cperl-mode
Goran
TIA
Gordon
Ilya Zakharevich
Goran
> Why cperl-mode in Emacs 22.1 is so slow?
As you see from the pletora of people rushing to help you:
E2LITTLECONTEXT
There may be many factors affecting the speed; the complexity of the
Perl script, the version of CPerl mode, the settings currently in use
in your customization files, the settings in the shipped customization
files, and, finally, the version of Emacs.
One way of reading your question above is when you exclude all other
factors (e.g., by running with -no-site-file, and loading
cperl-mode.el from a known location, and checking the same Perl
script), Emacs 22.1 is slower than some other (unspecified) version of
Emacs.
You need to provide enough information for people to start ESP on what
you meant exactly,
Hope this helps,
Ilya
Tim X
Just to add to the info for th eOP.
I'm using cperl-mode under the latest emacs CVS sources (22.1.50.1) on perl
scripts of an average size around 200+ lines and I've not observed any
speed/performance issues. However, I rarely find it necessary to use the
debugger (I think using trace/step etc is a poor substitute for
understanding and as a rule only use debuggers when the problem is likely
to be related to something like interaction with the kernel, memory,
foreign functions etc). Likewise, I only use the profiler at the end if
performance is an issue.
Just mentioning this to emphasise Illya's point that more information is
required to understand where the performance problem may be.
Tim
--
tcross (at) rapttech dot com dot au
Goran
Goran
<arg...@yahoo.com>], who wrote in article
<mailman.5390.118820551...@gnu.org>:
> > Why cperl-mode in Emacs 22.1 is so slow?
>As you see from the pletora of people rushing to help you:
> E2LITTLECONTEXT
Thanks for your replies Ilya and Tim.
Paging C-v, M-v and typing characters is most time
slow, no matter the size of file.
I have compiled Emacs 22.1.1 from source on Mandriva 2005
Linux, kernel 2.6, 512M memory and Debian 3.1 Linux kernel 2.6
384M memory.
cperl-mode is as shipped with Emacs 22.1.1
(defvar cperl-version
(let ((v "Revision: 5.22"))
arch-tag: 42e5b19b-e187-4537-929f-1a7408980ce6
Tried on both systems:
./configure --without-x
./configure --with-gtk
Also tried to compile under /opt directory
With
emacs --no-site-file and
emacs -Q
symptoms remains the same.
Thanks again and best wishes.
Gordon
gordon gordon
[A complimentary Cc of this posting was sent to
Goran
], who wrote in article :
> Why cperl-mode in Emacs 22.1 is so slow?
As you see from the pletora of people rushing to help you:
E2LITTLECONTEXT
Paging C-v, M-v and typing characters is most time
slow, no matter the size of file.
I have compiled Emacs 22.1.1 from source on Mandriva 2005
Linux, kernel 2.6, 512M memory and Debian 3.1 Linux kernel 2.6 384M memory.
cperl-mode is as shipped with Emacs 22.1.1
(defvar cperl-version
(let ((v "Revision: 5.22"))
arch-tag: 42e5b19b-e187-4537-929f-1a7408980ce6
Tried on both systems:
./configure --without-x
./configure --with-gtk
Also tried to compile under /opt directory
With
emacs --no-site-file and
emacs -Q
symptoms remains the same.
Thanks again and best wishes.
Gordon
Be a better Globetrotter. Get better travel answers from someone who knows.
Yahoo! Answers - Check it out.
Peter Dyballa
Am 31.08.2007 um 11:22 schrieb Goran:
> Tried on both systems:
> ./configure --without-x
> ./configure --with-gtk
> Also tried to compile under /opt directory
>
> With
> emacs --no-site-file and
> emacs -Q
> symptoms remains the same.
Can you check font-lock? It certainly plays no role which X widgets
you use. My largest Perl file is about 64 K. Scrolling becomes slow
when the lines are around 100 characters wide and get broken into the
next line.
(cond ((fboundp 'global-font-lock-mode)
;; Turn on font-lock in all modes that support it
(global-font-lock-mode 1)
; (setq font-lock-multiline 'undecided')
(setq jit-lock-stealth-verbose t)
(setq jit-lock-mode t)
; (setq jit-lock-stealth-load 80)
;; Maximum colors
; (setq font-lock-maximum-decoration t)
(setq font-lock-maximum-decoration '((tex-mode . 2) (latex-mode .
