map_ninepatch_axis() fix

[lucypero]

Not all control paths return in the function map_ninepatch_axis(). Removing an if fixes it.

[akien-mga]

What does this actually fix? Are you just trying to solve a compiler/static checker warning?

Those two uniforms are set via NinePatchRect's axis_stretch_{horizontal,vertical} properties, which indeed only take values 0 (stretch), 1 (tile) and 2 (tile fit):

godot/scene/gui/nine_patch_rect.cpp

Lines 82 to 84 in 97ebde4

	ADD_GROUP("Axis Stretch", "axis_stretch_");
	ADD_PROPERTY(PropertyInfo(Variant::INT, "axis_stretch_horizontal", PROPERTY_HINT_ENUM, "Stretch,Tile,Tile Fit"), "set_h_axis_stretch_mode", "get_h_axis_stretch_mode");
	ADD_PROPERTY(PropertyInfo(Variant::INT, "axis_stretch_vertical", PROPERTY_HINT_ENUM, "Stretch,Tile,Tile Fit"), "set_v_axis_stretch_mode", "get_v_axis_stretch_mode");

Please also try to write more explicit commit messages, "fix" and the name of a static method deep in the backend doesn't give us much information on what this commit actually does. A commit log should describe the bug that it fixes, so that we can easily understand it without having to read the code. See https://github.com/godotengine/godot/blob/master/CONTRIBUTING.md#format-your-commit-logs-with-readability-in-mind

[reduz]

Oh sorry, I asked him to submit this fix. Technically this is not a bug because it will never go via a different path, but lets say that in some mysterious platforms that some people may be porting code too, the compiler will complain because not all return paths are satisfied.

[akien-mga]

Well I think it's better to add an else with an error message if we really want to satisfy compilers, and keep the == 2 as is.

[akien-mga]

Superseded by e4907e5 (GLES3, master) and 6546735 (Vulkan, vulkan). Thanks for the contribution nevertheless!