Created attachment 318716 [details] emerge fail log dev-libs/popt fails tests with clang-3.1-r4
This bug does not show up when clang-3.1-r5 is used, 3.1-r4 makes the tests fail.
(In reply to comment #1) > This bug does not show up when clang-3.1-r5 is used, 3.1-r4 makes the tests > fail. Er, weird/semi-random. -r5 practically only adds additional targets, unless I'm missing something...
I agree that is weird or semi-random given the failure output, however, in order to be sure, I checked again inside the VM and on my host system too. In both cases when clang-r4 is used tests fail with error almost identical to the emerge log already attached (my host now fails at test 59 with clang-r4 instead of 58). When clang-r5 is used, there is no error on either system.
(In reply to comment #3) > I agree that is weird or semi-random given the failure output, however, in > order to be sure, I checked again inside the VM and on my host system too. > In both cases when clang-r4 is used tests fail with error almost identical > to the emerge log already attached (my host now fails at test 59 with > clang-r4 instead of 58). When clang-r5 is used, there is no error on either > system. Well, for me -r5 fails on test 58 ;).
Uhm... retested again on host system, bootstrapped again clang and llvm with themselves twice and did testing, this part went fine - no failures. -r4 fails tests. Tested popt with clang-3.1-r5, compiles and tests pass. Now the interesting/funny part: tested popt with -r5 outside of portage - manually reproducing the steps of portage in a folder... and it fails at test 58. I don't know what/where to look now :-/ Consider that I use an ~amd64 system for testing, if there's any further info I can provide I will. I use sane C/LDFLAGS btw.
Tests pass fine here on ~amd64 with clang 3.4, marking fixed then :) Don't hesitate to reopen if you still have the problem