2) (t . t)))
))
Check whether you can reduce fontification with faces and colours. In
my example it's maximal for all but tex- or latex-mode.
--
Greetings
Pete
Math illiteracy affects 7 out of every 5 Americans.
Ilya Zakharevich
Goran
> Paging C-v, M-v and typing characters is most time
> slow, no matter the size of file.
>
> I have compiled Emacs 22.1.1 from source on Mandriva 2005
> Linux, kernel 2.6, 512M memory and Debian 3.1 Linux kernel 2.6
> 384M memory.
You still did not answer "Slow with respect to WHAT?".
> cperl-mode is as shipped with Emacs 22.1.1
> (defvar cperl-version
> (let ((v "Revision: 5.22"))
This is not the latest version. ilyaz.org/software.
I saw bugs reports that v22.1 is practically unusable with CPerl (too
many bugs: syntaxification does not work correct), but not about slowdown.
Hope this helps,
Ilya
Tim X
> Subject: Re: cperl-mode
Interesting regarding your reference to reports that cperl-mode is almost
unusable under emacs 22. I've not found this. I have noticed some
fontification that doesn't look correct and do get considerable 'error
messages' reported in the mini-buffer concerning unmatched parens etc, but
yo don't get this if your inserting parens or other delimiters in
pairs. For exmaple, qw/ will give unmatched/unbalanced errors after every
character you insert. I guess its possible this could impact on
performance. However, if you do qw/ /, with point inbetween the /s, then
you don't get the warning/errors after each character you insert.
I tend to use some of the 'hairy' options, which means when I do something
like an if or for/foreach etc, most of the delimiters are automatically
added in matched pairs - maybe thats why I'm not having the problems others have reported.
I've not noticed any problems with scrolling of pages or any slowness - not
sure if thats because I'm not suffering the same problems or maybe I'm just
naturally slow and therefore don't notice it!
gordon gordon
> cperl-mode is as shipped with Emacs 22.1.1
> (defvar cperl-version
> (let ((v "Revision: 5.22"))
This is not the latest version. ilyaz.org/software.
Dowloaded v5.23 and M-x load-file
~/.emacs.d/cperl-mode.el-5.23 and everything work fine.
>You still did not answer "Slow with respect to WHAT?".
I can type 5-7 characters on keyboard without nothing showing under the point, and I am not so fast (300CPM).
Thanks to all.
Gordon
Choose the right car based on your needs. Check out Yahoo! Autos new Car Finder tool.
rahed
> Interesting regarding your reference to reports that cperl-mode is almost
> unusable under emacs 22. I've not found this. I have noticed some
> fontification that doesn't look correct and do get considerable 'error
> messages' reported in the mini-buffer concerning unmatched parens etc, but
> yo don't get this if your inserting parens or other delimiters in
> pairs. For exmaple, qw/ will give unmatched/unbalanced errors after every
> character you insert. I guess its possible this could impact on
> performance. However, if you do qw/ /, with point inbetween the /s, then
> you don't get the warning/errors after each character you insert.
>
> I tend to use some of the 'hairy' options, which means when I do something
> like an if or for/foreach etc, most of the delimiters are automatically
> added in matched pairs - maybe thats why I'm not having the problems others have reported.
>
> I've not noticed any problems with scrolling of pages or any slowness - not
> sure if thats because I'm not suffering the same problems or maybe I'm just
> naturally slow and therefore don't notice it!
I also hear this for the first time, I don't experience any of these
problems and don't use 'hairy' options.
--
Radek
Ilya Zakharevich
<ti...@nospam.dev.null>], who wrote in article <87tzqfa...@lion.rapttech.com.au>:
> Interesting regarding your reference to reports that cperl-mode is almost
> unusable under emacs 22.
Of course, "usability" is a relative notion. Some people would even
use perl-mode...
What I'm refering to is that CPerl happens to fontify "about 99%" of
constructs correct; which means that if see something
fishily-fontified, the probability that you mistyped something is much
higher than the probability that it is a CPerl's misunderstanding. In
short, "you can rely on fontification as on early-warning system".
(Of course, the exact proportion depends on the style of programming;
and whether you decide to rely on it or not is a personal choice...)
With 22.1, I get reports that something as simple as s(a)(b) is not
correctly fontified. This makes me worry a lot...
