Romain Francoise wrote:
> While doing some testing I found the following crash. It looks
> related to the copy/output mode rewrite:
> It can easily be reproduced by maximizing a terminal which contains
> a tmux window showing `list-keys' output.
Thanks, I believe I just fixed it in SourceForge.
Micah Cowan writes:
> Thanks, I believe I just fixed it in SourceForge.
Yes, your change fixes the issue for me, thanks.
> It looks to me like it predates the copy/output-mode merge; but of
> course it wouldn't have shown up for "list-keys" prior to the
> merge. I think it would still have cras
I wonder if to make it nicer we could unify the signal set functions
into one set_signals(void (*)(int, short, void *)). And similarly could
have one clear_signals function.
All the processes could ignore SIGINT, SIGPIPE, SIGUSR2, SIGTSTP,
SIGHUP, and all catch SIGCONT, SIGTERM, SIGWINCH, SIGCHLD,
Hmm. Not sure what is going on here then.
sendmsg is failing with EINVAL, I know that there is weirdness because
NetBSD have made CMSG_SPACE() non-constant but I thought gcc built
correctly anyway.
What version of NetBSD is this?
Can you try this, on the off-chance?
Index: compat/imsg.h
==
Thanks Nicholas for having a look at this.
Nicholas Marriott (Wednesday 28 April 2010, 19:08):
>
> What version of NetBSD is this?
$ uname -srm
NetBSD 4.0.1_PATCH alpha
(There are also AMD64 machines: happens the same.)
Probably something weird happened during the server move, as tmux
(from pk
Nicholas Marriott writes:
> I wonder if to make it nicer we could unify the signal set
> functions into one set_signals(void (*)(int, short, void *)). And
> similarly could have one clear_signals function.
> All the processes could ignore SIGINT, SIGPIPE, SIGUSR2, SIGTSTP,
> SIGHUP, and all catc
