Wedge009
Wedge009
I suspect it's related to the state of a system's memory at a given point in time. At the time of writing, I can 100% replicate this on my primary...
Interesting. I tried again on a third machine, release 1.17.18, Windows (as opposed to Linux), no add-ons aside from GUI Debug Tools, and debug-jumped to SotA S8 Carcyn. The problem...
At 1920x1200 window size? That's the only size I could reliably replicate it, on two machines as well (Windows 7 + Linux/X11).
Oh well, it was worth a shot, since I was surprised that I found a case for me where I could consistently replicate the issue on more than one machine....
Since I can consistently replicate this, I belatedly realised I can try to bisect. Assuming I'm replicating this consistently, I found the problem was introduced with 06254e5581cbb79b5eaf72bec0dbb3fc1566b7a4 (included in 1.17.6),...
Confirmed - on the assumption that my Carcyn save is still a valid replication scenario for the original issue - this is resolved in #7982.
May not necessarily be the same issue, but I just got the following while attempting to replicate 6879 (add-ons menu) on 1.17.24 release on Windows 10.  I am able...
The 'artefacts' looks like a slight misalignment of the hex images at certain zoom levels. I observe the issue in current master (314f6f7c78d), but not in 1.16 branch.
A git bisect suggests this regression started in 46750666b25704b97766bc691a199dee111b56f6, introduced with 1.17.5.
At an abstract level, likewise without knowing the technical details of the implementation, that does sound like a good approach. @mesilliac Sorry to bother you, but noticed a few comments...