Hope this helps,
Ilya
Ilya Zakharevich
<arg...@yahoo.com>], who wrote in article <mailman.133.118863647...@gnu.org>:
> Dowloaded v5.23 and M-x load-file
> ~/.emacs.d/cperl-mode.el-5.23 and everything work fine.
I do not think that this is due to difference between 5.22 and 5.23;
the more probable cause is that the shipped 5.22 is too defective
comparing to "my" 5.22. Could you check it with my 5.22 too?
Thanks,
Ilya
P.S. The reason for two versions is that I stand for
backward-compatibility, and the shipped version must conform to
certain guidelines which are not backward-compatibe. The
translation is done "by volunteers", and it is not always correct...
P.P.S. HTML posts are very frowned upon...
Goran
>> Dowloaded v5.23 and M-x load-file
>> ~/.emacs.d/cperl-mode.el-5.23 and everything work fine.
>
> I do not think that this is due to difference between 5.22 and 5.23;
> the more probable cause is that the shipped 5.22 is too defective
> comparing to "my" 5.22. Could you check it with my 5.22 too?
>
> Thanks,
> Ilya
>
> P.S. The reason for two versions is that I stand for
> backward-compatibility, and the shipped version must conform to
> certain guidelines which are not backward-compatibe. The
> translation is done "by volunteers", and it is not always correct...
>
> P.P.S. HTML posts are very frowned upon...
Ilya Zakharevich
<arg...@yahoo.com>], who wrote in article <mailman.222.118880540...@gnu.org>:
> >> Dowloaded v5.23 and M-x load-file
> >> ~/.emacs.d/cperl-mode.el-5.23 and everything work fine.
> > I do not think that this is due to difference between 5.22 and 5.23;
> > the more probable cause is that the shipped 5.22 is too defective
> > comparing to "my" 5.22. Could you check it with my 5.22 too?
> Tried "your" 5.22 and its fine too.
Thanks for checking. Can somebody in development team log this as a
bug in "translation" of the shipped version of 5.22?
Thanks,
Ilya
Ilya Zakharevich
> >I do not think that this is due to difference between 5.22 and 5.23;
> >the more probable cause is that the shipped 5.22 is too defective
> >comparing to "my" 5.22. Could you check it with my 5.22 too?
> Tried "your" 5.22 and its fine too.
Thanks for checking, yours
Ilya
Stefan Monnier
>> (defvar cperl-version
>> (let ((v "Revision: 5.22"))
> This is not the latest version. ilyaz.org/software.
> Dowloaded v5.23 and M-x load-file
> ~/.emacs.d/cperl-mode.el-5.23 and everything work fine.
Assuming the problem is in 5.22-Emacs but not in 5.22 (the two versions are
slightly different), it's likely that 5.23-Emacs (which I just installed in
the CVS repository) will suffer from the same problem. If you could post
a bug report via M-x report-emacs-bug and give a recipe starting from
"emacs -Q" that would be very helpful.
Stefan
Ilya Zakharevich
<mon...@iro.umontreal.ca>], who wrote in article <jwvmyvzrssv.fsf-mon...@gnu.org>:
Myself, I do not have access to 22 (and do not plan to). What I've
heard is that
(setq cperl-hairy t)
open a file with
s(a)(b);
and the second part is not properly highlighted (it should be
highlighted the same as the first).
Hope this helps,
Ilya
Eli Zaretskii
> Date: Fri, 7 Sep 2007 23:25:10 +0000 (UTC)
> Bcc: il...@gnu.org
> Originator: il...@powdermilk.math.berkeley.edu
>
> Myself, I do not have access to 22 (and do not plan to). What I've
> heard is that
>
> (setq cperl-hairy t)
>
> open a file with
>
> s(a)(b);
>
> and the second part is not properly highlighted (it should be
> highlighted the same as the first).
You heard wrong. I just tested this, and the highlighting is exactly
as you wanted it to be.
Ilya Zakharevich
<el...@gnu.org>], who wrote in article <mailman.483.118923679...@gnu.org>:
> > (setq cperl-hairy t)
> >
> > open a file with
> >
> > s(a)(b);
> >
> > and the second part is not properly highlighted (it should be
> > highlighted the same as the first).
>
> You heard wrong. I just tested this, and the highlighting is exactly
> as you wanted it to be.
... which means it is an intermittent bug. Yet worse than a
reproducible one...
