47695 2010-10-14 15:49:54 -0700 [Qt] WTR is sloooow 2010-10-15 01:27:13 -0700 1 1 1 Unclassified WebKit Tools / Tests 528+ (Nightly build) PC All RESOLVED FIXED P2 Normal --- 1 kbalazs webkit-unassigned ossy tonikitoo oldest_to_newest 294422 0 kbalazs 2010-10-14 15:49:54 -0700 Currently WTR is way slower than DRT. One of the reasons is that the timer interval in the RunUntilConditionLoop (that lives in TestControllerQt.cpp) is too big so we do not detect that the test had been finished at the web process side as soon as possible. With a 1 millisecond interval it runs passing tests almost a magnitude faster (I have tried it with LayoutTests/http and the result was 22.19 seconds against 146.34). 294425 1 70791 kbalazs 2010-10-14 15:55:30 -0700 Created attachment 70791 proposed patch 294463 2 tonikitoo 2010-10-14 16:49:09 -0700 Do other ports have the same timer mechanism? I assume there is no drawback with this patch ... 294597 3 70791 ossy 2010-10-14 23:32:12 -0700 Comment on attachment 70791 proposed patch Clearing flags on attachment: 70791 Committed r69839: <http://trac.webkit.org/changeset/69839> 294598 4 ossy 2010-10-14 23:32:21 -0700 All reviewed patches have been landed. Closing bug. 294609 5 ossy 2010-10-15 00:02:03 -0700 before the patch: 3684.45s total testing time after patch landed: 1415.56s total testing time Thanks for the speedup. ;) 294634 6 kbalazs 2010-10-15 01:27:13 -0700 (In reply to comment #2) > Do other ports have the same timer mechanism? I assume there is no drawback with this patch ... They have more direct API for that task. See the implementation of TestController::runUntil in TestControllerWin.cpp and TestControllerMac.cpp. If you have an idea about how can we do it more effectively, feel free to rewrite my stuff ;) 70791 2010-10-14 15:55:30 -0700 2010-10-14 23:32:12 -0700 proposed patch wtr_slow.diff text/plain 1209 kbalazs ZGlmZiAtLWdpdCBhL1dlYktpdFRvb2xzL0NoYW5nZUxvZyBiL1dlYktpdFRvb2xzL0NoYW5nZUxv ZwppbmRleCBiZjRmZjZjLi5mMWZjZjM4IDEwMDY0NAotLS0gYS9XZWJLaXRUb29scy9DaGFuZ2VM b2cKKysrIGIvV2ViS2l0VG9vbHMvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTUgQEAKKzIwMTAtMTAt MTQgIEJhbGF6cyBLZWxlbWVuICA8a2JhbGF6c0B3ZWJraXQub3JnPgorCisgICAgICAgIFJldmll d2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIFtRdF0gV1RSIGlzIHNsb29vb3cKKyAg ICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTQ3Njk1CisKKyAg ICAgICAgKiBXZWJLaXRUZXN0UnVubmVyL3F0L1Rlc3RDb250cm9sbGVyUXQuY3BwOgorICAgICAg ICBDaGFuZ2UgdGhlIHRpbWVyIGludGVydmFsIG9mIFJ1blVudGlsQ29uZGl0aW9uTG9vcCBmcm9t CisgICAgICAgIDUwIG1pbGxpc2Vjb25kcyB0byAxIHRvIGF2b2lkIHdhc3RpbmcgdGltZSBhZnRl ciB0aGUgCisgICAgICAgIHRlc3QgaGFkIGJlZW4gZmluaXNoZWQuCisKIDIwMTAtMTAtMTMgIExl YW5kcm8gUGVyZWlyYSAgPGxlYW5kcm9AcHJvZnVzaW9uLm1vYmk+CiAKICAgICAgICAgUmV2aWV3 ZWQgYnkgQ3NhYmEgT3N6dHJvZ29uw6FjLgpkaWZmIC0tZ2l0IGEvV2ViS2l0VG9vbHMvV2ViS2l0 VGVzdFJ1bm5lci9xdC9UZXN0Q29udHJvbGxlclF0LmNwcCBiL1dlYktpdFRvb2xzL1dlYktpdFRl c3RSdW5uZXIvcXQvVGVzdENvbnRyb2xsZXJRdC5jcHAKaW5kZXggNGFjY2RlNi4uZGU5OTM0ZiAx MDA2NDQKLS0tIGEvV2ViS2l0VG9vbHMvV2ViS2l0VGVzdFJ1bm5lci9xdC9UZXN0Q29udHJvbGxl clF0LmNwcAorKysgYi9XZWJLaXRUb29scy9XZWJLaXRUZXN0UnVubmVyL3F0L1Rlc3RDb250cm9s bGVyUXQuY3BwCkBAIC0zOCw3ICszOCw5IEBACiAKIG5hbWVzcGFjZSBXVFIgewogCi1zdGF0aWMg Y29uc3QgdW5zaWduZWQga1RpbWVySW50ZXJ2YWxNUyA9IDUwOworLy8gV2l0aCBhIGJpZ2dlciBp bnRlcnZhbCB3ZSB3b3VsZCB3YXN0ZSB0byBtdWNoIHRpbWUKKy8vIGFmdGVyIHRoZSB0ZXN0IGhh ZCBiZWVuIGZpbmlzaGVkLgorc3RhdGljIGNvbnN0IHVuc2lnbmVkIGtUaW1lckludGVydmFsTVMg PSAxOwogCiBjbGFzcyBSdW5VbnRpbENvbmRpdGlvbkxvb3AgOiBwdWJsaWMgUU9iamVjdCB7CiAg ICAgUV9PQkpFQ1QK