37830 2010-04-19 16:04:35 -0700 send-oncancel-event touch test should use js-test-post-function.js 2010-09-02 17:01:59 -0700 1 1 1 Unclassified WebKit Tools / Tests 528+ (Nightly build) Android Android RESOLVED FIXED P2 Normal --- 38084 1 benm benm android-webkit-unforking commit-queue gdk hausmann jamesr kim.1.gronholm yuzo oldest_to_newest 214303 0 benm 2010-04-19 16:04:35 -0700 When the touch layout tests were moved in http://trac.webkit.org/changeset/54253 send-oncancel-event was changed to not use js-test-post-function.js erroneously. This meant that the test threw an error as it used the isSuccessfullyParsed function. This failure was avoided in http://trac.webkit.org/changeset/55735 which removed the reference to isSuccessfullyParsed. Due to the asynchronous nature of touch events on Android, we need to use js-test-post-function to run an asynchronous test. Patch to follow to reinsert the call to isSuccesfullyParsed and make the test use js-test-post-function. 214333 1 53737 benm 2010-04-19 16:42:17 -0700 Created attachment 53737 Patch. 215596 2 53737 jorlow 2010-04-21 18:08:13 -0700 Comment on attachment 53737 Patch. r=me 215708 3 53737 commit-queue 2010-04-21 23:02:14 -0700 Comment on attachment 53737 Patch. Clearing flags on attachment: 53737 Committed r58058: <http://trac.webkit.org/changeset/58058> 215709 4 commit-queue 2010-04-21 23:02:19 -0700 All reviewed patches have been landed. Closing bug. 216149 5 54107 jamesr 2010-04-22 16:34:18 -0700 Created attachment 54107 Patch 216150 6 jamesr 2010-04-22 16:36:13 -0700 Hey guys. The first patch broke Chromium because in Chromium the touch stuff is not deferred and so the entire test body was running before hitting the end of send-oncancel-events.js. This means that js-test-post-function.js was not getting parsed, so isSuccessfullyParsed was not being defined, and the 'var successfullyParsed = true' line was not able to run. This patch moves the test body to the onload event which should work fine on all platforms. I don't have any way to verify this on Android. Thanks 216157 7 benm 2010-04-22 16:51:51 -0700 Ah, I see. Sorry about that. Alternatively, we could set isSuccessfullyParsed right before calling layoutTestController.notifyDone, like in: http://trac.webkit.org/browser/trunk/LayoutTests/fast/events/touch/script-tests/basic-single-touch-events.js#L23 ? Thanks, Ben 216164 8 benm 2010-04-22 16:59:58 -0700 Sorry, for that to work we'd also need to move the js-test-post-function link inside the send-oncancel-event.html file to above the link to the test sctipt, like it's done in basic-single-touch-events.html. What do you think? It would be nice to maintain consistency. 