[ESNext][BigInt] Implement "~" unary operation
[WebKit-https.git] / PerformanceTests / ChangeLog
1 2019-03-11  Caio Lima  <ticaiolima@gmail.com>
2
3         [ESNext][BigInt] Implement "~" unary operation
4         https://bugs.webkit.org/show_bug.cgi?id=182216
5
6         Reviewed by Keith Miller.
7
8         * BigIntBench/big-int-simple-bit-not.js: Added.
9
10 2019-03-04  Saam Barati  <sbarati@apple.com>
11
12         Add a detailed summary page for JetStream 2
13         https://bugs.webkit.org/show_bug.cgi?id=195014
14
15         Rubber-stamped by Darin Adler.
16
17         * JetStream2/JetStream.css:
18         (a:link, a:visited):
19         * JetStream2/in-depth.html: Added.
20
21 2019-02-08  Saam barati  <sbarati@apple.com>
22
23         Update JetStream2 CLI to not print "ms" since we now just print the score values
24         https://bugs.webkit.org/show_bug.cgi?id=193981
25
26         Reviewed by Yusuke Suzuki.
27
28         * JetStream2/JetStreamDriver.js:
29         (DefaultBenchmark.prototype.updateUIAfterRun):
30         (DefaultBenchmark):
31         (WSLBenchmark.prototype.updateUIAfterRun):
32         (WSLBenchmark):
33         (WasmBenchmark.prototype.updateUIAfterRun):
34         (WasmBenchmark):
35
36 2019-01-31  Sihui Liu  <sihui_liu@apple.com>
37
38         REGRESSION (r240358): IndexedDB/large-binary-keys.html and IndexedDB/large-string-keys.html perf tests failing
39         https://bugs.webkit.org/show_bug.cgi?id=193956
40         <rdar://problem/47623109>
41
42         Reviewed by Ryosuke Niwa.
43
44         r240358 sets default IDB quota as 50 MB for single test, we need to increase that for big tests. 
45
46         * IndexedDB/large-binary-keys.html:
47         * IndexedDB/large-string-keys.html:
48
49 2019-01-18  Saam Barati  <sbarati@apple.com>
50
51         Use scores everywhere in JetStream2's UI
52         https://bugs.webkit.org/show_bug.cgi?id=193593
53
54         Reviewed by Keith Miller.
55
56         * JetStream2/JetStreamDriver.js:
57         (DefaultBenchmark.prototype.processResults):
58         (DefaultBenchmark.prototype.get score):
59         (DefaultBenchmark.scoreDescription):
60         (WSLBenchmark.prototype.processResults):
61         (WSLBenchmark.prototype.get score):
62         (WSLBenchmark.scoreDescription):
63         (WasmBenchmark.prototype.processResults):
64         (WasmBenchmark.prototype.get score):
65         (WasmBenchmark.scoreDescription):
66         * JetStream2/index.html:
67
68 2019-01-18  Jer Noble  <jer.noble@apple.com>
69
70         SDK_VARIANT build destinations should be separate from non-SDK_VARIANT builds
71         https://bugs.webkit.org/show_bug.cgi?id=189553
72
73         Reviewed by Tim Horton.
74
75         * MediaTime/Configurations/Base.xcconfig:
76         * MediaTime/Configurations/SDKVariant.xcconfig: Added.
77
78 2019-01-17  Truitt Savell  <tsavell@apple.com>
79
80         Unreviewed, rolling out r240124.
81
82         This commit broke an internal build.
83
84         Reverted changeset:
85
86         "SDK_VARIANT build destinations should be separate from non-
87         SDK_VARIANT builds"
88         https://bugs.webkit.org/show_bug.cgi?id=189553
89         https://trac.webkit.org/changeset/240124
90
91 2019-01-17  Jer Noble  <jer.noble@apple.com>
92
93         SDK_VARIANT build destinations should be separate from non-SDK_VARIANT builds
94         https://bugs.webkit.org/show_bug.cgi?id=189553
95
96         Reviewed by Tim Horton.
97
98         * MediaTime/Configurations/Base.xcconfig:
99         * MediaTime/Configurations/SDKVariant.xcconfig: Added.
100
101 2018-12-20  Caio Lima  <ticaiolima@gmail.com>
102
103         [BigInt] We should enable CSE into arithmetic operations that speculate BigIntUse
104         https://bugs.webkit.org/show_bug.cgi?id=192723
105
106         Reviewed by Yusuke Suzuki.
107
108         * BigIntBench/big-int-cse.js: Added.
109         * BigIntBench/big-int-global-cse.js: Added.
110         * BigIntBench/big-int-licm.js: Added.
111
112 2018-12-19  Commit Queue  <commit-queue@webkit.org>
113
114         Unreviewed, rolling out r239377.
115         https://bugs.webkit.org/show_bug.cgi?id=192921
116
117         broke 32-bit JSC tests (Requested by keith_miller on #webkit).
118
119         Reverted changeset:
120
121         "[BigInt] We should enable CSE into arithmetic operations that
122         speculate BigIntUse"
123         https://bugs.webkit.org/show_bug.cgi?id=192723
124         https://trac.webkit.org/changeset/239377
125
126 2018-12-19  Caio Lima  <ticaiolima@gmail.com>
127
128         [BigInt] We should enable CSE into arithmetic operations that speculate BigIntUse
129         https://bugs.webkit.org/show_bug.cgi?id=192723
130
131         Reviewed by Saam Barati.
132
133         * BigIntBench/big-int-cse.js: Added.
134         * BigIntBench/big-int-global-cse.js: Added.
135         * BigIntBench/big-int-licm.js: Added.
136
137 2018-12-17  Suresh Koppisetty  <skoppisetty@apple.com>
138
139         Add "-o/--output" option to startup.py and new_tab.py benchmark scripts to save the results in json format.
140         https://bugs.webkit.org/show_bug.cgi?id=192385
141
142         Reviewed by Ryosuke Niwa.
143
144         Sample json output for new tab benchmark script after running for 2 iterations and 2 groups. Values are in milliseconds.
145         {
146                 "NewTabBenchmark": {
147                         "metrics": {
148                                 "Time": {
149                                         "current": [
150                                                 [
151                                                         410.2939453125,
152                                                         307.81494140625
153                                                 ],
154                                                 [
155                                                         340.616943359375,
156                                                         265.94384765625
157                                                 ]
158                                         ]
159                                 }
160                         }
161                 }
162         }
163
164         Sample json output for startup time benchmark script after running for 2 iterations. Values are in milliseconds.
