ninja
ninja
https://github.com/bambulab/BambuStudio/issues/2040 https://github.com/bambulab/BambuStudio/issues/2846 upstream reports
ah, I think my diagnosis was not correct, because AsyncWebSocketMultiMessage is created per-connection and deleted, and not the buffer itself. the crash under > 8 clients is coming from some...
I guess this will be fixed after manually applying this set of patches https://github.com/Aircoookie/WLED/issues/3382#issuecomment-1751363805, which I will try soon
see #1363 to see the code that causes mentioned crash
looks like the fork also suffers from this bug
after closer look, the fork will crash but after much higer load as you said...it might be related to active number of connections. maybe in between calls ws.cleanupClients which I...
I believe this will be fixed after applying this set of patches https://github.com/Aircoookie/WLED/issues/3382#issuecomment-1751363805, which I plan to do soon on top of nthd fork
I've done some more digging, which might be helpful to someone. After getting the exception decoder working, abort during stress tests was always coming down to lack of memory so...
@hoeken that's some really nice code - looks like we've been duplicating effort of porting the websockets demo code to c++, a quite slow and painful process :)
I've got caught up in a couple issues today: - Arduino library shipped suffers from bug where c_str() can return NULLPTR under memory pressure. Took me a while to debug...