Package: perl Version: 5.28.1-6 Since enabling the reprotest pipeline on salsa.d.o, we're seeing persistent test failures on that:
dist/threads-shared/t/waithires ................................ FAILED at test 7 This seems to be due to faketime(1). On current sid / amd64 (perl 5.28.1-6): % PERL_CORE=1 faketime yesterday perl t/waithires.t 1..63 ok 1 - Loaded ok 2 - Shared synchronization tests preparation ok 3 - cond_timedwait [simple]: obtained initial lock ok 4 - cond_timedwait [simple]: child before lock ok 5 - cond_timedwait [simple]: child obtained lock ok 6 - cond_timedwait [simple]: child signalled condition not ok 7 - test_type: child exceeded time # Failed test at line 181 # child time = 6.81877136230469e-05 not ok 8 - test_type: parent exceeded time # Failed test at line 183 # parent time = 1.50203704833984e-05 [...] The faketime '-m' (multi-threading) option makes test 7 just time out for me. This is not particularly surprising; faketime is known to cause such issues. See for instance https://alioth-lists.debian.net/pipermail/reproducible-builds/Week-of-Mon-20140922/000464.html I expect the best "fix" for this is to disable faketime usage in the reprotest pipeline, if possible. -- Niko Tyni nt...@debian.org