165         {
166                 "StartupBenchmark": {
167                         "metrics": {
168                                 "Time": {
169                                         "current": [
170                                                 [
171                                                         1415.2099609375,
172                                                         1439.552978515625
173                                                 ]
174                                         ]
175                                 }
176                         }
177                 }
178         }
179
180         * LaunchTime/launch_time.py:
181         * LaunchTime/new_tab.py:
182         (NewTabBenchmark.get_test_name):
183         * LaunchTime/startup.py:
184         (StartupBenchmark.get_test_name):
185
186 2018-12-17  Suresh Koppisetty  <skoppisetty@apple.com>
187
188         Import FeedbackServer only if "-f/--feedback-in-browser" option is enabled.
189         https://bugs.webkit.org/show_bug.cgi?id=192378
190
191         Reviewed by Ryosuke Niwa.
192
193         FeedbackServer currently depends on Tornado-5.1, which further adds
194         dependency of "singledispatch", "backports-abc" and "futures" python libraries.
195         Importing FeedbackServer only if "-f/--feedback-in-browser" option is enabled
196         will let us run the benchmark scripts without installing any new python libraries.
197
198         * LaunchTime/launch_time.py:
199         * LaunchTime/new_tab.py:
200
201 2018-12-13  Caio Lima  <ticaiolima@gmail.com>
202
203         [BigInt] Add ValueDiv into DFG
204         https://bugs.webkit.org/show_bug.cgi?id=186178
205
206         Reviewed by Yusuke Suzuki.
207
208         * BigIntBench/big-int-simple-div.js: Added.
209         * BigIntBench/value-div-type-propagation.js: Added.
210
211 2018-12-10  Caio Lima  <ticaiolima@gmail.com>
212
213         [BigInt] Add ValueMul into DFG
214         https://bugs.webkit.org/show_bug.cgi?id=186175
215
216         Reviewed by Yusuke Suzuki.
217
218         * BigIntBench/big-int-simple-mul.js: Added.
219         * BigIntBench/value-mul-type-propagation.js: Added.
220
221 2018-11-07  Caio Lima  <ticaiolima@gmail.com>
222
223         [BigInt] Add support to BigInt into ValueAdd
224         https://bugs.webkit.org/show_bug.cgi?id=186177
225
226         Reviewed by Keith Miller.
227
228         The idea of BigIntBench is to provide a set of microbenchmarks and
229         benchmarks to evaluate how fast BigInt computations are happening on
230         JSC implementation.
231
232         Now, we are adding microbenchmarks in this set,
233         but the plan is to move these tests to "JSTest/microbenchmarks" when
234         BigInt is enabled by default. After that, the focus of Bigint bench is
235         to provide a set of tests that represents real use cases of BigInt in
236         JS programs.
237
238         * BigIntBench/big-int-add-prediction-propagation.js: Added.
239         * BigIntBench/big-int-simple-add.js: Added.
240         * BigIntBench/big-int-simple-sub.js: Added.
241
242 2018-11-07  Tadeu Zagallo  <tzagallo@apple.com>
243
244         REGRESSION(r237547): Test failures on 32-bit JSC since the JIT was disabled
245         https://bugs.webkit.org/show_bug.cgi?id=191184
246
247         Reviewed by Saam Barati.
248
249         Skip test when the JIT is disabled, it takes too long to run on CLoop.
250
251         * JetStream/cdjs/main.js:
252
253 2018-11-05  Myles C. Maxfield  <mmaxfield@apple.com>
254
255         Cache glyph paths and share underline skipping code between all the ports
256         https://bugs.webkit.org/show_bug.cgi?id=191239
257
258         Reviewed by Alex Christensen.
259
260         Measures the performance of drawing a whole lot of underlines
261
262         * Layout/underline.html: Added.
263
264 2018-10-25  Saam Barati  <sbarati@apple.com>
265
266         Check in corresponding C code in JetStream 2
267         https://bugs.webkit.org/show_bug.cgi?id=190733
268
269         Rubber-stamped by Filip Pizlo.
270
271         * JetStream2/simple/float-mm.c: Added.
272         * JetStream2/wasm/HashSet.cpp: Added.
273         * JetStream2/wasm/gcc-loops.cpp: Added.
274         * JetStream2/wasm/quicksort.c: Added.
275         * JetStream2/wasm/TSF: Added.
276         * JetStream2/wasm/TSF/build.sh: Added.
277         * JetStream2/wasm/TSF/config.h: Added.
278         * JetStream2/wasm/TSF/gpc.h: Added.
279         * JetStream2/wasm/TSF/gpc_code_gen_util.c: Added.
280         * JetStream2/wasm/TSF/gpc_instruction.c: Added.
281         * JetStream2/wasm/TSF/gpc_instruction_dispatch.gen: Added.
282         * JetStream2/wasm/TSF/gpc_instruction_size.gen: Added.
283         * JetStream2/wasm/TSF/gpc_instruction_stack_effects.gen: Added.
284         * JetStream2/wasm/TSF/gpc_instruction_static_size.gen: Added.
285         * JetStream2/wasm/TSF/gpc_instruction_to_string.gen: Added.
286         * JetStream2/wasm/TSF/gpc_int_common.h: Added.
287         * JetStream2/wasm/TSF/gpc_intable.c: Added.
288         * JetStream2/wasm/TSF/gpc_internal.h: Added.
289         * JetStream2/wasm/TSF/gpc_interpreter.gen: Added.
290         * JetStream2/wasm/TSF/gpc_program.c: Added.
291         * JetStream2/wasm/TSF/gpc_proto.c: Added.
292         * JetStream2/wasm/TSF/gpc_stack_height.c: Added.
293         * JetStream2/wasm/TSF/gpc_threaded.c: Added.
294         * JetStream2/wasm/TSF/gpc_worklist.h: Added.
295         * JetStream2/wasm/TSF/tsf.h: Added.
296         * JetStream2/wasm/TSF/tsf_adaptive_reader.c: Added.
297         * JetStream2/wasm/TSF/tsf_asprintf.c: Added.
298         * JetStream2/wasm/TSF/tsf_atomics.h: Added.
299         * JetStream2/wasm/TSF/tsf_buf_reader.c: Added.
300         * JetStream2/wasm/TSF/tsf_buf_writer.c: Added.
301         * JetStream2/wasm/TSF/tsf_buffer.c: Added.
302         * JetStream2/wasm/TSF/tsf_build_defines.h: Added.
303         * JetStream2/wasm/TSF/tsf_config.h: Added.
304         * JetStream2/wasm/TSF/tsf_config_stub.h: Added.
305         * JetStream2/wasm/TSF/tsf_copier.c: Added.
306         * JetStream2/wasm/TSF/tsf_define_helpers.c: Added.
307         * JetStream2/wasm/TSF/tsf_define_helpers.h: Added.
308         * JetStream2/wasm/TSF/tsf_destructor.c: Added.
309         * JetStream2/wasm/TSF/tsf_error.c: Added.
310         * JetStream2/wasm/TSF/tsf_format.h: Added.
311         * JetStream2/wasm/TSF/tsf_fsdb.c: Added.
312         * JetStream2/wasm/TSF/tsf_fsdb_protocol.c: Added.
313         * JetStream2/wasm/TSF/tsf_fsdb_protocol.h: Added.
