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