[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
bug#20352: Factor 2 slow down of font-locking in c-mode
From: |
Alan Mackenzie |
Subject: |
bug#20352: Factor 2 slow down of font-locking in c-mode |
Date: |
Sat, 18 Apr 2015 17:32:34 +0000 |
User-agent: |
Mutt/1.5.23 (2014-03-12) |
Hello again, Stefan.
Apologies for some misinformation in my previous post. My timings for
the master branch were on an Emacs configured with no compiler
optimisation. When I reconfigure and rebuild with default optimisation
the figures look like....
On Sat, Apr 18, 2015 at 09:45:52AM +0000, Alan Mackenzie wrote:
> On Sat, Apr 18, 2015 at 12:35:19AM -0400, Stefan Monnier wrote:
> > > Twice as slow, eh? You should be so lucky! I've just tried it, and for
> > > me it's nearer a factor of 10. ;-(
> > My original case (M-x swiper RET when in src/xdisp.c, while running my
> > locally hacked Emacs on my desktop with all debugging enabled) seemed to
> > be also way past a factor 2 (it just seemed like it would never end and
> > did not have the patience to wait for it to finish), so
> > I made a measurement on a dedicated machine where I recompiled the
> > unadulterated Emacs with default flags and all and got a "disappointing"
> > factor of 2, which still seemed significant enough.
> Some figures: I tried running (with M-:)
> (let ((start (float-time)))
> (font-lock-fontify-region (point-min) (point-max))
> (- (float-time) start))
> on xdisp.c, freshly loaded before each attempt, with 2 versions of Emacs
> and 2 versions of CC Mode. Here are the timings (in seconds):
> Released Emacs-24.5
> Emacs-24.5's CC Mode: 29.82369565963745
> savannah master's CC Mode: 55.37644958496094
> savannah Master
> Emacs-24.5's CC Mode: 109.47807884216309
> savannah Master's CC Mode: 241.9050052165985
savannah Master compiled with default optimisation
Emacs-24.5's CC Mode: 28.840153694152832
savannah Master's CC Mode: 55.49798059463501
So you were right - the master branch's CC Mode is fontifying at half
speed. I'm doing some bisecting to try and find the commit which did
this.
> . So it seems that some changes in CC Mode and some other changes
> outside of CC Mode are both to blame, here.
No, it's a CC Mode thing only.
> > Stefan
--
Alan Mackenzie (Nuremberg, Germany).