On 02/26/2020 02:55:37 PM Wed, Peter Bloomfield wrote:
Hi Albrecht, On 02/26/2020 01:42:55 PM Wed, Albrecht Dreß wrote:Hi Peter! Am 25.02.20 18:39 schrieb(en) Peter Bloomfield:If you get a chance, could you get back traces for all threads?As it's actually easy to reproduce, here is the latest CPU hog:… The other threads all look well behaved, so it must be the main thread. Could you run with the attached patch? If the issue is one of the recent changes, you'll get messages spewed all over the console. If not, it must be some other idle handler that's somehow been impacted by them, which will be harder to find!
Running with that patch, I just saw a flood of messages from lbm_sort_idle_cb(), which waits for threading to complete before sorting. There were enough to scroll off the terminal buffer (10 000 lines), but the storm was over before I got back to the terminal window. Instead of an idle callback, it could be a timeout of, say, 10 milliseconds; that would be an imperceptible delay, and even 100 msec would probably not be noticeable. But it's not clear whether that's the problem you're seeing! Peter
Attachment:
pgpxNjtlGolwb.pgp
Description: PGP signature