#8007: [app_server] fully unresponsive when resizing window of big textfile -----------------------------+---------------------------- Reporter: ttcoder | Owner: axeld Type: bug | Status: new Priority: normal | Milestone: R1 Component: System/Kernel | Version: R1/Development Resolution: | Keywords: Blocked By: | Blocking: 7882, 8136 Has a Patch: 1 | Platform: All -----------------------------+---------------------------- Comment (by Pete): Replying to [comment:54 Pete]: I've done a little more investigation, and I probably shouldn't have been quite so pessimistic. In particular: > > Finally, I didn't find the hoped-for improvement in audio glitching. Moving a window still causes just as many 'crackles' as before, despite the supposed higher priority of audio. I was using Csound for this, which was not the best choice. I have to use a very small audio buffer (64 bytes) to keep the latency low. (Not the audio-chain latency -- this is Csound's fault, in the way it generates the audio.) My MusicWeaver/fluidsynth stuff can use a more sensible buffer size (960 bytes), and has much better behaviour. Most of the time... I find that I can fire up a live music configuration, and play fluidsynth with *no* audible glitches, even when I move windows around. This is vastly different from my experience without the patch, where window movement causes continual crackles. In this condition, it's the app_server (I guess) that slows down! Only when the screen contains windows belonging to the audio app, though; other workspaces seem normal. Window movement starts lagging behind the cursor. I can understand this if the audio threads are getting preference, except that the total load on the CPU (shown by 'top') doesn't seem excessive -- about 7-10%. And, as I say, it's only when related windows are visible. In any case it's much preferable to ugly audio! ''However''... things don't always seem to work this way. I seem to still be able to get into a situation where window movement interferes with audio, and it then continues this way, much like without the patch. I haven't been able to determine what causes the switch, but I'm going to run the patched system as standard for a while, to see if I can figure out what's happening. -- Ticket URL: <http://dev.haiku-os.org/ticket/8007#comment:55> Haiku <http://dev.haiku-os.org> Haiku - the operating system.