314         * JetStream2/wasm/TSF/tsf_generator.c: Added.
315         * JetStream2/wasm/TSF/tsf_gpc_code_gen.c: Added.
316         * JetStream2/wasm/TSF/tsf_indent.h: Added.
317         * JetStream2/wasm/TSF/tsf_internal.h: Added.
318         * JetStream2/wasm/TSF/tsf_internal_config.h: Added.
319         * JetStream2/wasm/TSF/tsf_internal_config_stub.h: Added.
320         * JetStream2/wasm/TSF/tsf_inttypes.h: Added.
321         * JetStream2/wasm/TSF/tsf_io.c: Added.
322         * JetStream2/wasm/TSF/tsf_io_utils.c: Added.
323         * JetStream2/wasm/TSF/tsf_ir.c: Added.
324         * JetStream2/wasm/TSF/tsf_ir.h: Added.
325         * JetStream2/wasm/TSF/tsf_ir_different.c: Added.
326         * JetStream2/wasm/TSF/tsf_ir_different.h: Added.
327         * JetStream2/wasm/TSF/tsf_ir_speed.c: Added.
328         * JetStream2/wasm/TSF/tsf_limits.c: Added.
329         * JetStream2/wasm/TSF/tsf_named_type.c: Added.
330         * JetStream2/wasm/TSF/tsf_native.c: Added.
331         * JetStream2/wasm/TSF/tsf_parser.c: Added.
332         * JetStream2/wasm/TSF/tsf_primitive.c: Added.
333         * JetStream2/wasm/TSF/tsf_ra_type_man.c: Added.
334         * JetStream2/wasm/TSF/tsf_reflect.c: Added.
335         * JetStream2/wasm/TSF/tsf_region.h: Added.
336         * JetStream2/wasm/TSF/tsf_serial_in_man.c: Added.
337         * JetStream2/wasm/TSF/tsf_serial_out_man.c: Added.
338         * JetStream2/wasm/TSF/tsf_serial_protocol.h: Added.
339         * JetStream2/wasm/TSF/tsf_sha1.c: Added.
340         * JetStream2/wasm/TSF/tsf_sha1.h: Added.
341         * JetStream2/wasm/TSF/tsf_sha1_writer.c: Added.
342         * JetStream2/wasm/TSF/tsf_sort.c: Added.
343         * JetStream2/wasm/TSF/tsf_st.c: Added.
344         * JetStream2/wasm/TSF/tsf_st.h: Added.
345         * JetStream2/wasm/TSF/tsf_st_typetable.c: Added.
346         * JetStream2/wasm/TSF/tsf_stream_file_input.c: Added.
347         * JetStream2/wasm/TSF/tsf_stream_file_output.c: Added.
348         * JetStream2/wasm/TSF/tsf_type.c: Added.
349         * JetStream2/wasm/TSF/tsf_type_in_map.c: Added.
350         * JetStream2/wasm/TSF/tsf_type_out_map.c: Added.
351         * JetStream2/wasm/TSF/tsf_type_table.c: Added.
352         * JetStream2/wasm/TSF/tsf_types.h: Added.
353         * JetStream2/wasm/TSF/tsf_util.h: Added.
354         * JetStream2/wasm/TSF/tsf_version.c: Added.
355         * JetStream2/wasm/TSF/tsf_zip_abstract.c: Added.
356         * JetStream2/wasm/TSF/tsf_zip_abstract.h: Added.
357         * JetStream2/wasm/TSF/tsf_zip_attr.c: Added.
358         * JetStream2/wasm/TSF/tsf_zip_reader.c: Added.
359         * JetStream2/wasm/TSF/tsf_zip_writer.c: Added.
360
361 2018-10-18  Saam Barati  <sbarati@apple.com>
362
363         Make JetStream 2
364         https://bugs.webkit.org/show_bug.cgi?id=187829
365
366         Rubber-stamped by Mark Lam.
367
368         This patch checks in the new JetStream 2 benchmark. JetStream 2 is
369         a new JavaScript and Web Assembly benchmark. JetStream 2's goal
370         is to measure the startup, worst case, and peak throughput performance
371         of the JavaScript engine. JetStream 2 incorporates these previous
372         benchmarks:
373         - JetStream
374         - ARES-6
375         - Kraken
376         - Web Tooling Benchmark
377         - WasmBench
378         - RexBench
379         
380         JetStream 2 also adds some new benchmarks:
381         - Two tests emphasizing web worker performance.
382         - One test emphasizing Promise, async iteration, and DataView performance.
383         - Two new code load tests.
384         - WSL: a test measuring all kinds of things, especially emphasizing exception performance.
385
386         * JetStream2: Added.
387
388 2018-10-02  Chris Dumez  <cdumez@apple.com>
389
390         Regression(r236613): Parser/html-parser.html performance test is failing
391         https://bugs.webkit.org/show_bug.cgi?id=190201
392
393         Unreviewed, apply the same fix as in r236699 to make sure the iframe is treated as
394         same origin, so that calls to document.open() / document.write() succeed.
395
396         * Parser/html-parser.html:
397
398 2018-10-01  Chris Dumez  <cdumez@apple.com>
399
400         Regression(r236613): Parser/HTML5-8266-ParseOnly.html performance tests no longer runs
401         https://bugs.webkit.org/show_bug.cgi?id=190174
402
403         Reviewed by Ryosuke Niwa.
404
405         Update iframe's sandbox to use 'allow-same-origin' so that the iframe is treated as same-origin and so that
406         the top frame can call document.open() / document.write() on the subframe. This test was also failing in
407         Firefox without this change.
408
409         * Parser/HTML5-8266-ParseOnly.html:
410
411 2018-08-27  Aditya Keerthi  <akeerthi@apple.com>
412
413         Consolidate ENABLE_INPUT_TYPE_COLOR and ENABLE_INPUT_TYPE_COLOR_POPOVER
414         https://bugs.webkit.org/show_bug.cgi?id=188931
415
416         Reviewed by Wenson Hsieh.
417
418         * StitchMarker/wtf/FeatureDefines.h: Removed ENABLE_INPUT_TYPE_COLOR_POPOVER.
419
420 2018-08-23  Don Olmstead  <don.olmstead@sony.com>
421
422         [CMake] Have checks are not getting set inside CMake properly
423         https://bugs.webkit.org/show_bug.cgi?id=188901
424
425         Reviewed by Michael Catanzaro.
426
427         Add notification that MallocBench is disabled.
428
429         * CMakeLists.txt:
430
431 2018-08-23  Don Olmstead  <don.olmstead@sony.com>
432
433         [CMake] Add HAVE_MALLOC_TRIM definition
434         https://bugs.webkit.org/show_bug.cgi?id=188897
435
436         Reviewed by Konstantin Tokarev.
437
438         MallocBench should only be built on Apple platforms and platforms that
439         have malloc_trim.
