I'm stupid, I don't get it.
What is the problem with Ikeb's last filter?

Ikeb,

Have you noticed something in 'TRegExpr' that might help you in another way....provided that the software that was calling the RegExp Engine in MWP were written to use it?

I am talking about the charactor position information that is returned with each match.

A program that was written to use that information properly would be able to specify the actual positional occurences of a RegExp match. Through the use of followup matches, by position, the kind of thing that you are looking for could be easily done.

MWP would have to re-write their filter routine to take advantage of this....but could you imagine the kind of filter power that such a re-write could provide to the user?

Think about it.....

Ikeb,

I am glad that I didn't mistate what you realized...

As far as a multi-level filter goes, the fact that the RegExp engine in MWP does not support 'negative lookahead' could be remedied by a filter that does a second comparison of any phrase matching the first level RegExp.

For instance....

If the first stage of the RegExp looked for:

you could have a follow on RegExp (as a subrule to the top level rule) that looked for the following RegExp within any matching phrase found in the top-level:

Code:

</?(br|blockquote)>

You could place all the 'legal' HTML in that subrule.

If the second level RegExp is found within any of the first level matches, the subrule could be designed to treat it as a positive or a negative.

In the above case, you would want to treat it as a negative...in other words...if the second level expression is found, the filter will not trigger on that first level match.

The filter could then continue its way through looking for any further matches of the top-level rule.

Would something like this be of any help to you?

By the way....Happy Thanksgiving all

Might want to look at the folks that created the regex code that mailwasher uses (if I understood that correctly) and getting them to add the code there at the source and then getting mailwasher to pick up the changes.

The fellow who receives credit by FireTrust is one Andrey V. Sorokin who lives in St-Petersburg, Russia. FireTrust's credit references a web site which was giving Perry some grief a couple of weeks ago. It looks like Andrey hasn't yet cleaned up the site because I still get an MS Security popup asking if it's OK to install a certain 'Precision Time' program from GAIN Publishing (Terms and Conditions reference the Gator Corporation, an infamous Spyware generator).

The web site offers the TRegExpr utility. The last version of TregExpr was issued in October 2001, so it would appear Andrey is no longer actively developing his software.

Thanks for the program install warning, I never see that type of thing since I don't use explorer unless I'm desperate.

Has anyone tried to contact Andrey to see if he is interested or if the site is truly abandonded? If it is the firetrust crew might want to take some action to duplicate the information there elsewhere in case the site goes away someday.

It would be much easier to modify the calling routines in MWP than it would be to modify the RegExp engine.

From the looks of things, MWP is currently only looking for a TRUE-FALSE return from the RegExp routines used.

From what I see when I run 'TRegExpr', the RegExp routines will return not only the text that matches the RegExp, but the start and end position within the search string that the matching text was found.

With that information returned to the filter routine (the calling routine) it would not be all that difficult to have that calling routine do a second stage of filtering...the 'sub-rule' that I mentioned previously.

The only thing that might hinder such an endeavor by MWP would be if the license agreement between them and Andrey V. Sorokin prohibited them from pursuing this.

If it does not prohibit this, all MWP would have to do would be to modify the 'Edit Filter' GUI to allow a sub-rule to any rule specified. That sub-rule would simply be: