Unreviewed, rolling in r241237 again
[WebKit-https.git] / PerformanceTests / ChangeLog
1 2019-02-08  Saam barati  <sbarati@apple.com>
2
3         Update JetStream2 CLI to not print "ms" since we now just print the score values
4         https://bugs.webkit.org/show_bug.cgi?id=193981
5
6         Reviewed by Yusuke Suzuki.
7
8         * JetStream2/JetStreamDriver.js:
9         (DefaultBenchmark.prototype.updateUIAfterRun):
10         (DefaultBenchmark):
11         (WSLBenchmark.prototype.updateUIAfterRun):
12         (WSLBenchmark):
13         (WasmBenchmark.prototype.updateUIAfterRun):
14         (WasmBenchmark):
15
16 2019-01-31  Sihui Liu  <sihui_liu@apple.com>
17
18         REGRESSION (r240358): IndexedDB/large-binary-keys.html and IndexedDB/large-string-keys.html perf tests failing
19         https://bugs.webkit.org/show_bug.cgi?id=193956
20         <rdar://problem/47623109>
21
22         Reviewed by Ryosuke Niwa.
23
24         r240358 sets default IDB quota as 50 MB for single test, we need to increase that for big tests. 
25
26         * IndexedDB/large-binary-keys.html:
27         * IndexedDB/large-string-keys.html:
28
29 2019-01-18  Saam Barati  <sbarati@apple.com>
30
31         Use scores everywhere in JetStream2's UI
32         https://bugs.webkit.org/show_bug.cgi?id=193593
33
34         Reviewed by Keith Miller.
35
36         * JetStream2/JetStreamDriver.js:
37         (DefaultBenchmark.prototype.processResults):
38         (DefaultBenchmark.prototype.get score):
39         (DefaultBenchmark.scoreDescription):
40         (WSLBenchmark.prototype.processResults):
41         (WSLBenchmark.prototype.get score):
42         (WSLBenchmark.scoreDescription):
43         (WasmBenchmark.prototype.processResults):
44         (WasmBenchmark.prototype.get score):
45         (WasmBenchmark.scoreDescription):
46         * JetStream2/index.html:
47
48 2019-01-18  Jer Noble  <jer.noble@apple.com>
49
50         SDK_VARIANT build destinations should be separate from non-SDK_VARIANT builds
51         https://bugs.webkit.org/show_bug.cgi?id=189553
52
53         Reviewed by Tim Horton.
54
55         * MediaTime/Configurations/Base.xcconfig:
56         * MediaTime/Configurations/SDKVariant.xcconfig: Added.
57
58 2019-01-17  Truitt Savell  <tsavell@apple.com>
59
60         Unreviewed, rolling out r240124.
61
62         This commit broke an internal build.
63
64         Reverted changeset:
65
66         "SDK_VARIANT build destinations should be separate from non-
67         SDK_VARIANT builds"
68         https://bugs.webkit.org/show_bug.cgi?id=189553
69         https://trac.webkit.org/changeset/240124
70
71 2019-01-17  Jer Noble  <jer.noble@apple.com>
72
73         SDK_VARIANT build destinations should be separate from non-SDK_VARIANT builds
74         https://bugs.webkit.org/show_bug.cgi?id=189553
75
76         Reviewed by Tim Horton.
77
78         * MediaTime/Configurations/Base.xcconfig:
79         * MediaTime/Configurations/SDKVariant.xcconfig: Added.
80
81 2018-12-20  Caio Lima  <ticaiolima@gmail.com>
82
83         [BigInt] We should enable CSE into arithmetic operations that speculate BigIntUse
84         https://bugs.webkit.org/show_bug.cgi?id=192723
85
86         Reviewed by Yusuke Suzuki.
87
88         * BigIntBench/big-int-cse.js: Added.
89         * BigIntBench/big-int-global-cse.js: Added.
90         * BigIntBench/big-int-licm.js: Added.
91
92 2018-12-19  Commit Queue  <commit-queue@webkit.org>
93
94         Unreviewed, rolling out r239377.
95         https://bugs.webkit.org/show_bug.cgi?id=192921
96
97         broke 32-bit JSC tests (Requested by keith_miller on #webkit).
98
99         Reverted changeset:
100
101         "[BigInt] We should enable CSE into arithmetic operations that
102         speculate BigIntUse"
103         https://bugs.webkit.org/show_bug.cgi?id=192723
104         https://trac.webkit.org/changeset/239377
105
106 2018-12-19  Caio Lima  <ticaiolima@gmail.com>
107
108         [BigInt] We should enable CSE into arithmetic operations that speculate BigIntUse
109         https://bugs.webkit.org/show_bug.cgi?id=192723
110
111         Reviewed by Saam Barati.
112
113         * BigIntBench/big-int-cse.js: Added.
114         * BigIntBench/big-int-global-cse.js: Added.
115         * BigIntBench/big-int-licm.js: Added.
116
117 2018-12-17  Suresh Koppisetty  <skoppisetty@apple.com>
118
119         Add "-o/--output" option to startup.py and new_tab.py benchmark scripts to save the results in json format.
120         https://bugs.webkit.org/show_bug.cgi?id=192385
121
122         Reviewed by Ryosuke Niwa.
123
124         Sample json output for new tab benchmark script after running for 2 iterations and 2 groups. Values are in milliseconds.
125         {
126                 "NewTabBenchmark": {
127                         "metrics": {
128                                 "Time": {
129                                         "current": [
130                                                 [
131                                                         410.2939453125,
132                                                         307.81494140625
133                                                 ],
134                                                 [
135                                                         340.616943359375,
136                                                         265.94384765625
137                                                 ]
138                                         ]
139                                 }
140                         }
141                 }
142         }
143
144         Sample json output for startup time benchmark script after running for 2 iterations. Values are in milliseconds.
145         {
146                 "StartupBenchmark": {
147                         "metrics": {
148                                 "Time": {
149                                         "current": [
150                                                 [
151                                                         1415.2099609375,
152                                                         1439.552978515625
153                                                 ]
154                                         ]
155                                 }
156                         }
157                 }
158         }
159
160         * LaunchTime/launch_time.py:
161         * LaunchTime/new_tab.py:
162         (NewTabBenchmark.get_test_name):
163         * LaunchTime/startup.py:
164         (StartupBenchmark.get_test_name):
165
166 2018-12-17  Suresh Koppisetty  <skoppisetty@apple.com>
167
168         Import FeedbackServer only if "-f/--feedback-in-browser" option is enabled.
169         https://bugs.webkit.org/show_bug.cgi?id=192378
170
171         Reviewed by Ryosuke Niwa.
172
173         FeedbackServer currently depends on Tornado-5.1, which further adds
174         dependency of "singledispatch", "backports-abc" and "futures" python libraries.
