In emacs 23 compile-mode doesn't recognize (c)perl error messages

In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

I did a fresh install of gnu emacs 23 on my ubuntu box and everything
seems to work fine, but when running mode-compile from cperl-mode the
error messages are not recognized.

That means the errors are not highlighted, clicking has no effect
("there is no error") and
M-x next-error only produces "Moved past last error".

Any idea?

Cheers
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

just wanted to add that I did an upgrade to cperl-mode 6.2, but the
problem remains.

The emacs 23 is a snapshot installation  parallel to my "normal" emacs
22, does this lead to conflicts?

Cheers
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 01.10.2009 um 02:21 schrieb LanX:

> The emacs 23 is a snapshot installation  parallel to my "normal" emacs
> 22, does this lead to conflicts?


No. Check load-path. Maybe re-organise the init files.

--
Greetings

   Pete

Indentation?! I will show you how to indent when I indent your skull!

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

[-- Attachment #1: Type: text/plain, Size: 2742 bytes --]

(This is a fwd to the mailinglist)

Well,did you try it out?

I've spend hours to isolate the problem without any success!

Starting emacs23 with

> emacs-snapshot -debug-init -no-site-file  tst.pl &

doesn't work, while taking emacs22 instead, does highlight the bugs after
calling "M-x mode-compile"!

The config .emacs links to debug.el which only (auto)loads the newest
versions of mode-compile and cperl-mode.

--------------------------------
lanx@nc10-ubuntu:~$ ll .emacs
lrwxrwxrwx 1 lanx lanx 23 2009-10-01 15:28 .emacs -> .emacs.d/debug/debug.el
lanx@nc10-ubuntu:~$ cat .emacs


(message "loading debug.el")

;--------------------
;        load-path
;--------------------
(normal-top-level-add-to-load-path
 '("~/.emacs.d/debug"))


;--------------------
;    cperl-mode
;--------------------

(load "~/.emacs.d/debug/cperl-mode.el")


;--------------------
;    mode-compile
;--------------------

(autoload 'mode-compile "mode-compile"
  "Command to compile current buffer file based on the major mode" t)

(add-to-list 'auto-mode-alist '("\\.pl\\'" . cperl-mode))




(message "leaving debug.el")lanx@nc10-ubuntu:~$
lanx@nc10-ubuntu:~$ ll .emacs.d/debug/
insgesamt 588
-r-------- 1 lanx lanx 402932 2009-10-01 15:16 cperl-mode.el
-rw-r--r-- 1 lanx lanx    520 2009-10-01 15:22 debug.el
-rw-r--r-- 1 lanx lanx 101476 2009-06-04 14:46 mode-compile.el
-rw-r--r-- 1 lanx lanx  74288 2009-10-01 13:58 mode-compile.elc
lanx@nc10-ubuntu:~$
--------------------------------


mode-compiling a bash file works without problems.





BTW: deleting .emacs and starting with

> emacs-snapshot -debug-init -no-site-file  tst.pl &

leads to a buffer *Local Variables* with the Message
--------------------------------
The local variables list in mode-compile.el
contains values that may not be safe (*).

Do you want to apply it?  You can type
y  -- to apply the local variables list.
n  -- to ignore the local variables list.
!  -- to apply the local variables list, and permanently mark these
      values (*) as safe (in the future, they will be set automatically.)

    outline-regexp : ";; @+"
  * eval : (outline-minor-mode 1)
------------------------------------

But no matter which option I chose the problem remains!



So what am I supposed to try now???


Please note: This is specific to cperl-mode and perl-mode. Bash and
python-files compile as expected.


2009/10/1 Peter Dyballa <Peter_Dyballa@web.de>


> Am 01.10.2009 um 02:21 schrieb LanX:
>
>  The emacs 23 is a snapshot installation  parallel to my "normal" emacs
>> 22, does this lead to conflicts?
>>
>
>
> No. Check load-path. Maybe re-organise the init files.
>
> --
> Greetings
>
>  Pete
>
> Indentation?! I will show you how to indent when I indent your skull!
>
>

[-- Attachment #2: Type: text/html, Size: 3793 bytes --]

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 01.10.2009 um 15:54 schrieb LanX:

> (autoload 'mode-compile "mode-compile"
>   "Command to compile current buffer file based on the major mode" t)
>
> (add-to-list 'auto-mode-alist '("\\.pl\\'" . cperl-mode))
>
>
>
>
> (message "leaving debug.el")lanx@nc10-ubuntu:~$
> lanx@nc10-ubuntu:~$ ll .emacs.d/debug/
> insgesamt 588
> -r-------- 1 lanx lanx 402932 2009-10-01 15:16 cperl-mode.el
> -rw-r--r-- 1 lanx lanx    520 2009-10-01 15:22 debug.el
> -rw-r--r-- 1 lanx lanx 101476 2009-06-04 14:46 mode-compile.el


Mode-compile is not part of GNU Emacs.
Cperl-mode is part of GNU Emacs, so there is no sense in installing  
some foreign file with the same name. Except looking for trouble...
You could try to make default value perl-mode in auto-mode-alist to  
point to cperl-mode (in your init file, for example):

	(defalias 'perl-mode 'cperl-mode)

--
Greetings

   Pete

They that can give up essential liberty to obtain a little temporary  
safety deserve neither liberty nor safety.
		-Benjamin Franklin, Historical Review of Pennsylvania.

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 01.10.2009 um 20:52 schrieb LanX:

> Alternatively I would appreciate a .emacs that fixes this problem.


Does compile-mode work without Ilya's latest, or, to be more exact,  
does it work with GNU Emacs' cperl-mode?

If it does not, then you might ask for help on an XEmacs list,  
because compile-mode comes from XEmacs. (Or you could try to use it  
in XEmacs.) Because it obviously isn't GNU Emacs' fault. Or try to  
adjust the external compile-mode with any of your cperl-mode's. There  
should be some ways of customisation...

Does it work without compile-mode? GNU Emacs has its own compile.el...

Does it work without compile-mode and without Ilya's latest, or, to  
be more exact, does it work with GNU Emacs' own tools?

--
Greetings

   Pete

We are usually convinced more easily by reasons we have found  
ourselves than by those which have occurred to others.
				– Blaise Pascal

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

[-- Attachment #1: Type: text/plain, Size: 1300 bytes --]

> Does compile-mode work without Ilya's latest, or, to be more exact, does
it
> work with GNU Emacs' cperl-mode?

in emacs 22 yes in emacs 23 no!

> If it does not, then you might ask for help on an XEmacs list, because
> compile-mode comes from XEmacs. (Or you could try to use it in XEmacs.)

I'm using Gnu Emacs for two years now w/o problems, and just try to upgrade.

> Because it obviously isn't GNU Emacs' fault.

I don't care who's "fault" it is, I just want to pin down that there is an
upgrade problem, such that we can search a solution.

> Or try to adjust the external
> compile-mode with any of your cperl-mode's. There should be some ways of
> customisation...

cperl-mode.el has over 9000 lines, before I intervene I'll wait what Ilya
says.

Well AFAIK he's still working with 19 ... but he's surely the best to know.

