Summary: | dev-python/pygobject-3.22.0 fails tests | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Paolo Pedroni <paolo.pedroni> |
Component: | Current packages | Assignee: | Gentoo Linux Gnome Desktop Team <gnome> |
Status: | RESOLVED OBSOLETE | ||
Severity: | normal | CC: | whissi |
Priority: | Normal | Keywords: | TESTFAILURE |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
pygobject-3.22.0:20170810-121052.log.gz
Remaining part of build log. build.log on x86 |
Description
Paolo Pedroni
2017-08-10 12:32:20 UTC
Created attachment 488502 [details]
Remaining part of build log.
This is the rest of the build log (taken from the console).
Is it the only application having unittest problems ? I am a bit worried about your CFLAGS. (In reply to Gilles Dartiguelongue from comment #2) > Is it the only application having unittest problems ? I am a bit worried > about your CFLAGS. Thanks for your feedback. What does concern you most? - "-march=sandybridge -msahf --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=8192" is just unbundling -march=native to allow distcc compilation - "-flto=8 -fuse-linker-plugin" is for link time optimization - "-floop-nest-optimize -ftree-loop-distribution -fgraphite-identity -ftree-vectorize" are the least "offensive" graphite optimizations I could find. I will try without LTO and graphite (and without both) and I'll report ASAP. You are partly right. Removing LTO optimizations fixes the "test_add_watch_no_data" failure, but the "test_python_calls_sync" stubbornly remains even disabling both LTO and graphite optimizations. The log still gets truncated, too. Created attachment 511000 [details]
build.log on x86
======================================================================
FAIL: test_cairo_font_options (test_cairo.TestPango)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/var/tmp/portage/dev-python/pygobject-3.22.0/work/pygobject-3.22.0-python2_7/tests/test_cairo.py", line 75, in test_cairo_font_options
self.assertEqual(type(font_opts.get_subpixel_order()), int)
AssertionError: <type 'cairo.SubpixelOrder'> != <type 'int'>
----------------------------------------------------------------------
Ran 1187 tests in 9.571s
FAILED (failures=1, skipped=4, expected failures=12)
Looks like the cairo.SubpixelOrder part was workarounded in 3.24 ebuilds via a sed. I'm going to remove most of these test workarounds in a 3.28+ bump, as things have changed there and seem fixed to me. I think the SubpixelOrder stuff was mainly with older pycairo, plus now the test uses isinstance() instead of type() in there. |