175         Importing FeedbackServer only if "-f/--feedback-in-browser" option is enabled
176         will let us run the benchmark scripts without installing any new python libraries.
177
178         * LaunchTime/launch_time.py:
179         * LaunchTime/new_tab.py:
180
181 2018-12-13  Caio Lima  <ticaiolima@gmail.com>
182
183         [BigInt] Add ValueDiv into DFG
184         https://bugs.webkit.org/show_bug.cgi?id=186178
185
186         Reviewed by Yusuke Suzuki.
187
188         * BigIntBench/big-int-simple-div.js: Added.
189         * BigIntBench/value-div-type-propagation.js: Added.
190
191 2018-12-10  Caio Lima  <ticaiolima@gmail.com>
192
193         [BigInt] Add ValueMul into DFG
194         https://bugs.webkit.org/show_bug.cgi?id=186175
195
196         Reviewed by Yusuke Suzuki.
197
198         * BigIntBench/big-int-simple-mul.js: Added.
199         * BigIntBench/value-mul-type-propagation.js: Added.
200
201 2018-11-07  Caio Lima  <ticaiolima@gmail.com>
202
203         [BigInt] Add support to BigInt into ValueAdd
204         https://bugs.webkit.org/show_bug.cgi?id=186177
205
206         Reviewed by Keith Miller.
207
208         The idea of BigIntBench is to provide a set of microbenchmarks and
209         benchmarks to evaluate how fast BigInt computations are happening on
210         JSC implementation.
211
212         Now, we are adding microbenchmarks in this set,
213         but the plan is to move these tests to "JSTest/microbenchmarks" when
214         BigInt is enabled by default. After that, the focus of Bigint bench is
215         to provide a set of tests that represents real use cases of BigInt in
216         JS programs.
217
218         * BigIntBench/big-int-add-prediction-propagation.js: Added.
219         * BigIntBench/big-int-simple-add.js: Added.
220         * BigIntBench/big-int-simple-sub.js: Added.
221
222 2018-11-07  Tadeu Zagallo  <tzagallo@apple.com>
223
224         REGRESSION(r237547): Test failures on 32-bit JSC since the JIT was disabled
225         https://bugs.webkit.org/show_bug.cgi?id=191184
226
227         Reviewed by Saam Barati.
228
229         Skip test when the JIT is disabled, it takes too long to run on CLoop.
230
231         * JetStream/cdjs/main.js:
232
233 2018-11-05  Myles C. Maxfield  <mmaxfield@apple.com>
234
235         Cache glyph paths and share underline skipping code between all the ports
236         https://bugs.webkit.org/show_bug.cgi?id=191239
237
238         Reviewed by Alex Christensen.
239
240         Measures the performance of drawing a whole lot of underlines
241
242         * Layout/underline.html: Added.
243
244 2018-10-25  Saam Barati  <sbarati@apple.com>
245
246         Check in corresponding C code in JetStream 2
247         https://bugs.webkit.org/show_bug.cgi?id=190733
248
249         Rubber-stamped by Filip Pizlo.
250
251         * JetStream2/simple/float-mm.c: Added.
252         * JetStream2/wasm/HashSet.cpp: Added.
253         * JetStream2/wasm/gcc-loops.cpp: Added.
254         * JetStream2/wasm/quicksort.c: Added.
255         * JetStream2/wasm/TSF: Added.
256         * JetStream2/wasm/TSF/build.sh: Added.
257         * JetStream2/wasm/TSF/config.h: Added.
258         * JetStream2/wasm/TSF/gpc.h: Added.
259         * JetStream2/wasm/TSF/gpc_code_gen_util.c: Added.
260         * JetStream2/wasm/TSF/gpc_instruction.c: Added.
261         * JetStream2/wasm/TSF/gpc_instruction_dispatch.gen: Added.
262         * JetStream2/wasm/TSF/gpc_instruction_size.gen: Added.
263         * JetStream2/wasm/TSF/gpc_instruction_stack_effects.gen: Added.
264         * JetStream2/wasm/TSF/gpc_instruction_static_size.gen: Added.
265         * JetStream2/wasm/TSF/gpc_instruction_to_string.gen: Added.
266         * JetStream2/wasm/TSF/gpc_int_common.h: Added.
267         * JetStream2/wasm/TSF/gpc_intable.c: Added.
268         * JetStream2/wasm/TSF/gpc_internal.h: Added.
269         * JetStream2/wasm/TSF/gpc_interpreter.gen: Added.
270         * JetStream2/wasm/TSF/gpc_program.c: Added.
271         * JetStream2/wasm/TSF/gpc_proto.c: Added.
272         * JetStream2/wasm/TSF/gpc_stack_height.c: Added.
273         * JetStream2/wasm/TSF/gpc_threaded.c: Added.
274         * JetStream2/wasm/TSF/gpc_worklist.h: Added.
275         * JetStream2/wasm/TSF/tsf.h: Added.
276         * JetStream2/wasm/TSF/tsf_adaptive_reader.c: Added.
277         * JetStream2/wasm/TSF/tsf_asprintf.c: Added.
278         * JetStream2/wasm/TSF/tsf_atomics.h: Added.
279         * JetStream2/wasm/TSF/tsf_buf_reader.c: Added.
280         * JetStream2/wasm/TSF/tsf_buf_writer.c: Added.
281         * JetStream2/wasm/TSF/tsf_buffer.c: Added.
282         * JetStream2/wasm/TSF/tsf_build_defines.h: Added.
283         * JetStream2/wasm/TSF/tsf_config.h: Added.
284         * JetStream2/wasm/TSF/tsf_config_stub.h: Added.
285         * JetStream2/wasm/TSF/tsf_copier.c: Added.
286         * JetStream2/wasm/TSF/tsf_define_helpers.c: Added.
287         * JetStream2/wasm/TSF/tsf_define_helpers.h: Added.
288         * JetStream2/wasm/TSF/tsf_destructor.c: Added.
289         * JetStream2/wasm/TSF/tsf_error.c: Added.
290         * JetStream2/wasm/TSF/tsf_format.h: Added.
291         * JetStream2/wasm/TSF/tsf_fsdb.c: Added.
292         * JetStream2/wasm/TSF/tsf_fsdb_protocol.c: Added.
293         * JetStream2/wasm/TSF/tsf_fsdb_protocol.h: Added.
294         * JetStream2/wasm/TSF/tsf_generator.c: Added.
295         * JetStream2/wasm/TSF/tsf_gpc_code_gen.c: Added.
296         * JetStream2/wasm/TSF/tsf_indent.h: Added.
297         * JetStream2/wasm/TSF/tsf_internal.h: Added.
298         * JetStream2/wasm/TSF/tsf_internal_config.h: Added.
299         * JetStream2/wasm/TSF/tsf_internal_config_stub.h: Added.
300         * JetStream2/wasm/TSF/tsf_inttypes.h: Added.
