Bug#869122: Bug#869373: perl: Please add workaround to fix FTBFS on sh4

2017-07-22 Thread John Paul Adrian Glaubitz
On 07/22/2017 09:38 PM, Niko Tyni wrote:
> Sure, I was just thinking of a manual "porter upload" to get the binNMUs
> going before the next source upload. But it looks like you (or someone
> else) did that already.

That was me ;-). I'm doing the same for hppa now.

>>> The failure mode looks very similar to the hppa one (#869122, cc'd)
>>> so I guess it should get similar treatment.
>>
>> I tried the same workaround on hppa, it didn't help unfortunately.
> 
> Too bad. Dropping that bug then.

Adding it back on. For some reason, adding "noopt" wasn't passed to
the build environment. It builds with the following hack added
to debian/config.debian:

hppa:*)
optimize="$optimize${optimize:+ }-O0";;

I will bootstrap perl_5.26 for hppa now.

>>> Would you be able to narrow this down to file level? Cf. #838613 and
>>> debian/patches/debian/hppa_op_optimize_workaround.diff
>>
>> This patch seems to fix a failure of the testsuite, no? The build on
>> hppa crashes much earlier though.
> 
> My point was just that file-level optimizations can be achieved by
> modifying cflags.SH as seen in that patch. I think (but haven't
> verified) that miniperl is built via cflags.SH too.

OK, I wasn't aware of that. That's the first time I am hacking
on the perl package ;).

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913



Bug#869122: Bug#869373: perl: Please add workaround to fix FTBFS on sh4

2017-07-22 Thread John Paul Adrian Glaubitz
On 07/22/2017 09:08 PM, Niko Tyni wrote:
> Thanks. In that case you should be able to work around the breakage
> (and hopefully catch up with the 5.26 transition) by building with
> DEB_BUILD_OPTIONS=noopt.

This is for the buildds. I know that I can build the package manually.

> The failure mode looks very similar to the hppa one (#869122, cc'd)
> so I guess it should get similar treatment.

I tried the same workaround on hppa, it didn't help unfortunately.

> Would you be able to narrow this down to file level? Cf. #838613 and
> debian/patches/debian/hppa_op_optimize_workaround.diff

This patch seems to fix a failure of the testsuite, no? The build on
hppa crashes much earlier though.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaub...@debian.org
`. `'   Freie Universitaet Berlin - glaub...@physik.fu-berlin.de
  `-GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913



Bug#869122: Bug#869373: perl: Please add workaround to fix FTBFS on sh4

2017-07-22 Thread Niko Tyni
On Sat, Jul 22, 2017 at 08:51:23PM +0200, John Paul Adrian Glaubitz wrote:
> Source: perl
> Version: 5.26.0-4
> Severity: normal
> Tags: patch
> User: debian-sup...@lists.debian.org
> Usertags: sh4
> 
> Hi!
> 
> src:perl currently fails to build from source on sh4 with:
> 
> ./miniperl -Ilib make_ext.pl cpan/Archive-Tar/pm_to_blib  
> MAKE="/usr/bin/make" LIBPERL_A=libperl.a
> qemu: uncaught target signal 11 (Segmentation fault) - core dumped
> Unsuccessful Makefile.PL(cpan/Archive-Tar): code=11 at make_ext.pl line 518.
> Makefile:586: recipe for target 'cpan/Archive-Tar/pm_to_blib' failed
> make[1]: *** [cpan/Archive-Tar/pm_to_blib] Error 2
> make[1]: Leaving directory '/<>'
> debian/rules:124: recipe for target 'perl.static' failed
> make: *** [perl.static] Error 2
> dpkg-buildpackage: error: debian/rules build-arch gave error exit status 2
> 
> I have not tracked down the actual reason for the bug yet, but it can
> be worked around by building perl without optimizations enabled on sh4.
> 
> The attached patch disables the optimizations on sh4 and makes the build
> succeed. Please consider applying it for the next upload.

Thanks. In that case you should be able to work around the breakage
(and hopefully catch up with the 5.26 transition) by building with
DEB_BUILD_OPTIONS=noopt.

The failure mode looks very similar to the hppa one (#869122, cc'd)
so I guess it should get similar treatment.

Would you be able to narrow this down to file level? Cf. #838613 and
debian/patches/debian/hppa_op_optimize_workaround.diff
-- 
Niko Tyni   nt...@debian.org