440
441         * CMakeLists.txt:
442
443 2018-08-13  Jon Lee  <jonlee@apple.com>
444
445         Update MotionMark version number
446         https://bugs.webkit.org/show_bug.cgi?id=188535
447         <rdar://problem/43254078>
448
449         Reviewed by Said Abou-Hallawa.
450
451         * MotionMark/resources/strings.js: To 1.1.
452
453 2018-08-13  Jon Lee  <jonlee@apple.com>
454
455         [MotionMark] Update Multiply test
456         https://bugs.webkit.org/show_bug.cgi?id=188532
457         <rdar://problem/43252151>
458
459         Reviewed by Said Abou-Hallawa.
460
461         Update Multiply test to have a larger upper limit of number of particles. Cycle through three
462         different ways of hiding an element. Add new Multiple suite that isolates those components for
463         debugging purposes.
464
465         * MotionMark/resources/debug-runner/tests.js:
466         * MotionMark/tests/dom/multiply.html: Added.
467         * MotionMark/tests/dom/resources/multiply.js: Added.
468         * MotionMark/tests/master/resources/multiply.js:
469
470 2018-08-13  Jon Lee  <jonlee@apple.com>
471
472         [MotionMark] Update Leaves test
473         https://bugs.webkit.org/show_bug.cgi?id=188530
474         <rdar://problem/43251862>
475
476         Reviewed by Said Abou-Hallawa.
477
478         Update Leaves test to include opacity and scale. Add new Leaves suite that isolate those components
479         for debugging purposes.
480
481         * MotionMark/resources/debug-runner/tests.js:
482         * MotionMark/tests/dom/leaves.html:
483         * MotionMark/tests/dom/resources/leaves.js:
484         * MotionMark/tests/master/resources/leaves.js:
485
486 2018-08-10  Ben Richards  <benton_richards@apple.com>
487
488         Add ability to ignore process prewarming for launch time benchmark
489         https://bugs.webkit.org/show_bug.cgi?id=188462
490
491         Reviewed by Ryosuke Niwa.
492
493         Added a flag to the new tab benchmark that will open a blank tab before the tab to be measured in order to ignore process prewarming.
494
495         * LaunchTime/launch_time.py:
496         (DefaultLaunchTimeHandler):
497         (DefaultLaunchTimeHandler.on_receive_stop_time): Deleted.
498         (DefaultLaunchTimeHandler.on_receive_stop_signal):
499         (DefaultLaunchTimeHandler.do_HEAD):
500         (DefaultLaunchTimeHandler.do_GET):
501         (DefaultLaunchTimeHandler.do_POST):
502         (LaunchTimeBenchmark):
503         (LaunchTimeBenchmark._standard_deviation): Fixed divide by zero bug when '-n' is set to 1
504         (LaunchTimeBenchmark.open_tab): Added option to open a blank tab
505         (LaunchTimeBenchmark.run):
506         * LaunchTime/new_tab.py:
507         (NewTabBenchmark.initialize):
508         (NewTabBenchmark.run_iteration):
509         (NewTabBenchmark.will_parse_arguments):
510         (NewTabBenchmark.did_parse_arguments):
511         (NewTabBenchmark.ResponseHandler.Handler.get_test_page):
512         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
513         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
514         (NewTabBenchmark):
515         * LaunchTime/startup.py:
516         (StartupBenchmark.ResponseHandler.Handler.get_test_page):
517         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
518         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
519         (StartupBenchmark):
520
521 2018-08-10  Truitt Savell  <tsavell@apple.com>
522
523         Unreviewed, rolling out r234750.
524
525         Caused 185 perf test failures.
526
527         Reverted changeset:
528
529         "Add ability to ignore process prewarming for launch time
530         benchmark"
531         https://bugs.webkit.org/show_bug.cgi?id=188462
532         https://trac.webkit.org/changeset/234750
533
534 2018-08-09  Ben Richards  <benton_richards@apple.com>
535
536         Add ability to ignore process prewarming for launch time benchmark
537         https://bugs.webkit.org/show_bug.cgi?id=188462
538
539         Reviewed by Ryosuke Niwa.
540
541         Added a flag to the new tab benchmark that will open a blank tab before the tab to be measured in order to ignore process prewarming.
542
543         * LaunchTime/launch_time.py:
544         (DefaultLaunchTimeHandler):
545         (DefaultLaunchTimeHandler.on_receive_stop_time): Deleted.
546         (DefaultLaunchTimeHandler.on_receive_stop_signal):
547         (DefaultLaunchTimeHandler.do_HEAD):
548         (DefaultLaunchTimeHandler.do_GET):
549         (DefaultLaunchTimeHandler.do_POST):
550         (LaunchTimeBenchmark):
551         (LaunchTimeBenchmark._standard_deviation): Fixed divide by zero bug when '-n' is set to 1
552         (LaunchTimeBenchmark.open_tab): Added option to open a blank tab
553         (LaunchTimeBenchmark.run):
554         * LaunchTime/new_tab.py:
555         (NewTabBenchmark.initialize):
556         (NewTabBenchmark.run_iteration):
557         (NewTabBenchmark.will_parse_arguments):
558         (NewTabBenchmark.did_parse_arguments):
559         (NewTabBenchmark.ResponseHandler.Handler.get_test_page):
560         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
561         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
562         (NewTabBenchmark):
563         * LaunchTime/startup.py:
564         (StartupBenchmark.ResponseHandler.Handler.get_test_page):
565         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
566         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
567         (StartupBenchmark):
568
569 2018-07-26  Andy VanWagoner  <andy@vanwagoner.family>
570
571         [INTL] Remove INTL sub-feature compile flags
572         https://bugs.webkit.org/show_bug.cgi?id=188081
573
574         Reviewed by Michael Catanzaro.
575
576         Removed old and unused i18n flag.
577
578         * StitchMarker/wtf/FeatureDefines.h:
579
580 2018-07-20  David Fenton  <david_fenton@apple.com>
581
582         REGRESSION (234006) ERROR: Blocked access to external URL http://code.jquery.com/jquery-1.9.1.min.js
583         https://bugs.webkit.org/show_bug.cgi?id=187858
584
585         Unreviewed test gardening, skipped test.
586
587         * Skipped: LaunchTime
588
589 2018-07-19  Ben Richards  <benton_richards@apple.com>
590
591         Add benchmark for WebKit process launch times
592         https://bugs.webkit.org/show_bug.cgi?id=186414
593
594         Reviewed by Ryosuke Niwa.
595
596         Added two benchmarks, one for measuring browser new tab launch time and one for browser startup time.
597
598         * LaunchTime/.gitignore: Added.
599         * LaunchTime/feedback_client.html: Added.
600         Displays benchmark progress in browser
601         * LaunchTime/feedback_server.py: Added.