> Does it work without compile-mode? GNU Emacs has its own compile.el...

yes it "works" but  limited. It doesn't and never did  support compile.el,
so one cannot compile.

> Does it work without compile-mode and without Ilya's latest, or, to be
more
> exact, does it work with GNU Emacs' own tools?


see above. Everything worked fine with Gnu Emacs 22.

Anyway I posted the problem on some perl sites asking for confirmation,
thanks for your help!

Cheers
  rolf

[-- Attachment #2: Type: text/html, Size: 1532 bytes --]

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-01, Peter Dyballa <Peter_Dyballa@Web.DE> wrote:
> If it does not, then you might ask for help on an XEmacs list,  
> because compile-mode comes from XEmacs. (Or you could try to use it  
> in XEmacs.) Because it obviously isn't GNU Emacs' fault.

In my experience, it usually is.  What makes you say so, and what
makes it particularly "obvious"?

Puzzled,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 01.10.2009 um 22:45 schrieb Ilya Zakharevich:

> In my experience, it usually is.  What makes you say so, and what
> makes it particularly "obvious"?


Because GNU Emacs 23.1 works with the files that come with it.  
Problems so often come from external files...

--
Greetings

   Pete       (:
         _    / __    -    -
       _/ \__/_/        -     -
      (´`)      (´`)   -    -
       `´        `´

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi guys

just got the information that my problem is not reproducebal on a
similar system.

http://groups.google.com/group/emacs-perl-intersection/browse_thread/thread/328823127d1f8b7

So my best guess is that my installation of v23 might be broken
somehow....

I will try to reinstall emacs 23 on wednesday when I'm back home and
I'll keep you informed.

Thanks, Danke and Spassiva 8)
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi guys

I had no luck so far.  8(

That's what I tried:

* actually emacs-snapshot reported an installation error due to eieio!
** so I deinstalled the eieieo package and reinstalled snapshot but
the described problem remained

* I did a cvs install as described in
  http://www.saltycrane.com/blog/2008/10/installing-emacs-23-cvs-ubuntu-hardy/
** installation without problems, but mode-compile bug remained

* I deinstalled emacs22 and xemacs21 by deinstalling emacs-common to
avoid side effects
** emacs-23 starts without problems but the bug remains

I can assure it's not a cperl-mode error, since I have exactly the
same problem when running mode-compile from perl-mode without even
load cperl-mode.el. So it's not Ilya's concern.

Mode-compile works perfectly with python or sh modes.

The only "error"-message I was able to spot till now was
"Loading `mode-compile': old-style backquotes detected!"

I'm currently running Ubuntu Intrepid 8.10, I'm going to upgrade to
Jaunty 9.04 tonight.

Maybe meanwhile some ubuntu-users can check if they are experiencing
the same problems?

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

So far I ran my test case on 3 different systems with emacs 23 freshly
installed and the newest mode-compile.el:

- Ubuntu Intrepid 8.10 it failed
- Xubuntu Hardy 8.04 it failed
+ Windows XP it worked fine.

I decided not to try upgrading to Jaunty 9.04, the risk of resulting
problems followed with long adjustments on my laptop are actually too
high. 8(

OK now  I think it's reasonable to send a bug report, but to whose
address?

Is this for the main developer team or do they have special sub
responsibilities for special OS ports?

Or should I ask this on gnu.emacs.dev ?

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 09.10.2009 um 04:09 schrieb LanX:

> OK now  I think it's reasonable to send a bug report, but to whose
> address?
>
> Is this for the main developer team or do they have special sub
> responsibilities for special OS ports?


In the Help menu you can find an entry: Send Bug Report... You can  
invoke this function on both systems and either send two bug reports  
with quite the same subject or combine the two system descriptions in  
one bug report. You might also think of adding a system description  
from the working system.

--
Greetings

   Pete                                           0
                                            %-/\_//
                                             (*)(*)

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

I found a workaround by extending the regexes to match the whole
line!

Changing in mode-compile.el from

------------------------------------------------
(defvar perl-compilation-error-regexp-alist
  ;; Contributed by Martin Jost
  '(
    ;; PERL 4
    ("in file \\([^ ]+\\) at line \\([0-9]+\\).*" 1 2)
    ;; PERL 5   Blubber at FILE line XY, <XY> line ab.
    ("at \\([^ ]+\\) line \\([0-9]+\\)," 1 2)
    ;; PERL 5   Blubber at FILE line XY.
    ("at \\([^ ]+\\) line \\([0-9]+\\)." 1 2)
    )
  ;; This look like a paranoiac regexp: could anybody find a better
one? (which WORK).
  ;;'(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\) [^\n]*line \\([0-9]+\\)
[\\.,]" 2 3))
  "Alist that specifies how to match errors in perl output.

See variable compilation-error-regexp-alist for more details.")
------------------------------------------------

to
------------------------------------------------
(defvar perl-compilation-error-regexp-alist
  ;; Contributed by Martin Jost
  '(
    ;; PERL 4
    ("^[^\n]* in file \\([^ ]+\\) at line \\([0-9]+\\).*" 1 2)
    ;; PERL 5   Blubber at FILE line XY, <XY> line ab.
    ("^[^\n]* at \\([^ ]+\\) line \\([0-9]+\\)," 1 2)
    ;; PERL 5   Blubber at FILE line XY.
    ("^[^\n]* at \\([^ ]+\\) line \\([0-9]+\\)." 1 2)
    )
  ;; This look like a paranoiac regexp: could anybody find a better
one? (which WORK).
  ;;'(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\) [^\n]*line \\([0-9]+\\)
[\\.,]" 2 3))
  "Alist that specifies how to match errors in perl output.

See variable compilation-error-regexp-alist for more details.")
\f------------------------------------------------

fixes the problem in emacs 23! 8)

(Plz note: the WHOLE error line gets underlined now)


Have there been changes to compile.el???

(I'm not sure if this problem might be caused by the German
localization of my Ubuntu installation since I had encoding problems
in the past. But I never changed the default settings .
In mode-compile.el I occasionally see `^L´ between paragraphs...
)

Anyway I will send a bug report and mail to the authors of mode-
compile.el and compile.el.

Cheers
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 09.10.2009 um 14:56 schrieb LanX:

> In mode-compile.el I occasionally see `^L´ between paragraphs...

These are OK: they send a form-feed to the printer that the new  
paragraph is started on new page.

--
Greetings

   Pete

It isn't pollution that's harming the environment. It's the  
impurities in our air and water that are doing it.

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

[-- Attachment #1: Type: text/plain, Size: 268 bytes --]

Here my ticket

4689: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=4689
Emacs Bug Tracking System

I wonder if the change in unicode support in emacs 23 broke the way regexes
work in a subtle way...or if ubuntu somehow mixed up the unicode settings
???

[-- Attachment #2: Type: text/html, Size: 412 bytes --]

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

[-- Attachment #1: Type: text/plain, Size: 671 bytes --]

Sorry the patch I posted only works for perl-mode

Here a patch for cperl-mode

(defvar cperl-compilation-error-regexp-alist-lanx-patch
  '(
    ;; PERL 4
    ("^[^\n]* in file \\([^ ]+\\) at line \\([0-9]+\\).*" 1 2)
    ;; PERL 5   Blubber at FILE line XY, <XY> line ab.
    ("^[^\n]* at \\([^ ]+\\) line \\([0-9]+\\)," 1 2)
    ;; PERL 5   Blubber at FILE line XY.
    ("^[^\n]* at \\([^ ]+\\) line \\([0-9]+\\)." 1 2)
    )
  "Alist that specifies how to match errors in perl output.")

