The JavaScriptCore offline assembler makes heavy use of regular expressions when parsing. This would be a good test for RexBench.
<rdar://problem/33898475>
Created attachment 318138 [details] Patch
Attachment 318138 [details] did not pass style-queue: ERROR: PerformanceTests/RexBench/OfflineAssembler/convert_asm_to_js.py:48: multiple statements on one line (semicolon) [pep8/E702] [5] ERROR: PerformanceTests/RexBench/OfflineAssembler/convert_asm_to_js.py:59: no spaces around keyword / parameter equals [pep8/E251] [5] ERROR: PerformanceTests/RexBench/OfflineAssembler/convert_asm_to_js.py:60: no spaces around keyword / parameter equals [pep8/E251] [5] Total errors found: 3 in 23 files If any of these errors are false positives, please file a bug against check-webkit-style.
Comment on attachment 318138 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=318138&action=review r=me > PerformanceTests/ChangeLog:16 > + The input data is the current tip of tree LLInt .asm source files. Those file > + were converted to JavaScript using the python script, convert_asm_to_js.py. > + There is also a new JavaScript file, generate_expected.js, to create the > + expected output JavaScript file. How much faster is this than the ruby version? > PerformanceTests/RexBench/offline_assembler_benchmark.js:44 > +var results = []; > +var benchmark = new OfflineAssemblerBenchmark(); > +var numIterations = 20; I vote for `let` here > PerformanceTests/RexBench/OfflineAssembler/parser.js:27 > + It would be cool if you extended it to allow for unicode in some way (maybe in the future)
(In reply to Saam Barati from comment #4) > Comment on attachment 318138 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=318138&action=review > > r=me > > > PerformanceTests/ChangeLog:16 > > + The input data is the current tip of tree LLInt .asm source files. Those file > > + were converted to JavaScript using the python script, convert_asm_to_js.py. > > + There is also a new JavaScript file, generate_expected.js, to create the > > + expected output JavaScript file. > > How much faster is this than the ruby version? I didn't measure. > > PerformanceTests/RexBench/offline_assembler_benchmark.js:44 > > +var results = []; > > +var benchmark = new OfflineAssemblerBenchmark(); > > +var numIterations = 20; > > I vote for `let` here That is the style for all the other tests and also the form used in ARES-6. > > PerformanceTests/RexBench/OfflineAssembler/parser.js:27 > > + > > It would be cool if you extended it to allow for unicode in some way (maybe > in the future) I'm planning on writing one more subtest focused on just Unicode.
Committed r220752: <http://trac.webkit.org/changeset/220752>