602         (FeedbackServer): Sends data to feedback_client via websocket
603         (FeedbackServer.__init__):
604         (FeedbackServer._create_app):
605         (FeedbackServer._start_server):
606         (FeedbackServer._send_all_messages):
607         (FeedbackServer.start):
608         (FeedbackServer.stop):
609         (FeedbackServer.send_message): Send a message to the feedback_client
610         (FeedbackServer.wait_until_client_has_loaded): Wait until the feedback_client has opened a websocket connection to the feedback_server
611         (FeedbackServer.MainHandler): Handler factory to create handler that serves feedback_client.html
612         (FeedbackServer.MainHandler.Handler):
613         (FeedbackServer.MainHandler.Handler.get):
614         (FeedbackServer.WSHandler): Handler factory to create handler that sends data to feedback client
615         (FeedbackServer.WSHandler.Handler):
616         (FeedbackServer.WSHandler.Handler.open): On websocket connection opened
617         (FeedbackServer.WSHandler.Handler.on_close): On websocket connection closed
618         * LaunchTime/launch_time.py: Added.
619         (DefaultLaunchTimeHandler): Abstract HTTP request handler for launch time benchmarks
620         (DefaultLaunchTimeHandler.get_test_page): Default test page to be overridden by benchmarks
621         (DefaultLaunchTimeHandler.get_blank_page):
622         (DefaultLaunchTimeHandler.on_receive_stop_time):
623         (DefaultLaunchTimeHandler.do_HEAD):
624         (DefaultLaunchTimeHandler.do_GET):
625         (DefaultLaunchTimeHandler.do_POST):
626         (DefaultLaunchTimeHandler.log_message): Suppresses HTTP logs from SimpleHTTPRequestHandler
627         (LaunchTimeBenchmark): Abstract class which launch time benchmarks inherit from and override methods desired to customize
628         (LaunchTimeBenchmark.__init__):
629         (LaunchTimeBenchmark._parse_browser_bundle_path): Parser for bundle path option
630         (LaunchTimeBenchmark._parse_args):
631         (LaunchTimeBenchmark._run_server): Target function for main server thread
632         (LaunchTimeBenchmark._setup_servers):
633         (LaunchTimeBenchmark._clean_up):
634         (LaunchTimeBenchmark._exit_due_to_exception):
635         (LaunchTimeBenchmark._geometric_mean):
636         (LaunchTimeBenchmark._standard_deviation):
637         (LaunchTimeBenchmark._compute_results): Returns mean and std dev of list of results and pretty prints if should_print=True is specified
638         (LaunchTimeBenchmark._wait_times): Mimic numpy.linspace
639         (LaunchTimeBenchmark.open_tab): Open a browser tab with the html given by self.response_handler.get_test_page
640         (LaunchTimeBenchmark.launch_browser): Open a broser to either the feedback client (if option is set) or a blank page
641         (LaunchTimeBenchmark.quit_browser):
642         (LaunchTimeBenchmark.quit_browser.quit_app):
643         (LaunchTimeBenchmark.quit_browser.is_app_closed):
644         (LaunchTimeBenchmark.close_tab):
645         (LaunchTimeBenchmark.wait):
646         (LaunchTimeBenchmark.log): Print to console and send to feedback client if --feedback-in-browser flag is used
647         (LaunchTimeBenchmark.log_verbose): Only logs if --verbose flag is used
648         (LaunchTimeBenchmark.run):
649         (LaunchTimeBenchmark.group_init): Initialization done before each round of iterations
650         (LaunchTimeBenchmark.run_iteration):
651         (LaunchTimeBenchmark.initialize): Convenience method to be overriden by subclasses which is called at the end of __init__
652         (LaunchTimeBenchmark.will_parse_arguments): Called before argparse.parse_args to let subclasses add new command line arguments
653         (LaunchTimeBenchmark.did_parse_arguments): Called after argparse.parse_args to let subclass initialize based on command line arguments
654         * LaunchTime/new_tab.py: Added
655         (NewTabBenchmark):
656         (NewTabBenchmark._parse_wait_time): Parser for wait time option
657         (NewTabBenchmark.initialize):
658         (NewTabBenchmark.run_iteration):
659         (NewTabBenchmark.group_init):
660         (NewTabBenchmark.will_parse_arguments):
661         (NewTabBenchmark.did_parse_arguments):
662         (NewTabBenchmark.ResponseHandler):
663         (NewTabBenchmark.ResponseHandler.Handler):
664         (NewTabBenchmark.ResponseHandler.Handler.get_test_page):
665         * LaunchTime/startup.py: Added
666         (StartupBenchmark): This benchmark measures browser startup time and initial page load time
667         (StartupBenchmark.initialize):
668         (StartupBenchmark.run_iteration):
669         (StartupBenchmark.ResponseHandler):
670         (StartupBenchmark.ResponseHandler.Handler):
671         (StartupBenchmark.ResponseHandler.Handler.get_test_page):
672         * LaunchTime/thirdparty/__init__.py: Added.
673         (AutoinstallImportHook): Auto installs tornado package for feedback server
674         (AutoinstallImportHook.__init__):
675         (AutoinstallImportHook._ensure_autoinstalled_dir_is_in_sys_path):
676         (AutoinstallImportHook.find_module):
677         (AutoinstallImportHook._install_tornado):
678         (AutoinstallImportHook.greater_than_equal_to_version):
679         (AutoinstallImportHook._install):
680         (AutoinstallImportHook.get_latest_pypi_url):
681         (AutoinstallImportHook.install_binary):
682         (autoinstall_everything):
683         (get_os_info):
684
685 2018-06-25  Jon Lee  <jonlee@apple.com>
686
687         [MotionMark] Add support for version numbers
688         https://bugs.webkit.org/show_bug.cgi?id=186479
689
690         Reviewed by Said Abou-Hallawa.
691
692         Add support for displaying the version number as well as including it in the JSON results.
693
694         When loading the front page, script replaces any element with classname version with the
695         version number of the benchmark, which is stored in Strings.version.
696
697         The JSON structure for the results includes a new version property:
698             {
699               "version": "1.0",
700               "options": { ... },
701               "data": [ ... ]
702             }
703
704         When dragging a results file, the version listed will come from the JSON file. Older
705         results will not have had the version property, in which case it will default to "1.0".
706
707         * MotionMark/index.html: Update title to some other default. Script will update it.
708         Include the version number in the logo title.
709         * MotionMark/developer.html: Ditto.
710         * MotionMark/about.html: Ditto.
711
712         * MotionMark/resources/runner/motionmark.js:
713         (ResultsDashboard): Update constructor to include version. This is used when serializing
714         results out to JSON, and displaying the results panel in developer mode.
715         (ResultsDashboard._processData): When running the benchmark, include benchmark version string
716         in the results object.
717         (ResultsDashboard.version):
718         (window.benchmarkRunnerClient.willStartFirstIteration): When running the benchmark, pass the
719         benchmark version string to the dashboard, which holds the results.
720         (window.sectionsManager.setSectionVersion): Helper function to update the element in the
721         section with the class name version.
722         (window.benchmarkController.initialize): Populate all DOM elements with class name "version"
723         with the version string. Update the page title.
