Re: Undo glyphs.c change

I'm not sure why you would expect memcmp to be faster. When I spent time optimizing Feval and related thing to make the interpreter faster, the biggest gains came from inlining things that were previously function call. Function call overhead can be a real killer. You have to do register save/restore plus pushing and popping args off the stack. And you still have to compare the same bits that the old inlined code did. If there's a machine instruction that lets you do memcmp in an instruction or two, you will save some time, but given chip speeds today, most of the time is probably spent in the memory fetches. You can't get around the fetches.