Open mattdm opened 1 year ago
For what it's worth, it seemed OK when run with sdl12-compat on Debian 12, when I tried to test all the remaining SDL 1.2 games in Debian.
FWIW, whatever is happening is happening very early, in the call to SDL_SetVideoMode()
.
I would like to emphasize again the possibility that I did something very stupid in the code, or in the hand-crafted makefile.
making everything up and not knowing how to do memory management in C so it's just a bunch of arrays. :)
Honestly, I have such an affection for these sort of programs, because you powered through it and built a whole game anyway.
The icebrea:disk$0
thread is the Mesa shader cache, so the stack trace is probably a red herring, and memory is corrupting elsewhere.
I couldn't reproduce it on Ubuntu (and valgrind reports no incorrect memory accesses), but I'll try to get this running on Fedora a little later on and see if something goes south for me.
Over 20 years ago, I wrote a game called Icebreaker. People still like it! Someone already addressed one issue with it here (amazing!) — #68
However, with Fedora Linux 35 it started having hard-to-reproduce crash-on-startup issues. As of F38, it seems to crash on startup all the time, so... progress?
I will freely admit this is not good code. I should re-write it all -- it still might not be good code, but at least I'll be using Bevy Engine and Rust rather than making everything up and not knowing how to do memory management in C so it's just a bunch of arrays. :)
But, until that time, which may never come.... it's not working, and figuring it out alone is beyond me.
I get...
User reports at https://discussion.fedoraproject.org/t/icebreaker-on-fedora-35-and-sadly-up/35254/7 and https://bugzilla.redhat.com/show_bug.cgi?id=2048689
If someone could help, that would be amazing.