Re: [emacs-announce] Emacs pretest 25.0.93

[Eli Zaretskii <eliz@gnu.org>]

> From: Achim Gratz <Stromeko@nexgo.de>
> Date: Sat, 23 Apr 2016 15:29:42 +0200
> 
> Eli Zaretskii writes:
> > Run the faulting command under a debugger, put a breakpoint in
> > Fsignal, and show the backtrace.  That might produce enough hints to
> > find out what is failing and why.
> 
> It doesn't really fault, at least not at the lisp level.

Fsignal is on the C level, so a breakpoint set there in GDB will catch
any error.

> There actually is an auto-save file that seems to have all the
> extracted loaddefs in, but somehow it then doesn't finish.  For the
> autoloads entries there are lots of differences like these (I'm
> comparing with the installed version of my last build):
> 
> -;;;### (autoloads nil "cal-iso" "cal-iso.el" t)
> +;;;### (autoloads nil "cal-iso" "cal-iso.el" (22225 49314 854982
> +;;;;;;  791000))

AFAIU, these differences are meaningless.  But I have no autoloads
whose time stamp is 't'.

The problem is still the error that is signaled, and the way to
investigate it is to figure out why the error is signaled.