Source: gscan2pdf
Severity: serious
Version: 2.13.4-4
Tags: ftbfs
Justification: fails to build from source (but built successfully in the past)
Dear Maintainer,
During a ratt run for src:ossp-uuid, I see
2024/09/26 16:46:59 Building package 68 of 123: gscan2pdf
It's currently 2024-09-26T17:03:48+02:00,
and the process tree looks like this:
2992217 1628 S 0.0 0.0 0:00.00 └─ sh /usr/bin/xvfb-run -a dh_auto_test
2992230 5284 S 0.0 0.0 0:00.11 ├─ perl /usr/bin/dh_auto_test
2992237 2064 S 0.0 0.0 0:00.02 │ └─ make -j24 test TEST_VERBOSE=1
2992252 5336 S 0.0 0.0 0:00.63 │ └─ perl -MExtUtils::Command::MM -MTest::Harness -e undef *Test::Harness::Switches; test_harness(1,
3001120 75204 S 0.0 0.2 0:02.01 │ └─ perl t/1113_save_pdf_with_error.t
3001252 75204 S 0.0 0.2 0:00.14 │ └─ perl t/1113_save_pdf_with_error.t
2992227 50892 S 0.0 0.1 0:02.35 └─ Xvfb :99 -screen 0 1280x1024x24 -nolisten tcp -auth /tmp/xvfb-run.wk10k0/Xauthority
and hasn't changed the last few times I looked (3001252 is a thread).
strace says
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
ppoll([{fd=4, events=POLLIN}, {fd=10, events=POLLIN}], 2, {tv_sec=0, tv_nsec=99179000}, NULL, 8) = 0 (Timeout)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
ppoll([{fd=4, events=POLLIN}, {fd=10, events=POLLIN}], 2, {tv_sec=0, tv_nsec=99457000}, NULL, 8) = 0 (Timeout)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
ppoll([{fd=4, events=POLLIN}, {fd=10, events=POLLIN}], 2, {tv_sec=0, tv_nsec=99636000}, NULL, 8) = 0 (Timeout)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
recvmsg(4, {msg_namelen=0}, 0) = -1 EAGAIN (Resource temporarily unavailable)
ppoll([{fd=4, events=POLLIN}, {fd=10, events=POLLIN}], 2, {tv_sec=0, tv_nsec=99280000}, NULL, 8^Cstrace: Process 3001120 detached
futex(0x7f7054000ca4, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, NULL, FUTEX_BITSET_MATCH_ANY^Cstrace: Process 3001252 detached
epoll_wait(3, ^Cstrace: Process 2992227 detached
This is a ratt run, so the build is in a schroot.
I created the chroot this morning, so no funny business there.
I SIGINTed the test, and the tests continued.
The same happened with t/133_save_tiff_with_error.t later.
The same happened with t/1612_import_TIFF_with_error.t later.
The same happened with t/1626_import_PDF_with_error.t later.
The same happened with t/1632_import_ppm_with_error.t later.
The same happened with t/213_rotate_with_error.t later.
The same happened with t/243_threshold_with_error.t later.
The same happened with t/253_negate_with_error.t later.
The same happened with t/263_unsharp_mask_with_error.t later.
The same happened with t/273_crop_with_error.t later.
The same happened with t/283_to_png_with_error.t later.
The same happened with t/377_user_defined_with_error.t later.
One has to wonder if the error is that it hangs.
I'm attaching the build log.
Don't think this is the same as #1012250 since that's autopkgtests.
Naturally,
2024/09/26 17:21:32 building gscan2pdf failed: exit status 2
but I'm forced to assume this is due to whatever flakiness in the test?
I reproduce this behaviour with a plain sbuild -d unstable gscan2pdf
so it's not the fault of the new src:ossp-uuid.
Best,
-- System Information:
Debian Release: 12.4
APT prefers stable-updates
APT policy: (500, 'stable-updates'), (500, 'stable-security'), (500, 'stable-debug'), (500, 'stable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 6.1.0-12-amd64 (SMP w/24 CPU threads; PREEMPT)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_FIRMWARE_WORKAROUND, TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8), LANGUAGE=en_GB:en
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Acknowledgement sent
to Santiago Vila <[email protected]>:
Extra info received and forwarded to list. Copy sent to Jeffrey Ratcliffe <[email protected]>.
(Sat, 28 Sep 2024 11:15:01 GMT) (full text, mbox, link).
Subject: Re: Bug#1082810: gscan2pdf: deadlock in tests
Date: Sat, 28 Sep 2024 13:10:48 +0200
El 28/9/24 a las 12:31, Jeff escribió:
> I can't reproduce this. The tests pass elsewhere in chroots:
>
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/gscan2pdf.html
>
> What is different about your setup?
Most probably, the fact that he's building the package as root.
This is from the build log:
LOGNAME=root
Hi наб, you should probably say this in your reports.
Thanks.
retitle 1082815 analizo: FTBFS (tests fail) on sid if building as root
retitle 1082810 gscan2pdf: deadlock in tests if building as root
# 1082804 is already ledgersmb: FTBFS (test failure) if building as root
thanks
On Sat, Sep 28, 2024 at 01:10:48PM +0200, Santiago Vila wrote:
> El 28/9/24 a las 12:31, Jeff escribió:
> > What is different about your setup?
> LOGNAME=root
>
> Hi наб, you should probably say this in your reports.
Yeah, I didn't so much as consider running as root being the issue.
Best,
Changed Bug title to 'gscan2pdf: deadlock in tests if building as root' from 'gscan2pdf: deadlock in tests'.
Request was from наб <[email protected]>
to [email protected].
(Sat, 28 Sep 2024 14:43:34 GMT) (full text, mbox, link).
On 28/09/2024 16:41, наб wrote:
> retitle 1082815 analizo: FTBFS (tests fail) on sid if building as root
> retitle 1082810 gscan2pdf: deadlock in tests if building as root
> # 1082804 is already ledgersmb: FTBFS (test failure) if building as root
> thanks
>
> On Sat, Sep 28, 2024 at 01:10:48PM +0200, Santiago Vila wrote:
>> El 28/9/24 a las 12:31, Jeff escribió:
>>> What is different about your setup?
>> LOGNAME=root
If that is the case, I wonder if this really should be severity=serious.
Maybe severity=normal?
Acknowledgement sent
to Petter Reinholdtsen <[email protected]>:
Extra info received and forwarded to list. Copy sent to Jeffrey Ratcliffe <[email protected]>.
(Sat, 05 Oct 2024 07:45:01 GMT) (full text, mbox, link).
Subject: Re: Bug#1082810: gscan2pdf: deadlock in tests if building as root
Date: Sat, 5 Oct 2024 09:42:25 +0200
If the build should not be done as root, perhaps debian/rules should
be extended with a call to ! dh_testroot, to ensure the build fail
with root privileges?
--
Happy hacking
Petter Reinholdtsen
Debbugs is free software and licensed under the terms of the GNU General
Public License version 2. The current version can be obtained
from https://bugs.debian.org/debbugs-source/.