On Aug 16, 2006, at 9:16 AM, Michel Fortin wrote:
Le 16 août 2006 à 9:39, Paul M Jones a écrit :
So the functional/procedural PHP-Markdown is a little more than
twice as fast as plugin-capable Solar_Markdown. If you transform
the source text anew on every page load, this might be an issue;
if you cache the transformed text for subsequent page loads, it's
probably not.
Interesting numbers. PHP Markdown has been optimized quite much,
it'd probably be hard to beat, speed-wise, for another PHP
implementation.
Yeah, the use of multiple classes in separate files does create more
overhead than using only functions in one file.
I'm preparing a version of PHP Markdown which is more extendable
too, and it'll probably be as fast as the current version.
We can certainly benchmark it once it's ready; I'd be happy to
include your final product in the Solar_Markdown benchmarks for
comparison.
I'm also changing the way span elements are parsed to prevent
**things *like** this* from generating invalid markup. It may or
may not be slower once ready since I'm also removing other code
obsoleted by the process.
This would be very nice; it's been a common request for Text_Wiki.
Interesting: I just found out that Solar Markdown is using the HTML
Tidy library (built into PHP 5) to fix bad XHTML at the end of the
process. That's clever. How come you've not mentioned it before? :-)
Heh. :-) It's optional, and turned off by default iirc. I have it
mostly to standardize whitespace, not to fix HTML. (Hm, might also
be a solution to the "invalid markup" problem you note above.)
-- pmj_______________________________________________
Markdown-Discuss mailing list
[email protected]
http://six.pairlist.net/mailman/listinfo/markdown-discuss