724         (window.benchmarkController.showResults): When showing results, update the version string
725         based on what is included in the JSON results, which would be the same as the benchmark version.
726         * MotionMark/resources/runner/motionmark.css: Include missing copyright. Wrap the SVG logo
727         in a div and include the version string.
728         * MotionMark/resources/strings.js: Add strings for the page title template, and the version.
729
730         * MotionMark/resources/debug-runner/motionmark.css:
731         * MotionMark/resources/debug-runner/motionmark.js:
732         (window.benchmarkRunnerClient.willStartFirstIteration): When running the benchmark, pass the
733         benchmark version string to the dashboard, which holds the results.
734         (window.benchmarkController.initialize): Populate all DOM elements with class name "version"
735         with the version string. Update the page title. When dragging in JSON results, look for
736         version to pass to the dashboard. If it doesn't exist, default to "1.0".
737         (window.benchmarkController.showResults): When showing results, update the version string
738         based on what is included in the JSON results, instead of the current benchmark version.
739         * MotionMark/resources/debug-runner/tests.js: Update page title template.
740
741 2018-06-08  Jon Lee  <jonlee@apple.com>
742
743         [MotionMark] Rename Suits test files
744         https://bugs.webkit.org/show_bug.cgi?id=186447
745
746         Reviewed by Said Abou-Hallawa.
747
748         * MotionMark/resources/runner/tests.js:
749         * MotionMark/tests/master/resources/suits.js: Renamed from PerformanceTests/MotionMark/tests/master/resources/svg-particles.js.
750         * MotionMark/tests/master/suits.html: Renamed from PerformanceTests/MotionMark/tests/master/svg-particles.html.
751         * MotionMark/tests/svg/suits.html:
752
753 2018-06-08  Jon Lee  <jonlee@apple.com>
754
755         Add sub-tests based on Suits
756         https://bugs.webkit.org/show_bug.cgi?id=186260
757
758         Reviewed by Said Abou-Hallawa.
759
760         Add a new developer Suits suite with sub-tests that isolate parts of each particle.
761         The sub-tests are:
762             - Particles using only clip paths
763             - Particles using only shapes
764             - Particles that have no gradients
765             - Particles that have no rotation
766             - Particles that do not move around at all (but all of the physics calculations are
767             still performed)
768
769         * MotionMark/resources/debug-runner/tests.js:
770         * MotionMark/tests/master/resources/svg-particles.js: Rename the stage, particle, and
771         benchmark to "Suits" from "SVG". Change SuitsParticle so that we guarantee 50% clip and
772         shape paths rather than relying on random chance. Wrap a check around creation of the
773         gradient element, and use a simple fill color when we aren't using a gradient.
774         * MotionMark/tests/svg/suits.html: Added.
775         * MotionMark/tests/svg/suits.js: Added.
776         Look for the query string and set the particle based on what is selected.
777
778 2018-06-06  Jon Lee  <jonlee@apple.com>
779
780         Allow the ramp controller to run tests that take less time than the initial ramp-up phase
781         https://bugs.webkit.org/show_bug.cgi?id=186257
782
783         Reviewed by Said Abou-Hallawa.
784
785         * MotionMark/tests/resources/main.js: Extend the desired end timestamp when ramp-up occurs.
786         Previously the controller assumed that the duration of the test was always longer than the
787         time it took to figure out the order of magnitude of particles to render. When that range
788         is determined, the end timestamp is then extended. If the duration is too short, the controller
789         would throw an exception. The fix is to continue extending the timestamp every time the order of
790         magnitude tiers up.
791
792         In general testing with really short durations isn't practical, but sometimes it is
793         desired for debugging.
794
795         To avoid confusion of the private variables, I renamed Benchmark._startTimestamp to
796         Benchmark._benchmarkStartTimestamp, to disambiguate from Controller._startTimestamp.
797         Benchmark._startTimestamp is set once and never changed, after the initial 100ms warm-up.
798         Controller._startTimestamp is initialized to 0 (which means "not initialized"), and then after
799         the warm up is complete, set in Controller.start() to the same wall clock timestamp, and never
800         changed afterwards.
801
802 2018-06-06  Jon Lee  <jonlee@apple.com>
803
804         Remove unneeded data processing in MotionMark
805         https://bugs.webkit.org/show_bug.cgi?id=186256
806
807         Reviewed by Said Abou-Hallawa.
808
809         Remove calculation of average frame rate based on multiple samples at a given complexity.
810         It is not used in the calculation of the score, and is just extra work to do at the end
811         of a test.
812
813         * MotionMark/developer.html: Remove options to show average data and regression based on
814         that data.
815         * MotionMark/resources/debug-runner/graph.js: Remove data. Update the time graph that
816         shows the relationship between the break point in the regression against the complexity
817         graph.
818         * MotionMark/resources/debug-runner/motionmark.css: Make the confidence intervals easier
819         to see.
820         * MotionMark/resources/runner/motionmark.js: Simplify score calculation now that we are
821         only calculating regressions on the raw data. We were calculating it based on the averages,
822         but that was just for analysis, and not used in the actual score. Get rid of the
823         |complexitySamples| local variable.
824         * MotionMark/resources/strings.js: Remove unneeded string constant.
825         * MotionMark/tests/resources/main.js:
826         (processSamples): Refactor so that more of it can be shared between Controller and
827         RampController. It also avoids processing the raw samples (controllerSamples) twice for
828         the ramp controller.
829         (_processComplexitySamples): Remove logic to calculate averages based on complexity.
830         (_processMarks): Split this out from the old processSamples.
831         (_processControllerSamples): Split this out from the old processSamples.
832
833 2018-06-06  Jon Lee  <jonlee@apple.com>
834
835         Strip out more characters when creating permalinks
836         https://bugs.webkit.org/show_bug.cgi?id=186259
837
838         Reviewed by Said Abou-Hallawa.
839
840         Tests and suites may have names that make permalinks to the tests less readable when
841         shared. Expand the set of characters that are stripped out of the names.
842
843         * MotionMark/resources/debug-runner/motionmark.js: Refactor to call
844         stripUnwantedCharactersForURL. For existing links, decode the suite and test names.
845         * MotionMark/resources/extensions.js:
846         (Utilities.stripUnwantedCharactersForURL): Rename from stripNonASCIICharacters.
847         Remove any non-alphanumeric character.
848
849 2018-06-04  Jon Lee  <jonlee@apple.com>
850
851         Remove unnecessary MotionMark controllers
852         https://bugs.webkit.org/show_bug.cgi?id=186255
853
854         Reviewed by Said Abou-Hallawa.
855
856         Get rid of the ramp30 and the fixed-with-a-step controllers, which aren't used for testing
857         or debugging.
858
859         * MotionMark/developer.html: Move the "ramp" option as first, since that is the default
860         controller for the benchmark.
