bug#45658: Infinite loop in run loop on macOS

[Alan Third]

On Wed, Oct 20, 2021 at 10:02:23AM +0800, Po Lu wrote:
> Alan Third <alan@idiocy.org> writes:
> 
> > ns_send_appdefined somewhere. Can you switch to the thread that's
> > running fd_handler and see what it's doing? There are only two
> > branches, so it should be easy to work out which one is in use.
> 
> >>     frame #8: 0x00007fff2d14c039 AppKit`_DPSNextEvent + 883
> >>     frame #9: 0x00007fff2d14a880 AppKit`-[NSApplication(NSEvent)
> >> _nextEventMatchingEventMask:untilDate:inMode:dequeue:] + 1352
> >>     frame #10: 0x00007fff2d13c58e AppKit`-[NSApplication run] + 658
> >>     frame #11: 0x000000010036c5fa Emacs`-[EmacsApp
> >> run](self=0x000000010525e180, _cmd="run") at nsterm.m:5602:9
> >>     frame #12: 0x000000010036a8eb Emacs`ns_select(nfds=30,
> >> readfds=0x00007ffeefbfd1f0, writefds=0x00007ffeefbfd170,
> >> exceptfds=0x0000000000000000, timeout=0x00007ffeefbfd148,
> >> sigmask=0x0000000000000000) at nsterm.m:4690:3
> 
> Does it also make sense for this to occur on GNUstep?  I have just seen
> a very similar bug there, but I lost the backtrace :(

It doesn't make any sense to me for it to happen anywhere. ;)

If it happens on macOS I don't see why it shouldn't happen on GNUstep,
unless it's a bug in Cocoa and not in GNUstep. I wasn't able to
replicate the crash and I don't have any real clue what was causing it
for Jimmy.
-- 
Alan Third