On 04/12/11 12:32, MZMcBride wrote:
This may be a stupid question as I don't understand the mechanics particularly well, but... as far as I understand it, there's a Squid cache layer that contains the HTML output of parsed and rendered wikitext pages. This stored HTML is what most anonymous viewers receive when they access the site. Why can't that be dumped into a output file rather than running expensive and timely HTML dump generation scripts?
In other words, it's not as though the HTML doesn't exist already. It's served millions and millions of times each day. Why is it so painful to make it available as a dump?
Most of the code would be the same, it's just a bit more flexible to do the parsing in the extension, it makes it easier to change some details of the generated HTML, and lets you avoid polluting the caches with rarely-viewed pages. It's not especially painful either way.
-- Tim Starling