(if (fboundp 'eval-after-load)
    (eval-after-load
    "mode-compile"
      '(setq perl-compilation-error-regexp-alist
         cperl-compilation-error-regexp-alist-lanx-patch)))

[-- Attachment #2: Type: text/html, Size: 821 bytes --]

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-01, Peter Dyballa <Peter_Dyballa@Web.DE> wrote:
>> In my experience, it usually is.  What makes you say so, and what
>> makes it particularly "obvious"?

> Because GNU Emacs 23.1 works with the files that come with it.  
> Problems so often come from external files...

This is a MAJORLY remarkable attitude.  I'm discussing framing this
and putting it on my wall...  Maybe it is more clear after minor
rewording:

 Our program does not have problems.  If you have problems, they are yours.

===

Let me repeat: in my experience, if older Emacs "works", and newer one
"does not", in about 95% of cases it indicates a bug in newer Emacs,
not in Elisp code calling Emacs...

Hope this helps,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Lennart Borgman]

On Sat, Oct 10, 2009 at 12:15 AM, Ilya Zakharevich
<nospam-abuse@ilyaz.org> wrote:
>
> Let me repeat: in my experience, if older Emacs "works", and newer one
> "does not", in about 95% of cases it indicates a bug in newer Emacs,
> not in Elisp code calling Emacs...

Maybe that is correct for the code you are writing, but I have seen
plenty of other code where the problem is in the external code.

The main problem is of course how to handle this. What about using
unit tests? When I take time to write them for nXhtml I think they
helps a lot. (But I am mostly not fighting problems with getting it to
fit with changes in Emacs. I am fighting my own mistakes. And mistakes
in external libraries.)

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

> The main problem is of course how to handle this. What about using
> unit tests? When I take time to write them for nXhtml I think they
> helps a lot. (But I am mostly not fighting problems with getting it to
> fit with changes in Emacs. I am fighting my own mistakes. And mistakes
> in external libraries.)

Sure testing is crucial!

let's see
- cperl-mode depends  on
  - mode-compile which depends on
    - compile.el which depends on
      - ???.

ATM me with my reduced knowledge of elisp have to PROVE the bug I'm
trying to report, instead of getting testtools or at least help to
isolate the problem.

But with an existing test framework every new bug would eventually
lead to a new test added.

I even was able to produce patches which solve the problem for me...
but this knowledge, which I was able to accumulate after 3 days, will
now mostly get lost, because the support still denies existence of and
responsibility for a problem, which involves external code and I'm now
loosing interest to help.

But with a test-set for compile.el I would have been in a position to
either prove the problem very fast  or I would at least have a
starting point to add new tests, improving the test coverage.

I have to say that even today after using emacs for more than 15 years
now my knowledge of elisp is quite reduced, so I can't even tell what
kind of testing is realistically possible.

From a perlish point of view (excuse me ;-) someting like
www::mechanize or selenium used to test web-apps would be nice.

This way even elisp hating hackers would be able to supply test-code
in a scripting language of their choice!

Well, just some brainstorming ...

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Joseph Brenner]

LanX <rolf.langsdorf@googlemail.com> writes:

>> The main problem is of course how to handle this. What about using
>> unit tests? When I take time to write them for nXhtml I think they
>> helps a lot. (But I am mostly not fighting problems with getting it to
>> fit with changes in Emacs. I am fighting my own mistakes. And mistakes
>> in external libraries.)
>
> Sure testing is crucial!

[...]

> But with a test-set for compile.el I would have been in a position to
> either prove the problem very fast  or I would at least have a
> starting point to add new tests, improving the test coverage.
>
> I have to say that even today after using emacs for more than 15 years
> now my knowledge of elisp is quite reduced, so I can't even tell what
> kind of testing is realistically possible.

As it happens, I'm also enamored of suites of automated tests.
Oh wait, I'm a perl programmer.

There are something like a dozen elisp test frameworks at this point.
None of them have been endorsed by the emacs-devel team.  Anyone
interested in this approach would have to evaluate the existing test
packages, or write yet-another-one... and I'm afraid someone coming
out of the perl culture is not likely to be impressed by any of them
(e.g. to my knowledge, none of them respect the TAP standard).

I have a vague impression that ert.el may have a little more momentum
than the rest.  But then, the only way to get a copy of it is as a
side-effect of installing the nxhtml packages.

I gather that the reason we're stuck with this state of affairs is that
the emacs-devel team (including rms) isn't really interested.  They like
the idea of having a tool to do higher level testing of the entire
interface, but don't see the point of lower level tests... which is to
say they'd like to solve the hardest problem first, and aren't too
likely to do this soon.

> From a perlish point of view (excuse me ;-) someting like
> www::mechanize or selenium used to test web-apps would be nice.

Sure, but it might be a good idea to think about an analog of
Test::More first.

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Stefan Monnier]

> I gather that the reason we're stuck with this state of affairs is that
> the emacs-devel team (including rms) isn't really interested.  They like
> the idea of having a tool to do higher level testing of the entire
> interface, but don't see the point of lower level tests... which is to
> say they'd like to solve the hardest problem first, and aren't too
> likely to do this soon.

That is actually not the case.  It's more like we have plenty of things
to do and we're waiting for someone else to setup such a testsuite.
We do have a "test" subdirectory in the CVS repository, although it
currently doesn't hold much.

I have no experience with such test rigs, nor with any the many Emacs
test frameworks, and I have time neither to review them, nor to write
my own.  But I would very much welcome some test framework to be
installed in Emacs's CVS repository such that I can easily add tests in
the "test" subdirectory for simple low-level tests.

Being able to have higher-level tests would be nice as well and is
indeed a hard problem in general, but it's not a prerequisite at all.

Any help would be greatly welcome.  E.g. help to review of the various
Elisp test frameworks out there, highlighting their advantages and
shortcoming (to help choose one), and then help to set it up for Emacs
(e.g. adding Makefile rules to run the suite).


        Stefan

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Lennart Borgman]

On Fri, Oct 30, 2009 at 7:04 AM, Stefan Monnier
<monnier@iro.umontreal.ca> wrote:
>
> I have no experience with such test rigs, nor with any the many Emacs
> test frameworks, and I have time neither to review them, nor to write
> my own.  But I would very much welcome some test framework to be
> installed in Emacs's CVS repository such that I can easily add tests in
> the "test" subdirectory for simple low-level tests.
>
> Being able to have higher-level tests would be nice as well and is
> indeed a hard problem in general, but it's not a prerequisite at all.


I have added some things to ert.el so that it can be used for testing
commands and also fontification. (Some lisp expert could probably
transform my additions to something better, but the funcitonality is
there.)

