> >
> > Bunny ears of shame will be deployed in the morning.
>
> You should still be able to drop linkage of libvirt.la, just keep
> the gnulib bits. The gnulib addition is harmless, because it is
> a static library the linker will drop all functions which are not
> actually used by the mock.o, so it will end up being a no-op on
> Linux, and will only pull in rpl_strup on mingw32. We should never
> link libvirt.la to mock libs though.
Looks like it is complaining about virFileClose(), plus
virFilePrint() in the full output, as well. So maybe linking against
libvirt.la is needed after all?
That's a bug in virportallocatortest.c that was introduced in
0ee90812. The init_syms() fnuction should *not* be calling
VIR_FORCE_CLOSE - it must use close(), because this is the
mock-override and so should not be using libvirt.so symbols.
If there's no harm in that, how about defining
MOCKLIBS_LDADD = \
$(GNULIB_LIBS) \
../src/libvirt.la
and using it for every mock library, like the existing
$(MOCKLIB_LDFLAGS)?
Nope as above, we should not be linking libvirt.la to the mock lib,
because the mock lib is supposed to be replacing libvirt.la symbols
Regards,
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
4:17 a.m.
On Thu, 2016-02-11 at 09:49 +0000, Daniel P. Berrange wrote:
On Thu, Feb 11, 2016 at 09:54:45AM +0100, Andrea Bolognani wrote:
> On Wed, 2016-02-10 at 18:09 +0000, Daniel P. Berrange wrote:
> > On Wed, Feb 10, 2016 at 06:33:21PM +0100, Andrea Bolognani wrote:
> > > This reverts commit 6aa90452aa63cb1e1ffa84ff5f93f5873bf810a0.
> > >
> > > Turns out that not linking against libvirt and gnulib is okay for
> > > regular Linux (and FreeBSD) builds, but makes mingw very unhappy.
> > >
> > > .../virnetserverclientmock_la-virnetserverclientmock.o:
> > > In function `virNetSocketGetSELinuxContext':
> > > .../virnetserverclientmock.c:61: undefined reference to
`rpl_strdup'
> > > .../libvirportallocatormock_la-virportallocatortest.o:
> > > In function `init_syms':
> > > .../virportallocatortest.c:61: undefined reference to
`virFileClose'
> > > ---
> > > Pushed as build breaker.
> > >
> > > Bunny ears of shame will be deployed in the morning.
> >
> > You should still be able to drop linkage of libvirt.la, just keep
> > the gnulib bits. The gnulib addition is harmless, because it is
> > a static library the linker will drop all functions which are not
> > actually used by the mock.o, so it will end up being a no-op on
> > Linux, and will only pull in rpl_strup on mingw32. We should never
> > link libvirt.la to mock libs though.
>
> Looks like it is complaining about virFileClose(), plus
> virFilePrint() in the full output, as well. So maybe linking against
> libvirt.la is needed after all?
That's a bug in virportallocatortest.c that was introduced in
0ee90812. The init_syms() fnuction should *not* be calling
VIR_FORCE_CLOSE - it must use close(), because this is the
mock-override and so should not be using libvirt.so symbols.
Okay, I will change it to use plain close().
> If there's no harm in that, how about defining
>
> MOCKLIBS_LDADD = \
> $(GNULIB_LIBS) \
> ../src/libvirt.la
>
> and using it for every mock library, like the existing
> $(MOCKLIB_LDFLAGS)?
Nope as above, we should not be linking libvirt.la to the mock lib,
because the mock lib is supposed to be replacing libvirt.la symbols
I will define $(MOCKLIBS_LDADD) to contain just $(GNULIB_LIBS)
then, and make sure mock libraries link against it.
Cheers.
--
Andrea Bolognani
Software Engineer - Virtualization Team
3460
days inactive
3461
days old
4 comments
2 participants
participants (2)
-
Andrea Bolognani -
Daniel P. Berrange