GLES2: Fix VersionKey comparison in ShaderGLES2::bind()

[akien-mga]

This was comparing arrays, GCC 12 raises a warning for it:

drivers/gles2/shader_gles2.cpp: In member function 'bool ShaderGLES2::bind()':
drivers/gles2/shader_gles2.cpp:80:71: error: comparison between two arrays [-Werror=array-compare]
   80 |         if (active != this || !version || new_conditional_version.key != conditional_version.key) {
      |                                           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gles2/shader_gles2.cpp:80:71: note: use unary '+' which decays operands to pointers or '&'component_ref' not supported by dump_decl<declaration error>[0] != &'component_ref' not supported by dump_decl<declaration error>[0]' to compare the addresses

For context:

godot/drivers/gles2/shader_gles2.h

Lines 126 to 134 in b28eea6

	union VersionKey {
	struct {
	uint64_t version;
	uint32_t code_version;
	};
	unsigned char key[12];
	bool operator==(const VersionKey &p_key) const { return version == p_key.version && code_version == p_key.code_version; }
	bool operator<(const VersionKey &p_key) const { return version < p_key.version || (version == p_key.version && code_version < p_key.code_version); }
	};

GLES3 is somewhat different and spells things out more in detail:

godot/drivers/gles3/shader_gles3.cpp

Lines 96 to 109 in b28eea6

	bool ShaderGLES3::_bind(bool p_binding_fallback) {
	// Same base shader and version valid version?
	if (active == this && version) {
	if (new_conditional_version.code_version == conditional_version.code_version) {
	if (new_conditional_version.version == conditional_version.version) {
	return false;
	}
	// From ubershader to ubershader of the same code?
	if ((conditional_version.version & VersionKey::UBERSHADER_FLAG) && (new_conditional_version.version & VersionKey::UBERSHADER_FLAG)) {
	conditional_version.version = new_conditional_version.version;
	return false;
	}
	}
	}

Ubershader logic aside, it should amount to the same as what my GLES2 fix does.

Tested on Jetpaca, didn't spot any obvious issue.

Would need review from rendering team, I just fixed the C++ bug, and that might fix some actual GLES2 shader issues or introduce some.

[RandomShaper]

Looks good, and to me it looks as it can only do good.

[akien-mga]

Cherry-picked for 3.4.4.