cperl-mode

cperl-mode

[Stefan Monnier]

What was the reason for the change below:

2004-03-25  Juanma Barranquero  <lektu@terra.es>

	* progmodes/cperl-mode.el (cperl-indent-alist, cperl-where-am-i):
	Comment out (it's unused and unfinished code).

I suggest we try and refrain from such changes because they make it more
difficult to merge in changes from the author in the future.


        Stefan

Re: cperl-mode

[Juanma Barranquero]

On 26 Mar 2004 09:47:49 -0500, Stefan Monnier <monnier@iro.umontreal.ca> wrote:

> What was the reason for the change below:

Well, is a bunch of complex, completely unused code and has been so for
a long time (at least two years), and it generates a warning.

> I suggest we try and refrain from such changes because they make it more
> difficult to merge in changes from the author in the future.

I know you've been merging Ilya's version about once a year (judging by
the ChangeLog). I didn't do the change lightly; I thought about it, and
in the end I concluded it wouldn't represent too much of a problem. I'll
revert the change if you feel I'm mistaken.

                                                           /L/e/k/t/u

Re: cperl-mode

[Stefan Monnier]

> I know you've been merging Ilya's version about once a year (judging by
> the ChangeLog). I didn't do the change lightly; I thought about it, and
> in the end I concluded it wouldn't represent too much of a problem.

Fair enough.

> I'll revert the change if you feel I'm mistaken.

Not worth the trouble,


        Stefan

cperl-mode

[Goran]

Why cperl-mode in Emacs 22.1 is so slow?
TIA
Gordon

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Goran 
<argoxo@yahoo.com>], who wrote in article <mailman.5390.1188205514.32220.help-gnu-emacs@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

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

Re: cperl-mode

[Tim X]

Ilya Zakharevich <nospam-abuse@ilyaz.org> writes:

> [A complimentary Cc of this posting was sent to
> Goran 
> <argoxo@yahoo.com>], who wrote in article <mailman.5390.1188205514.32220.help-gnu-emacs@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
>
> 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,
>

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

Re: cperl-mode

[gordon gordon]

[-- Attachment #1.1: Type: text/plain, Size: 1037 bytes --]

Ilya Zakharevich <nospam-abuse@ilyaz.org> wrote: [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

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




       
---------------------------------
Be a better Globetrotter. Get better travel answers from someone who knows.
Yahoo! Answers - Check it out.

[-- Attachment #1.2: Type: text/html, Size: 1697 bytes --]

[-- Attachment #2: Type: text/plain, Size: 152 bytes --]

_______________________________________________
help-gnu-emacs mailing list
help-gnu-emacs@gnu.org
http://lists.gnu.org/mailman/listinfo/help-gnu-emacs

Re: cperl-mode

[Goran]

[A complimentary Cc of this posting was sent to
Goran
<argoxo@yahoo.com>], who wrote in article 
<mailman.5390.1188205514.32220.help-gnu-emacs@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

Re: cperl-mode

[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.

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Goran 
<argoxo@yahoo.com>], who wrote in article <mailman.109.1188567372.18990.help-gnu-emacs@gnu.org>:
> 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

Re: cperl-mode

[Tim X]

Ilya Zakharevich <nospam-abuse@ilyaz.org> writes:

> Subject: Re: cperl-mode
>
> [A complimentary Cc of this posting was sent to
> Goran 
> <argoxo@yahoo.com>], who wrote in article <mailman.109.1188567372.18990.help-gnu-emacs@gnu.org>:
>> 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

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!

Tim
-- 
tcross (at) rapttech dot com dot au

Re: cperl-mode

[gordon gordon]

[-- Attachment #1.1: Type: text/plain, Size: 585 bytes --]

Problem solved 

> 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.

[-- Attachment #1.2: Type: text/html, Size: 1061 bytes --]

[-- Attachment #2: Type: text/plain, Size: 152 bytes --]

_______________________________________________
help-gnu-emacs mailing list
help-gnu-emacs@gnu.org
http://lists.gnu.org/mailman/listinfo/help-gnu-emacs

Re: cperl-mode

[rahed]

Tim X <timx@nospam.dev.null> writes:

> 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

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Tim X 
<timx@nospam.dev.null>], who wrote in article <87tzqfatlh.fsf@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

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
gordon gordon 
<argoxo@yahoo.com>], who wrote in article <mailman.133.1188636475.18990.help-gnu-emacs@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...

Re: cperl-mode

[Goran]

Ilya Zakharevich wrote:
>> 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,
> 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...
Sorry for HTML, hope its ok now.

Re: cperl-mode

[Ilya Zakharevich]

On Mon, Sep 03, 2007 at 09:41:46AM +0200, Goran wrote:
> >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

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Goran 
<argoxo@yahoo.com>], who wrote in article <mailman.222.1188805400.18990.help-gnu-emacs@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

Re: cperl-mode

[Stefan Monnier]

>> 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.

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

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Stefan Monnier 
<monnier@iro.umontreal.ca>], who wrote in article <jwvmyvzrssv.fsf-monnier+gnu.emacs.help@gnu.org>:
> >> 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.
> 
> 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.

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

Re: cperl-mode

[Eli Zaretskii]

> From: Ilya Zakharevich <nospam-abuse@ilyaz.org>
> Date: Fri, 7 Sep 2007 23:25:10 +0000 (UTC)
> Bcc: ilya@gnu.org
> Originator: ilya@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.

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Eli Zaretskii 
<eliz@gnu.org>], who wrote in article <mailman.483.1189236797.18990.help-gnu-emacs@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

Re: cperl-mode

[Eli Zaretskii]

> From: Ilya Zakharevich <nospam-abuse@ilyaz.org>
> Date: Mon, 10 Sep 2007 08:33:53 +0000 (UTC)
> Bcc: ilya@gnu.org
> Originator: ilya@powdermilk.math.berkeley.edu
> 
> [A complimentary Cc of this posting was sent to
> Eli Zaretskii 
> <eliz@gnu.org>], who wrote in article <mailman.483.1189236797.18990.help-gnu-emacs@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.

Or no bug at all.

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Eli Zaretskii 
<eliz@gnu.org>], who wrote in article <mailman.654.1189453685.18990.help-gnu-emacs@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?

Possible.  But a bug in Emacs is MUCH more probable...

Hope this helps,
Ilya

Re: cperl-mode

[Eli Zaretskii]

> From: Ilya Zakharevich <nospam-abuse@ilyaz.org>
> Date: Tue, 11 Sep 2007 01:08:57 +0000 (UTC)
> Bcc: ilya@gnu.org
> Originator: ilya@powdermilk.math.berkeley.edu
> 
> [A complimentary Cc of this posting was sent to
> Eli Zaretskii 
> <eliz@gnu.org>], who wrote in article <mailman.654.1189453685.18990.help-gnu-emacs@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.

Re: cperl-mode

[Tim X]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Ilya Zakharevich <nospam-abuse@ilyaz.org>
>> Date: Mon, 10 Sep 2007 08:33:53 +0000 (UTC)
>> Bcc: ilya@gnu.org
>> Originator: ilya@powdermilk.math.berkeley.edu
>> 
>> [A complimentary Cc of this posting was sent to
>> Eli Zaretskii 
>> <eliz@gnu.org>], who wrote in article <mailman.483.1189236797.18990.help-gnu-emacs@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.
>
> Or no bug at all.
>
>
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


) 
-- 
tcross (at) rapttech dot com dot au

Re: cperl-mode

[Eli Zaretskii]

> From: Tim X <timx@nospam.dev.null>
> 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?

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Tim X 
<timx@nospam.dev.null>], who wrote in article <87d4wpfwov.fsf@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

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Eli Zaretskii 
<eliz@gnu.org>], who wrote in article <mailman.673.1189481287.18990.help-gnu-emacs@gnu.org>:
> > > 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

Re: cperl-mode

[Eli Zaretskii]

> From: Ilya Zakharevich <nospam-abuse@ilyaz.org>
> Date: Tue, 11 Sep 2007 19:58:23 +0000 (UTC)
> Bcc: ilya@gnu.org
> Originator: ilya@powdermilk.math.berkeley.edu
> 
> > 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...

