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