Strange results from lwarn + match-string

Strange results from lwarn + match-string

[Lennart Borgman (gmail)]

Can anyone see anything wrong with this line:

   (lwarn t :warning "1: %s" (match-string-no-properties 1))

I get some very strange result sometimes. If I use (message ...) instead 
the result looks ok.

Re: Strange results from lwarn + match-string

[Andreas Schwab]

"Lennart Borgman (gmail)" <lennart.borgman@gmail.com> writes:

> Can anyone see anything wrong with this line:
>
>   (lwarn t :warning "1: %s" (match-string-no-properties 1))
>
> I get some very strange result sometimes.

Please define "strange".

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Strange results from lwarn + match-string

[Lennart Borgman (gmail)]

Andreas Schwab wrote:
> "Lennart Borgman (gmail)" <lennart.borgman@gmail.com> writes:
> 
>> Can anyone see anything wrong with this line:
>>
>>   (lwarn t :warning "1: %s" (match-string-no-properties 1))
>>
>> I get some very strange result sometimes.
> 
> Please define "strange".

I was running the function below (I am trying to export to FreeMind). I 
am in the .notes file buffer when I exectute the function. The first 
lines in this file contains

   * Notes
   ** Wed Oct 24 22:13:56 2007 (Output filter for Thunderbird?)

This is what I get:

   Warning (t): 0: * Notes
   Warning (t): 1: t 24
   Warning (t): 2: nil
   Warning (t): 3: nil


Below is the function. I also get some crashes, I believe that was when 
I had `anything' instead of `nonl' in the pattern. There was some other 
strange errors too, but maybe this is enough? I have tested this with 
unpatched Emacs too:

   GNU Emacs 23.0.50.1 (i386-mingw-nt5.1.2600) of 2008-01-27




(defun org2mm-buffer-to-freemind ()
   (interactive)
   (unless (derived-mode-p 'org-mode)
     (error "Not an org-mode buffer"))
   (save-match-data
     (let* ((default-mm-file (concat (file-name-sans-extension
                                      (file-name-nondirectory
                                       (buffer-file-name)))
                                     ".mm"))
            (mm-file (read-file-name "Output FreeMind file: " nil nil 
nil default-mm-file
                                     (lambda (fn)
                                       (string-match "^mm$"
 
(file-name-extension fn)))))
            (node-pattern (rx bol
                              (submatch (1+ "*"))
                              (1+ space)
                              (submatch (*? nonl))
                              eol)))
       (message "%s" mm-file)
       ;;(save-excursion
       (goto-char (point-min))
       (when (re-search-forward node-pattern)
         (lwarn t :warning "0: %s" (match-string-no-properties 0))
         (lwarn t :warning "1: %s" (match-string-no-properties 1))
         (lwarn t :warning "2: %s" (match-string-no-properties 2))
         (lwarn t :warning "3: %s" (match-string-no-properties 3))
         )
       ;;)
       )))

Re: Strange results from lwarn + match-string

[Stefan Monnier]

>         (lwarn t :warning "0: %s" (match-string-no-properties 0))
>         (lwarn t :warning "1: %s" (match-string-no-properties 1))
>         (lwarn t :warning "2: %s" (match-string-no-properties 2))
>         (lwarn t :warning "3: %s" (match-string-no-properties 3))

Since you don't know for a fact that lwarn is among the few functions
that do not affect match-data, you need to call all
4 `match-string-no-properties' before you call the 4 `lwarn'.


        Stefan

Re: Strange results from lwarn + match-string

[Lennart Borgman (gmail)]

Stefan Monnier wrote:
>>         (lwarn t :warning "0: %s" (match-string-no-properties 0))
>>         (lwarn t :warning "1: %s" (match-string-no-properties 1))
>>         (lwarn t :warning "2: %s" (match-string-no-properties 2))
>>         (lwarn t :warning "3: %s" (match-string-no-properties 3))
> 
> Since you don't know for a fact that lwarn is among the few functions
> that do not affect match-data, you need to call all
> 4 `match-string-no-properties' before you call the 4 `lwarn'.

Hm. It cured the problem, but that was really a surprise.

Re: Strange results from lwarn + match-string

[Stefan Monnier]

>>> (lwarn t :warning "0: %s" (match-string-no-properties 0))
>>> (lwarn t :warning "1: %s" (match-string-no-properties 1))
>>> (lwarn t :warning "2: %s" (match-string-no-properties 2))
>>> (lwarn t :warning "3: %s" (match-string-no-properties 3))
>> 
>> Since you don't know for a fact that lwarn is among the few functions
>> that do not affect match-data, you need to call all
>> 4 `match-string-no-properties' before you call the 4 `lwarn'.

> Hm. It cured the problem, but that was really a surprise.

There are very few functions which guarantee that they preserve
match-data, so if there's the slightest doubt, assume that match-data
will be destroyed,


        Stefan