861         * MotionMark/resources/debug-runner/motionmark.js:
862         * MotionMark/resources/runner/motionmark.js: Added missing copyright notice.
863         * MotionMark/tests/resources/main.js:
864
865 2018-05-25  Myles C. Maxfield  <mmaxfield@apple.com>
866
867         Improve the performance of Font::canRenderCombiningCharacterSequence()
868         https://bugs.webkit.org/show_bug.cgi?id=185933
869
870         Reviewed by Ryosuke Niwa.
871
872         * Layout/ComplexLongUnique.html: Added.
873
874 2018-05-25  Saam Barati  <sbarati@apple.com>
875
876         Have a memory test where we can validate JSCs mini memory mode
877         https://bugs.webkit.org/show_bug.cgi?id=185932
878
879         Reviewed by Mark Lam.
880
881         We add a directory here with the contents of the testmem benchmark.
882         To run it, use `Tools/Scripts/run-testmem`. To add new tests in the future,
883         you just need to add JS files to this directory.
884
885         * testmem: Added.
886         * testmem/air.js: Added.
887         * testmem/base64.js: Added.
888         * testmem/basic.js: Added.
889         * testmem/box2d.js: Added.
890         * testmem/crypto-md5.js: Added.
891         * testmem/date-format-tofte.js: Added.
892         * testmem/earley-boyer.js: Added.
893         * testmem/hash-map.js: Added.
894         * testmem/regex-dna.js: Added.
895         * testmem/splay.js: Added.
896         * testmem/tagcloud.js: Added.
897
898 2018-05-22  Ryan Haddad  <ryanhaddad@apple.com>
899
900         Unreviewed, rolling out r232052.
901
902         Breaks internal builds.
903
904         Reverted changeset:
905
906         "Use more C++17"
907         https://bugs.webkit.org/show_bug.cgi?id=185176
908         https://trac.webkit.org/changeset/232052
909
910 2018-05-21  Yusuke Suzuki  <utatane.tea@gmail.com>
911
912         Use more C++17
913         https://bugs.webkit.org/show_bug.cgi?id=185176
914
915         Reviewed by JF Bastien.
916
917         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
918         * MediaTime/Configurations/Base.xcconfig:
919         * MediaTime/MediaTime.xcodeproj/project.pbxproj:
920
921 2018-04-30  Myles C. Maxfield  <mmaxfield@apple.com>
922
923         Improve the performance of FontCascadeDescription's effectiveFamilies
924         https://bugs.webkit.org/show_bug.cgi?id=184720
925
926         Reviewed by Simon Fraser.
927
928         This performance test calls SystemFontDatabase::systemFontCascadeList() around 2,000,000 times (before
929         this patch is applied), which is roughly equivalent to the page we found the performance problem on.
930         The calling pattern is roughly equivalent in this test.
931
932         * Layout/system-ui.html: Added.
933
934 2018-03-08  Antti Koivisto  <antti@apple.com>
935
936         Update StyleBench version number in page title to 0.3
937
938         * StyleBench/index.html:
939
940         There has been a bunch of changes, most notably the addition of attribute selector test step.
941
942 2018-02-16  Dean Jackson  <dino@apple.com>
943
944         Use OPENGL macros to be more clear about which OpenGL/ES WebGL uses on Cocoa
945         https://bugs.webkit.org/show_bug.cgi?id=182894
946
947         Reviewed by Tim Horton.
948
949         Rename OPENGL_ES_2 to OPENGL_ES.
950
951         * StitchMarker/wtf/Platform.h:
952
953 2018-02-09  Ryosuke Niwa  <rniwa@webkit.org>
954
955         Make run-perf-tests work with Speedometer 2.0 and re-enable the test
956         https://bugs.webkit.org/show_bug.cgi?id=182088
957
958         Reviewed by Antti Koivisto.
959
960         This patch updates the test harness to report new metrics of Speedometer 2.0.
961
962         Replaced the total time by the geometric mean of the time spent in each suite, and added scores
963         as a metric in addition to time. Also, updated the PerfTestRunner harness and the results template
964         to support "Score" metric with "pt" as its unit.
965
966         * Skipped: Unskipped the test.
967         * Speedometer/resources/benchmark-report.js:
968         (window.benchmarkClient.willStartFirstIteration..createTest): Added the unit as an optional argument
969         to report the score in addition to time. Also report the geomean of time instead of the total time.
970         (window.benchmarkClient.willStartFirstIteration): Report the geomean of time.
971         (window.benchmarkClient.didRunSuites): Record the entire measuredValues instead of just tests.
972         (window.benchmarkClient.didFinishLastIteration): Report the scores.
973
974 2018-02-09  Ross Kirsling  <ross.kirsling@sony.com>
975
976         Use REQUEST_ANIMATION_FRAME_DISPLAY_MONITOR on any non-Windows port.
977         https://bugs.webkit.org/show_bug.cgi?id=182623
978
979         Reviewed by Alex Christensen.
980
981         * StitchMarker/wtf/Platform.h:
982         Simplify #if.
983
984 2018-02-01  Antti Koivisto  <antti@apple.com>
985
986         StyleBench: Attribute selectors and other improvements
987         https://bugs.webkit.org/show_bug.cgi?id=182387
988
989         Reviewed by Joseph Pecoraro.
990
991         - Add some attributes to elements in all tests
992         - Add some attribute selectors to stylesheets in all tests
993         - Also add some * selectors to all stylesheets.
994         - Add attribute mutation step to all suites
995         - Make test steps do more mutations (25->100) and reduce the number of steps to keep testing time in check.
996           Too fast steps were running into timer resolution limits.
997
998         * StyleBench/resources/style-bench.js:
999         (defaultConfiguration):
1000         (prototype.randomAttributeName):
1001         (prototype.randomAttributeValue):
1002         (prototype.randomAttributeSelector):
1003         (prototype.makeCompoundSelector):
1004         (prototype.makeElement):
1005         (prototype.addClasses):
1006         (prototype.removeClasses):
1007         (prototype.mutateAttributes):
1008         (prototype.async.runForever):
1009         * StyleBench/resources/tests.js:
1010         (makeSteps):
1011         (makeSuite):
1012
1013 2018-02-01  Geoffrey Garen  <ggaren@apple.com>
1014
1015         Make MallocBench easier for non-WebKit engineers to run
1016         https://bugs.webkit.org/show_bug.cgi?id=182415
1017
1018         Reviewed by Saam Barati.
1019
1020         * MallocBench/MallocBench.xcodeproj/project.pbxproj: Use c++14 so we
1021         can make_unique.
1022
1023         Specify that we support all Darwin platforms so you can test them.
1024
1025         * MallocBench/run-malloc-benchmarks: Specify the path to MallocBench
1026         and libmbmalloc explicitly, rather than computing them implicitly
1027         using webkitdirs. Non-WebKit folks don't have the directory structure
1028         required by webkitdirs.
1029
1030         Remove Linux-specific and cmake-specific behaviors because we only
1031         needed them in the world of implicit path computation.