301         * JetStream2/wasm/TSF/tsf_io.c: Added.
302         * JetStream2/wasm/TSF/tsf_io_utils.c: Added.
303         * JetStream2/wasm/TSF/tsf_ir.c: Added.
304         * JetStream2/wasm/TSF/tsf_ir.h: Added.
305         * JetStream2/wasm/TSF/tsf_ir_different.c: Added.
306         * JetStream2/wasm/TSF/tsf_ir_different.h: Added.
307         * JetStream2/wasm/TSF/tsf_ir_speed.c: Added.
308         * JetStream2/wasm/TSF/tsf_limits.c: Added.
309         * JetStream2/wasm/TSF/tsf_named_type.c: Added.
310         * JetStream2/wasm/TSF/tsf_native.c: Added.
311         * JetStream2/wasm/TSF/tsf_parser.c: Added.
312         * JetStream2/wasm/TSF/tsf_primitive.c: Added.
313         * JetStream2/wasm/TSF/tsf_ra_type_man.c: Added.
314         * JetStream2/wasm/TSF/tsf_reflect.c: Added.
315         * JetStream2/wasm/TSF/tsf_region.h: Added.
316         * JetStream2/wasm/TSF/tsf_serial_in_man.c: Added.
317         * JetStream2/wasm/TSF/tsf_serial_out_man.c: Added.
318         * JetStream2/wasm/TSF/tsf_serial_protocol.h: Added.
319         * JetStream2/wasm/TSF/tsf_sha1.c: Added.
320         * JetStream2/wasm/TSF/tsf_sha1.h: Added.
321         * JetStream2/wasm/TSF/tsf_sha1_writer.c: Added.
322         * JetStream2/wasm/TSF/tsf_sort.c: Added.
323         * JetStream2/wasm/TSF/tsf_st.c: Added.
324         * JetStream2/wasm/TSF/tsf_st.h: Added.
325         * JetStream2/wasm/TSF/tsf_st_typetable.c: Added.
326         * JetStream2/wasm/TSF/tsf_stream_file_input.c: Added.
327         * JetStream2/wasm/TSF/tsf_stream_file_output.c: Added.
328         * JetStream2/wasm/TSF/tsf_type.c: Added.
329         * JetStream2/wasm/TSF/tsf_type_in_map.c: Added.
330         * JetStream2/wasm/TSF/tsf_type_out_map.c: Added.
331         * JetStream2/wasm/TSF/tsf_type_table.c: Added.
332         * JetStream2/wasm/TSF/tsf_types.h: Added.
333         * JetStream2/wasm/TSF/tsf_util.h: Added.
334         * JetStream2/wasm/TSF/tsf_version.c: Added.
335         * JetStream2/wasm/TSF/tsf_zip_abstract.c: Added.
336         * JetStream2/wasm/TSF/tsf_zip_abstract.h: Added.
337         * JetStream2/wasm/TSF/tsf_zip_attr.c: Added.
338         * JetStream2/wasm/TSF/tsf_zip_reader.c: Added.
339         * JetStream2/wasm/TSF/tsf_zip_writer.c: Added.
340
341 2018-10-18  Saam Barati  <sbarati@apple.com>
342
343         Make JetStream 2
344         https://bugs.webkit.org/show_bug.cgi?id=187829
345
346         Rubber-stamped by Mark Lam.
347
348         This patch checks in the new JetStream 2 benchmark. JetStream 2 is
349         a new JavaScript and Web Assembly benchmark. JetStream 2's goal
350         is to measure the startup, worst case, and peak throughput performance
351         of the JavaScript engine. JetStream 2 incorporates these previous
352         benchmarks:
353         - JetStream
354         - ARES-6
355         - Kraken
356         - Web Tooling Benchmark
357         - WasmBench
358         - RexBench
359         
360         JetStream 2 also adds some new benchmarks:
361         - Two tests emphasizing web worker performance.
362         - One test emphasizing Promise, async iteration, and DataView performance.
363         - Two new code load tests.
364         - WSL: a test measuring all kinds of things, especially emphasizing exception performance.
365
366         * JetStream2: Added.
367
368 2018-10-02  Chris Dumez  <cdumez@apple.com>
369
370         Regression(r236613): Parser/html-parser.html performance test is failing
371         https://bugs.webkit.org/show_bug.cgi?id=190201
372
373         Unreviewed, apply the same fix as in r236699 to make sure the iframe is treated as
374         same origin, so that calls to document.open() / document.write() succeed.
375
376         * Parser/html-parser.html:
377
378 2018-10-01  Chris Dumez  <cdumez@apple.com>
379
380         Regression(r236613): Parser/HTML5-8266-ParseOnly.html performance tests no longer runs
381         https://bugs.webkit.org/show_bug.cgi?id=190174
382
383         Reviewed by Ryosuke Niwa.
384
385         Update iframe's sandbox to use 'allow-same-origin' so that the iframe is treated as same-origin and so that
386         the top frame can call document.open() / document.write() on the subframe. This test was also failing in
387         Firefox without this change.
388
389         * Parser/HTML5-8266-ParseOnly.html:
390
391 2018-08-27  Aditya Keerthi  <akeerthi@apple.com>
392
393         Consolidate ENABLE_INPUT_TYPE_COLOR and ENABLE_INPUT_TYPE_COLOR_POPOVER
394         https://bugs.webkit.org/show_bug.cgi?id=188931
395
396         Reviewed by Wenson Hsieh.
397
398         * StitchMarker/wtf/FeatureDefines.h: Removed ENABLE_INPUT_TYPE_COLOR_POPOVER.
399
400 2018-08-23  Don Olmstead  <don.olmstead@sony.com>
401
402         [CMake] Have checks are not getting set inside CMake properly
403         https://bugs.webkit.org/show_bug.cgi?id=188901
404
405         Reviewed by Michael Catanzaro.
406
407         Add notification that MallocBench is disabled.
408
409         * CMakeLists.txt:
410
411 2018-08-23  Don Olmstead  <don.olmstead@sony.com>
412
413         [CMake] Add HAVE_MALLOC_TRIM definition
414         https://bugs.webkit.org/show_bug.cgi?id=188897
415
416         Reviewed by Konstantin Tokarev.
417
418         MallocBench should only be built on Apple platforms and platforms that
419         have malloc_trim.
420
421         * CMakeLists.txt:
422
423 2018-08-13  Jon Lee  <jonlee@apple.com>
424
425         Update MotionMark version number
426         https://bugs.webkit.org/show_bug.cgi?id=188535
427         <rdar://problem/43254078>
428
429         Reviewed by Said Abou-Hallawa.
430
431         * MotionMark/resources/strings.js: To 1.1.
