Bug#724469: FTBFS on all big-endian architectures

[Lennart Sorensen]

https://rt.cpan.org/Ticket/Display.html?id=89552

Sure. Debian porters, I'll let you subscribe to the RT ticket, and
hopefully take care of this porting problem.

I'd like to see this RC bug fixed eventually, and I still hope this
can be done without dropping support for too many architectures in
this package.
AFAICT the latest patch proposed by upstream on February 9 [1] has
been tested on mips only. My understanding is that upstream has been
waiting for more test results since then. Can anyone please test this
on other big-endian architectures?
It would good if we could at least fix this for the 32-bit ones.
[1] https://rt.cpan.org/Ticket/Attachment/1324475/702426/0001-Fix-return-value-handling-on-big-endian-architecture.patch

Patch works for powerpc and someone else already reported it working
for powerpcspe.

As expected it does NOT work on ppc64. I am not currently awake enough
to try and figure out why.

Since mips was already tested with the patch originally, that probably
just leaves sparc and s390 to test (I can't tell if s390 tested it or not,
only that s390x does not work yet due to being 64bit).

[Adam Conrad]

AFAICT the latest patch proposed by upstream on February 9 [1] has
been tested on mips only. My understanding is that upstream has been
waiting for more test results since then. Can anyone please test this
on other big-endian architectures?
It would good if we could at least fix this for the 32-bit ones.
[1] https://rt.cpan.org/Ticket/Attachment/1324475/702426/0001-Fix-return-value-handling-on-big-endian-architecture.patch

Works fine for me on powerpc, but fails miserably on s390x:

t/00-basic-types.t ............ ok
t/arg-checks.t ................ ok
t/arrays.t .................... 1/29
# Failed test at t/arrays.t line 14.
# got: '0'
# expected: '6'
Out of memory!
# Looks like you planned 29 tests but ran 2.
# Looks like you failed 1 test of 2 run.
# Looks like your test exited with 1 just after 2.
t/arrays.t .................... Failed 28/29 subtests
t/boxed.t ..................... ok
t/cairo-integration.t ......... ok
t/callbacks.t ................. 1/25
# Failed test at t/callbacks.t line 14.
# got: '6941192'
# expected: '23'

# Failed test at t/callbacks.t line 16.
# got: '894'
# expected: '23'

# Failed test at t/callbacks.t line 17.
# got: '894'
# expected: '23'

# Failed test at t/callbacks.t line 18.
# got: '-1071533088'
# expected: '46'

# Failed test at t/callbacks.t line 22.
# got: '0'
# expected: '23'

# Failed test at t/callbacks.t line 26.
# got: '-1071861040'
# expected: '23'
# Looks like you failed 6 tests of 25.
t/callbacks.t ................. Dubious, test returned 6 (wstat 1536, 0x600)
Failed 6/25 subtests
t/closures.t .................. ok
t/constants.t ................. ok
t/enums.t ..................... Failed 3/4 subtests
t/hashes.t .................... ok
t/interface-implementation.t .. ok
t/objects.t ................... ok
t/structs.t ................... ok
t/values.t .................... ok
t/vfunc-chaining.t ............ ok
t/vfunc-ref-counting.t ........ ok

Test Summary Report
-------------------
t/arrays.t (Wstat: 9 Tests: 2 Failed: 1)
Failed test: 2
Non-zero wait status: 9
Parse errors: Bad plan. You planned 29 tests but ran 2.
t/callbacks.t (Wstat: 1536 Tests: 25 Failed: 6)
Failed tests: 3, 6, 9, 14, 19, 25
Non-zero exit status: 6
t/enums.t (Wstat: 11 Tests: 1 Failed: 0)
Non-zero wait status: 11
Parse errors: Bad plan. You planned 4 tests but ran 1.
Files=16, Tests=297, 222 wallclock secs ( 0.07 usr 0.03 sys + 11.27 cusr 39.60 csys = 50.97 CPU)
Result: FAIL
Failed 3/16 test programs. 7/297 subtests failed.

[intrigeri]

Hi,

Post by Lennart Sorensen
Patch works for powerpc and someone else already reported it working
for powerpcspe.
Works fine for me on powerpc,

Thanks a lot for testing!

I've uploaded libglib-object-introspection-perl 0.020-2 with this
patch applied. This should at least fix the problem for 32-bit
big-endian architectures.

Any s390x porter planning to work on this? (And if so, ETA?)

I'd rather not drop s390x from the list of architectures this package
is built for, but this RC bug has now been around for 6 months, and at
some point I'll want to get rid of it.

Cheers,

[intrigeri]

Hi,

Care to share the results of your own findings?

Unfortunately, I don't have the skills needed to work on this problem
myself, so there is no such thing to share. The best I can do is to go
on forwarding patches and test results between upstream and
Debian porters.