1032
1033 2018-01-31  Ryosuke Niwa  <rniwa@webkit.org>
1034
1035         Make run-perf-tests work with StyleBench and re-enable it
1036         https://bugs.webkit.org/show_bug.cgi?id=182322
1037
1038         Reviewed by Antti Koivisto.
1039
1040         This patch updates the test harness to report new metrics of StyleBench.
1041
1042         Replaced the total time by the geometric mean of the time spent in each suite, and added scores
1043         as a metric in addition to time. Also, updated the PerfTestRunner harness and the results template
1044         to support "Score" metric with "pt" as its unit.
1045
1046         * Skipped: Unskipped the test.
1047         * StyleBench/resources/benchmark-report.js:
1048         (window.benchmarkClient.willStartFirstIteration..createTest): Added the unit as an optional argument
1049         to report the score in addition to time. Also report the geomean of time instead of the total time.
1050         (window.benchmarkClient.willStartFirstIteration): Report the geomean of time.
1051         (window.benchmarkClient.didRunSuites): Record the entire measuredValues instead of just tests.
1052         (window.benchmarkClient.didFinishLastIteration): Report the scores.
1053         * resources/results-template.html: Added "Score" as a metric and "pt" as its unit.
1054         * resources/runner.js:
1055         (PerfTestRunner..finish): Ditto.
1056
1057 2018-01-30  Ryosuke Niwa  <rniwa@webkit.org>
1058
1059         StyleBench: Generate more mutation steps
1060         https://bugs.webkit.org/show_bug.cgi?id=182291
1061
1062         Rubber-stamped by Antti Koivisto
1063
1064         Use zero-based index instead of one-based index.
1065
1066         * StyleBench/resources/tests.js:
1067         (makeSteps):
1068
1069 2018-01-30  Antti Koivisto  <antti@apple.com>
1070
1071         StyleBench: Generate more mutation steps
1072         https://bugs.webkit.org/show_bug.cgi?id=182291
1073
1074         Reviewed by Ryosuke Niwa.
1075
1076         The test should be doing multiple rounds of mutations in a single document to test what is supposed to test (like it did before r227756).
1077         This also makes test results more stable and less dependent on possible first-time costs.
1078
1079         * StyleBench/resources/benchmark-runner.js:
1080
1081         Adjust correctionFactor to match the new steps count. With 1.5 Firefox is ~72rpm.
1082
1083         * StyleBench/resources/tests.js:
1084         (makeSteps):
1085
1086         Generate 10 rounds of class/element mutation steps.
1087         Give each generated step unique name so reporting works.
1088         Make individual steps do fewer mutations.
1089
1090 2018-01-29  Antti Koivisto  <antti@apple.com>
1091
1092         StyleBench: Remove : from a test name
1093         https://bugs.webkit.org/show_bug.cgi?id=182258
1094
1095         Reviewed by Ryosuke Niwa.
1096
1097         It breaks reporting.
1098
1099         * StyleBench/resources/style-bench.js:
1100         (nthPseudoClassConfiguration):
1101
1102 2018-01-29  Ryosuke Niwa  <rniwa@webkit.org>
1103
1104         Temporarily skip StyleBench until webkit.org/b/182088 is fixed.
1105
1106         * Skipped:
1107
1108 2018-01-29  Ryosuke Niwa  <rniwa@webkit.org>
1109
1110         Make StyleBench compatible with run-benchmark and run-perf-tests
1111         https://bugs.webkit.org/show_bug.cgi?id=182262
1112
1113         Reviewed by Antti Koivisto.
1114
1115         Copied resource files referenced from Speedometer directory since run-benchmark needs to be able
1116         to checkout each benchmark separately.
1117
1118         Removed the code to create tests of the same name five times in makeSteps since this
1119         won't be compatible with either run-benchmark or run-perf-tests.
1120
1121         * StyleBench/index.html: Removed the code to show warnings for local files since run-benchmark
1122         doesn't use HTTP server in WebDriver mode.
1123         * StyleBench/resources/benchmark-report.js: Copied from resources/benchmark-report.js.
1124         * StyleBench/resources/benchmark-runner.js: Copied from resources/benchmark-runner.js.
1125         (BenchmarkRunner.prototype._finalize): Use the correction factor of 8 instead of 5 in StyleBench.
1126         * StyleBench/resources/gauge.png: Copied from resources/gauge.png.
1127         * StyleBench/resources/gauge@2x.png: Copied from resources/gauge@2x.png.
1128         * StyleBench/resources/main.css: Copied from resources/main.css.
1129         * StyleBench/resources/main.js: Copied from resources/main.js.
1130         * StyleBench/resources/tests.js:
1131         (makeSteps): Only make each test once.
1132
1133 2018-01-26  Chris Nardi  <cnardi@chromium.org>
1134
1135         Addressing post-review comments after r226614
1136         https://bugs.webkit.org/show_bug.cgi?id=182151
1137
1138         Reviewed by Myles C. Maxfield.
1139
1140         * StitchMarker/wtf/text/StringImpl.h:
1141         (WTF::isSpaceOrNewline):
1142         * StitchMarker/wtf/text/TextBreakIterator.cpp:
1143         (WTF::numCodeUnitsInGraphemeClusters):
1144         * StitchMarker/wtf/text/TextBreakIterator.h:
1145
1146 2018-01-22  Antti Koivisto  <antti@apple.com>
1147
1148         StyleBench: Separate test for :nth pseudo classes
1149         https://bugs.webkit.org/show_bug.cgi?id=181932
1150
1151         Reviewed by Sam Weinig.
1152
1153         Optimization strategies for nth are different from other structural pseudo classes (:first-child etc). Move them to a separate test.
1154
1155         * StyleBench/resources/style-bench.js:
1156         (structuralPseudoClassConfiguration):
1157         (nthPseudoClassConfiguration):
1158
1159             Add a new configuration.
1160
1161         (predefinedConfigurations):
1162         (pseudoClassConfiguration): Deleted.
1163
1164 2018-01-17  Ryosuke Niwa  <rniwa@webkit.org>
1165
1166         Temporarily skip Speedometer until we fix the metrics.
1167
1168         * Skipped:
1169
1170 2018-01-16  Alex Christensen  <achristensen@webkit.org>
1171
1172         Fix Mac CMake build after r222900
1173
1174         * MallocBench/MallocBench/CMakeLists.txt:
1175         The mbmalloc target doesn't exist if we're not in developer mode, so only use it if we are.
1176
1177 2018-01-09  Ryosuke Niwa  <rniwa@webkit.org>
1178
1179         Speedometer 2.0: Update the About page.
1180         https://bugs.webkit.org/show_bug.cgi?id=175918
1181
1182         Reviewed by Keith Miller.
1183
1184         Updated the about page for Speedometer 2.0.
1185
1186         * Speedometer/index.html:
1187
1188 == Rolled over to ChangeLog-2018-01-01 ==