[ruby-core:88491] Re: [ruby-cvs:71466] k0kubun:r64374 (trunk): test_function.rb: skip running test

[ruby-core:88491] Re: [ruby-cvs:71466] k0kubun:r64374 (trunk): test_function.rb: skip running test

[Eric Wong]

k0kubun@ruby-lang.org wrote:
>   Log:
>     test_function.rb: skip running test
>     
>     that times out with test-all w/ --jit-wait.
>     
>     I'm running the following command on Wecker CI everyday:
>     ```
>     make test-all TESTOPTS="--color=never --job-status=normal" RUN_OPTS="--disable-gems --jit-wait --jit-warnings" RUBY_FORCE_TEST_JIT=1
>     ```
>     
>     By running yesterday's all commits, r64354 ran successfully but r64355
>     didn't. So the test should be fixed to run with --jit-wait at first.

Fwiw, I don't think test_nogvl_poll can ever work reliably
without EINTR when spawning processes for JIT.

Maybe the test is unnecessary, nowadays; because I only wrote it
to prove fiddle could release GVL for parallelism.  Fwiw, I
don't think we have parallelism test for most IO methods,
either.

[ruby-core:88495] Re: [ruby-cvs:71466] k0kubun:r64374 (trunk): test_function.rb: skip running test

[Takashi Kokubun]

I see. Please remove the test if the test is unnecessary.

The skip commit didn't let the --jit-wait CI succeed anyway. So I'm
searching which commit of yesterday caused the failure. The cause
could be overall slowdown or hanging on one test, but not sure yet
because it's silently finished by timeout.
2018年8月15日(水) 14:26 Eric Wong <normalperson@yhbt.net>:
>
> k0kubun@ruby-lang.org wrote:
> >   Log:
> >     test_function.rb: skip running test
> >
> >     that times out with test-all w/ --jit-wait.
> >
> >     I'm running the following command on Wecker CI everyday:
> >     ```
> >     make test-all TESTOPTS="--color=never --job-status=normal" RUN_OPTS="--disable-gems --jit-wait --jit-warnings" RUBY_FORCE_TEST_JIT=1
> >     ```
> >
> >     By running yesterday's all commits, r64354 ran successfully but r64355
> >     didn't. So the test should be fixed to run with --jit-wait at first.
>
> Fwiw, I don't think test_nogvl_poll can ever work reliably
> without EINTR when spawning processes for JIT.
>
> Maybe the test is unnecessary, nowadays; because I only wrote it
> to prove fiddle could release GVL for parallelism.  Fwiw, I
> don't think we have parallelism test for most IO methods,
> either.

[ruby-core:88496] Re: [ruby-cvs:71466] k0kubun:r64374 (trunk): test_function.rb: skip running test

[Eric Wong]

Takashi Kokubun <takashikkbn@gmail.com> wrote:
> The skip commit didn't let the --jit-wait CI succeed anyway. So I'm
> searching which commit of yesterday caused the failure. The cause
> could be overall slowdown or hanging on one test, but not sure yet
> because it's silently finished by timeout.

I guess --jit-wait failure is my fault :x   Will investigate

[ruby-core:88498] Re: [ruby-cvs:71466] k0kubun:r64374 (trunk): test_function.rb: skip running test

[Eric Wong]

Eric Wong <normalperson@yhbt.net> wrote:
> Takashi Kokubun <takashikkbn@gmail.com> wrote:
> > The skip commit didn't let the --jit-wait CI succeed anyway. So I'm
> > searching which commit of yesterday caused the failure. The cause
> > could be overall slowdown or hanging on one test, but not sure yet
> > because it's silently finished by timeout.
> 
> I guess --jit-wait failure is my fault :x   Will investigate

Once in a blue moon, I get this, not sure why, yet:

IMAPTest#test_exception_during_idle:
Test::Unit::ProxyError: Attempt to unlock a mutex which is locked by another thread
    /foo/ruby/lib/monitor.rb:200:in `unlock'
    /foo/ruby/lib/monitor.rb:200:in `mon_exit'
    /foo/ruby/lib/monitor.rb:228:in `ensure in mon_synchronize'
    /foo/ruby/lib/monitor.rb:228:in `mon_synchronize'
    /foo/ruby/lib/net/imap.rb:957:in `idle'
    /foo/ruby/test/net/imap/test_imap.rb:246:in `test_exception_during_idle'

Finished tests in 29.648353s, 0.0337 tests/s, 0.0000 assertions/s.
1 tests, 0 assertions, 1 failures, 1 errors, 0 skips

[ruby-core:88522] Re: [ruby-cvs:71466] k0kubun:r64374 (trunk): test_function.rb: skip running test

[Takashi Kokubun]

Thanks to confirm.

Not sure why, but recent 2 daily builds (r64397, r64417) worked fine.
2018年8月16日(木) 8:16 Eric Wong <normalperson@yhbt.net>:
>
> Eric Wong <normalperson@yhbt.net> wrote:
> > Takashi Kokubun <takashikkbn@gmail.com> wrote:
> > > The skip commit didn't let the --jit-wait CI succeed anyway. So I'm
> > > searching which commit of yesterday caused the failure. The cause
> > > could be overall slowdown or hanging on one test, but not sure yet
> > > because it's silently finished by timeout.
> >
> > I guess --jit-wait failure is my fault :x   Will investigate
>
> Once in a blue moon, I get this, not sure why, yet:
>
> IMAPTest#test_exception_during_idle:
> Test::Unit::ProxyError: Attempt to unlock a mutex which is locked by another thread
>     /foo/ruby/lib/monitor.rb:200:in `unlock'
>     /foo/ruby/lib/monitor.rb:200:in `mon_exit'
>     /foo/ruby/lib/monitor.rb:228:in `ensure in mon_synchronize'
>     /foo/ruby/lib/monitor.rb:228:in `mon_synchronize'
>     /foo/ruby/lib/net/imap.rb:957:in `idle'
>     /foo/ruby/test/net/imap/test_imap.rb:246:in `test_exception_during_idle'
>
> Finished tests in 29.648353s, 0.0337 tests/s, 0.0000 assertions/s.
> 1 tests, 0 assertions, 1 failures, 1 errors, 0 skips