moonburnt / fortune_crawler Goto Github PK

Currently Python-like formatting is blocked by this issue: https://reviews.llvm.org/D109557

While testing ec9aa38, I've discovered a bug that seems to be related to chests looting - sometimes after doing so, game segfaults. It doesn't seem to be caused by lockpicking minigame implemented in this patch, but rather by something related to non-locked chests. Thus far, I've managed to catch this bug 3 times, always about 10+ levels in game. And thus far, I couldn't pin-point the exact cause of it

https://reviews.llvm.org/D37143

Yesterday I've updated my arch linux setup and now its impossible to compile the game with gcc due to errors in gcc's string implementation - it starts fine, but fails on compilation of mapgen.
So... yeah - don't use gcc 12.1.0 for now. Or at least the one from arch repos

Related issue: raysan5/raylib#2415

Right now, toggling fullscreen make things go bananas, and amount of issues vary depending on platform (on windows it may not work at all, or require about 10 seconds to switch to this mode. And on all platforms GetScreenWidth() and GetScreenHeight() break after fullscreen mode has been toggled once).
I'm not sure if I can fix it by myself, since its an engine issue. Workaround can be implemented tho

If png used as map template has different width and height values, mapgen goes wild and produces broken in-game map.

For example - the following map template (7x13):

Produces following results:

The issue may be caused by incorrect transition from map's width/height to map_size.x/map_size.y... somewhere. But, honestly - I'm struggling to determine the exact piece of code that causes trouble

With commit 2c275f0, I've first encountered a random corrupted size vs. prev_size in fastbins segfault on exit or scene switch.
This may or may not be related to handling of level's restart() functionality, but since it does not happen consistently - I call shenanigans and ask for help.