This is, as Joseph noted, part of nXhtml. There is no harm in
downloading nXhtml and installing that ( ;-) ), but you can also grab
the files directly from the nXhtml repository at Launchpad. In some
moment of extraordinary orderlineness I happened to put them in the
nxhtml/test directory.

If you install nXhtml you can run the tests for nXhtml with

   M-x nxhtmltest-run-Q

to see how it works.


> Any help would be greatly welcome.  E.g. help to review of the various
> Elisp test frameworks out there, highlighting their advantages and
> shortcoming (to help choose one), and then help to set it up for Emacs
> (e.g. adding Makefile rules to run the suite).
>
>
>        Stefan
>

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Joseph Brenner]

Lennart Borgman <lennart.borgman@gmail.com> writes:
> Stefan Monnier <monnier@iro.umontreal.ca> wrote:

>> I have no experience with such test rigs, nor with any the many Emacs
>> test frameworks, and I have time neither to review them, nor to write
>> my own.  But I would very much welcome some test framework to be
>> installed in Emacs's CVS repository such that I can easily add tests in
>> the "test" subdirectory for simple low-level tests.
>>
>> Being able to have higher-level tests would be nice as well and is
>> indeed a hard problem in general, but it's not a prerequisite at all.
>
> I have added some things to ert.el so that it can be used for testing
> commands and also fontification. (Some lisp expert could probably
> transform my additions to something better, but the funcitonality is
> there.)
>
> This is, as Joseph noted, part of nXhtml. There is no harm in
> downloading nXhtml and installing that ( ;-) ),

I would certainly hope not, but I've come to be somewhat leery of emacs
extensions that consist of large numbers of interelated elisp files.

I'm afraid that the lack of a package manager combined with Stallman's
fearsome reputation as guardian of the cathedral has made gnu emacs
prone to these things... they often seem to me like forks-in-disguise.

> but you can also grab the files directly from the nXhtml repository at
> Launchpad. In some moment of extraordinary orderlineness I happened to put
> them in the nxhtml/test directory.
>
> If you install nXhtml you can run the tests for nXhtml with
>
>    M-x nxhtmltest-run-Q
>
> to see how it works.

I guess it's good to have a live example at hand, like this, but the immediate
question (in my mind at least) would be is ert.el an entirely stand-alone
package, or has some some dependencies with your other packages crept in?

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Lennart Borgman]

On Tue, Nov 3, 2009 at 3:01 AM, Joseph Brenner <doom@kzsu.stanford.edu> wrote:
>
>> If you install nXhtml you can run the tests for nXhtml with
>>
>>    M-x nxhtmltest-run-Q
>>
>> to see how it works.
>
> I guess it's good to have a live example at hand, like this, but the immediate
> question (in my mind at least) would be is ert.el an entirely stand-alone
> package, or has some some dependencies with your other packages crept in?


Not that I remember of now. Perhaps some minor left over, but nothing important.

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Joseph Brenner]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> I gather that the reason we're stuck with this state of affairs is that
>> the emacs-devel team (including rms) isn't really interested.  They like
>> the idea of having a tool to do higher level testing of the entire
>> interface, but don't see the point of lower level tests... which is to
>> say they'd like to solve the hardest problem first, and aren't too
>> likely to do this soon.
>
> That is actually not the case.

Cool.  That was my interpretation of this old thread:

  http://lists.gnu.org/archive/html/emacs-devel/2007-09/msg00752.html

> It's more like we have plenty of things to do and we're waiting for
> someone else to setup such a testsuite.  We do have a "test"
> subdirectory in the CVS repository, although it currently doesn't hold
> much.
>
> I have no experience with such test rigs, nor with any the many Emacs
> test frameworks, and I have time neither to review them, nor to write
> my own.  But I would very much welcome some test framework to be
> installed in Emacs's CVS repository such that I can easily add tests in
> the "test" subdirectory for simple low-level tests.
>
> Being able to have higher-level tests would be nice as well and is
> indeed a hard problem in general, but it's not a prerequisite at all.

> Any help would be greatly welcome.  E.g. help to review of the various
> Elisp test frameworks out there, highlighting their advantages and
> shortcoming (to help choose one), and then help to set it up for Emacs
> (e.g. adding Makefile rules to run the suite).

I'll put it no my lists of things to do.  If there's any interest at all
in the results of the research, I'm more likely to take the trouble to
do it.

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Lennart Borgman]

On Tue, Nov 3, 2009 at 3:11 AM, Joseph Brenner <doom@kzsu.stanford.edu> wrote:
>
> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>
>> I have no experience with such test rigs, nor with any the many Emacs
>> test frameworks, and I have time neither to review them, nor to write
>> my own.  But I would very much welcome some test framework to be
>> installed in Emacs's CVS repository such that I can easily add tests in
>> the "test" subdirectory for simple low-level tests.
>>
>> Being able to have higher-level tests would be nice as well and is
>> indeed a hard problem in general, but it's not a prerequisite at all.
>
>> Any help would be greatly welcome.  E.g. help to review of the various
>> Elisp test frameworks out there, highlighting their advantages and
>> shortcoming (to help choose one), and then help to set it up for Emacs
>> (e.g. adding Makefile rules to run the suite).
>
> I'll put it no my lists of things to do.  If there's any interest at all
> in the results of the research, I'm more likely to take the trouble to
> do it.

Of course there is. The most difficult thing seems sometimes to be the
authors of different packages to work together so we can get the best
out of it without the burden of many similar packages. Getting things
together and comparing is very difficult and takes time however so it
is not surprising.

There is a page on EmacsWiki about the available packages. Maybe
cleaning up and rewriting that page can be a good start?

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-09, Lennart Borgman <lennart.borgman@gmail.com> wrote:
>> Let me repeat: in my experience, if older Emacs "works", and newer one
>> "does not", in about 95% of cases it indicates a bug in newer Emacs,
>> not in Elisp code calling Emacs...
>
> Maybe that is correct for the code you are writing, but I have seen
> plenty of other code where the problem is in the external code.

At startup, I load MEGABYTES of external code (IMO, Emacs UI is
untolerable without a lot of tweaking - at least it was a decade ago,
when my configuration got more or less frozen).  Some of it is written
by me, but most is other people code.  So I base my judgement not only
on my code - although "the other code" is also "chosen by me", so might
be not very representative. ;-)

> The main problem is of course how to handle this. What about using
> unit tests?

This is not the first (and maybe even not the third :-() time Emacs
breaks compile.el.  Which means that REx engine can't even handle
quite primitive RExen.  I would have expected that already on the
first time these would become items of the test suite...

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

HI Ilya

> At startup, I load MEGABYTES of external code (IMO, Emacs UI is
> untolerable without a lot of tweaking - at least it was a decade ago,
> when my configuration got more or less frozen).  Some of it is written
> by me, but most is other people code.  So I base my judgement not only
> on my code - although "the other code" is also "chosen by me", so might
> be not very representative. ;-)

Did you ever consider breaking up cperl-mode.el into different
submoduls hosted on CPAN?

It would be much easier for the perl-crowd to support patches and
maintainance if we had something like

cperl-fontlock.el
cperl-compilation.el
cperl-expansion.el
cperl-perldoc.el
...

together with a testsuite written in perl?

