[OOT] Fix wrong segments on import #275
Conversation
| if name == "gEmptyDL": | ||
| return None | ||
| return name | ||
| else: | ||
| for attr_name in dir(self.materialContext.ootMaterial.opaque): | ||
| if re.match("segment[89ABCD]", attr_name): | ||
| setattr(self.materialContext.ootMaterial.opaque, attr_name, False) | ||
| for attr_name in dir(self.materialContext.ootMaterial.transparent): | ||
| if re.match("segment[89ABCD]", attr_name): | ||
| setattr(self.materialContext.ootMaterial.transparent, attr_name, False) | ||
| return name |
There was a problem hiding this comment.
would like an explanation
There was a problem hiding this comment.
I have to admit, I didn't look at it 4Head
There was a problem hiding this comment.
the loop is different tho, should I zip it?
There was a problem hiding this comment.
I'm pretty sure this "fix" breaks this comment
dynamic material calls aka self.materialContext.ootMaterial is exactly what we want to clear under certain conditions (this PR implementing "clear every dlist jump", which is pretty rough imo) but this comment says they shouldn't be cleared ever for skeletons
|
last commit broke importing entirely |
|
I tried and works on my machine™ |
There was a problem hiding this comment.
OOTF3DContext.processDLName(name) has the following purpose to hook into the handling of gsSPDisplayList({name}):
if name is a raw segment offset like 0x08000000, don't try to find a dlist named "0x08000000" and instead mark the material as calling segment 8 (with materialContext.ootMaterial.{opaque,transparent}.segment8 = True)
(btw I don't know what is the issue this PR fixes)
The suggested fix here would mean a dlist like
gsSPDisplayList(0x08000000),
gsSPDisplayList(gSomeSymbolDL),
(vertices & triangles),
would be imported with the resulting material not calling segment 8, which sounds undesirable
Note that this:
gsSPDisplayList(gSomeSymbolDL),
gsSPDisplayList(0x08000000),
(vertices & triangles),
would have the imported material be set to call segment 8
|
closing this for now, it seems this fix needs more testing/better implementation (also I don't really understand how materials works so I won't be able to be efficient on this one anyway) |
|
This was attempting to fix #252 btw |
Thanks to kenton and cdi-fails for figuring this out