Aug. 5, 2009
12:07 p.m.
Will do! Submitted Issue 1041 as a simple case where an (unneccessary) primitive failure happens. If a similar slowdown is experienced in f.ex. the fallback bitblt code used in font rendering, it might be worth investigating what exactly causes it to fail. Fallback code isn't always a good thing when everything still works as expected, but cause performance hits that add up over time, and are hard to track down :) Cheers, Henry On Aug 5, 2009, at 9:42 27AM, Stéphane Ducasse wrote:
Hi henrik
what I suggest is: cut your messages into smaller ones so that we can get multiple and clear thread. Then enter bug items when needed.
Stef