Would facilitate the collaboration of perl&emacs devs a lot!

For instance I'm trying to extend the compilation-regexes and some
projects like sepia.el (a perl REPL) could profit from cperl-
fontlock.el.

http://cpansearch.perl.org/src/SEANO/Sepia-0.97/Sepia.html

And sorry to say so - some aspects of cperl-mode are very outdated,
e.g there is no up-to-date info-file for perldocs.

Maybe you're interested to join at

emacs-perl-intersection@googlegroups.com

Cheers
  Rolf

PS: BTW in xemacs I had a context-menu on right-click, offering the
cperl-mode menu do you know how to activate this in gnu emacs?

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-15, LanX <lanx.perl@googlemail.com> wrote:
> Did you ever consider breaking up cperl-mode.el into different
> submoduls hosted on CPAN?

> It would be much easier for the perl-crowd to support patches and
> maintainance if we had something like
>
> cperl-fontlock.el
> cperl-compilation.el
> cperl-expansion.el
> cperl-perldoc.el
> ...

???  Just use Emacs with a good code viewer (I use my imenu-go.el and
a handful of other tricks).  Do not see how breaking into parts would
simplify anything...

> together with a testsuite written in perl?

Supporting a testsuite usually means 1.5x increase of time.  Assume an
estimate of extra 250hours. Who is going to invest in it?  Given that
"mine" cperl-mode.el is "already working", definitely not me...

> And sorry to say so - some aspects of cperl-mode are very outdated,
> e.g there is no up-to-date info-file for perldocs.

???  Make it.

> PS: BTW in xemacs I had a context-menu on right-click, offering the
> cperl-mode menu do you know how to activate this in gnu emacs?

Bind right-click to [menubar cperl] or whatever it is?  The docs of
easymenu may have a chance to mention it...

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> ???  Just use Emacs with a good code viewer (I use my imenu-go.el and
> a handful of other tricks).  Do not see how breaking into parts would
> simplify anything...

Well it wasn't meant just for me but for all users. How many perl-
developers know elisp? And how many of them can handle 9000 lines of
code?

> > together with a testsuite written in perl?
>
> Supporting a testsuite usually means 1.5x increase of time.  Assume an
> estimate of extra 250hours. Who is going to invest in it?  Given that
> "mine" cperl-mode.el is "already working", definitely not me...

My idea is that the perl&&emacs-community  could/would supply
testcases for every problem/bug they encounter.


> > And sorry to say so - some aspects of cperl-mode are very outdated,
> > e.g there is no up-to-date info-file for perldocs.
>
> ???  Make it.

Actually I'm trying to, but I'm not really sure if it's worth it,
since there are already good html-versions of perldoc, even solutions
which are rendering just-in-time any module in INC.

I didn't mean to demand of you making a new info-file but to suggest
(for the sake of simplicity) to skip an outdated feature from the
menu.

> Bind right-click to [menubar cperl] or whatever it is?  The docs of
> easymenu may have a chance to mention it...

yeah already found a workaround, other thread.

Thanx for your help and don't mind, I was just brainstorming.

Cheers
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-16, LanX <rolf.langsdorf@googlemail.com> wrote:
> Hi
>
>> ???  Just use Emacs with a good code viewer (I use my imenu-go.el and
>> a handful of other tricks).  Do not see how breaking into parts would
>> simplify anything...
>
> Well it wasn't meant just for me but for all users. How many perl-
> developers know elisp?

How would breaking into parts fix this?

> And how many of them can handle 9000 lines of code?

???  Myself, I see no difference between working with 1000 LOC and
9000 LOC - as far as I can jump between functions without effort.

>> > together with a testsuite written in perl?
>>
>> Supporting a testsuite usually means 1.5x increase of time.  Assume an
>> estimate of extra 250hours. Who is going to invest in it?  Given that
>> "mine" cperl-mode.el is "already working", definitely not me...
>
> My idea is that the perl&&emacs-community  could/would supply
> testcases for every problem/bug they encounter.

I do not think this will change my estimate much...  (And I do not have
so much believe that communities work the way you think they do...)

>> > And sorry to say so - some aspects of cperl-mode are very outdated,
>> > e.g there is no up-to-date info-file for perldocs.
>>
>> ???  Make it.
>
> Actually I'm trying to, but I'm not really sure if it's worth it,
> since there are already good html-versions of perldoc, even solutions
> which are rendering just-in-time any module in INC.

> I didn't mean to demand of you making a new info-file but to suggest
> (for the sake of simplicity) to skip an outdated feature from the
> menu.

a) I do not find it outdated.  (IMO, HTML (and PDF) are not suitable
			        for documentation.)

b) How would removing it help somebody?

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> How would breaking into parts fix this?

Well do you really question the benefits of modularizations and
plugins? :-)

Modularizing would lower the danger to break the whole package making
changes and plugins easier.
People would be able to assemble the moduls and plugins they need with
the help of a test-suite.

Actually only some people dare touching cperl-mode. But having small
moduls as templates facilitates a try and error approach.

Some moduls like sepia try to hook on cperl, but Sean doesn't know how
to make your font-lock usable for his REPL.


But I clearly see it would be to much to ask from you to invest more
work in it...

(it would mean loosing control over the oeuvre.. ;-)

I asked many questions about customizing and extending cperl at
perlmonks ...
The actual trend is to develop perl-editors like Padre or Kephra which
are extendable with perl.

> a) I do not find it outdated.  (IMO, HTML (and PDF) are not suitable
>                                 for documentation.)

IIRC the info files for perl were very outdated (2001), very hard to
find and had even problems.

> b) How would removing it help somebody?

I've spend two days to find the (depressing) answer for a)

http://www.perlmonks.org/?node_id=782153

That's too hard for a beginner...

Suggestion: you may check if the info file is accessible and skip the
entry if not.

Cheers
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-17, LanX <lanx.perl@googlemail.com> wrote:
> Well do you really question the benefits of modularizations and
> plugins? :-)

> Modularizing would lower the danger to break the whole package

To modularize, one needs to define boundaries

> making changes and plugins easier.

Given the amount of people's interest in changes and plugins, I doubt
that the total of the work will be "easier".

Given enough incentive, yes, I might spend 100 hours (estimated) to
define demarkation lines, and document/interface all the stuff which
crosses the lines.  This would slow CPerl down, but with progress in
CPUs, this may be forgotten.

I just do not see how these hours could save 100s of hours of work for
other people - to justify my investment.  Without this, the incentive
must be pretty tempting...

> The actual trend is to develop perl-editors like Padre or Kephra which
> are extendable with perl.

Keep in mind that perl has - in a certain sense - very limited
usability when working with strings.  (See my interview with Perl
journal for more details.)

>> a) I do not find it outdated.  (IMO, HTML (and PDF) are not suitable
>>                                 for documentation.)
>
> IIRC the info files for perl were very outdated (2001), very hard to
> find and had even problems.

The info files are as old as you make them.

>> b) How would removing it help somebody?
>
> I've spend two days to find the (depressing) answer for a)
>
> http://www.perlmonks.org/?node_id=782153
>
> That's too hard for a beginner...

