GNU bug report logs - #50629
28.0.50; hard to debug an uncaught error with ert

Previous Next

Full log

Message #49 received at 50629 <at> debbugs.gnu.org (full text, mbox):

From: Gerd Möllmann <gerd.moellmann <at> gmail.com>
To: Lars Ingebrigtsen <larsi <at> gnus.org>
Cc: Mike Kupfer <mkupfer <at> alum.berkeley.edu>,
 Ihor Radchenko <yantar92 <at> gmail.com>, 50629 <at> debbugs.gnu.org
Subject: Re: bug#50629: 28.0.50; hard to debug an uncaught error with ert
Date: Sun, 28 Aug 2022 12:31:20 +0200

Lars Ingebrigtsen <larsi <at> gnus.org> writes:

> Gerd Möllmann <gerd.moellmann <at> gmail.com> writes:
>
>> Interestingly with condition-case-unless I get one less error with make
>> check:
>
> Are you sure you removed all the .elc files under test first?  Since
> this is a change in an ert macro, you have to do that to get meaningful
> results.

I did a make -C test clean, but now that you say it I checked, and make
clean does not remove elcs.  Which I personally find surprising.

I'll check with elcs removed between runs later.

> Anyway, that condition-case is presumably there to allow checking for
> failures?  I.e., the :expected stuff.

Ok.

> Instead of altering removing the condition-case there, wouldn't it be
> possible to output the backtrace explicitly in the handler?

You mean the signal-hook-function?  It is invoked, so one can do
something there.  If it would invoke the debugger, that would also work,
I think, but I haven't checked.  Directly outputting something there is
probably not a good idea.  If anything it should record the backtrace in
ERT result structs, like ERT's debugger function.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.