The 2.X code is much cleaner.

I went more of the "plugin" route with Junct, that collab tool I mentioned. By that I basically mean you write a class instead of a function for new behavior. Example signature:

class TableRule(RegexRule):\n    def process(self, topic, content, UI):

Point being I tried to make it easy to write a new module when you wanted new parsing functionality. Still just a callable. But then it was expected with heavyweight parsers (like Calendar, Progress tracking, etc) that the "plugin" module would also include the UI code for new pages.

In addition, I had a need to do multiple passes on the content, and allow some handlers to "lock down" the portions of the content which they altered, so they couldn't then be munged by a later handler. I ended up making a LockableString class with finditer() and sub() methods (a la RE). Plus there were some handlers which a simple regex couldn't handle--Wiki-style formatting stuff.

Hmm... that's been running stably for a month now here at work. I should probably put it up somewhere...

Post #148,392 by admin 3/24/04 10:48:22 PM Reply	The 2.X code is much cleaner. I've got older versions for 2.1 and 2.2 around here somewhere. The code is a Python rewrite (plus additions) of Ben Tilly's perl HTML filter. The code is somewhat messy, since it was just done to prove a point to Ben. :-) Come to think of it, I've got a much nicer ObjectiveC version too. The nice thing about it is the ability to pass in your own set of handlers. Regards, -scott anderson "Welcome to Rivendell, Mr. Anderson..."
Post #148,399 by FuManChu 3/24/04 11:08:14 PM Reply	Dig that funky extensibility groove. :) I went more of the "plugin" route with Junct, that collab tool I mentioned. By that I basically mean you write a class instead of a function for new behavior. Example signature: class TableRule(RegexRule):\n def process(self, topic, content, UI): Point being I tried to make it easy to write a new module when you wanted new parsing functionality. Still just a callable. But then it was expected with heavyweight parsers (like Calendar, Progress tracking, etc) that the "plugin" module would also include the UI code for new pages. In addition, I had a need to do multiple passes on the content, and allow some handlers to "lock down" the portions of the content which they altered, so they couldn't then be munged by a later handler. I ended up making a LockableString class with finditer() and sub() methods (a la RE). Plus there were some handlers which a simple regex couldn't handle--Wiki-style formatting stuff. Hmm... that's been running stably for a month now here at work. I should probably put it up somewhere...
Post #148,404 by admin 3/24/04 11:27:34 PM Reply	Why functions and regexes (new thread) Created as new thread #148403 titled [link\|/forums/render/content/show?contentid=148403\|Why functions and regexes] Regards, -scott anderson "Welcome to Rivendell, Mr. Anderson..."
Post #148,402 by ben_tilly 3/24/04 11:26:24 PM Reply	And the point was proven... "good ideas and bad code build communities, the other three combinations do not" - [link\|http://archives.real-time.com/pipermail/cocoon-devel/2000-October/003023.html\|Stefano Mazzocchi]

Welcome to IWETHEY!