"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".

Re: cperl-mode

[Tim X]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Tim X <timx@nospam.dev.null>
>> 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?
>
>

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

-- 
tcross (at) rapttech dot com dot au

Re: cperl-mode

[Tim X]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Tim X <timx@nospam.dev.null>
>> 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?
>
>

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

-- 
tcross (at) rapttech dot com dot au

Re: cperl-mode

[Tim X]

Ilya Zakharevich <nospam-abuse@ilyaz.org> writes:

> [A complimentary Cc of this posting was sent to
> Tim X 
> <timx@nospam.dev.null>], who wrote in article <87d4wpfwov.fsf@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
-- 
tcross (at) rapttech dot com dot au

Re: cperl-mode

[David Combs]

In article <fbdrus$29qp$1@agate.berkeley.edu>,
Ilya Zakharevich  <nospam-abuse@ilyaz.org> wrote:
>[A complimentary Cc of this posting was sent to
>Tim X 
><timx@nospam.dev.null>], who wrote in article <87tzqfatlh.fsf@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

Re: cperl-mode

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
David Combs
<dkcombs@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

cperl-mode

[Terrence Brannon]

Hi, would someone share their cperl-mode settings with me that will
format a hash like this
 ( http://pastebin.com/dfQKjXkS )

my %default = (
    amount      => 0,
    balance     => 0,
    total_items => 0,
    trade_disc  => 0,
    debit       => 1
);

Re: cperl-mode

[LanX]

> > are you not familar with cperl-set-style?
>
> I am quite new to Emacs, so I am not familiar with a lot of things. But
> is there a style that does this automatically?
>
> Or else: what do you want to tell me?

John,  better don't care about the FUD metaperl spreads. 8)

Feeding the trolls is a waste of time...

Re: cperl-mode

[Andreas Politz]

Terrence Brannon <metaperl@gmail.com> writes:

> Hi, would someone share their cperl-mode settings with me that will
> format a hash like this
>  ( http://pastebin.com/dfQKjXkS )
>
> my %default = (
>     amount      => 0,
>     balance     => 0,
>     total_items => 0,
>     trade_disc  => 0,
>     debit       => 1
> );

I don't know about cperl, but align seams to have a ready to use rule
for this.  Try `align-current'.

-ap

Re: cperl-mode

[John Bokma]

LanX <lanx.perl@googlemail.com> writes:

>> > are you not familar with cperl-set-style?
>>
>> I am quite new to Emacs, so I am not familiar with a lot of things. But
>> is there a style that does this automatically?
>>
>> Or else: what do you want to tell me?
>
> John,  better don't care about the FUD metaperl spreads. 8)
>
> Feeding the trolls is a waste of time...

OK, thanks for the heads up. I am still an Emacs newbie, and thought I
was missing something.

-- 
John Bokma                                                               j3b

Hacking & Hiking in Mexico -  http://johnbokma.com/
http://castleamber.com/ - Perl & Python Development

Re: cperl-mode

[LanX]

> OK, thanks for the heads up. I am still an Emacs newbie, and thought I
> was missing something.

The only thing that you were missing is that Terrence aka Metaperl aka
Princepawn seems far to cool for a serious discussion.

People are trying here to figure out what he exactly wants in this
thread simply named cperl-mode.

Is it the line indentation in a multiline hash literal ???

Or is it the alignment of the fat comma ("=>") ???

cperl-set-style doesn't affect any of these, you have to look into the
documentation or try them out to see it.

And IMHO J.Rockway is just publishing a fork to support modules like
Moose and Devel::REPL, he is not the new maintainer.

Don't know if this "splitted personality" really intends to be a
troll, but he certainly has a talent in spreading confusion.
(just go to perlmonks and look at his list of worst posts)

So better ignore him as long he doesn't posts precise informations, if
you don't wanna waste your time.

