[Pkg-zsh-devel] zsh FTBFS under Salsa CI's reprotest: 'KEY_EVENT' undeclared / blhc CI check disabled / bullseye freeze: zsh is considered a key package

[Axel Beckert]

Hi again,

Axel Beckert wrote:
> Daniel Shahaf wrote:
> > Axel Beckert wrote on Mon, 12 Oct 2020 15:45 +00:00:
> > > 1. zsh FTBFS inside reprotest on Salsa CI:
> > > https://salsa.debian.org/debian/zsh/-/jobs/1012123#L2931
> > > In file included from ../../../Src/Modules/curses.c:210:
> > > ./curses_keys.h:93:15: error: ‘KEY_EVENT’ undeclared here (not in a 
> > > function); did you mean ‘KEY_RESET’?
> > >    93 |     {"EVENT", KEY_EVENT},
> > >       |               ^~~~~~~~~
> > >       |               KEY_RESET
> > > make[5]: *** [Makefile:240: curses..o] Error 1
> […]
> > Looks just like the issue fixed upstream in zsh-5.8-251-gc6a851636.
> 
> Indeed, that looks a lot like it, thanks!
[...]
> I'd cherry-pick that commit, too, to fix this. But first I'd like to
> understand this. Which I currently don't.

Tried it in the feature branch abe/reprotest-cherry-pick and it indeed
fixed this issue. Very strange.

Now we've got another issue inside reprotest which I can't reproduce
locally either:

  52 successful test scripts, 2 failures, 0 skipped

(from https://salsa.debian.org/debian/zsh/-/jobs/1066149)

The failed ones seem to be these (grepping for "failed" in the build
log):

Test ../../Test/D02glob.ztst failed: test was expected to fail, but passed.
../../Test/D02glob.ztst: test failed.
Test ../../Test/P01privileged.ztst failed: error output differs from expected as shown above for:
../../Test/P01privileged.ztst: test failed.


Details:

Running test: unreadable directories can be globbed (users/24619, users/24626)
Test ../../Test/D02glob.ztst failed: test was expected to fail, but passed.
Was testing: unreadable directories can be globbed (users/24619, users/24626)

This might be an test environment issue, maybe some kernel features
not available or so? Not sure.


Running test: PRIVILEGED automatically enabled when RUID != EUID
--- /tmp/zsh.ztst.15097/ztst.err        2020-10-12 05:43:42.052156929 -1200
+++ /tmp/zsh.ztst.15097/ztst.terr       2020-10-12 05:43:42.064159863 -1200
@@ -0,0 +1 @@
+ERROR: ld.so: object 'libeatmydata.so' from LD_PRELOAD cannot be preloaded (cannot open shared object file): ignored.
Test ../../Test/P01privileged.ztst failed: error output differs from expected as shown above for:
  re_zsh $ruid $ruid -1 -1 'echo $UID/$EUID $options[privileged]'
  re_zsh $euid $euid -1 -1 'echo $UID/$EUID $options[privileged]'
  re_zsh $ruid $euid -1 -1 'echo $UID/$EUID $options[privileged]'
Was testing: PRIVILEGED automatically enabled when RUID != EUID
../../Test/P01privileged.ztst: test failed.

This issue is clearly caused by reprotest running under eatmydata.
Files an issue to get an opt in for not using eatmydata here:
https://salsa.debian.org/salsa-ci-team/pipeline/-/issues/181

And at least for the second issue it's now also clear why I can't
reproduce it: My pbuilder chroots (which reprotest uses) always have
eatmydata installed on the inside, too. :-)


P.S.: I wonder if the other, more obscure test suite failure — which
only happens inside the reprotest check on Salsa CI, too — is also
related to eatmydata. Unfortunately the debug output is not as verbose
as in P01privileged.ztst. Like maybe like it is just checked for some
empty output, but it actually contains "ERROR: ld.so: object
'libeatmydata.so' from LD_PRELOAD cannot be preloaded (cannot open
shared object file): ignored." and the test suite thinks the test
failed because there was any output from some globbing routine.

P.P.S.: I also wonder if that missing KEY_EVENT thingy could also be
triggered by not being able to load some arbitrary .so file.

		Regards, Axel
-- 
 ,''`.  |  Axel Beckert <abe at debian.org>, https://people.debian.org/~abe/
: :' :  |  Debian Developer, ftp.ch.debian.org Admin
`. `'   |  4096R: 2517 B724 C5F6 CA99 5329  6E61 2FF9 CD59 6126 16B5
  `-    |  1024D: F067 EA27 26B9 C3FC 1486  202E C09E 1D89 9593 0EDE