432
433 2018-08-13  Jon Lee  <jonlee@apple.com>
434
435         [MotionMark] Update Multiply test
436         https://bugs.webkit.org/show_bug.cgi?id=188532
437         <rdar://problem/43252151>
438
439         Reviewed by Said Abou-Hallawa.
440
441         Update Multiply test to have a larger upper limit of number of particles. Cycle through three
442         different ways of hiding an element. Add new Multiple suite that isolates those components for
443         debugging purposes.
444
445         * MotionMark/resources/debug-runner/tests.js:
446         * MotionMark/tests/dom/multiply.html: Added.
447         * MotionMark/tests/dom/resources/multiply.js: Added.
448         * MotionMark/tests/master/resources/multiply.js:
449
450 2018-08-13  Jon Lee  <jonlee@apple.com>
451
452         [MotionMark] Update Leaves test
453         https://bugs.webkit.org/show_bug.cgi?id=188530
454         <rdar://problem/43251862>
455
456         Reviewed by Said Abou-Hallawa.
457
458         Update Leaves test to include opacity and scale. Add new Leaves suite that isolate those components
459         for debugging purposes.
460
461         * MotionMark/resources/debug-runner/tests.js:
462         * MotionMark/tests/dom/leaves.html:
463         * MotionMark/tests/dom/resources/leaves.js:
464         * MotionMark/tests/master/resources/leaves.js:
465
466 2018-08-10  Ben Richards  <benton_richards@apple.com>
467
468         Add ability to ignore process prewarming for launch time benchmark
469         https://bugs.webkit.org/show_bug.cgi?id=188462
470
471         Reviewed by Ryosuke Niwa.
472
473         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.
474
475         * LaunchTime/launch_time.py:
476         (DefaultLaunchTimeHandler):
477         (DefaultLaunchTimeHandler.on_receive_stop_time): Deleted.
478         (DefaultLaunchTimeHandler.on_receive_stop_signal):
479         (DefaultLaunchTimeHandler.do_HEAD):
480         (DefaultLaunchTimeHandler.do_GET):
481         (DefaultLaunchTimeHandler.do_POST):
482         (LaunchTimeBenchmark):
483         (LaunchTimeBenchmark._standard_deviation): Fixed divide by zero bug when '-n' is set to 1
484         (LaunchTimeBenchmark.open_tab): Added option to open a blank tab
485         (LaunchTimeBenchmark.run):
486         * LaunchTime/new_tab.py:
487         (NewTabBenchmark.initialize):
488         (NewTabBenchmark.run_iteration):
489         (NewTabBenchmark.will_parse_arguments):
490         (NewTabBenchmark.did_parse_arguments):
491         (NewTabBenchmark.ResponseHandler.Handler.get_test_page):
492         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
493         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
494         (NewTabBenchmark):
495         * LaunchTime/startup.py:
496         (StartupBenchmark.ResponseHandler.Handler.get_test_page):
497         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
498         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
499         (StartupBenchmark):
500
501 2018-08-10  Truitt Savell  <tsavell@apple.com>
502
503         Unreviewed, rolling out r234750.
504
505         Caused 185 perf test failures.
506
507         Reverted changeset:
508
509         "Add ability to ignore process prewarming for launch time
510         benchmark"
511         https://bugs.webkit.org/show_bug.cgi?id=188462
512         https://trac.webkit.org/changeset/234750
513
514 2018-08-09  Ben Richards  <benton_richards@apple.com>
515
516         Add ability to ignore process prewarming for launch time benchmark
517         https://bugs.webkit.org/show_bug.cgi?id=188462
518
519         Reviewed by Ryosuke Niwa.
520
521         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.
522
523         * LaunchTime/launch_time.py:
524         (DefaultLaunchTimeHandler):
525         (DefaultLaunchTimeHandler.on_receive_stop_time): Deleted.
526         (DefaultLaunchTimeHandler.on_receive_stop_signal):
527         (DefaultLaunchTimeHandler.do_HEAD):
528         (DefaultLaunchTimeHandler.do_GET):
529         (DefaultLaunchTimeHandler.do_POST):
530         (LaunchTimeBenchmark):
531         (LaunchTimeBenchmark._standard_deviation): Fixed divide by zero bug when '-n' is set to 1
532         (LaunchTimeBenchmark.open_tab): Added option to open a blank tab
533         (LaunchTimeBenchmark.run):
534         * LaunchTime/new_tab.py:
535         (NewTabBenchmark.initialize):
536         (NewTabBenchmark.run_iteration):
537         (NewTabBenchmark.will_parse_arguments):
538         (NewTabBenchmark.did_parse_arguments):
539         (NewTabBenchmark.ResponseHandler.Handler.get_test_page):
540         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
541         (NewTabBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
542         (NewTabBenchmark):
543         * LaunchTime/startup.py:
544         (StartupBenchmark.ResponseHandler.Handler.get_test_page):
545         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_time): Deleted.
546         (StartupBenchmark.ResponseHandler.Handler.on_receive_stop_signal):
547         (StartupBenchmark):
548
549 2018-07-26  Andy VanWagoner  <andy@vanwagoner.family>
550
551         [INTL] Remove INTL sub-feature compile flags
552         https://bugs.webkit.org/show_bug.cgi?id=188081
553
554         Reviewed by Michael Catanzaro.
555
556         Removed old and unused i18n flag.
557
558         * StitchMarker/wtf/FeatureDefines.h:
559
560 2018-07-20  David Fenton  <david_fenton@apple.com>
561
562         REGRESSION (234006) ERROR: Blocked access to external URL http://code.jquery.com/jquery-1.9.1.min.js
563         https://bugs.webkit.org/show_bug.cgi?id=187858
564
565         Unreviewed test gardening, skipped test.
566
567         * Skipped: LaunchTime
568
569 2018-07-19  Ben Richards  <benton_richards@apple.com>
570
571         Add benchmark for WebKit process launch times
572         https://bugs.webkit.org/show_bug.cgi?id=186414
573
574         Reviewed by Ryosuke Niwa.
575
576         Added two benchmarks, one for measuring browser new tab launch time and one for browser startup time.
577
578         * LaunchTime/.gitignore: Added.
579         * LaunchTime/feedback_client.html: Added.
580         Displays benchmark progress in browser
581         * LaunchTime/feedback_server.py: Added.
582         (FeedbackServer): Sends data to feedback_client via websocket
583         (FeedbackServer.__init__):
584         (FeedbackServer._create_app):
585         (FeedbackServer._start_server):
586         (FeedbackServer._send_all_messages):
587         (FeedbackServer.start):
588         (FeedbackServer.stop):
589         (FeedbackServer.send_message): Send a message to the feedback_client
590         (FeedbackServer.wait_until_client_has_loaded): Wait until the feedback_client has opened a websocket connection to the feedback_server
591         (FeedbackServer.MainHandler): Handler factory to create handler that serves feedback_client.html
592         (FeedbackServer.MainHandler.Handler):
593         (FeedbackServer.MainHandler.Handler.get):
594         (FeedbackServer.WSHandler): Handler factory to create handler that sends data to feedback client
595         (FeedbackServer.WSHandler.Handler):
596         (FeedbackServer.WSHandler.Handler.open): On websocket connection opened
597         (FeedbackServer.WSHandler.Handler.on_close): On websocket connection closed
598         * LaunchTime/launch_time.py: Added.
