Bug#1000896: Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2022-01-22 Thread gregor herrmann
On Sat, 22 Jan 2022 18:57:30 +0200, Niko Tyni wrote:

> Given  % 2^32 == 3567587327 it's pretty clear that the
> requested size does not fit into 32 bits (shmget(2) takes a size_t)
> so just its lower bits get passed in. Not sure if perl should protest
> somehow there.

Thanks for the analysis!
 
> Anyway, seems like the test could just be skipped on 32-bit or something.

As I'm not sure there's a reliable way detect the bitness from the
test, I've marked it as TODO for the time being (and added your
findings to the upstream issue).


Cheers,
gregor

-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   NP: Bob Dylan: Mr Tambourine Man


signature.asc
Description: Digital Signature


Bug#997829: Bug#1000896: Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2022-01-22 Thread Niko Tyni
On Sat, Jan 15, 2022 at 11:20:51PM +0100, gregor herrmann wrote:
> On Mon, 25 Oct 2021 19:59:39 +0200, gregor herrmann wrote:
 
> No reaction so far at https://github.com/stevieb9/ipc-shareable/issues/14
> 
> Failures are visible at
> https://qa.debian.org/excuses.php?package=libipc-shareable-perl
> 
> And can easily be reproduced by building in an i386 chroot.
> 
> After some looking at the the code I have no idea what's broken
> there.
> Anyone else?

It boils down to

  perl -e 'shmget(0, , 0666) or die'

which dies on amd64 but not in i386.

Looking further:

  # strace -e shmget perl -we 'shmget(0, , 0666) or die'
  shmget(IPC_PRIVATE, 3567587327, 0666)   = 5177377
  +++ exited with 0 +++

Given  % 2^32 == 3567587327 it's pretty clear that the
requested size does not fit into 32 bits (shmget(2) takes a size_t)
so just its lower bits get passed in. Not sure if perl should protest
somehow there.

Anyway, seems like the test could just be skipped on 32-bit or something.
-- 
Niko Tyni   nt...@debian.org



Bug#1000896: Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2022-01-21 Thread gregor herrmann
On Sat, 15 Jan 2022 23:20:51 +0100, gregor herrmann wrote:

> And can easily be reproduced by building in an i386 chroot.

With some debugging statements:

lib/IPC/Shareable/SharedMem.pm
31  my $id = shmget($key, $size, $flags);

On amd64 $id is undef, as expected, but contains a value on i386.


Cheers,
gregor

-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   NP: Beatles


signature.asc
Description: Digital Signature


Bug#1000896: Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2022-01-15 Thread gregor herrmann
On Mon, 25 Oct 2021 19:59:39 +0200, gregor herrmann wrote:

> I'm going the forward the bug upstream.

No reaction so far at https://github.com/stevieb9/ipc-shareable/issues/14

Failures are visible at
https://qa.debian.org/excuses.php?package=libipc-shareable-perl

And can easily be reproduced by building in an i386 chroot.

After some looking at the the code I have no idea what's broken
there.
Anyone else?


Cheers,
gregor

-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   


signature.asc
Description: Digital Signature


Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2021-10-25 Thread gregor herrmann
Control: tags -1 + ftbfs sid bookworm

On Mon, 25 Oct 2021 16:51:00 +0200, Paul Gevers wrote:

> With a recent upload of libipc-shareable-perl the autopkgtest of
> libipc-shareable-perl fails in testing when that autopkgtest is run with
> the binary packages of libipc-shareable-perl from unstable. It passes
> when run with only packages from testing. In tabular form:

Thanks for noticing.
It also fails to build on the same architectures:
https://tests.reproducible-builds.org/debian/rb-pkg/unstable/i386/libipc-shareable-perl.html
https://tests.reproducible-builds.org/debian/rb-pkg/unstable/armhf/libipc-shareable-perl.html
with the same error.
 
CPAN smoke testers also look suspicious:
http://matrix.cpantesters.org/?dist=IPC-Shareable%201.06;os=linux;perl=5.32.0;reports=1


I'm going the forward the bug upstream.


Cheers,
gregor

-- 
 .''`.  https://info.comodo.priv.at -- Debian Developer https://www.debian.org
 : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D  85FA BB3A 6801 8649 AA06
 `. `'  Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe
   `-   


signature.asc
Description: Digital Signature


Processed: Re: Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2021-10-25 Thread Debian Bug Tracking System
Processing control commands:

> tags -1 + ftbfs sid bookworm
Bug #997829 [src:libipc-shareable-perl] libipc-shareable-perl: autopkgtest 
regression on armhf and i386
Added tag(s) ftbfs, sid, and bookworm.

-- 
997829: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=997829
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems



Bug#997829: libipc-shareable-perl: autopkgtest regression on armhf and i386

2021-10-25 Thread Paul Gevers
Source: libipc-shareable-perl
Version: 1.06-1
X-Debbugs-CC: debian...@lists.debian.org
Severity: serious
User: debian...@lists.debian.org
Usertags: regression

Dear maintainer(s),

With a recent upload of libipc-shareable-perl the autopkgtest of
libipc-shareable-perl fails in testing when that autopkgtest is run with
the binary packages of libipc-shareable-perl from unstable. It passes
when run with only packages from testing. In tabular form:

   passfail
libipc-shareable-perl  from testing1.06-1
all others from testingfrom testing

I copied some of the output at the bottom of this report. The issue
seems to be related to 32 bit archs.

Currently this regression is blocking the migration to testing [1]. Can
you please investigate the situation and fix it?

More information about this bug and the reason for filing it can be found on
https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation

Paul

[1] https://qa.debian.org/excuses.php?package=libipc-shareable-perl

https://ci.debian.net/data/autopkgtest/testing/armhf/libi/libipc-shareable-perl/16170185/log.gz

t/65-seg_size.t .. ok 1 - size larger than MAX croaks ok
ok 2 - ...and error is sane
IPC_MEM env var not set, skipping the exhaust memory test
not ok 3 - We croak if size is greater than max RAM

#   Failed test 'We croak if size is greater than max RAM'
#   at t/65-seg_size.t line 52.
#  got: '1'
# expected: undef
not ok 4 - ...and error is sane

#   Failed test '...and error is sane'
#   at t/65-seg_size.t line 54.
#   ''
# doesn't match '(?^:Cannot allocate memory|Out of memory|Invalid
argument)'
ok 5 - size param is the same as the segment size
ok 6 - actual size in bytes ok if sending in custom size
1..6
# Looks like you failed 2 tests of 6.
Dubious, test returned 2 (wstat 512, 0x200)
Failed 2/6 subtests

[...]

t/85-spawn_object_contains_data.t  skipped: IPC_SPAWN_TEST env var
not set
t/86-unspawn_object_contains_data.t .. skipped: IPC_SPAWN_TEST env var
not set
t/91-pod_linkcheck.t . skipped: Author test:
RELEASE_TESTING not set
t/93-manifest.t .. skipped: Author test:
RELEASE_TESTING not set

Test Summary Report
---
t/65-seg_size.t(Wstat: 512 Tests: 6 Failed: 2)
  Failed tests:  3-4
  Non-zero exit status: 2
Files=36, Tests=412, 16 wallclock secs ( 0.20 usr  0.36 sys +  4.66 cusr
 4.92 csys = 10.14 CPU)
Result: FAIL
autopkgtest [10:47:54]: test autodep8-perl-build-deps




OpenPGP_signature
Description: OpenPGP digital signature