216166 9 jamesr 2010-04-22 17:02:58 -0700 I really dislike the pattern in basic-single-touch-events.js to be honest. It's too easy to add code below the touchTargets() call that runs on some platforms (that do touch events synchronously) and does not run on others (which do touch events asynchronously). To be consistent we should make sure that all platforms have a chance to evaluate the same javascript and then run the test. The simplest way to do that is to just put the actual test body in the onload handler or later. 216180 10 benm 2010-04-22 17:31:00 -0700 I can confirm that the window.onload approach works on Android. Looking through the current touch tests and the TEMPLATE.html for those tests, there seems to be some tidying up needed. I think that future tests should make use of TEMPLATE.html (which incidentally also needs fixing as it was a copy from another directory where asynchronous tests were not needed) rather than relying on remembering to run the test code inside window.onload. We can make TEMPLATE.html link to the scripts in the correct order. I don't mind doing this if you'd like to submit this now as-is. Cheers, Ben 216208 11 yuzo 2010-04-22 18:34:11 -0700 *** Bug 38020 has been marked as a duplicate of this bug. *** 216209 12 yuzo 2010-04-22 18:35:51 -0700 Changed LayoutTests/platform/chromium/test_expectations.txt to mark the test failing. Please change it again once the patch is landed. 220626 13 benm 2010-05-04 09:06:04 -0700 With the resolution of https://bugs.webkit.org/show_bug.cgi?id=38084 I think this should now be fixed for Chromium. James, would you mind syncing past r58644 to verify? Thanks, Ben 237403 14 54107 darin 2010-06-12 18:52:29 -0700 Comment on attachment 54107 Patch Looks OK. r=me 273675 15 54107 commit-queue 2010-09-02 17:01:52 -0700 Comment on attachment 54107 Patch Clearing flags on attachment: 54107 Committed r66695: <http://trac.webkit.org/changeset/66695> 273676 16 commit-queue 2010-09-02 17:01:59 -0700 All reviewed patches have been landed. Closing bug. 53737 2010-04-19 16:42:17 -0700 2010-04-22 16:34:14 -0700 Patch. 