Rob Kendrick e7c5e16b9b Performance improvements: rather than calling content_clean() every poll, we now call it no more frequently than once every 5 seconds. Additionally, we cache the result of talloc_total_size() in content_clean() rather than calculating it twice. On large documents, this function took 25% of CPU time. This makes the fetching/rendering/scrolling/redrawing of large documents over twice as fast. ...

svn path=/trunk/netsurf/; revision=4527

2008-07-07 14:05:29 +00:00

..

fetchers

…

content_type.h

…

content.c

Performance improvements: rather than calling content_clean() every poll, we now call it no more frequently than once every 5 seconds. Additionally, we cache the result of talloc_total_size() in content_clean() rather than calculating it twice. On large documents, this function took 25% of CPU time. This makes the fetching/rendering/scrolling/redrawing of large documents over twice as fast.

2008-07-07 14:05:29 +00:00

content.h

Performance improvements: rather than calling content_clean() every poll, we now call it no more frequently than once every 5 seconds. Additionally, we cache the result of talloc_total_size() in content_clean() rather than calculating it twice. On large documents, this function took 25% of CPU time. This makes the fetching/rendering/scrolling/redrawing of large documents over twice as fast.

2008-07-07 14:05:29 +00:00

fetch.c

…

fetch.h

…

fetchcache.c

…

fetchcache.h

…

urldb.c

…

urldb.h

…