Cheers,

[intrigeri]

Hi,

Not fixing a bug isn't the way to get rid of it.

I agree. My wording was not appropriate, and I'm sorry for the bad
feelings I may have caused.

Hoping to clarify a bit:

1. I'll try to go on playing the intermediate between the relevant
parties (upstream and the Debian porters) and pinging people as
needed, as I've been doing in the last 2.5 months. But if I cause
communication problems again, then I'll ask for someone else on
the Perl team to take over this task from me.

2. I want Jessie to be released with this package (and its
reverse-dependencies) working on as many supported architectures
as possible. Given I don't have the skills needed to port it to
big-endian 64-bit architectures myself, all I can do is #1. So,
whether Jessie ships this package on these architectures does not
depend much on me.

Cheers,

[Hiroyuki Yamamoto]

Post by Lennart Sorensen

https://rt.cpan.org/Ticket/Display.html?id=89552

Sure. Debian porters, I'll let you subscribe to the RT ticket, and
hopefully take care of this porting problem.

I'd like to see this RC bug fixed eventually, and I still hope this
can be done without dropping support for too many architectures in
this package.
AFAICT the latest patch proposed by upstream on February 9 [1] has
been tested on mips only. My understanding is that upstream has been
waiting for more test results since then. Can anyone please test this
on other big-endian architectures?
It would good if we could at least fix this for the 32-bit ones.
[1] https://rt.cpan.org/Ticket/Attachment/1324475/702426/0001-Fix-return-value-handling-on-big-endian-architecture.patch

Patch works for powerpc and someone else already reported it working
for powerpcspe.
As expected it does NOT work on ppc64. I am not currently awake enough
to try and figure out why.
Since mips was already tested with the patch originally, that probably
just leaves sparc and s390 to test (I can't tell if s390 tested it or not,
only that s390x does not work yet due to being 64bit).

On ppc64, failing to build the source package patched was confirmed.

--

make[1]: Leaving directory `/«PKGBUILDDIR»'
dh_auto_test -a
make[1]: Entering directory `/«PKGBUILDDIR»'
LD_LIBRARY_PATH=:build PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/00-basic-types.t ............ ok
t/arg-checks.t ................ ok

# Failed test at t/arrays.t line 14.
# got: '0'
# expected: '6'
t/arrays.t ....................
Failed 28/29 subtests
t/boxed.t ..................... ok
t/cairo-integration.t ......... ok

# Failed test at t/callbacks.t line 14.
# got: '-40706304'
# expected: '23'

# Failed test at t/callbacks.t line 16.
# got: '7395392'
# expected: '23'

# Failed test at t/callbacks.t line 17.
# got: '7395392'
# expected: '23'

# Failed test at t/callbacks.t line 18.
# got: '-33534596'
# expected: '46'

# Failed test at t/callbacks.t line 22.
# got: '7171660'
# expected: '23'

# Failed test at t/callbacks.t line 26.
# got: '-40706256'
# expected: '23'
# Looks like you failed 6 tests of 25.
t/callbacks.t .................
Dubious, test returned 6 (wstat 1536, 0x600)
Failed 6/25 subtests
t/closures.t .................. ok
t/constants.t ................. ok
t/enums.t .....................
Failed 3/4 subtests
t/hashes.t .................... ok
t/interface-implementation.t .. ok
t/objects.t ................... ok
t/structs.t ................... ok
t/values.t .................... ok
t/vfunc-chaining.t ............ ok
t/vfunc-ref-counting.t ........ ok
Failed 3/16 test programs. 7/299 subtests failed.

Test Summary Report
-------------------
t/arrays.t (Wstat: 9 Tests: 2 Failed: 1)
Failed test: 2
Non-zero wait status: 9
Parse errors: Bad plan. You planned 29 tests but ran 2.
t/callbacks.t (Wstat: 1536 Tests: 25 Failed: 6)
Failed tests: 3, 6, 9, 14, 19, 25
Non-zero exit status: 6
t/enums.t (Wstat: 11 Tests: 1 Failed: 0)
Non-zero wait status: 11
Parse errors: Bad plan. You planned 4 tests but ran 1.
Files=16, Tests=299, 835 wallclock secs ( 0.25 usr 0.09 sys + 74.57 cusr 19.01 csys = 93.92 CPU)
Result: FAIL
make[1]: *** [test_dynamic] Error 255
make[1]: Leaving directory `/«PKGBUILDDIR»'
dh_auto_test: make -j1 test returned exit code 2
make: *** [build-arch] Error 2
dpkg-buildpackage: error: debian/rules build-arch gave error exit status 2
────────────────────────────────────────────────────────────────────────────────
Build finished at 20140323-1923

Finished