93636 2012-08-09 11:46:49 -0700 (r125185) http/tests/images/jpg-img-partial-load.html timeout on Lion 2012-11-07 00:40:10 -0800 1 1 1 Unclassified WebKit Images 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED LayoutTestFailure P2 Normal --- 93171 1 dino hclam dino dongseong.hwang eric.carlson noel.gordon simon.fraser skyul webkit.review.bot oldest_to_newest 691648 0 dino 2012-08-09 11:46:49 -0700 The test http/tests/images/jpg-img-partial-load.html added in r125185 is failing on lion. 691659 1 dino 2012-08-09 11:53:53 -0700 Expectation timeout on mac: http://trac.webkit.org/changeset/125191 756071 2 noel.gordon 2012-11-01 04:15:15 -0700 I've replaced this test with fast/images/jpeg-partial-load.html http://test-results.appspot.com/dashboards/flakiness_dashboard.html#group=%40ToT%20-%20webkit.org&showAllRuns=true&tests=http%2Ftests%2Fimages%2Fjpg-img-partial-load.html%20%20http%2Ftests%2Fimages%2Fjpeg-partial-load.html%20%20 756077 3 171810 noel.gordon 2012-11-01 04:18:51 -0700 Created attachment 171810 Patch 756482 4 hclam 2012-11-01 11:31:47 -0700 What was the reason that jpg-img-partial-load.html fails on Lion? Was that because of a difference jpeg image and stall offset? 756922 5 noel.gordon 2012-11-01 23:51:48 -0700 I don't know why your test timed-out on LIon. I do know that your test also fails with a timed-out on Snow Leopard and Mountain Lion, so it's not just Lion. > Was that because of a difference jpeg image and stall offset? Nope. The test times-out regardless of image and regardless of stall offset. RenderImage::paintIntoRect() is never called. 757384 6 hclam 2012-11-02 11:30:39 -0700 This test and the one I wrote are using the same technique but the stopping conditions are different. Your test is using a timeout of 500m while mine is using availability of width and height as a signal. It is best to avoid using timeout as it depends and the running time (debug vs release vs speed), have you explored other options? 757398 7 hclam 2012-11-02 11:36:54 -0700 It's good that we remove this test by the way as it times out on Mac anyway I think it's because of the difference in platform decoder that requires more bytes to decode header. I just want to know how we can avoid using timeout for test. 758469 8 noel.gordon 2012-11-05 06:11:59 -0800 (In reply to comment #7) > It's good that we remove this test by the way as it times out on Mac That's the point of this bug, so let's remove it. > Anyway I think it's because of the difference in platform decoder that requires more bytes to decode header. Maybe, maybe not. You could test your theory to find out. > I just want to know how we can avoid using timeout for test. Not sure you can avoid it for partial load tests. I also choose timeout values to account for debug builds on all ports. The values seem to work: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#group=%40ToT%20-%20webkit.org&showAllRuns=true&tests=http%2Ftests%2Fimages 760199 9 noel.gordon 2012-11-06 16:36:50 -0800 Ping? 760206 10 171810 webkit.review.bot 2012-11-06 16:47:18 -0800 Comment on attachment 171810 Patch Clearing flags on attachment: 171810 Committed r133695: <http://trac.webkit.org/changeset/133695> 760207 11 webkit.review.bot 2012-11-06 16:47:21 -0800 All reviewed patches have been landed. Closing bug. 760420 12 dongseong.hwang 2012-11-06 23:23:04 -0800 I