Current HLE push buffer processing code is very rough and messed up.
especially the code for method 0x1808. From the NV2A LLE code, it should be a 32bit element method.
but the current code is meant for 16 bit index data. and there are obvious bugs in the code such as copy IB buffer into a small WORD array with only 4 elements which obviously will introduce overflow issues.
we'll need a test case first to speed up the debugging and code revising.
Tip: Add a LOG_TEST_CASE invocation at the appropriate place(s) - once in master, testers will give feedback upon it
Most helpful comment
Tip: Add a LOG_TEST_CASE invocation at the appropriate place(s) - once in master, testers will give feedback upon it