It seems for unknown reasons in miral::WindowInfo::constrain_resize the ar.width == 0 (pretty sure due to default value). Is getting past the error bit which seems from my read can only happen is the new_size.width < 0. So its pretty much doing 0 - -width * max_uint which will get past the error check.
From the logs them selfs I dont see *any* crazy numbers that would cause the width to become negative.
A fix is up, and a test that confirms that negative width bit but unsure how it got to be in that state.
miral-kiosk crashing specificity with a mir-kiosk-apps fancybrowser (or any webkit example)
Stacktrace: paste.ubuntu. com/25523431/
http://
miral logging: paste.ubuntu. com/25530436/
http://
qt logging: paste.ubuntu. com/25530449/
http://
It seems for unknown reasons in miral:: WindowInfo: :constrain_ resize the ar.width == 0 (pretty sure due to default value). Is getting past the error bit which seems from my read can only happen is the new_size.width < 0. So its pretty much doing 0 - -width * max_uint which will get past the error check.
From the logs them selfs I dont see *any* crazy numbers that would cause the width to become negative.
A fix is up, and a test that confirms that negative width bit but unsure how it got to be in that state.
https:/ /code.launchpad .net/~brandonts chaefer/ mir/dont- divide- by-zero- miral/+ merge/330702