37830-patch.txt text/plain 2058 benm SW5kZXg6IExheW91dFRlc3RzL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBMYXlvdXRUZXN0cy9D aGFuZ2VMb2cJKHJldmlzaW9uIDU3ODUxKQorKysgTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCSh3b3Jr aW5nIGNvcHkpCkBAIC0xLDMgKzEsMTggQEAKKzIwMTAtMDQtMTkgIEJlbiBNdXJkb2NoICA8YmVu bUBnb29nbGUuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisg ICAgICAgIHNlbmQtb25jYW5jZWwtZXZlbnQgdG91Y2ggdGVzdCBzaG91bGQgdXNlIGpzLXRlc3Qt cG9zdC1mdW5jdGlvbi5qcworICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1 Zy5jZ2k/aWQ9Mzc4MzAKKworICAgICAgICBSZXN0b3JlIHJlZmVyZW5jZSB0byBqcy10ZXN0LXBv c3QtZnVuY3Rpb24gcmVtb3ZlZCBlcnJvbmVvdXNseSBpbiBjb3B5IAorICAgICAgICB3aGVuIHRo aXMgdGVzdCB3YXMgbW92ZSBpbiBodHRwOi8vdHJhYy53ZWJraXQub3JnL2NoYW5nZXNldC81NDI1 MworICAgICAgICBhbmQgdGhlIGNvcnJlc3BvbmRpbmcgY2FsbCB0byB0aGUgaXNTdWNjZXNzZnVs bHlQYXJzZWQgZnVuY3Rpb24uCisKKyAgICAgICAgKiBmYXN0L2V2ZW50cy90b3VjaC9zY3JpcHQt dGVzdHMvc2VuZC1vbmNhbmNlbC1ldmVudC5qczoKKyAgICAgICAgKHRvdWNoY2FuY2VsSGFuZGxl cik6IENhbGwgaXNTdWNjZXNzZnVsbHlQYXJzZWQuCisgICAgICAgICogZmFzdC9ldmVudHMvdG91 Y2gvc2VuZC1vbmNhbmNlbC1ldmVudC5odG1sOiBBZGQganMtdGVzdC1wb3N0LWZ1bmN0aW9uLmpz IGJhY2suCisKIDIwMTAtMDQtMTkgIENocmlzIEZsZWl6YWNoICA8Y2ZsZWl6YWNoQGFwcGxlLmNv bT4KIAogICAgICAgICBSZXZpZXdlZCBieSBCZXRoIERha2luLgpJbmRleDogTGF5b3V0VGVzdHMv ZmFzdC9ldmVudHMvdG91Y2gvc2VuZC1vbmNhbmNlbC1ldmVudC5odG1sCj09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0t IExheW91dFRlc3RzL2Zhc3QvZXZlbnRzL3RvdWNoL3NlbmQtb25jYW5jZWwtZXZlbnQuaHRtbAko cmV2aXNpb24gNTc4MTQpCisrKyBMYXlvdXRUZXN0cy9mYXN0L2V2ZW50cy90b3VjaC9zZW5kLW9u Y2FuY2VsLWV2ZW50Lmh0bWwJKHdvcmtpbmcgY29weSkKQEAgLTgsNiArOCw2IEBACiA8cCBpZD0i ZGVzY3JpcHRpb24iPjwvcD4KIDxkaXYgaWQ9ImNvbnNvbGUiPjwvZGl2PgogPHNjcmlwdCBzcmM9 InNjcmlwdC10ZXN0cy9zZW5kLW9uY2FuY2VsLWV2ZW50LmpzIj48L3NjcmlwdD4KLTxzY3JpcHQg c3JjPSIuLi8uLi9qcy9yZXNvdXJjZXMvanMtdGVzdC1wb3N0LmpzIj48L3NjcmlwdD4KKzxzY3Jp cHQgc3JjPSIuLi8uLi9qcy9yZXNvdXJjZXMvanMtdGVzdC1wb3N0LWZ1bmN0aW9uLmpzIj48L3Nj cmlwdD4KIDwvYm9keT4KIDwvaHRtbD4KSW5kZXg6IExheW91dFRlc3RzL2Zhc3QvZXZlbnRzL3Rv dWNoL3NjcmlwdC10ZXN0cy9zZW5kLW9uY2FuY2VsLWV2ZW50LmpzCj09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIExh eW91dFRlc3RzL2Zhc3QvZXZlbnRzL3RvdWNoL3NjcmlwdC10ZXN0cy9zZW5kLW9uY2FuY2VsLWV2 ZW50LmpzCShyZXZpc2lvbiA1NzgxNCkKKysrIExheW91dFRlc3RzL2Zhc3QvZXZlbnRzL3RvdWNo L3NjcmlwdC10ZXN0cy9zZW5kLW9uY2FuY2VsLWV2ZW50LmpzCSh3b3JraW5nIGNvcHkpCkBAIC0x Myw2ICsxMyw3IEBAIGZ1bmN0aW9uIHRvdWNoY2FuY2VsSGFuZGxlcigpIHsKICAgICBzaG91bGRC ZSgiY2FuY2VsRXZlbnQucGFnZVkiLCB0b3VjaFkudG9TdHJpbmcoKSk7CiAgICAgaWYgKHdpbmRv dy5sYXlvdXRUZXN0Q29udHJvbGxlcikgewogICAgICAgICBsYXlvdXRUZXN0Q29udHJvbGxlci5u b3RpZnlEb25lKCk7CisgICAgICAgIGlzU3VjY2Vzc2Z1bGx5UGFyc2VkKCk7IAogICAgIH0KIH0K ICAgICAK 54107 2010-04-22 16:34:18 -0700 2010-09-02 17:01:52 -0700 Patch bug-37830-20100422163417.patch