599         (DefaultLaunchTimeHandler): Abstract HTTP request handler for launch time benchmarks
600         (DefaultLaunchTimeHandler.get_test_page): Default test page to be overridden by benchmarks
601         (DefaultLaunchTimeHandler.get_blank_page):
602         (DefaultLaunchTimeHandler.on_receive_stop_time):
603         (DefaultLaunchTimeHandler.do_HEAD):
604         (DefaultLaunchTimeHandler.do_GET):
605         (DefaultLaunchTimeHandler.do_POST):
606         (DefaultLaunchTimeHandler.log_message): Suppresses HTTP logs from SimpleHTTPRequestHandler
607         (LaunchTimeBenchmark): Abstract class which launch time benchmarks inherit from and override methods desired to customize
608         (LaunchTimeBenchmark.__init__):
609         (LaunchTimeBenchmark._parse_browser_bundle_path): Parser for bundle path option
610         (LaunchTimeBenchmark._parse_args):
611         (LaunchTimeBenchmark._run_server): Target function for main server thread
612         (LaunchTimeBenchmark._setup_servers):
613         (LaunchTimeBenchmark._clean_up):
614         (LaunchTimeBenchmark._exit_due_to_exception):
615         (LaunchTimeBenchmark._geometric_mean):
616         (LaunchTimeBenchmark._standard_deviation):
617         (LaunchTimeBenchmark._compute_results): Returns mean and std dev of list of results and pretty prints if should_print=True is specified
618         (LaunchTimeBenchmark._wait_times): Mimic numpy.linspace
619         (LaunchTimeBenchmark.open_tab): Open a browser tab with the html given by self.response_handler.get_test_page
620         (LaunchTimeBenchmark.launch_browser): Open a broser to either the feedback client (if option is set) or a blank page
621         (LaunchTimeBenchmark.quit_browser):
622         (LaunchTimeBenchmark.quit_browser.quit_app):
623         (LaunchTimeBenchmark.quit_browser.is_app_closed):
624         (LaunchTimeBenchmark.close_tab):
625         (LaunchTimeBenchmark.wait):
626         (LaunchTimeBenchmark.log): Print to console and send to feedback client if --feedback-in-browser flag is used
627         (LaunchTimeBenchmark.log_verbose): Only logs if --verbose flag is used
628         (LaunchTimeBenchmark.run):
629         (LaunchTimeBenchmark.group_init): Initialization done before each round of iterations
630         (LaunchTimeBenchmark.run_iteration):
631         (LaunchTimeBenchmark.initialize): Convenience method to be overriden by subclasses which is called at the end of __init__
632         (LaunchTimeBenchmark.will_parse_arguments): Called before argparse.parse_args to let subclasses add new command line arguments
633         (LaunchTimeBenchmark.did_parse_arguments): Called after argparse.parse_args to let subclass initialize based on command line arguments
634         * LaunchTime/new_tab.py: Added
635         (NewTabBenchmark):
636         (NewTabBenchmark._parse_wait_time): Parser for wait time option
637         (NewTabBenchmark.initialize):
638         (NewTabBenchmark.run_iteration):
639         (NewTabBenchmark.group_init):
640         (NewTabBenchmark.will_parse_arguments):
641         (NewTabBenchmark.did_parse_arguments):
642         (NewTabBenchmark.ResponseHandler):
643         (NewTabBenchmark.ResponseHandler.Handler):
644         (NewTabBenchmark.ResponseHandler.Handler.get_test_page):
645         * LaunchTime/startup.py: Added
646         (StartupBenchmark): This benchmark measures browser startup time and initial page load time
647         (StartupBenchmark.initialize):
648         (StartupBenchmark.run_iteration):
649         (StartupBenchmark.ResponseHandler):
650         (StartupBenchmark.ResponseHandler.Handler):
651         (StartupBenchmark.ResponseHandler.Handler.get_test_page):
652         * LaunchTime/thirdparty/__init__.py: Added.
653         (AutoinstallImportHook): Auto installs tornado package for feedback server
654         (AutoinstallImportHook.__init__):
655         (AutoinstallImportHook._ensure_autoinstalled_dir_is_in_sys_path):
656         (AutoinstallImportHook.find_module):
657         (AutoinstallImportHook._install_tornado):
658         (AutoinstallImportHook.greater_than_equal_to_version):
659         (AutoinstallImportHook._install):
660         (AutoinstallImportHook.get_latest_pypi_url):
661         (AutoinstallImportHook.install_binary):
662         (autoinstall_everything):
663         (get_os_info):
664
665 2018-06-25  Jon Lee  <jonlee@apple.com>
666
667         [MotionMark] Add support for version numbers
668         https://bugs.webkit.org/show_bug.cgi?id=186479
669
670         Reviewed by Said Abou-Hallawa.
671
672         Add support for displaying the version number as well as including it in the JSON results.
673
674         When loading the front page, script replaces any element with classname version with the
675         version number of the benchmark, which is stored in Strings.version.
676
677         The JSON structure for the results includes a new version property:
678             {
679               "version": "1.0",
680               "options": { ... },
681               "data": [ ... ]
682             }
683
684         When dragging a results file, the version listed will come from the JSON file. Older
685         results will not have had the version property, in which case it will default to "1.0".
686
687         * MotionMark/index.html: Update title to some other default. Script will update it.
688         Include the version number in the logo title.
689         * MotionMark/developer.html: Ditto.
690         * MotionMark/about.html: Ditto.
691
692         * MotionMark/resources/runner/motionmark.js:
693         (ResultsDashboard): Update constructor to include version. This is used when serializing
694         results out to JSON, and displaying the results panel in developer mode.
695         (ResultsDashboard._processData): When running the benchmark, include benchmark version string
696         in the results object.
697         (ResultsDashboard.version):
698         (window.benchmarkRunnerClient.willStartFirstIteration): When running the benchmark, pass the
699         benchmark version string to the dashboard, which holds the results.
700         (window.sectionsManager.setSectionVersion): Helper function to update the element in the
701         section with the class name version.
702         (window.benchmarkController.initialize): Populate all DOM elements with class name "version"
703         with the version string. Update the page title.
704         (window.benchmarkController.showResults): When showing results, update the version string
705         based on what is included in the JSON results, which would be the same as the benchmark version.
706         * MotionMark/resources/runner/motionmark.css: Include missing copyright. Wrap the SVG logo
707         in a div and include the version string.
