Greg.mpls / gmail.com wrote:
> FYI, both mswin & mingw are building.  I don't believe mswin
> runs JIT tests, but when I run them locally on MinGW, I just
> get a frozen flashing cursor.  No output at all.

Right, the fallback code for non-SIGCHLD doesn't benefit MJIT.
However, maybe polling will work, since timer-thread in
thread_win32.c is always polling anyways...

> I'm happy to help with the problem, but I'm not quite sure
> what to do...

Can you try the following patch?  It is a newer version of
[ruby-core:87682] and needs r63793[*]

https://80x24.org/spew/20180629224058.20652-1-e / 80x24.org/raw


[*] ("signal.c: use ATOMIC_EXCHANGE for sigchld_hit")
    r63793 may solve OSX problem with CI machines, as I suspect
    pthread_sigmask is broken on that platform.  Still waiting
    on results from rubyci.org.


Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>