Thanks,
Ilya
Eli Zaretskii
> Date: Mon, 10 Sep 2007 08:33:53 +0000 (UTC)
> Bcc: il...@gnu.org
> Originator: il...@powdermilk.math.berkeley.edu
>
Or no bug at all.
Ilya Zakharevich
Eli Zaretskii
> > > > s(a)(b);
> > > >
> > > > and the second part is not properly highlighted (it should be
> > > > highlighted the same as the first).
> > > You heard wrong. I just tested this, and the highlighting is exactly
> > > as you wanted it to be.
> > ... which means it is an intermittent bug.
> Or no bug at all.
??? You mean whoever sent the bug report to me was a nut case?
Possible. But a bug in Emacs is MUCH more probable...
Hope this helps,
Ilya
Eli Zaretskii
> Date: Tue, 11 Sep 2007 01:08:57 +0000 (UTC)
> Bcc: il...@gnu.org
> Originator: il...@powdermilk.math.berkeley.edu
>
> Eli Zaretskii
> <el...@gnu.org>], who wrote in article <mailman.654.118945368...@gnu.org>:
> > > > > s(a)(b);
> > > > >
> > > > > and the second part is not properly highlighted (it should be
> > > > > highlighted the same as the first).
>
> > > > You heard wrong. I just tested this, and the highlighting is exactly
> > > > as you wanted it to be.
>
> > > ... which means it is an intermittent bug.
>
> > Or no bug at all.
>
> ??? You mean whoever sent the bug report to me was a nut case?
No, I mean that it could be a cockpit error. Or incomplete bug report
(i.e. the problem does exist, but shows up in a situation other than
just "emacs -Q"). In any case, the recipe as stated does not exhibit
the bug.
Tim X
Eli, I just grabbed your latest version
($Id: cperl-mode.el,v 5.23 2007/02/15 11:34:23 vera Exp vera $) and when I
byte compile it, I get a lot of warnings, many of them about it using
depricated features, such as the VC header alist, backquote format (which
it claims has bee depricated for 10 years, which seems odd), new calling
args for error or align or something (sorry, can't remember specifics just
now) and some other things - all of them just warnings, no errors..
this was with emacs 22.1.50.1
Do you want to be informed about these, do you want patches that fix the
depricated warnings or is this all stuff you know about and are working on
or will get to later?
Tim
)
Eli Zaretskii
> Date: Tue, 11 Sep 2007 18:11:44 +1000
>
> Eli, I just grabbed your latest version
``My'' latest version? I don't develop cperl-mode, so it's certainly
not ``mine''.
> ($Id: cperl-mode.el,v 5.23 2007/02/15 11:34:23 vera Exp vera $)
Where did you get this version? This thread is (AFAIK) about the
version of cperl-mode that comes with the stock Emacs 22.1. Are you
talking about the version available from some other place?
Ilya Zakharevich
<ti...@nospam.dev.null>], who wrote in article <87d4wpf...@lion.rapttech.com.au>:
> Eli, I just grabbed your latest version
I think you are mixing me and Eli. I was not talking to myself... ;-)
> ($Id: cperl-mode.el,v 5.23 2007/02/15 11:34:23 vera Exp vera $) and when I
Unfortunately (as explained earlier), this may be not a full specification
(unless the official version has $Id stripped).
> byte compile it, I get a lot of warnings, many of them about it using
> depricated features, such as the VC header alist, backquote format (which
> it claims has bee depricated for 10 years, which seems odd), new calling
> args for error or align or something (sorry, can't remember specifics just
> now) and some other things - all of them just warnings, no errors..
"New calling args" are interesting. "Deprecated features" are not.
Thanks,
Ilya
Ilya Zakharevich
Eli Zaretskii
> > > Or no bug at all.
> >
> > ??? You mean whoever sent the bug report to me was a nut case?
>
> No, I mean that it could be a cockpit error. Or incomplete bug report
> (i.e. the problem does exist, but shows up in a situation other than
> just "emacs -Q").
It was checked with emacs -q only...
[I think that anything which shows with
emacs -q
on a "shipped system" is a bug in Emacs. A packager of a distribution
is a part of developpers team, you want it or not...]
Yours,
Ilya
Eli Zaretskii
> Date: Tue, 11 Sep 2007 19:58:23 +0000 (UTC)
> Bcc: il...@gnu.org
> Originator: il...@powdermilk.math.berkeley.edu
>
> > (i.e. the problem does exist, but shows up in a situation other than
> > just "emacs -Q").