(seems like that is what Ilya actually does)


Now, to give this discussion a productive turn here is the part of
cperl-mode.el  which IMHO affects the line indentation:

----------
	 ((eq 'in-parens (elt i 0))
	  ;; in-parens char-after old-indent-point is-brace containing-sexp

	  ;; group is an expression, not a block:
	  ;; indent to just after the surrounding open parens,
	  ;; skip blanks if we do not close the expression.
	  (+ (progn
	       (goto-char (elt i 2))		; old-indent-point
	       (current-column))
	     (if (and (elt i 3)		; is-brace
		      (eq (elt i 1) ?\})) ; char-after
		 ;; Correct indentation of trailing ?\}
		 (+ cperl-indent-level cperl-close-paren-offset)
	       0)))
--------

Would be nice if Ilya (or someone else) could add a new  option to not
indent after surrounding parens, I have only limited resources to do
it right now.

for the moment an extra newline is the best workaround for a dense
indentation.

------------
my %default =
  (
   amount	 => 0,
   balance	 => 0,
   total_items	 => 0,
   trade_disc	 => 0,
   debit	 => 1
  );
-------------

cheers
  rolf

Re: cperl-mode

[Ilya Zakharevich]

On 2010-05-25, LanX <lanx.perl@googlemail.com> wrote:
> Would be nice if Ilya (or someone else) could add a new  option to not
> indent after surrounding parens, I have only limited resources to do
> it right now.

I'll gladly accept well-coded patches.  (This is a recurring request
from the users.)

Yours,
Ilya

Re: cperl-mode

[LanX]

Hi Ilya

> I'll gladly accept well-coded patches.  (This is a recurring request
> from the users.)


talking about "well coded", I'm still a beginner in elisp and
sometimes wondering about your code.

You use and handle around quite often array like lists where every
index represents something specific which needs to be documented each
time you access it.

for instance in the posted code:
-----------
 ((eq 'in-parens (elt i 0))
...
 (goto-char (elt i 2))            ; old-indent-point
...
 (if (and (elt i 3)         ; is-brace
-----------
where i is the return value from  cperl-sniff-for-indent'

So why don't you use alists or even hashes instead of documenting each
numerical index?

http://www.gnu.org/s/emacs/manual/html_node/elisp/Association-Lists.html

Are there technical reasons for avoiding alists?

And wouldn't calling the variable something like "sniff"  improve the
readability?

Cheers
  Rolf

Re: cperl-mode

[Ilya Zakharevich]

On 2010-05-26, LanX <lanx.perl@googlemail.com> wrote:
> So why don't you use alists or even hashes instead of documenting each
> numerical index?
>
> http://www.gnu.org/s/emacs/manual/html_node/elisp/Association-Lists.html
>
> Are there technical reasons for avoiding alists?

I have no idea.  All I know is that elisp is very much pessimized; for
the speed reasons, I would prefer to use simplest constructs (the
initial version could have been hundreds times slower than the current
one; I was forced to change the whole framework how stuff operates on
the C level of Emacs).

There may be some benchmarks around; I just do not know about them.
(But this late in development cycle, I doubt I would change such
things now.)

Yours,
Ilya

Re: cperl-mode

[LanX]

Hi

> I have no idea.  All I know is that elisp is very much pessimized; for
> the speed reasons, I would prefer to use simplest constructs (the
> initial version could have been hundreds times slower than the

OK, just wanna understand not criticize.

> (But this late in development cycle, I doubt I would change such
> things now.)

IMHO refactoring gradually to a macro construct should do it safely
and fast, something like (value i keyname) translating to (elt i
keyindex)


> There may be some benchmarks around; I just do not know about them.

Hmm do you have any automated testing framework, e.g. for syntax
parsing?

Cheers
  Rolf

Re: cperl-mode

[Ilya Zakharevich]

On 2010-05-26, LanX <lanx.perl@googlemail.com> wrote:
>> There may be some benchmarks around; I just do not know about them.

> Hmm do you have any automated testing framework, e.g. for syntax
> parsing?

x) I meant array vs vector vs hash etc benchmarks;

x) No.  ;-)

Yours,
Ilya