GNU bug report logs - #56682
Fix the long lines font locking related slowdowns

Previous Next

Full log

Message #1147 received at 56682 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Gregory Heytings <gregory <at> heytings.org>
Cc: 56682 <at> debbugs.gnu.org, Eli Zaretskii <eliz <at> gnu.org>,
 monnier <at> iro.umontreal.ca
Subject: Re: bug#56682: Fix the long lines font locking related slowdowns
Date: Sun, 7 Aug 2022 14:14:24 +0300

On 07.08.2022 02:29, Gregory Heytings wrote:
> 
>>> Especially given that the said correctness has always been relative. 
>>> So it's not like we're leaving the heaven of 100% correctness and 
>>> falling into the hell of 10% correctness.  A more reasonable view of 
>>> the situation is that we had 90% correctness and now have,
>>
>> As someone who worked on different major mode and syntax-ppss itself a 
>> little, that feels moderately insulting.
>>
>> No: we strive to close to 100% correctness in supporting language 
>> syntax and can often reach it with moderate effort (programming-wise).
>>
> 
> It isn't meant to be insulting, and I think you understood that.  My 
> experience is simply that highlighting in Emacs is definitely not "close 
> to 100% correct", whatever be the mode.  The most correct one is perhaps 
> emacs-lisp-mode (unsurprisingy).

Have you ever managed to trigger incorrect highlighting in a JSON file?

Without the recent changes, that is.

>>> in files with "loo long lines" and only in those files, 60% correctness.
>>
>> Does the attached screenshot look like 60% correctness to you?
>>
> 
> A small sample of a big file is definitely not representative, and I 
> think you understood that.

Did I need to send the screenshots of every screen? Like I said, more 
than half of the file was like that.

And then I restarted with 'emacs -Q', and the result looked even worse.

>> To me, it's more like -60%. Or at least, that's what the utility of 
>> such highlighting will be (negative).
>>
> 
> In which case you should turn highlighting off in such files.  That's 
> what all other editors do anyway.

That's the same as saying the user should turn off font-lock if they 
experience performance problems.

Except in this case they would still have choice to suffer a little in 
the performance department (and the amount of said suffering will highly 
depend on their use case) but keep good syntax highlighting.

Previous Next

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