Ryan C. Gordon <icculus@icculus.org> [Thu, 10 Feb 2011 18:37:29 -0500] rev 991
More work on overloaded intrinsic functions.
I really dislike this whole piece of code. Would really like to not flood the
symbol table with thousands of overloads like this.
Ryan C. Gordon <icculus@icculus.org> [Thu, 10 Feb 2011 00:05:36 -0800] rev 990
More work on fixing up overloaded function selection.
Ryan C. Gordon <icculus@icculus.org> [Thu, 10 Feb 2011 00:05:10 -0800] rev 989
Add more intrinsic function overloads. This is getting out of control now.
Ryan C. Gordon <icculus@icculus.org> [Wed, 09 Feb 2011 04:32:00 -0500] rev 988
More work on overloaded function matching.
This works more like Microsoft's compiler appears to now, but I've still got
failing cases to shake out.
Ryan C. Gordon <icculus@icculus.org> [Sun, 06 Feb 2011 04:01:43 -0500] rev 987
Allow constructors with sub-vectors (or whatever you'd call this).
Now "float4(float3(1,2,3),1)" works as expected.
Ryan C. Gordon <icculus@icculus.org> [Wed, 19 Jan 2011 01:38:01 -0800] rev 986
Fixed some uninitialized memory reads, compliments of valgrind.
Ryan C. Gordon <icculus@icculus.org> [Wed, 19 Jan 2011 01:24:08 -0800] rev 985
Vectors and matrices can be dereferenced like arrays.
Ryan C. Gordon <icculus@icculus.org> [Wed, 19 Jan 2011 01:04:40 -0800] rev 984
Fixed dropping pieces of structs with comma-separated members.
Ryan C. Gordon <icculus@icculus.org> [Tue, 18 Jan 2011 13:49:42 -0800] rev 983
Let type coercion deal with vectors and matrices.
Ryan C. Gordon <icculus@icculus.org> [Tue, 18 Jan 2011 13:44:52 -0800] rev 982
Moved assignment into the scope where it's actually used.