text/plain 1840 jamesr ZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL0NoYW5nZUxvZyBiL0xheW91dFRlc3RzL0NoYW5nZUxv ZwppbmRleCA5ZDM1Nzc0Li44OGI1MmM3IDEwMDY0NAotLS0gYS9MYXlvdXRUZXN0cy9DaGFuZ2VM b2cKKysrIGIvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTUgQEAKKzIwMTAtMDQt MjIgIEphbWVzIFJvYmluc29uICA8amFtZXNyQGNocm9taXVtLm9yZz4KKworICAgICAgICBSZXZp ZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcv c2hvd19idWcuY2dpP2lkPTM3ODMwCisKKyAgICAgICAgRml4IGFuIG9yZGVyaW5nIGlzc3VlIGlu IHRoaXMgdGVzdC4gIFRoZSBpc1N1Y2Nlc3NmdWxseVBhcnNlZAorICAgICAgICBmdW5jdGlvbiBp cyBub3QgZGVmaW5lZCB1bnRpbCB0aGUgb25sb2FkIGV2ZW50LCBzbyBhdm9pZCBydW5uaW5nCisg ICAgICAgIHRoZSBzY3JpcHQgYm9keSB1bnRpbCB0aGlzIHBvaW50LgorCisgICAgICAgICogZmFz dC9ldmVudHMvdG91Y2gvc2NyaXB0LXRlc3RzL3NlbmQtb25jYW5jZWwtZXZlbnQuanM6CisgICAg ICAgICh3aW5kb3cub25sb2FkKToKKwogMjAxMC0wNC0yMiAgTU9SSVRBIEhhamltZSA8bW9ycml0 YUBnb29nbGUuY29tPgogCiAgICAgICAgIE5vdCByZXZpZXdlZC4KZGlmZiAtLWdpdCBhL0xheW91 dFRlc3RzL2Zhc3QvZXZlbnRzL3RvdWNoL3NjcmlwdC10ZXN0cy9zZW5kLW9uY2FuY2VsLWV2ZW50 LmpzIGIvTGF5b3V0VGVzdHMvZmFzdC9ldmVudHMvdG91Y2gvc2NyaXB0LXRlc3RzL3NlbmQtb25j YW5jZWwtZXZlbnQuanMKaW5kZXggYzgzZGFhNC4uYWY1Mjc1ZCAxMDA2NDQKLS0tIGEvTGF5b3V0 VGVzdHMvZmFzdC9ldmVudHMvdG91Y2gvc2NyaXB0LXRlc3RzL3NlbmQtb25jYW5jZWwtZXZlbnQu anMKKysrIGIvTGF5b3V0VGVzdHMvZmFzdC9ldmVudHMvdG91Y2gvc2NyaXB0LXRlc3RzL3NlbmQt b25jYW5jZWwtZXZlbnQuanMKQEAgLTIwLDEzICsyMCwxNSBAQCBmdW5jdGlvbiB0b3VjaGNhbmNl bEhhbmRsZXIoKSB7CiBpZiAod2luZG93LmxheW91dFRlc3RDb250cm9sbGVyKQogICAgIHdpbmRv dy5sYXlvdXRUZXN0Q29udHJvbGxlci53YWl0VW50aWxEb25lKCk7CiAKLWlmICh3aW5kb3cuZXZl bnRTZW5kZXIpIHsKLSAgICBkb2N1bWVudC5hZGRFdmVudExpc3RlbmVyKCJ0b3VjaGNhbmNlbCIs IHRvdWNoY2FuY2VsSGFuZGxlciwgZmFsc2UpOwotICAgIGV2ZW50U2VuZGVyLmFkZFRvdWNoUG9p bnQodG91Y2hYLCB0b3VjaFkpOwotICAgIGV2ZW50U2VuZGVyLnRvdWNoU3RhcnQoKTsKLSAgICBl dmVudFNlbmRlci5jYW5jZWxUb3VjaFBvaW50KDApOwotICAgIGV2ZW50U2VuZGVyLnRvdWNoQ2Fu Y2VsKCk7Ci19IGVsc2UKLSAgICBkZWJ1ZygiVGhpcyB0ZXN0IHJlcXVpcmVzIER1bXBSZW5kZXJU cmVlLiIpOword2luZG93Lm9ubG9hZCA9IGZ1bmN0aW9uKCkgeworICAgIGlmICh3aW5kb3cuZXZl bnRTZW5kZXIpIHsKKyAgICAgICAgZG9jdW1lbnQuYWRkRXZlbnRMaXN0ZW5lcigidG91Y2hjYW5j ZWwiLCB0b3VjaGNhbmNlbEhhbmRsZXIsIGZhbHNlKTsKKyAgICAgICAgZXZlbnRTZW5kZXIuYWRk VG91Y2hQb2ludCh0b3VjaFgsIHRvdWNoWSk7CisgICAgICAgIGV2ZW50U2VuZGVyLnRvdWNoU3Rh cnQoKTsKKyAgICAgICAgZXZlbnRTZW5kZXIuY2FuY2VsVG91Y2hQb2ludCgwKTsKKyAgICAgICAg ZXZlbnRTZW5kZXIudG91Y2hDYW5jZWwoKTsKKyAgICB9IGVsc2UKKyAgICAgICAgZGVidWcoIlRo aXMgdGVzdCByZXF1aXJlcyBEdW1wUmVuZGVyVHJlZS4iKTsKK30KIAogdmFyIHN1Y2Nlc3NmdWxs eVBhcnNlZCA9IHRydWU7Cg==