Re: signalling errors in redisplay

[Dave Love]

>>>>> Eli Zaretskii writes:

 > AFAIK, it's a bug for redisplay code to signal an error, since that
 > triggers another redisplay.  

I don't see why it has to be, if it doesn't trigger an uninterruptible
storm of redisplays.

This case is definitely a significant bug.  Is it getting fixed?

 > If nothing else, it should put the message text into *Messages*
 > without signaling.

... where you probably never see it, and you get the same error
repeated anyhow.  I've wasted time on things which only used
*Messages* and I didn't realize there was a message written there.

Anyhow, to reproduce the problem I've seen in Gnus under X:

M-x gnus then G f and give the name of the file you've saved the
following attachment under, hit RET on the group and RET on the
article in the summary to display it.  The error occurs displaying the
mode line.  If you have a second frame open, you can iconify the
offending one and kill the buffer *Article* to recover.

redisplay-bug-article
Description: Binary data