C-g quitting wrong code?

[alex2daniel]

In my experiments for the new compile, I'm running high load tests where
font-lock is very busy fontifying huge buffers.

I find that when I'm in the mini-buffer and quit I have to do it maybe 2-3
times before I actually leave the mini-buffer.  Apparently the asynchronous
background job is what gets quit.

This is doubly bad:

- it is certainly confusing to C-g in the minibuffer and still be in it.

- maybe some part of the buffer has not been fontified, hence the error
messages in that part have not been detected?  I don't know how robust font-lock
is on that point.

My feeling is that background jobs should only be interruptible this way,
when there is nothing running or waiting for input in the foreground.

-- 
+++ GMX - die erste Adresse für Mail, Message, More +++
Neu: Preissenkung für MMS und FreeMMS! http://www.gmx.net