when profiling window activate, i found that 6% of the time to process the activate event is spent peristing size and position of the window. do we really need to do it on _every_ activate event?
Why don't we do this at window close time?
Doing it at close is not what users expect or want, but we should be able to avoid a write to disk every time. ->moz1.0/perf
Target Milestone: --- → mozilla1.0
This is one of the worst responsiveness bugs we have, and is a dependency of one of the most-voted-for nscatfood bugs. Any chance we could make progress on it for 0.9.1?
Priority: -- → P2
Whiteboard: wanted for 0.9.1/Mojo beta
This didn't show up on my Windows window activate profiling at all. Could this really be mac only? pink, how'd you test window activation?
Status: NEW → ASSIGNED
Target Milestone: mozilla1.0 → mozilla0.9.1
i used the MW profiler, testing the time spent processing the activate event.
See patch and time comparisons in bug 79060
Patch from bug 79060 is in (0.9.2). So now PersistPositionAndSize should (if times on the Mac and on Windows are relatively similar) be about 1% of window activation, unless some rude person has already sped up the rest of that process.
Status: ASSIGNED → RESOLVED
Last Resolved: 18 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.