Now that we only traverse the AST for bytecode generation and printing, the JSC_FAST_CALL intrinsic is likely unnecessary. We should remove it if it is not a performance regression.
Created attachment 23239 [details] Proposed patch
Created attachment 23240 [details] SunSpider results for last patch against r36264
Comment on attachment 23239 [details] Proposed patch Oliver pointed out that I should test the V8 benchmark suite and non-JIT build configurations.
This happened in August 2009, so we can close the bug.