Re: Deadlock when receiving kill-signal from child process

[Mathias Fredriksson <mafredri@xxxxxxxxx>]

On Wed, Aug 5, 2015 at 9:53 AM, Bart Schaefer wrote:
}
} Mathias, if you could apply that patch and try again to reproduce the
} deadlock, it might tell us something.

Of course, here's the output (as I see it in the terminal) with the
applied patch:

TRAPUSR1:1: no job table entry for pid 61571
TRAPUSR1:1: no job table entry for pid 61570
.TRAPUSR1:1: no job table entry for pid 61575
.TRAPUSR1:1: no job table entry for pid 61573
../deadlock.zsh:19: no job table entry for pid 61577
TRAPUSR1:1: no job table entry for pid 61581
.TRAPUSR1:1: no job table entry for pid 61595
TRAPUSR1:1: no job table entry for pid 61582
../deadlock.zsh:20: no job table entry for pid 61585
./deadlock.zsh:20: no job table entry for pid 61589
./deadlock.zsh:20: no job table entry for pid 61590

Furthermore, here is the gdb output from the last script I posted and
zsh built with --enable-zsh-debug:

#0  0x00007fff8abfe72a in __sigsuspend ()
#1  0x0000000108fc05df in signal_suspend ()
#2  0x0000000108f97b05 in zwaitjob ()
#3  0x0000000108f97958 in waitjobs ()
#4  0x0000000108f7a90f in execpline ()
#5  0x0000000108f79c42 in execlist ()
#6  0x0000000108f79a6a in execode ()
#7  0x0000000108f920bb in loop ()
#8  0x0000000108f94d0c in zsh_main ()
#9  0x00007fff8610c5c9 in start ()