How should I fix it?  (IIRC, the hard part is getting the word
`pod2texi'.  I expect that googling for this would quickly lead one to
an answer.)

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

> How should I fix it?  (IIRC, the hard part is getting the word
> `pod2texi'.  I expect that googling for this would quickly lead one to
> an answer.)

Of course I already tried it !!!

(without luck, like mentioned in the perlmonks thread I linked to. )

The problem is more fundamental: info-files are static!

For every new installation of a perl-modul one needs to build the info
file again.

Calling something like POD2HTML builds dynamically full hyperlinked
HTML-pages of all installed moduls, which might also be displayed
within emacs.

Or is it possible to produce and display info-files dynamically???

Cheers
  Rolf

PS: Ilya to be clear I do certainly not demand any further work of
you, I really appreciate the work you already did!!!
I'm just wanna discuss the best approach for further development ...

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-19, LanX <lanx.perl@googlemail.com> wrote:
>
>> How should I fix it?  (IIRC, the hard part is getting the word
>> `pod2texi'.  I expect that googling for this would quickly lead one to
>> an answer.)
>
> Of course I already tried it !!!
>
> (without luck, like mentioned in the perlmonks thread I linked to. )
>
> The problem is more fundamental: info-files are static!

As any file...

> For every new installation of a perl-modul one needs to build the info
> file again.

I do not see how switching to a different format might help here...

> Or is it possible to produce and display info-files dynamically???

I do not think you meant something like "in a moving vehicle".  ;-)
Otherwise, this does not make a lot of sense...

You may install a module. AND you may install its docs...

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> > The problem is more fundamental: info-files are static!
>
> As any file...
>
> > For every new installation of a perl-modul one needs to build the info
> > file again.
>
> I do not see how switching to a different format might help here...
>
> > Or is it possible to produce and display info-files dynamically???
>
> I do not think you meant something like "in a moving vehicle".  ;-)
> Otherwise, this does not make a lot of sense...

whatever perlversion I update or CPAN-Module I install man or perldoc
always show the most recent documentation.

If I need colors and xrefs I can run one of the many pod2html
solutions to "dynamically" produce html and display them using w3m.

pod2texi is a seperate process which has to be run seperately to
update the tex and info files, with the result that the files produced
are most of the time outdated.

AFAIK emacs needs these files to be installed in a special directory
and parsed at start up. Thats what I call "static".

A "dynamic" solution would allow to update a CPAN-module while emacs
is running and next time you call info the _new_ documentation is
displayed!

No idea if this is even possible with the actual Info implementation
in emacs.

Hope I was clearer
  Cheers Rolf

Re: perl TeXinfo docs emacs

[Ilya Zakharevich]

On 2009-10-24, LanX <lanx.perl@googlemail.com> wrote:
> whatever perlversion I update or CPAN-Module I install man or perldoc
> always show the most recent documentation.

... because somebody (installbin and/or installman scripts) do the
work for you.

> pod2texi is a seperate process which has to be run seperately to
> update the tex and info files, with the result that the files produced
> are most of the time outdated.

If you want pod2texi to be run automatically when a module or Perl
distribution is installed, raise the issue on Perl lists.  You need
also to specify conditions in which the script would recognize the
need to install TeXinfo.

  [I added c.l.p.m to newsgroups]

> AFAIK emacs needs these files to be installed in a special directory

... same as with any other format...

> and parsed at start up.

If true, this may be easily circumvented.

Yours,
Ilya

Re: perl TeXinfo docs emacs

[LanX]

Sometimes I have the impression you want to misunderstand me ....

...or is it just a passion for sophism? ;-)

this is _dynamic_:

modul="Data::Dumper";                                    # whatever
installed modul  you want
perldoc -oHTML $modul >/tmp/doc.html;          # no need for
precompiled files
w3m /tmp/doc.html;                                           # or
firefox or lynx or ...emacs


> ... same as with any other format...

nope!

> If true, this may be easily circumvented.

Plz show me how!

IIRC the newest sepia.el already implemented browsing of pod2html-docs
this way.

Cheers
  rolf

PS: can't post on "comp.lang.misc.perl"
--------------
Google Groups unterstützt zurzeit keine Veröffentlichungen in
folgenden Usenet-Gruppen: "comp.lang.misc.perl"
---------

Re: perl TeXinfo docs emacs

[LanX]

> PS: can't post on "comp.lang.misc.perl"
> --------------
> Google Groups unterstützt zurzeit keine Veröffentlichungen in
> folgenden Usenet-Gruppen: "comp.lang.misc.perl"
> ---------

did you mean comp.lang.perl.misc?

Re: perl TeXinfo docs emacs

[Ilya Zakharevich]

On 2009-10-24, LanX <lanx.perl@googlemail.com> wrote:
> this is _dynamic_:
>
> modul="Data::Dumper";                                    # whatever
> installed modul  you want

Above, you use precompiled files.  Probably, you just do not know you
use them?

> perldoc -oHTML $modul >/tmp/doc.html;          # no need for
> precompiled files

You just precompiled the file...

> w3m /tmp/doc.html;                                           # or
> firefox or lynx or ...emacs

Let me repeat it from the preceeding message (which had messed

>> ... same as with any other format...

> nope!

???

I give up.  After 3 or 4 messages, you still did not manage to explain
what it is exactly what you want.

  is it: Perl module and Perl-proper installations would run pod2texi
  for you transparently, and install them where info-mode can find them?

>> If true, this may be easily circumvented.

> Plz show me how!

Look where info.el stores parsed 'dir' files, and reset the cache.

Yours,
Ilya

Re: perl TeXinfo docs emacs

[LanX]

Hi Ilya

> I give up.  After 3 or 4 messages, you still did not manage to explain
> what it is exactly what you want.

We had a discussion if it's worth it to convert POD to INFO:

--------
> > And sorry to say so - some aspects of cperl-mode are very outdated,
> > e.g there is no up-to-date info-file for perldocs.

> ???  Make it.

Actually I'm trying to, but I'm not really sure if it's worth it,
since there are already good html-versions of perldoc, even solutions
which are rendering just-in-time any module in INC.
----------

I tried to explain that one can produce nice HTML for any POD or PL-
file dynamically just-in-time without any extra installations while
pod2texi is very wearisome.

Your happy with pod2texi, I prefer alternatives, that's all!

So no need to quarrel! 8)

Bye
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> ???  Myself, I see no difference between working with 1000 LOC and
> 9000 LOC - as far as I can jump between functions without effort.

Your certainly right!

But I suppose that more than the half is used for parsing perl, making
the rest smaller. 8)

take a look at regex-tool.el with 180 lines I have no fear to add
features.

http://www.cims.nyu.edu/~stucchio/software/elisp/regex-tool.el.html

But if you don't like the idea it's OK, do you mind if others try to
modularize your code?

(don't know if this is covered by the GPL)

cheers
  rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-17, LanX <lanx.perl@googlemail.com> wrote:
> But if you don't like the idea it's OK, do you mind if others try to
> modularize your code?

What I mind is people *breaking* the code.  And the moment they know
the code well enough so that the probability of breakage is low
enough, their incentive to modularize would disappear...  :-( ;-)

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi Ilya

> What I mind is people *breaking* the code.  And the moment they know
> the code well enough so that the probability of breakage is low
> enough, their incentive to modularize would disappear...  :-( ;-)

well that's why I was talking about a testsuite...