708         * MotionMark/resources/strings.js: Add strings for the page title template, and the version.
709
710         * MotionMark/resources/debug-runner/motionmark.css:
711         * MotionMark/resources/debug-runner/motionmark.js:
712         (window.benchmarkRunnerClient.willStartFirstIteration): When running the benchmark, pass the
713         benchmark version string to the dashboard, which holds the results.
714         (window.benchmarkController.initialize): Populate all DOM elements with class name "version"
715         with the version string. Update the page title. When dragging in JSON results, look for
716         version to pass to the dashboard. If it doesn't exist, default to "1.0".
717         (window.benchmarkController.showResults): When showing results, update the version string
718         based on what is included in the JSON results, instead of the current benchmark version.
719         * MotionMark/resources/debug-runner/tests.js: Update page title template.
720
721 2018-06-08  Jon Lee  <jonlee@apple.com>
722
723         [MotionMark] Rename Suits test files
724         https://bugs.webkit.org/show_bug.cgi?id=186447
725
726         Reviewed by Said Abou-Hallawa.
727
728         * MotionMark/resources/runner/tests.js:
729         * MotionMark/tests/master/resources/suits.js: Renamed from PerformanceTests/MotionMark/tests/master/resources/svg-particles.js.
730         * MotionMark/tests/master/suits.html: Renamed from PerformanceTests/MotionMark/tests/master/svg-particles.html.
731         * MotionMark/tests/svg/suits.html:
732
733 2018-06-08  Jon Lee  <jonlee@apple.com>
734
735         Add sub-tests based on Suits
736         https://bugs.webkit.org/show_bug.cgi?id=186260
737
738         Reviewed by Said Abou-Hallawa.
739
740         Add a new developer Suits suite with sub-tests that isolate parts of each particle.
741         The sub-tests are:
742             - Particles using only clip paths
743             - Particles using only shapes
744             - Particles that have no gradients
745             - Particles that have no rotation
746             - Particles that do not move around at all (but all of the physics calculations are
747             still performed)
748
749         * MotionMark/resources/debug-runner/tests.js:
750         * MotionMark/tests/master/resources/svg-particles.js: Rename the stage, particle, and
751         benchmark to "Suits" from "SVG". Change SuitsParticle so that we guarantee 50% clip and
752         shape paths rather than relying on random chance. Wrap a check around creation of the
753         gradient element, and use a simple fill color when we aren't using a gradient.
754         * MotionMark/tests/svg/suits.html: Added.
755         * MotionMark/tests/svg/suits.js: Added.
756         Look for the query string and set the particle based on what is selected.
757
758 2018-06-06  Jon Lee  <jonlee@apple.com>
759
760         Allow the ramp controller to run tests that take less time than the initial ramp-up phase
761         https://bugs.webkit.org/show_bug.cgi?id=186257
762
763         Reviewed by Said Abou-Hallawa.
764
765         * MotionMark/tests/resources/main.js: Extend the desired end timestamp when ramp-up occurs.
766         Previously the controller assumed that the duration of the test was always longer than the
767         time it took to figure out the order of magnitude of particles to render. When that range
768         is determined, the end timestamp is then extended. If the duration is too short, the controller
769         would throw an exception. The fix is to continue extending the timestamp every time the order of
770         magnitude tiers up.
771
772         In general testing with really short durations isn't practical, but sometimes it is
773         desired for debugging.
774
775         To avoid confusion of the private variables, I renamed Benchmark._startTimestamp to
776         Benchmark._benchmarkStartTimestamp, to disambiguate from Controller._startTimestamp.
777         Benchmark._startTimestamp is set once and never changed, after the initial 100ms warm-up.
778         Controller._startTimestamp is initialized to 0 (which means "not initialized"), and then after
779         the warm up is complete, set in Controller.start() to the same wall clock timestamp, and never
780         changed afterwards.
781
782 2018-06-06  Jon Lee  <jonlee@apple.com>
783
784         Remove unneeded data processing in MotionMark
785         https://bugs.webkit.org/show_bug.cgi?id=186256
786
787         Reviewed by Said Abou-Hallawa.
788
789         Remove calculation of average frame rate based on multiple samples at a given complexity.
790         It is not used in the calculation of the score, and is just extra work to do at the end
791         of a test.
792
793         * MotionMark/developer.html: Remove options to show average data and regression based on
794         that data.
795         * MotionMark/resources/debug-runner/graph.js: Remove data. Update the time graph that
796         shows the relationship between the break point in the regression against the complexity
797         graph.
798         * MotionMark/resources/debug-runner/motionmark.css: Make the confidence intervals easier
799         to see.
800         * MotionMark/resources/runner/motionmark.js: Simplify score calculation now that we are
801         only calculating regressions on the raw data. We were calculating it based on the averages,
802         but that was just for analysis, and not used in the actual score. Get rid of the
803         |complexitySamples| local variable.
804         * MotionMark/resources/strings.js: Remove unneeded string constant.
805         * MotionMark/tests/resources/main.js:
806         (processSamples): Refactor so that more of it can be shared between Controller and
807         RampController. It also avoids processing the raw samples (controllerSamples) twice for
808         the ramp controller.
809         (_processComplexitySamples): Remove logic to calculate averages based on complexity.
810         (_processMarks): Split this out from the old processSamples.
811         (_processControllerSamples): Split this out from the old processSamples.
812
813 2018-06-06  Jon Lee  <jonlee@apple.com>
814
815         Strip out more characters when creating permalinks
816         https://bugs.webkit.org/show_bug.cgi?id=186259
817
818         Reviewed by Said Abou-Hallawa.
819
820         Tests and suites may have names that make permalinks to the tests less readable when
821         shared. Expand the set of characters that are stripped out of the names.
822
823         * MotionMark/resources/debug-runner/motionmark.js: Refactor to call
824         stripUnwantedCharactersForURL. For existing links, decode the suite and test names.
825         * MotionMark/resources/extensions.js:
826         (Utilities.stripUnwantedCharactersForURL): Rename from stripNonASCIICharacters.
827         Remove any non-alphanumeric character.
828
829 2018-06-04  Jon Lee  <jonlee@apple.com>
830
831         Remove unnecessary MotionMark controllers
832         https://bugs.webkit.org/show_bug.cgi?id=186255
833
834         Reviewed by Said Abou-Hallawa.
835
836         Get rid of the ramp30 and the fixed-with-a-step controllers, which aren't used for testing
837         or debugging.
838
839         * MotionMark/developer.html: Move the "ramp" option as first, since that is the default
840         controller for the benchmark.
841         * MotionMark/resources/debug-runner/motionmark.js:
842         * MotionMark/resources/runner/motionmark.js: Added missing copyright notice.
