| Summary: | glib-2.0: glib/codegen.py.test ptest failure | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [QA/Testing] Package Testing (ptest) | Reporter: | Luca Ceresoli <luca.ceresoli> | ||||
| Component: | ptest | Assignee: | Unassigned <unassigned> | ||||
| Status: | RESOLVED DUPLICATE | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | Medium | CC: | alex.kanavin, alexandre.belloni, joaomarcos.costa, JPEWhacker, randy.macleod, richard.purdie, ross.burton, tgamblin | ||||
| Version: | unspecified | ||||||
| Target Milestone: | 5.3 | ||||||
| Hardware: | x86 | ||||||
| OS: | Multiple | ||||||
| Whiteboard: | AB-INT | ||||||
| OS type for building Yocto: | --- | Type of Regression: | --- | ||||
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |||||
| Attachments: |
|
||||||
|
Description
Luca Ceresoli
2023-06-23 12:29:52 UTC
Moved the builddir to /home/pokybuild/qemuarm64-ptest-5071-bug15154-glib-codegen on the worker (ubuntu2004-arm-1) Created attachment 4958 [details]
ptest log
My horrible fear is that this test case is notable for writing a lot to the console, so the recent changes to ptest-runner have introduced another race/crash. Seems to happen about 50% of the time when I'm running the ptest in a local build of core-image-ptest-glib-2.0:
for i in {1..100}; do ptest-runner glib-2.0 2>&1 | tee -a mountdir/ptest-$i.log; done
...
root@qemux86-64:~# grep -r "FAIL: 1" mountdir/ | wc -l
47
The test does indeed write a lot of text to the console (I captured 22640 lines in one test result, and 22734 in another). When it fails, I see output like these examples:
# static void
# org_project_bar_frobnicator_skeleton_iface_init (OrgProjectBarFrobnicatorIface *iface G_GNUC_UNUSED)
# {
# }
#
# /**
# * org_project_bar_frobnicator_skeleton_new:
# *
# * Creates a skeleton object for the D-Bus interface <link linkendFAIL: glib/codegen.py.test (Child process exited with code 1)
Running test: glib/markup-parse.test
TAP version 13
# random seed: R02S524b5572c18b7a8b31b4e94a3251b322
1..71
# Start of markup tests
...
# * The calling thread is blocked until a reply is received.
# *
# * See com_acme_coyote_proxy_new() for the asynchronous version of this constructor.
# *
# * Returns: (transfer full) (type ComAcmeCoyoteProxy): The constructed proxy object or %NULL if @error is set.
# */
# ComAcmeCoyote *
# com_acme_coyote_proxy_new_sync (
# GDBusConnection *connection,
# GDBusProxyFlags flags,
# const gchar *name,
# const gchar *object_path,
# GCancellable *cancellable,
# GError **error)
# {
# GInitable *ret;
# reFAIL: glib/codegen.py.test (Child process exited with code 1)
Running test: glib/gobject-private.test
TAP version 13
# random seed: R02Sd5f5e2618a8b992c5b1ce1ce197541ec
1..3
# Start of private tests
I can reliably reproduce this, and I can confirm that reverting ptest-runner changes make the issue disappear :( The ptest change changed the operations from blocking to non-blocking which probably has something to do with this... Alex, if you could find time to work on this soonish, that would be great. YP bug review I reworked this to eliminate the threads and race conditions, which should fix the problem if it's in ptest-runner: https://lists.yoctoproject.org/g/yocto/message/60479 Thanks Joshua, this does seem to eliminate the fails here. https://autobuilder.yoctoproject.org/typhoon/#/builders/81/builds/5231/steps/13/logs/stdio qemux86-64-ptest ubuntu1804-ty-3 https://autobuilder.yoctoproject.org/typhoon/#/builders/82/builds/5091/steps/12/logs/stdio qemuarm64-ptest ubuntu2204-arm-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/82/builds/5063/steps/12/logs/stdio qemuarm64-ptest ubuntu2004-arm-1 https://autobuilder.yoctoproject.org/typhoon/#/builders/81/builds/5256/steps/12/logs/stdio qemux86-64-ptest fedora38-ty-3 https://autobuilder.yoctoproject.org/typhoon/#/builders/81/builds/5239/steps/12/logs/stdio qemux86-64-ptest ubuntu2004-ty-1 Almost certainly caused by https://git.yoctoproject.org/poky/commit/?h=master-next&id=191bb1e95fda12067df4c707e5409ecf5c8e370f I sent a revert. NVM, these newly added links are all for old failures. Fixed by: https://git.openembedded.org/openembedded-core/commit/?id=8de47e5f3837a9c87c3cbf8dc45f9e90110eda1e It happened again, this time in riscv64. https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/184/steps/12/logs/stdio Failed ptests:{'glib-2.0': ['glib/codegen.py.test']} qemuriscv64-ptest alma9-vk-2 I see the same pattern of enormous amounts of text as in the previous logs. Reopening. Now happening on RISCV only so far. Why? This may be a different failure that is specific to risc-v. Can you link to the logs where you see the enormous output? https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/267/steps/12/logs/stdio Failed ptests:{'glib-2.0': ['glib/codegen.py.test', 'glib/gi-compile-repository.py.test']} qemuriscv64-ptest stream9-vk-1 https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/355/steps/13/logs/stdio qemuriscv64-ptest opensuse156-vk-1 The logs: https://valkyrie.yocto.io/pub/non-release/20250815-8/testresults/qemuriscv64-ptest/glib-2.0.log https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/379/steps/12/logs/stdio qemuriscv64-ptest fedora42-vk-1 Duplicate of 15891. See there for new discussion. *** This bug has been marked as a duplicate of bug 15891 *** |