A testsuite could be very simple to extend for outsiders, by defining
a) input-text
b) action
c) expected output

BTW: I just realized that Padre has a plugin system hosted on CPAN
where different people supply code.

However ...I'm gonna try to write some simple module add-ons to cperl
instead of larger suites like perlysense or sepia.

Maybe you can help me:

What's the best way to extend POD and comment handling?
----------------------------------------------
a) I'd like to render  markup like "B<word>" gets a bold face
b) links become active refs where I can click?

see http://www.perlmonks.org/index.pl?node_id=745098

Do you have hooks to extend your font-locking or is it better to parse
after cperl has finished for text-property "in-pod"?



(BTW: is cperl-6.2 buggy for POD detection?
...tested in emacs22 and emacs23
I'm gonna open a new thread with details or do you prefer another
channel?)

cheers
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-19, LanX <lanx.perl@googlemail.com> wrote:
> What's the best way to extend POD and comment handling?

Highlighters (facification and syntaxification) form a list.  Just
append to this list.

> a) I'd like to render  markup like "B<word>" gets a bold face

Do not forget to check for presence of "in POD" text-property.

> b) links become active refs where I can click?

Here you want to check for "in POD" and/or comment's face.

> see http://www.perlmonks.org/index.pl?node_id=745098

> Do you have hooks to extend your font-locking or is it better to parse
> after cperl has finished for text-property "in-pod"?

IIRC, syntaxification (which sents "in POD") is done before
facification.  So put your stuff in the list of facification recipies,
and you can check for this text-property.  There is no need to have
CPerl hooks, since facification is performed via normal Emacs lists.

> (BTW: is cperl-6.2 buggy for POD detection?
> ...tested in emacs22 and emacs23
> I'm gonna open a new thread with details or do you prefer another
> channel?)

Thread is best.  (Although Google search of groups looks quickly
deteriorating...  Anyone else noticing this?  Are there alternatives?)

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> ???  Just use Emacs with a good code viewer (I use myimenu-go.eland
> a handful of other tricks).  Do not see how breaking into parts would
> simplify anything...

Where do I get imenu-go.el?

ftp://ftp.math.ohio-state.edu/pub/users/ilya/emacs/imenu-go.el

is still offline

cheers
  rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> Where do I get imenu-go.el?

ok .... with the help of http://www.google.com/codesearch I was able
to localize it within one of your zip files.

 http://math.berkeley.edu/~ilya/software/emacs/emacs-el-src-2005-09.zip

bye

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-16, LanX <lanx.perl@googlemail.com> wrote:
> Hi
>
>> ???  Just use Emacs with a good code viewer (I use myimenu-go.eland
>> a handful of other tricks).  Do not see how breaking into parts would
>> simplify anything...
>
> Where do I get imenu-go.el?
>
> ftp://ftp.math.ohio-state.edu/pub/users/ilya/emacs/imenu-go.el
>
> is still offline

I see that

  ilyaz.org/software/emacs/imenu-go.el

leads there; sorry...  I'm afraid that currently one may need to
download a ZIP bundle from this directory...

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Peter Dyballa]

Am 10.10.2009 um 00:15 schrieb Ilya Zakharevich:

> On 2009-10-01, Peter Dyballa wrote:
>>> In my experience, it usually is.  What makes you say so, and what
>>> makes it particularly "obvious"?
>
>> Because GNU Emacs 23.1 works with the files that come with it.
>> Problems so often come from external files...
>
> This is a MAJORLY remarkable attitude.  I'm discussing framing this
> and putting it on my wall...  Maybe it is more clear after minor
> rewording:
>
>  Our program does not have problems.  If you have problems, they  
> are yours.
>
> ===
>
> Let me repeat: in my experience, if older Emacs "works", and newer one
> "does not", in about 95% of cases it indicates a bug in newer Emacs,
> not in Elisp code calling Emacs...
>
> Hope this helps,


Yes, for me this is helpful. It helps me strengthening my opinion.  
LanX yesterday posted patches for compile-mode.el...

--
Greetings

   Pete

Some day we may discover how to make magnets that can point in any  
direction.

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

Of course Ilya is right, if a change produces a bug, in at least 95%
of the cases the bug is in the changed code.

The "external" modules didn't change, only the core emacs!

Furthermore it was shown to depend on the OS (and maybe the
localization) and tested on different machines.


> Yes, for me this is helpful. It helps me strengthening my opinion.  
> LanX yesterday posted patches for compile-mode.el...

Not exactly, I posted *workarounds* for mode-compile.el  (!) and cperl-
mode.el by altering regexes!

Definitely plain emacs modules don't work on my system, running

M-x compile  "perl -w ~/tst.pl" in  *perl-mode*

Activates errors with emacs 22 and xemacs 21, but doesn't do so with
emacs 23.

Saying so I'm not sure if maybe the Ubuntu team or perl 5.10 messed it
up, my bet is a strange interference between unicode and regexes.

Anyway I'm tired of this Dogbert vs Kafka situation...

My IDE-System is working again and  everybody else has no problems. 8)

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-01, Peter Dyballa <Peter_Dyballa@web.de> wrote:
> Cperl-mode is part of GNU Emacs, so there is no sense in installing  
> some foreign file with the same name. Except looking for trouble...

Quite the opposite (discussed here again and again).  (Unless things
changed in the last week; one will need to sit and look...)

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

OK guys it seems the miracle why this bug only happens with me is
resolved!

I did CVS-trunk Installations of 23.1.50 while the other tests done
where with the stable 23.1.1.

To be sure, I just installed GNU Emacs 23.1.1, which doesn't show any
of these problems.

Sorry, Googling for "emacs 23 install" predominantly leads to advices
for "CVS trunk" installations, even including complete ubuntu
packages.

http://www.saltycrane.com/blog/2008/10/installing-emacs-23-cvs-ubuntu-hardy/
http://emacs.orebokech.com/

I have trouble to find any ubuntu packages for the stable release, any
advice?

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

The problem has been fixed in 23.1.50.1 from the CVS trunk. The ticket
was closed and I just checked it.

http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=3928

Apparently the problem was that '^ *' has been prefixed to all
regexes, but now this change was reverted.

Anyway my patches were not useless, because there is obviously another
new ...ehm .. "effect" I found.

cperl-mode tries to adjust "compilation-error-regexp-alist" (a
variable defined in `compile.el') to it's needs... but without
success, only the original regex-pattern from compile.el or mode-
compile are active.

AFAI can see this already didn't work in emacs22!

putting these lines into .emacs reactivates it:

(defvar cperl-compilation-error-regexp-alist-lanx-patch
  ;; This look like a paranoiac regexp: could anybody find a better
one? (which WORKS).
  '(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\) [^\n]*line \\([0-9]+\\)[\
\., \n]"
     2 3)
   )
  "Alist that specifies how to match errors in perl output.")