843         * MotionMark/tests/resources/main.js:
844
845 2018-05-25  Myles C. Maxfield  <mmaxfield@apple.com>
846
847         Improve the performance of Font::canRenderCombiningCharacterSequence()
848         https://bugs.webkit.org/show_bug.cgi?id=185933
849
850         Reviewed by Ryosuke Niwa.
851
852         * Layout/ComplexLongUnique.html: Added.
853
854 2018-05-25  Saam Barati  <sbarati@apple.com>
855
856         Have a memory test where we can validate JSCs mini memory mode
857         https://bugs.webkit.org/show_bug.cgi?id=185932
858
859         Reviewed by Mark Lam.
860
861         We add a directory here with the contents of the testmem benchmark.
862         To run it, use `Tools/Scripts/run-testmem`. To add new tests in the future,
863         you just need to add JS files to this directory.
864
865         * testmem: Added.
866         * testmem/air.js: Added.
867         * testmem/base64.js: Added.
868         * testmem/basic.js: Added.
869         * testmem/box2d.js: Added.
870         * testmem/crypto-md5.js: Added.
871         * testmem/date-format-tofte.js: Added.
872         * testmem/earley-boyer.js: Added.
873         * testmem/hash-map.js: Added.
874         * testmem/regex-dna.js: Added.
875         * testmem/splay.js: Added.
876         * testmem/tagcloud.js: Added.
877
878 2018-05-22  Ryan Haddad  <ryanhaddad@apple.com>
879
880         Unreviewed, rolling out r232052.
881
882         Breaks internal builds.
883
884         Reverted changeset:
885
886         "Use more C++17"
887         https://bugs.webkit.org/show_bug.cgi?id=185176
888         https://trac.webkit.org/changeset/232052
889
890 2018-05-21  Yusuke Suzuki  <utatane.tea@gmail.com>
891
892         Use more C++17
893         https://bugs.webkit.org/show_bug.cgi?id=185176
894
895         Reviewed by JF Bastien.
896
897         * MallocBench/MallocBench.xcodeproj/project.pbxproj:
898         * MediaTime/Configurations/Base.xcconfig:
899         * MediaTime/MediaTime.xcodeproj/project.pbxproj:
900
901 2018-04-30  Myles C. Maxfield  <mmaxfield@apple.com>
902
903         Improve the performance of FontCascadeDescription's effectiveFamilies
904         https://bugs.webkit.org/show_bug.cgi?id=184720
905
906         Reviewed by Simon Fraser.
907
908         This performance test calls SystemFontDatabase::systemFontCascadeList() around 2,000,000 times (before
909         this patch is applied), which is roughly equivalent to the page we found the performance problem on.
910         The calling pattern is roughly equivalent in this test.
911
912         * Layout/system-ui.html: Added.
913
914 2018-03-08  Antti Koivisto  <antti@apple.com>
915
916         Update StyleBench version number in page title to 0.3
917
918         * StyleBench/index.html:
919
920         There has been a bunch of changes, most notably the addition of attribute selector test step.
921
922 2018-02-16  Dean Jackson  <dino@apple.com>
923
924         Use OPENGL macros to be more clear about which OpenGL/ES WebGL uses on Cocoa
925         https://bugs.webkit.org/show_bug.cgi?id=182894
926
927         Reviewed by Tim Horton.
928
929         Rename OPENGL_ES_2 to OPENGL_ES.
930
931         * StitchMarker/wtf/Platform.h:
932
933 2018-02-09  Ryosuke Niwa  <rniwa@webkit.org>
934
935         Make run-perf-tests work with Speedometer 2.0 and re-enable the test
936         https://bugs.webkit.org/show_bug.cgi?id=182088
937
938         Reviewed by Antti Koivisto.
939
940         This patch updates the test harness to report new metrics of Speedometer 2.0.
941
942         Replaced the total time by the geometric mean of the time spent in each suite, and added scores
943         as a metric in addition to time. Also, updated the PerfTestRunner harness and the results template
944         to support "Score" metric with "pt" as its unit.
945
946         * Skipped: Unskipped the test.
947         * Speedometer/resources/benchmark-report.js:
948         (window.benchmarkClient.willStartFirstIteration..createTest): Added the unit as an optional argument
949         to report the score in addition to time. Also report the geomean of time instead of the total time.
950         (window.benchmarkClient.willStartFirstIteration): Report the geomean of time.
951         (window.benchmarkClient.didRunSuites): Record the entire measuredValues instead of just tests.
952         (window.benchmarkClient.didFinishLastIteration): Report the scores.
953
954 2018-02-09  Ross Kirsling  <ross.kirsling@sony.com>
955
956         Use REQUEST_ANIMATION_FRAME_DISPLAY_MONITOR on any non-Windows port.
957         https://bugs.webkit.org/show_bug.cgi?id=182623
958
959         Reviewed by Alex Christensen.
960
961         * StitchMarker/wtf/Platform.h:
962         Simplify #if.
963
964 2018-02-01  Antti Koivisto  <antti@apple.com>
965
966         StyleBench: Attribute selectors and other improvements
967         https://bugs.webkit.org/show_bug.cgi?id=182387
968
969         Reviewed by Joseph Pecoraro.
970
971         - Add some attributes to elements in all tests
972         - Add some attribute selectors to stylesheets in all tests
973         - Also add some * selectors to all stylesheets.
974         - Add attribute mutation step to all suites
975         - Make test steps do more mutations (25->100) and reduce the number of steps to keep testing time in check.
976           Too fast steps were running into timer resolution limits.
977
978         * StyleBench/resources/style-bench.js:
979         (defaultConfiguration):
980         (prototype.randomAttributeName):
981         (prototype.randomAttributeValue):
982         (prototype.randomAttributeSelector):
983         (prototype.makeCompoundSelector):
984         (prototype.makeElement):
985         (prototype.addClasses):
986         (prototype.removeClasses):
987         (prototype.mutateAttributes):
988         (prototype.async.runForever):
989         * StyleBench/resources/tests.js:
990         (makeSteps):
991         (makeSuite):
992
993 2018-02-01  Geoffrey Garen  <ggaren@apple.com>
994
995         Make MallocBench easier for non-WebKit engineers to run
996         https://bugs.webkit.org/show_bug.cgi?id=182415
997
998         Reviewed by Saam Barati.
999
1000         * MallocBench/MallocBench.xcodeproj/project.pbxproj: Use c++14 so we
1001         can make_unique.
1002
1003         Specify that we support all Darwin platforms so you can test them.
1004
1005         * MallocBench/run-malloc-benchmarks: Specify the path to MallocBench
1006         and libmbmalloc explicitly, rather than computing them implicitly
1007         using webkitdirs. Non-WebKit folks don't have the directory structure
1008         required by webkitdirs.
1009
1010         Remove Linux-specific and cmake-specific behaviors because we only
1011         needed them in the world of implicit path computation.
