Re: usr1-signal, usr2-signal, etc.

[Kim F. Storm]

YAMAMOTO Mitsuharu <address@hidden> writes:

>>>>>> On Mon, 18 Dec 2006 11:38:21 -0500, Chong Yidong <address@hidden> said:
>
>> address@hidden (Kim F. Storm) writes:
>>> So I would rather redefine them to single events (named sigusr1 and
>>> sigusr2 as those names are already known to the signal-process
>>> function), and handle them through special-event-map as you
>>> suggested.
>
>> In that case, is there a rationale for the rather large changes
>> checked into emacs.c, keyboard.c, and process.c over the last two
>> weeks?  Should they not be reverted?
>
> Do the "large changes" you think should be reverted include my
> 2006-12-14 changes?  These changes are not for the event format but
> for fixing a problem that has existed for a long time.  That is, about
> calling a non-reentrant function in a signal handler context while it
> is executed in a normal context.

Yes.  Those changes are very important on their own.

It still think that handling signals through special-event-map is
the proper (or the least problematic) thing to do.  So unless I
hear any objections, I'll make that change later this week.

-- 
Kim F. Storm <address@hidden> http://www.cua.dk