On 7/22/09 6:08 AM, Christian Boos wrote:
>> >  === wiki ===
>> >
>> >  The current wiki parser design is slow, doing the same large regex over
>> >  each line.  I think a general redesign to use re.finditer, rather than
>> >  line split then re.match, would help improve performance here.  However,
>> >  post-caching the content would be better.  I'm experimenting with
>> >  partial caching of the wiki content and have reduced my request timing
>> >  on WikiStart from 120ms to 75ms while still getting correct content.
>> >  The patch I have doesn't cache macro's unless the macro arguments have
>> >  'cache' in them (good for page outline macro, which btw seems to
>> >  re-parse the entire page, macros included).  There may be other issues
>> >  with the approach I have taken, I haven't tested it much.  Once I get
>> >  further, I might actually pre-process when the wiki page is saved and
>> >  stash the result into the database, avoiding the parse in most requests.
>> >
>
> Interesting as well, be sure to read the discussions about caching wiki
> content (#7739 and #1216).
>

I've attached (to #1216) the cache idea I am playing with right now for 
others to see.

Shane

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Trac 
Development" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/trac-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to