@CrackedPixel Please, could you take a look to this PR? It surprises me that RGFW_init() was not called before... maybe it's not being tested against right RGFW version?

@CrackedPixel Please, could you take a look to this PR? It surprises me that RGFW_init() was not called before... maybe it's not being tested against right RGFW version?

it usually automatically does it internally on createwindow for convenience
but we can't rely on that if we want the display size before that, it wouldn't be initialized yet and would segfault
so we need the explicit call

I don't really like a lot of this PR but i will look at it today and compare to glfw and other platforms to see about a good solution

@keks137 i made a PR for suggestions to your branch: keks137#1

the changes seemed reasonable
I also just noticed something, that's what cbc5a1f is for
I missed that I was doing the FLAG_FULLSCREEN_MODE check twice, because I just copied glfw's version and worked from there, but the rgfw path already had a check above
they definitely should just be one check

LGTM! nice work 😎

actually one thing to note is that the return value of RGFW_init is rather weird:

from RGFW's source code:
/**!

• @brief Initializes the RGFW library.
• @return 0 on success, or a negative error code on failure.
• @note This is automatically called when the first window is created.
  */

but if you look at the internal code, when you call it but it is already initialized it returns 1 - which is kind of weird but technically still correct within the api

where it gets truly weird is that it actually doesn't always return a negative number on error
for example in the wayland init there's this bit of code

if (_RGFW->compositor == NULL) {
		RGFW_sendDebugInfo(RGFW_typeError, RGFW_errWayland, "Can't find compositor.");
		return 1;
	}

this means we can't just check < 0, unlike the comment implies

for the purposes of raylib, breaking on multiple windows is technically not an issue, since that's not supported anyways, right?
#1989 (comment)

this might however still be an issue on
InitWindow
CloseWindow
and then
InitWindow
again
I'm not sure if that is something raylib supports
(one thing of note, this currently works in glfw, but segfaults in rgfw even without this specific issue, meaning I also can't check if this specifically actually doesn't work, but I'll assume it doesn't)

so maybe it would be better to ignore the return value entirely? that's what rgfw does whenever calling it internally
this return value is definitely a janky part of rgfw, maybe I should also open an issue there

@keks137 thanks for the improvement and thanks @CrackedPixel for the review

it seems we will be able to remove the RGFW_init call entirely in a future update of RGFW, the fact that we need it right now is a bug ColleagueRiley/RGFW#523 (comment)