1012
1013 2018-01-31  Ryosuke Niwa  <rniwa@webkit.org>
1014
1015         Make run-perf-tests work with StyleBench and re-enable it
1016         https://bugs.webkit.org/show_bug.cgi?id=182322
1017
1018         Reviewed by Antti Koivisto.
1019
1020         This patch updates the test harness to report new metrics of StyleBench.
1021
1022         Replaced the total time by the geometric mean of the time spent in each suite, and added scores
1023         as a metric in addition to time. Also, updated the PerfTestRunner harness and the results template
1024         to support "Score" metric with "pt" as its unit.
1025
1026         * Skipped: Unskipped the test.
1027         * StyleBench/resources/benchmark-report.js:
1028         (window.benchmarkClient.willStartFirstIteration..createTest): Added the unit as an optional argument
1029         to report the score in addition to time. Also report the geomean of time instead of the total time.
1030         (window.benchmarkClient.willStartFirstIteration): Report the geomean of time.
1031         (window.benchmarkClient.didRunSuites): Record the entire measuredValues instead of just tests.
1032         (window.benchmarkClient.didFinishLastIteration): Report the scores.
1033         * resources/results-template.html: Added "Score" as a metric and "pt" as its unit.
1034         * resources/runner.js:
1035         (PerfTestRunner..finish): Ditto.
1036
1037 2018-01-30  Ryosuke Niwa  <rniwa@webkit.org>
1038
1039         StyleBench: Generate more mutation steps
1040         https://bugs.webkit.org/show_bug.cgi?id=182291
1041
1042         Rubber-stamped by Antti Koivisto
1043
1044         Use zero-based index instead of one-based index.
1045
1046         * StyleBench/resources/tests.js:
1047         (makeSteps):
1048
1049 2018-01-30  Antti Koivisto  <antti@apple.com>
1050
1051         StyleBench: Generate more mutation steps
1052         https://bugs.webkit.org/show_bug.cgi?id=182291
1053
1054         Reviewed by Ryosuke Niwa.
1055
1056         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).
1057         This also makes test results more stable and less dependent on possible first-time costs.
1058
1059         * StyleBench/resources/benchmark-runner.js:
1060
1061         Adjust correctionFactor to match the new steps count. With 1.5 Firefox is ~72rpm.
1062
1063         * StyleBench/resources/tests.js:
1064         (makeSteps):
1065
1066         Generate 10 rounds of class/element mutation steps.
1067         Give each generated step unique name so reporting works.
1068         Make individual steps do fewer mutations.
1069
1070 2018-01-29  Antti Koivisto  <antti@apple.com>
1071
1072         StyleBench: Remove : from a test name
1073         https://bugs.webkit.org/show_bug.cgi?id=182258
1074
1075         Reviewed by Ryosuke Niwa.
1076
1077         It breaks reporting.
1078
1079         * StyleBench/resources/style-bench.js:
1080         (nthPseudoClassConfiguration):
1081
1082 2018-01-29  Ryosuke Niwa  <rniwa@webkit.org>
1083
1084         Temporarily skip StyleBench until webkit.org/b/182088 is fixed.
1085
1086         * Skipped:
1087
1088 2018-01-29  Ryosuke Niwa  <rniwa@webkit.org>
1089
1090         Make StyleBench compatible with run-benchmark and run-perf-tests
1091         https://bugs.webkit.org/show_bug.cgi?id=182262
1092
1093         Reviewed by Antti Koivisto.
1094
1095         Copied resource files referenced from Speedometer directory since run-benchmark needs to be able
1096         to checkout each benchmark separately.
1097
1098         Removed the code to create tests of the same name five times in makeSteps since this
1099         won't be compatible with either run-benchmark or run-perf-tests.
1100
1101         * StyleBench/index.html: Removed the code to show warnings for local files since run-benchmark
1102         doesn't use HTTP server in WebDriver mode.
1103         * StyleBench/resources/benchmark-report.js: Copied from resources/benchmark-report.js.
1104         * StyleBench/resources/benchmark-runner.js: Copied from resources/benchmark-runner.js.
1105         (BenchmarkRunner.prototype._finalize): Use the correction factor of 8 instead of 5 in StyleBench.
1106         * StyleBench/resources/gauge.png: Copied from resources/gauge.png.
1107         * StyleBench/resources/gauge@2x.png: Copied from resources/gauge@2x.png.
1108         * StyleBench/resources/main.css: Copied from resources/main.css.
1109         * StyleBench/resources/main.js: Copied from resources/main.js.
1110         * StyleBench/resources/tests.js:
1111         (makeSteps): Only make each test once.
1112
1113 2018-01-26  Chris Nardi  <cnardi@chromium.org>
1114
1115         Addressing post-review comments after r226614
1116         https://bugs.webkit.org/show_bug.cgi?id=182151
1117
1118         Reviewed by Myles C. Maxfield.
1119
1120         * StitchMarker/wtf/text/StringImpl.h:
1121         (WTF::isSpaceOrNewline):
1122         * StitchMarker/wtf/text/TextBreakIterator.cpp:
1123         (WTF::numCodeUnitsInGraphemeClusters):
1124         * StitchMarker/wtf/text/TextBreakIterator.h:
1125
1126 2018-01-22  Antti Koivisto  <antti@apple.com>
1127
1128         StyleBench: Separate test for :nth pseudo classes
1129         https://bugs.webkit.org/show_bug.cgi?id=181932
1130
1131         Reviewed by Sam Weinig.
1132
1133         Optimization strategies for nth are different from other structural pseudo classes (:first-child etc). Move them to a separate test.
1134
1135         * StyleBench/resources/style-bench.js:
1136         (structuralPseudoClassConfiguration):
1137         (nthPseudoClassConfiguration):
1138
1139             Add a new configuration.
1140
1141         (predefinedConfigurations):
1142         (pseudoClassConfiguration): Deleted.
1143
1144 2018-01-17  Ryosuke Niwa  <rniwa@webkit.org>
1145
1146         Temporarily skip Speedometer until we fix the metrics.
1147
1148         * Skipped:
1149
1150 2018-01-16  Alex Christensen  <achristensen@webkit.org>
1151
1152         Fix Mac CMake build after r222900
1153
1154         * MallocBench/MallocBench/CMakeLists.txt:
1155         The mbmalloc target doesn't exist if we're not in developer mode, so only use it if we are.
1156
1157 2018-01-09  Ryosuke Niwa  <rniwa@webkit.org>
1158
1159         Speedometer 2.0: Update the About page.
1160         https://bugs.webkit.org/show_bug.cgi?id=175918
1161
1162         Reviewed by Keith Miller.
1163
1164         Updated the about page for Speedometer 2.0.
1165
1166         * Speedometer/index.html:
1167
1168 == Rolled over to ChangeLog-2018-01-01 ==