autopkgtests are failing on valgrind

[Iain Lane]

This happens now

autopkgtest [07:42:07]: test make-check: -----------------------]
autopkgtest [07:42:08]: test make-check: - - - - - - - - - - results - - - - - - - - - -
make-check FAIL stderr: ==9257== 8 bytes in 1 blocks are definitely lost in loss record 67 of 368
autopkgtest [07:42:08]: test make-check: - - - - - - - - - - stderr - - - - - - - - - -
==9257== 8 bytes in 1 blocks are definitely lost in loss record 67 of 368
==9257== at 0x4C2FB0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==9257== by 0x18EF2D: xmalloc (in /bin/bash)
==9257== by 0x18898A: set_default_locale (in /bin/bash)
==9257== by 0x137E66: main (in /bin/bash)
==9257==
==9505== 8 bytes in 1 blocks are definitely lost in loss record 67 of 368
==9505== at 0x4C2FB0F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==9505== by 0x18EF2D: xmalloc (in /bin/bash)
==9505== by 0x18898A: set_default_locale (in /bin/bash)
==9505== by 0x137E66: main (in /bin/bash)
==9505==
autopkgtest [07:42:08]: @@@@@@@@@@@@@@@@@@@@ summary
make-check FAIL stderr: ==9257== 8 bytes in 1 blocks are definitely lost in loss record 67 of 368

From http://autopkgtest.ubuntu.com/packages/c/colord/bionic/amd64 it looks like it's cause by the new glib2.0, but I just tried on my machine without that package and it still happened.

Could you have a look please? Looks like it might not actually be colord's fault here, but maybe we want to reconsider running under valgrind at least for the autopktests?

[Jeremy Bícha (jbicha)]

autopkgtests pass now in cosmic and in disco (well once my 1.4.3-4 upload autosyncs in a few hours).

Do you think it's worth doing a bionic SRU if we can identify what fixed the autopkgtests in newer versions?

[Jeremy Bícha (jbicha)]

I asked bdmurray in #ubuntu-devel and he'd rather not have an SRU just to fix an autopkgtest like this.

I'm closing since this issue is fixed in the current development release.