This is an auto-filed bug because dev-lang/elixir fails tests. The issue was originally discovered on amd64, but it may be reproducible on other arches as well. If you think that a different summary clarifies the issue better, feel free to change it. Attached build log and emerge --info. NOTE: If you need further logs, feel free to ask.
Created attachment 643700 [details] build.log build log and emerge --info
Single test failure: """ ... 1) test compiles code (Kernel.CLI.CompileTest) test/elixir/kernel/cli_test.exs:120 Assertion with == failed code: assert :code.which(CompileSample) |> List.to_string() == Path.expand(context[:beam_file_path]) left: "" right: "/var/tmp/portage/dev-lang/elixir-1.10.3/work/elixir-1.10.3/lib/elixir/tmp/beams/120/Elixir.CompileSample.beam" stacktrace: test/elixir/kernel/cli_test.exs:126: (test) ... Finished in 30.4 seconds (14.6s on load, 15.7s on tests) 1574 doctests, 3401 tests, 1 failure, 8 excluded """ Is it on dev-lang/erlang-22.3.2? Is it consistently reproducible?
(In reply to Sergei Trofimovich from comment #2) > Is it on dev-lang/erlang-22.3.2? > > Is it consistently reproducible? Sorry for the delay, yes in both cases
Reproduced locally in -match=native chroot. Will poke at the details.
For some reason test failure happens only on the first attempt of building elixir when I enter the chroot, but not on subsequent ones. As if something is left after a build that prevents failure (maybe there is .beam cache somewhere?).
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=a6b90c93b8cc5539bf35db34fe472c72c7175fc1 commit a6b90c93b8cc5539bf35db34fe472c72c7175fc1 Author: Sergei Trofimovich <slyfox@gentoo.org> AuthorDate: 2020-06-17 07:29:11 +0000 Commit: Sergei Trofimovich <slyfox@gentoo.org> CommitDate: 2020-06-17 07:29:20 +0000 dev-lang/elixir: take src_test and src_install more verbose Print exact command to eare rerunning subset of test suite. Bug: https://bugs.gentoo.org/727344 Package-Manager: Portage-2.3.101, Repoman-2.3.22 Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org> dev-lang/elixir/elixir-1.10.3.ebuild | 5 +---- dev-lang/elixir/files/elixir-1.10.3-no-Q.patch | 12 ++++++++++++ 2 files changed, 13 insertions(+), 4 deletions(-)
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=7c2a12d218faca7eec9e62f0b9491a5832b80d2b commit 7c2a12d218faca7eec9e62f0b9491a5832b80d2b Author: Sergei Trofimovich <slyfox@gentoo.org> AuthorDate: 2020-06-18 22:09:21 +0000 Commit: Sergei Trofimovich <slyfox@gentoo.org> CommitDate: 2020-06-18 22:09:21 +0000 dev-lang/elixir: start erlang via 'empd -daemon' Original bug exposes some kind of a race in erlang process. It's not clear if 'epmd & exit' an unsafe construct or not. I'm assuming it's not safe. Let's see if it helps. Reported-by: Agostino Sarubbo Bug: https://bugs.gentoo.org/727344 Package-Manager: Portage-2.3.101, Repoman-2.3.22 Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org> dev-lang/elixir/elixir-1.10.3.ebuild | 1 + dev-lang/elixir/files/elixir-1.10.3-epmd-daemon.patch | 13 +++++++++++++ 2 files changed, 14 insertions(+)
With #comment7 I hope to catch race between partially started daemon and it's users. Chances are it's not the real problem. But maybe? Can you check it you can reproduce the failure for you?
excellent, it works for me..