Re: Can't debug without source code.

On Fri, Sep 24, 2010 at 12:56:59PM -0400, John E. Davis wrote:
> Brian Murray <[email protected]> wrote:
> > I was sort of wondering how I could proceed in debugging it myself but
> > of course I am happy to have others look at it too.
> 
> One of the things that I would like to do is to add support for the
> sldb debugger to slrn.  I believe that it can be accomplished without
> too much effort.
> 
> > You'll find it attached.  I've also added in comment that is an example
> > of a subject that was not threaded.
> > define subject_compare_hook ( subject1, subject2 )
> >{
> > variable bracket_pattern;
> > variable parens_pattern;
> > variable hyphen_image_pat;
> > variable bracket_matches, bracket_matches2;
> > variable parens_matches, parens_matches2;
> > variable hyphen_image_matches, hyphen_image_matches2;
> > variable replacement, replacement2;
> >
> > bracket_pattern = "\\[[0-9]+ ?\\(?:of\\) ?/\\([0-9]+\\)\\]";
> > parens_pattern = "([0-9]+/[0-9]+)";
> > hyphen_image_pat = " ?\\([0-9]+[ -_]?\\)[0-9]+\\(.[jJ][pP][gG]\\)";
> 
>    Are these slang regular expressions?  I suspect that
>    bracket_pattern is trying to match strings such as
> 
>       "[17 of 89]"
>       "[17/89]"
> 
>    If so, then bracket_pattern will not work when interpreted as a
>    slang RE.  You might want to use the onigurama or pcre modules for
>    more powerful REs.  Alternatively, you can try:
> 
>       bracket_pattern = "\[[0-9]+ ?[o/]f? ?\([0-9]+\)\]"R;
> 
>    The above pattern will match
> 
>       "[17 of 89]"
>       "[17/89]"
> 
>    as well as the false positives:
> 
>       "[17 /f89]"
>       "[17o89]"
> 
>    It looks like you also want to match the forms using "(".  If so,
>    use:
> 
>       bracket_pattern = "[[(][0-9]+ ?[o/]f? ?\([0-9]+\)[])]"R;
> 
> > bracket_matches = string_matches(subject1, bracket_pattern, 1);
> > if ( bracket_matches[0] != NULL )
> 
> The above works only because upon failure, bracket_matches will be
> NULL, which is a scalar that will be interpreted in the above context
> as an array with a single element.  It is better to simply use
> 
>    if (bracket_matches != NULL) ...
> 
> >{
> >     ( subject1, ) = strreplace(subject1, bracket_matches[0], bracket_matches[1], 1);
> 
>   Here, you are replacing 1 occurance of bracket_matches[0] in
>   subject1 with bracket_matches[1].  So you are replacing the entire
>   string of characters matched by bracket_pattern with the characters
>   matched by the "?:of" bit of the pattern.  Since "?:of" looks a bit
>   odd I suspect that you are not using a slang RE.
> 
> [...]
> > % check and see if the stripped subjects now match
> > if ( 0 == strcmp(subject1, subject2) )
> 
>   You can also use:
> 
>     if (subject1 == subject2)
> 
> I hope this helps.

Yes, it did some.

Debugging this further the issue occurs with subjects that have 2 sets
of ([0-9]+/[0-9]+) in them.  Consider the following subjects:

Austro.Pop.Show-2006-Zehn.(10/52) 'Austro.Pop.Show-2006-Zehn - 105 -
Rainhard Fendrich - Schickeria.mp3' yEnc (22/23)
Austro.Pop.Show-2006-Zehn.(10/52) 'Austro.Pop.Show-2006-Zehn - 105 -
Rainhard Fendrich - Schickeria.mp3' yEnc (23/23)

Because I am matching on the 10/52 section and stripping that out the
rest of the section won't match.  What I really want is to be able to
pass string_match a -1 so the string is scanned from the end not the
front.

I've also tried using pcre and pcre_matches and that seems to only
return the first match not multiple matches.

pcre_parens = pcre_compile("\\(\\d+/\\d+\\)");
slsh> pcre_matches(pcre_parens, subject1);          
String_Type[1]
slsh> pcre_matches(pcre_parens, subject1)[0];
(10/52)

Any ideas are extremely welcome.

Thanks,
-- 
Brian Murray

------------------------------------------------------------------------------
Virtualization is moving to the mainstream and overtaking non-virtualized
environment for deploying applications. Does it make network security 
easier or more difficult to achieve? Read this whitepaper to separate the 
two and get a better understanding.
http://p.sf.net/sfu/hp-phase2-d2d

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)

iEYEARECAAYFAkyqAJ8ACgkQDTAwc5ER+zXzrACgoEqVUkGQWMtdXQFxhsLytZiq
yZMAn1g5rLarpzqAkJU5+6vZ213YArlo
=u4B0
-----END PGP SIGNATURE-----