https://blogs.gentoo.org/ago/2020/07/04/gentoo-tinderbox/ Issue: dev-libs/date-3.0.3 fails tests. Discovered on: amd64 (internal ref: gcc15_tinderbox) System: GCC-15-SYSTEM (https://wiki.gentoo.org/wiki/Project:Tinderbox/Common_Issues_Helper#GCC-15) Info about the issue: https://wiki.gentoo.org/wiki/Project:Tinderbox/Common_Issues_Helper#CF0015
Created attachment 916170 [details] build.log build log and emerge --info
Error(s) that match a know pattern: ./testit: line 102: 81 Aborted ./$TEST_EXE ./testit: line 102: 93 Aborted ./$TEST_EXE ./testit: line 102: 105 Aborted ./$TEST_EXE ./testit: line 102: 127 Aborted ./$TEST_EXE ./testit: line 102: 858 Aborted ./$TEST_EXE ./testit: line 102: 874 Aborted ./$TEST_EXE ./testit: line 102: 931 Aborted ./$TEST_EXE ./testit: line 102: 943 Aborted ./$TEST_EXE
maybe related to gcc-15 - needs inspection
I *suspect* this one is a general test setup problem (but you're right to double check): terminate called after throwing an instance of 'std::runtime_error' what(): Timezone database not found at "/var/tmp/portage/dev-libs/date-3.0.3/homedir/Downloads/tzdata" I will verify though to be sure
Well this is weird. I have also successfully run the tests on my machine with gcc-15.0.0_pre20250105 (binhost provided the package :D) so I do not think it is a gcc-15 issue. Same with gcc-14-14.2.1_p20241221, which I used while writing the ebuild The error is indeed caused by the tzdata not being found: Timezone database not found at "/var/tmp/portage/dev-libs/date-3.0.3/homedir/Downloads/tzdata" That seems to be the download "~/Downloads/tzdata". However since we are setting # date: USE_SYSTEM_TZ_DB ON so it should not look there: "When compiling with USE_OS_TZDB == 1 INSTALL can not be used." https://howardhinnant.github.io/date/tz.html#Installation This might have something to do with this here: https://github.com/HowardHinnant/date/blob/8a1102fd08f3afff394512b453566cb971319917/CMakeLists.txt#L149 Here, and in the build logs you can see that `INSTALL=.`, which makes me wonder why the project is using the default path... This seems impossible, I am unable to reproduce this and yet it happens.