It makes sense to add debug code to compare the results from JIT regular expressions to interpreted regular expressions.
Created attachment 77850 [details] Patch Adding Debug Code to Compare RegExp JIT with RegExp Interpreter.
Comment on attachment 77850 [details] Patch Adding Debug Code to Compare RegExp JIT with RegExp Interpreter. Hi Michael, Functionally the code all looks great. I'd like to suggest three stylistic changes. 1027 #else 1028 #define ENABLE_YARR_JIT_DEBUG 0 The ENABLE() macro returns false for flags that aren't defined, so there should be no need to define this id YARR_JIT is not enabled. I think we should remove this redundant code. (Explicitly setting the flag to 0 in the block that enables YARR_JIT is good through, to broadcast that the setting exists.) 97 #if ENABLE(YARR_JIT_DEBUG) 98 res = JITCode; 99 else 100 res = ByteCode; 101 #else 102 return JITCode; 103 #endif The return on line 102 is mis-indented. I also think this code would be slightly clearer if you copied the if statement on the line just before this block into both the if and the else. Finally, I think it would be great to move the #if ENABLE(YARR_JIT_DEBUG) code from RegExp::match out into a new function. It's great to land this tool, but it would also be nice to keep the actual release code in match nice and clear. cheers, G.
Created attachment 78167 [details] Updated patch addressing reviewer's concerns Cleaned up some other extraneous trailing and whole line white space.
Comment on attachment 78167 [details] Updated patch addressing reviewer's concerns Looks great Michael
Committed r75208: <http://trac.webkit.org/changeset/75208>