rootston: fix segfault in view_at#1341
Conversation
|
Thanks! SDL2 still doesn't really work, but rootston doesn't segfault anymore, so I guess #1333 is fixed. I'll open a new ticket ;) |
|
Tried with 0AD which also makes rootston crash, and it fixes the issue. |
|
Oh, I spoke too early - it just segfaulted. Let me recheck. |
|
|
FWIW, SDL also outputs |
|
Hmm, can you run |
|
Segfaults are non-deterministic now :( Now got this one: |
|
Another one: |
|
@dos1 can you try compiling with ASAN enabled, then redirect stdout & stderr to a file and try to reproduce the crash? |
|
Make sure the view is mapped.
4296cb2 to
0c73712
Compare
|
Pushed a fix, can you try again? You'll now likely hit the error, which is probably a SDL issue. We're more strict than KWin when it comes to Wayland protocol conformance. |
|
ASAN doesn't complain anymore and rootston hasn't segfaulted so far 👍 The window doesn't appear at all and SDL shows that xdg_surface error as predicted, but yeah, that's a different issue. |
|
While SDL2 apps work fine on KWin and Mutter, they fail the same way like on Sway/rootston also on Weston, so I guess you're right and it's likely a protocol violation on SDL2 side. Other than that, the segfaults seem to be fixed now, thanks! :) |
|
Turns out SDL 2.0.9 already fixes that as well, so everything works great now \o/ |
|
Thanks! |
Make sure the view is mapped.
@dos1: can you try this patch?
Fixes #1333