libobs-opengl: Fix incompatible pointer type error for 32-bit arches

[Conan-Kudo]

Description

Fix a small pointer type error in libobs-opengl

Motivation and Context

It fixes a build for the library on 32-bit architectures.

How Has This Been Tested?

OBS Studio packages in Fedora has been shipping this since Fedora 40.

Types of changes

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• My code has been run through clang-format.
• I have read the contributing document.
• My code is not on the master branch.
• The code has been tested.
• All commit messages are properly formatted and commits squashed where appropriate.
• I have included updates to all appropriate documentation.

[tytan652]

I'm quiet sure that 32 bit support was dropped since the complete switch to Qt6.

[kkartaltepe]

Our api expects a size_t pointer so callers will be sending in pointers to size_t. You will need to copy the size_t into a correctly sized variable first for this to be correct.

[Lain-B]

Is there any way to exclude 32bit builds? We generally just don't support 32bit right now. Because this cast is technically erroneous on 32bit and will cast a 32bit integer (size_t on 32bit) to 64bit (EGLuint64KHR which presumably is 64bit even on 32bit)

[Conan-Kudo]

Honestly, the only reason I care is because of the vkcapture plugin... It needs a 32-bit libobs because it needs a 32-bit build for 32-bit games. ☹️

[norihiro]

beside if 32-bit architecture is supported or not, casting a 32-bit integer pointer to a 64-bit integer would result in a buffer overrun. You should have a 64-bit integer local variable and pass the pointer of it. Then, cast the value of the variable to n_modifiers.