GD-32: Unable to mock script which uses snake_case naming convention about gdunit3 HOT 8 CLOSED

Thanks for submiting this issue, i will have a deeper look to understand the problem.

from gdunit3.

After deeper investigation It seams to be it is not a snake case issue.
The doubler try to extend by class name but the class has no class_name defined.

So i need to deside to extend by existing class_name or class_path

from gdunit3.

It's working fine now! Thanks!

from gdunit3.

I made a small and quick fix, but I don't know if it's the best way. Right after the clazz_name definition, on GdUnitMockBuilder.build I've added the following:

	if "_" in clazz_name or clazz_name == clazz_name.to_lower():
		var result = PoolStringArray()
		var prev_is_underscore = false
		for ch in clazz_name:
			if ch=='_':
				prev_is_underscore = true
			else:
				if prev_is_underscore:
					result.append(ch.to_upper())
				else:
					result.append(ch)
				prev_is_underscore = false

		result[0] = result[0].to_upper()
		clazz_name = result.join('')

The logic I got from here. It is working fine for my tests cases.

from gdunit3.

@afonsolage you can alternative use the class name to mock the class instead by path.

mock(main_player)

It needs onyl to add the class name definition in your class

class_name main_player

The fix is now in master and you can mock by class path or name as you want.

from gdunit3.

I'll try it, before even using the class_name (which was in PascalCase - MainPlayer) didn't worked, but I'll give a try and post here.

from gdunit3.

It worked for main_class.gd, but not for mob.gd, since mob.gd is snake_case but doesn't have underscore. Something like that would fix:

	# handle snake_case class names, only if clazz path available and is a GdScript
	if "_" in clazz_name and not clazz_path.empty() and (clazz_path[0].find(".gd") != -1):
		return "'%s'" % clazz_path[0]
	elif not ClassDB.class_exists(clazz_name):
		return "'%s'" % clazz_path[0]
	return clazz_name

Add a new test case using a lowercaseonly.gd test file and you should get the problem

from gdunit3.

pushed new fix to master

from gdunit3.