>
> It was checked with emacs -q only...
"emacs -q" and "emacs -Q" are the same thing on my system, since
there's no site-init file; the only difference is that -Q doesn't show
the splash screen.
Anyway, I just rechecked with -q, just to be sure, and I still don't
see the problem.
> [I think that anything which shows with
>
> emacs -q
>
> on a "shipped system" is a bug in Emacs.
Yes.
> A packager of a distribution is a part of developpers team, you want
> it or not...]
On my system, _I_ am the packager. Emacs was produced from stock
ftp.gnu.org sources and installed by "make install".
Tim X
My stuff up - teach me to post when tired and not really paying
attention. I meant the message for Ilya Zakharevich. Actually, I can
partially blame it on my speech synthesiser as Ilya Zakharevich and
Eli Zaretskii can sound very similar, partially because I have speech rates
very high to try and get through at a speed similar to a fast reader (which
I use to be once upon a time). - well at least similar enough to make
a tired user confused.
My apologies for transferring my confusion to others!
Tim
Tim X
oh, and yes, I was talking about the version available from Ilya's site -
it is a later version and he suggested it may work better than the earlier
version shipped with emacs 22 (partialy because the version as shipped is
'modified' slightly.
Tim
Tim X
> [A complimentary Cc of this posting was sent to
> Tim X
> <ti...@nospam.dev.null>], who wrote in article <87d4wpf...@lion.rapttech.com.au>:
>> Eli, I just grabbed your latest version
>
> I think you are mixing me and Eli. I was not talking to myself... ;-)
>
yep, realised that afterwards. Sorry about that.
>> ($Id: cperl-mode.el,v 5.23 2007/02/15 11:34:23 vera Exp vera $) and when I
>
> Unfortunately (as explained earlier), this may be not a full specification
> (unless the official version has $Id stripped).
>
The official version does have it stripped and its cperl-version states it
is based on your 5.22
This is the version I grabbed from your site.
>> byte compile it, I get a lot of warnings, many of them about it using
>> depricated features, such as the VC header alist, backquote format (which
>> it claims has bee depricated for 10 years, which seems odd), new calling
>> args for error or align or something (sorry, can't remember specifics just
>> now) and some other things - all of them just warnings, no errors..
>
> "New calling args" are interesting. "Deprecated features" are not.
>
> Thanks,
> Ilya
OK, I'll post the warnings regarding changes in calling args.
Tim
David Combs
Ilya Zakharevich <nospam...@ilyaz.org> wrote:
>[A complimentary Cc of this posting was sent to
>Tim X
><ti...@nospam.dev.null>], who wrote in article <87tzqfa...@lion.rapttech.com.au>:
>> Interesting regarding your reference to reports that cperl-mode is almost
>> unusable under emacs 22.
>
>Of course, "usability" is a relative notion. Some people would even
>use perl-mode...
Yesterday I went looking for the newest cperl.el -- my (oldish) record of where to look was":
5361:<LI><a href="http://math.berkeley.edu/~ilya/software/emacs/">Index of /~ilya/software/emacs/ ilya's own version of cperl-mode</a>
What I found (here and there) was dated 2005, I think.
Where's the newer stuff (does anything even exist?)?
(I'm currently using 21.3.2 on sparc)
And its *current* set of pros and cons?
AND, you said that some people prefer old perl-mode?
WHY?
(maybe stupid question, but what do *you* use?)
Thanks,
David
Ilya Zakharevich
<dkc...@panix.com>], who wrote in article <fdm90e$gm7$1...@reader1.panix.com>:
> 5361:<LI><a href="http://math.berkeley.edu/~ilya/software/emacs/">Index of /~ilya/software/emacs/ ilya's own version of cperl-mode</a>
>
> What I found (here and there) was dated 2005, I think.
Feb 15 2007
> And its *current* set of pros and cons?
I update the pro list (in minidocs) when needed. Cons - I do not know
about them.
> AND, you said that some people prefer old perl-mode?
>
> WHY?
Well, the last such sentiment I saw might have been about 5 years
ago. But I'm quite sure some people still do - they do not need any
reason to prefer one thing to another, right?
> (maybe stupid question, but what do *you* use?)
Emacs 19.33.
Hope this helps,
Ilya