5:23 p.m.
Hi,
Makefile.maint assumes this in some places. This doesn't make the world
perfect but maybe a bit better?
Cheers,
-- Guido
10:52 a.m.
On Sat, Dec 13, 2008 at 12:23:30AM +0100, Guido G?nther wrote:
Hi,
Makefile.maint assumes this in some places. This doesn't make the world
perfect but maybe a bit better?
Seems reasonable to me.
I think it is about time I changed the 'autobuild.sh' to do a VPATH
build in the nightly builds, so we don't keep making regressions for
the srcdir != builddir scenario
Daniel
--
|: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :|
|: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|
2:31 a.m.
On Sat, Dec 13, 2008 at 04:52:10PM +0000, Daniel P. Berrange wrote:
On Sat, Dec 13, 2008 at 12:23:30AM +0100, Guido G?nther wrote:
> Hi,
> Makefile.maint assumes this in some places. This doesn't make the world
> perfect but maybe a bit better?
Seems reasonable to me.
Applied now.
-- Guido
8:18 a.m.
"Daniel P. Berrange" <berrange(a)redhat.com> wrote:
On Sat, Dec 13, 2008 at 12:23:30AM +0100, Guido G?nther wrote:
> Hi,
> Makefile.maint assumes this in some places. This doesn't make the world
> perfect but maybe a bit better?
Seems reasonable to me.
I think it is about time I changed the 'autobuild.sh' to do a VPATH
build in the nightly builds, so we don't keep making regressions for
the srcdir != builddir scenario
FYI, running "make distcheck" does such a build.
2:23 a.m.
On Wed, Dec 17, 2008 at 03:18:16PM +0100, Jim Meyering wrote:
"Daniel P. Berrange" <berrange(a)redhat.com> wrote:
> On Sat, Dec 13, 2008 at 12:23:30AM +0100, Guido G?nther wrote:
>> Hi,
>> Makefile.maint assumes this in some places. This doesn't make the world
>> perfect but maybe a bit better?
>
> Seems reasonable to me.
>
> I think it is about time I changed the 'autobuild.sh' to do a VPATH
> build in the nightly builds, so we don't keep making regressions for
> the srcdir != builddir scenario
FYI, running "make distcheck" does such a build.
"make
distcheck" itself fails for me when run with $srcdir != $buildir.
-- Guido
8:31 a.m.
Guido Günther <agx(a)sigxcpu.org> wrote:
Makefile.maint assumes this in some places. This doesn't make the
world
perfect but maybe a bit better?
Um, not really, since a pathological $(abs_srcdir) may contain "; rm -rf /;"
Best to avoid it and use $(srcdir) instead.
Were you able to construct a situation in which this change helped?
I wonder, since in a non-srcdir builds,
GNUmakefile and Makefile.maint don't exist, so it should be impossible
to run their rules without either copying them into the build directory
or using make's -f option.
[copying the files into place, to test this ...]
It looks like if you use VC_LIST's -C option,
then that invalidates the exempted names in the .x-sc* files.
FWIW, I did "make distclean && mkdir .j && cd .j &&
./configure".
From there, none of the Makefile.maint rules are available.
Eventually, I'll probably sync from coreutils/gnulib, where this
file is now named maint.mk.
2:29 a.m.
On Wed, Dec 17, 2008 at 03:31:45PM +0100, Jim Meyering wrote:
Guido Günther <agx(a)sigxcpu.org> wrote:
> Makefile.maint assumes this in some places. This doesn't make the world
> perfect but maybe a bit better?
Um, not really, since a pathological $(abs_srcdir) may contain "; rm -rf /;"
Best to avoid it and use $(srcdir) instead.
O.k.
Were you able to construct a situation in which this change helped?
Yes. Running:
make -f ../GNUmakefile <target>
starts working.
I wonder, since in a non-srcdir builds,
GNUmakefile and Makefile.maint don't exist, so it should be impossible
to run their rules without either copying them into the build directory
or using make's -f option.
[copying the files into place, to test this ...]
It looks like if you use VC_LIST's -C option,
then that invalidates the exempted names in the .x-sc* files.
But that's the
only problem left to fix to make "make distcheck" work
with $srcdir != builddir. Before that it wouldn't even run the tests.
Fixing up the pattern matching for the paths in the .x-sc* files still
on my todo list.
Cheers,
-- Guido
5853
days inactive
5859
days old
6 comments
3 participants
participants (3)
-
Daniel P. Berrange -
Guido Günther -
Jim Meyering