[squeak-dev] MNU and then segfault

Lauren Pullen drurowin at gmail.com
Wed Jun 29 05:00:19 UTC 2022


Hi Marcel, Karl,

On 6/27/22 07:49, Marcel Taeumel wrote:
> Hi Lauren --
> 
> Can you reproduce this issue? If so, please report here:
> https://github.com/squeak-smalltalk/squeak-object-memory/issues [https://github.com/squeak-smalltalk/squeak-object-memory/issues]
> 
> 
> Hmm... maybe some incorrect values reported via
> DisplayScreen class >> #actualScreenSize (primitive 106)
> DisplayScreen class >> #actualScreenScaleFactor (#primitiveScreenScaleFactor)
> 
> 
> I don't now. Such screen/scaling-related issues benefit from reporting:
> 
> - source screen size
> - source screen scale factor
> - target screen size
> - target screen scale factor
> 
> The crash.dmp is not enough.
> 
> Best,
> MarcelI cannot reproduce this by changing scale factor & screen size.  I have
a morph reporting those DisplayScreen values live while I plug and
unplug the external display, and, while not happy about being switched
between 1024x718 @ 1x and 2880x1708 @ 2x, it doesn't crash.

I saw it tried scavenging a weak array recently, as well.  Maybe that
was the cuplrit... ...  Testing round two.

I think there's an issue with Mac programs. I've noticed a poor HDMI
connection can toggle single and multiple monitor mode rapidly and that
randomly crashes programs.  I also reliably crash VLC with multiple
monitor mode... I have to wait to fullscreen a video to the external
monitor a second or so after playback begins or it will crash.

Am 26.06.2022 10:28:25 schrieb karl ramberg <karlramberg at gmail.com>:
> Damage rectangle bounds are wrong.
> They have weird values and they are not a rectangle.
> bounds: -9007198183096320 corner: -9007198183096320 at 308
> 
> 
> Best,
> 
> Karl

I noticed that as well.  I presume that was a side effect of it being
about to crash.


More information about the Squeak-dev mailing list