GNU bug report logs - #67837
29.1.90; inhibit-interaction breaks keyboard macros

Previous Next

Full log

View this message in rfc822 format

From: Eli Zaretskii <eliz <at> gnu.org>
To: sbaugh <at> catern.com
Cc: sbaugh <at> janestreet.com, larsi <at> gnus.org, 67837 <at> debbugs.gnu.org, monnier <at> iro.umontreal.ca
Subject: bug#67837: 29.1.90; inhibit-interaction breaks keyboard macros
Date: Sat, 16 Dec 2023 15:57:42 +0200

> From: sbaugh <at> catern.com
> Date: Sat, 16 Dec 2023 13:22:23 +0000 (UTC)
> Cc: Spencer Baugh <sbaugh <at> janestreet.com>, larsi <at> gnus.org,
> 	67837 <at> debbugs.gnu.org, monnier <at> iro.umontreal.ca
> 
> Eli Zaretskii <eliz <at> gnu.org> writes:
> > Why do you need to do that when inhibit-interaction is non-nil in the
> > first place?  Code that needs interaction shouldn't be run or tested
> > in those conditions.
> 
> As I said before: because otherwise read-char outside a keyboard macro
> will hang, and I want my test to fail instead of hang in that case.

The usual way of dealing with this in test frameworks is to mock
read-char (or whatever function that causes trouble when running
tests).  As long as read-char is not the API you are testing, you
don't really need read-char as it works in Emacs, you need something
that will provide the same output.

> Let me phrase the use case differently:
> 
> I have some tests which I'd like to run in batch Emacs.  By default, if
> any of the code under test runs read-char, Emacs will simply hang
> forever (that's what read-char does in batch mode).
> 
> I would prefer instead that my tests to fail immediately if any code
> runs read-char.  How would you suggest I do that?

Define a replacement read-char in the test harness that accepts the
same arguments and signals an error.

Previous Next

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