(if (fboundp 'eval-after-load)
    (eval-after-load
	"mode-compile"
      '(setq perl-compilation-error-regexp-alist
	     cperl-compilation-error-regexp-alist-lanx-patch)))

Looking into the source of the bundled cperl-mode reveals that it must
have been changed in this respect.

Oddly it has the same revision number like the newest release from
Ilya - 6.2!

Please note that the regex from cperl-mode shouldn't be effected by
the now fixed bug, since it already starts at the beginning of the
line.


@Ilya: I changed the regex slightly, since the errorline can get
broken into two lines when "use diagnostics" is enabled in perl.

Cheers
  Rolf





On 1 Okt., 01:36, LanX <lanx.p...@googlemail.com> wrote:
> Hi
>
> I did a fresh install of gnu emacs 23 on my ubuntu box and everything
> seems to work fine, but when running mode-compile from cperl-mode the
> error messages are not recognized.
>
> That means the errors are not highlighted, clicking has no effect
> ("there is no error") and
> M-x next-error only produces "Moved past last error".
>
> Any idea?
>
> Cheers
>   Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-13, LanX <rolf.langsdorf@googlemail.com> wrote:
> cperl-mode tries to adjust "compilation-error-regexp-alist" (a
> variable defined in `compile.el') to it's needs... but without
> success, only the original regex-pattern from compile.el or mode-
> compile are active.

Could you be more specific about what happens?  It looks like your
patch is doing exactly what my copy of cperl-mode.el does...

> @Ilya: I changed the regex slightly, since the errorline can get
> broken into two lines when "use diagnostics" is enabled in perl.

I will look into it later...

Thanks,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-13, LanX <rolf.langsdorf@googlemail.com> wrote:

>   '(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\) [^\n]*line \\([0-9]+\\)[\
> \., \n]"

> @Ilya: I changed the regex slightly, since the errorline can get
> broken into two lines when "use diagnostics" is enabled in perl.

I tried to find a difference with what is in cperl-mode.el, and
failed.  Where is Waldo?

(For best result, answer using `diff -pu'. ;-)

Yours,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> I tried to find a difference with what is incperl-mode.el, and
> failed.  Where is Waldo?

Oh...sorry!

 I'm very far from being a lisp expert and after 2 weeks of debugging
and quarreling with "emacs-experts" my system is cluttered with emacs
versions and config-files... and furthermore with all these
backslashes elisp-regexes are hard to read at a glance:

---------------
(defvar cperl-compilation-error-regexp-alist2
  ;; This look like a paranoiac regexp: could anybody find a better
one? (which WORKS).
  '(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\)[ \n\t]+line \\([0-9]+\\)
[\\., \n]"
     2 3 nil 0)
   )
  "Alist that specifies how to match errors in perl output.")

(if (fboundp 'eval-after-load)
    (eval-after-load
	"mode-compile"
      '(setq perl-compilation-error-regexp-alist
	     cperl-compilation-error-regexp-alist2)))
------------

"Waldo" is between "filename" and "at" because diagnostics tries to
fit the message into less then 80 characters.

And IMHO the variable needs to be renamed because of conflicts with
the patched version of cperl-mode which already uses the old variable
name.


My testsuite with  emacs22

---------------------------------
lanx@nc10-ubuntu:~$  emacs .emacs error.pl -f cperl-mode -l /home/
lanx/.emacs.d/debug/mode-compile.el -f mode-compile -eval "(describe-
variable 'compilation-error-regexp-alist)"

lanx@nc10-ubuntu:~$ emacs .emacs error.pl -f cperl-mode -l /home/
lanx/.emacs.d/debug/mode-compile.el -f mode-compile -eval "(describe-
variable 'compilation-error-regexp-alist)"
^C
lanx@nc10-ubuntu:~$ cat error.pl
use strict;
use diagnostics;

my $a=A;
my $a=B;
$b=B;
-----------------------------

here the corresponding code in diagnostics.pm which shortens the
errormessage:

----------------------------
lanx@nc10-ubuntu:~$ perldoc -m diagnostics |tail -15

sub shorten {
    my $line = $_[0];
    if (length($line) > 79 and index($line, "\n") == -1) {
	my $space_place = rindex($line, ' ', 79);
	if ($space_place != -1) {
	    substr($line, $space_place, 1) = "\n\t";
	}
    }
    return $line;
}
---------------------------------

Bye
  Rolf

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Sorry again, better take this setup, the last one had an experimental
twistfor marking the severity as "info".

It's now more anal, but works with and without disgnostics, if you
have mode-compile in your load path you don't need the -l option.

------------------------------------------------------------------------------------------
lanx@nc10-ubuntu:~$ cat .emacs
(autoload 'mode-compile "mode-compile"
  "Command to compile current buffer file based on the major mode" t)

(defvar cperl-compilation-error-regexp-alist
  ;; This look like a paranoiac regexp: could anybody find a better
one? (which WORKS).
  '(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\)\\( \\|\n\t\\)line \\
([0-9]+\\)[\\., \n]"
     2 4 )
   )
  "Alist that specifies how to match errors in perl output.")

(if (fboundp 'eval-after-load)
    (eval-after-load
	"mode-compile"
      '(setq perl-compilation-error-regexp-alist
	     cperl-compilation-error-regexp-alist)))
lanx@nc10-ubuntu:~$ cat error.pl
use strict;
use diagnostics;

my $a=A;
my $a=B;
$b=B;

lanx@nc10-ubuntu:~$ emacs error.pl -f cperl-mode -l /home/
lanx/.emacs.d/debug/mode-compile.el -f mode-compile
------------------------------------------------------------------------------------------

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-15, LanX <lanx.perl@googlemail.com> wrote:
> (defvar cperl-compilation-error-regexp-alist
>   ;; This look like a paranoiac regexp: could anybody find a better
> one? (which WORKS).
>   '(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\)\\( \\|\n\t\\)line \\
> ([0-9]+\\)[\\., \n]"
>      2 4 )
>    )
>   "Alist that specifies how to match errors in perl output.")

So the only difference is the group before "line", right?

Thanks,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[Ilya Zakharevich]

On 2009-10-15, LanX <lanx.perl@googlemail.com> wrote:
>   '(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\)[ \n\t]+line \\([0-9]+\\)
> [\\., \n]"

This STILL looks the same as in CPerl's cperl-compilation-error-regexp-alist.

> "Waldo" is between "filename" and "at" because diagnostics tries to
> fit the message into less then 80 characters.

Why not do what I asked and post `diff -pu' between two RExen?

Thanks,
Ilya

Re: In emacs 23 compile-mode doesn't recognize (c)perl error messages

[LanX]

Hi

> This STILL looks the same as in CPerl's cperl-compilation-error-regexp-alist.
...
> Why not do what I asked and post `diff -pu' between two RExen?

lanx@nc10-ubuntu:~/tmp$ diff -pu 1 2
--- 1	2009-10-16 02:43:28.000000000 +0200
+++ 2	2009-10-16 02:43:55.000000000 +0200
@@ -1 +1 @@
-'(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\) [^\n]*line \\([0-9]+\\)[\
\., \n]"
+'(("^[^\n]* \\(file\\|at\\) \\([^ \t\n]+\\)[ \n\t]+line \\([0-9]+\\)[\
\., \n]

cheers
  rolf