feel sorry for removing this awesome test. It is because the following reason. Currently, we can not pause a git animation image at the specific frame by javascript. I think this approach makes us check whether the specific frame of the gif animation image is rendered or not. However, we can not use this approach in mac so we can not use this approach for gif animations too. 760422 13 noel.gordon 2012-11-06 23:52:25 -0800 (In reply to comment #12). > Currently, we can not pause a git animation image at the specific frame by javascript. I think this approach makes us check whether the specific frame of the gif animation image is rendered or not. However, we can not use this approach in mac so we can not use this approach for gif animations too. This bug is about JPEG images. I'm not sure why you are talking about GIF images. 760440 14 dongseong.hwang 2012-11-07 00:40:10 -0800 (In reply to comment #13) > (In reply to comment #12). > > Currently, we can not pause a git animation image at the specific frame by javascript. I think this approach makes us check whether the specific frame of the gif animation image is rendered or not. However, we can not use this approach in mac so we can not use this approach for gif animations too. > > This bug is about JPEG images. I'm not sure why you are talking about GIF images. I skipped the detailed explanation. I'm sorry. This test intentionally pauses loading an image. When the image is a gif animation image, the default imageDecoder decodes the image until the last frame of the data received. If we use this technique properly, we can pause a git animation image at the specific frame. I had a plan to make some tests for gif animations using this technique. However, I knew this technique has a problem from this bug and this situation made me feel sorry. 171810 2012-11-01 04:18:51 -0700 2012-11-06 16:47:17 -0800 Patch bug-93636-20121101221712.patch text/plain 5585 noel.gordon U3VidmVyc2lvbiBSZXZpc2lvbjogMTMzMTIzCmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggMjIxNzc0N2IxODcwNzkxMGJkZTc3 YzUwMDg0Y2JmOWY4MzUzN2ZkNS4uYjRmYzA2MWJjOWVhM2ZmODc3YTI3MDIzYjk0ZTU5OGUwMzBj MTk4MSAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDIwIEBACisyMDEyLTExLTAxICBOb2VsIEdvcmRvbiAgPG5v ZWwuZ29yZG9uQGdtYWlsLmNvbT4KKworICAgICAgICAocjEyNTE4NSkgaHR0cC90ZXN0cy9pbWFn ZXMvanBnLWltZy1wYXJ0aWFsLWxvYWQuaHRtbCB0aW1lb3V0IG9uIExpb24KKyAgICAgICAgaHR0 cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTkzNjM2CisKKyAgICAgICAgUmV2 aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgUmVtb3ZlIHRoaXMgdGVzdDogZmFz dC9pbWFnZXMvanBlZy1wYXJ0aWFsLWxvYWQuaHRtbCBpcyBub3cgdXNlZCB0byB0ZXN0CisgICAg ICAgIHBhcnRpYWwganBlZyBpbWFnZSBsb2Fkcy4KKworICAgICAgICAqIGh0dHAvdGVzdHMvaW1h Z2VzL2pwZy1pbWctcGFydGlhbC1sb2FkLWV4cGVjdGVkLnBuZzogUmVtb3ZlZC4KKyAgICAgICAg KiBodHRwL3Rlc3RzL2ltYWdlcy9qcGctaW1nLXBhcnRpYWwtbG9hZC1leHBlY3RlZC50eHQ6IFJl bW92ZWQuCisgICAgICAgICogaHR0cC90ZXN0cy9pbWFnZXMvanBnLWltZy1wYXJ0aWFsLWxvYWQu aHRtbDogUmVtb3ZlZC4KKyAgICAgICAgKiBwbGF0Zm9ybS9tYWMvVGVzdEV4cGVjdGF0aW9uczoK KyAgICAgICAgKiBwbGF0Zm9ybS93aW4vVGVzdEV4cGVjdGF0aW9uczoKKyAgICAgICAgKiBwbGF0 Zm9ybS93aW5jYWlyby9UZXN0RXhwZWN0YXRpb25zOgorCiAyMDEyLTEwLTMxICBEaXJrIFByYW5r ZSAgPGRwcmFua2VAY2hyb21pdW0ub3JnPgogCiAgICAgICAgIFVucmV2aWV3ZWQsIHJlbW92ZSBl bnRyaWVzIGZvciB0ZXN0cyBubyBsb25nZXIgZmFpbGluZy4KZGlmZiAtLWdpdCBhL0xheW91dFRl c3RzL2h0dHAvdGVzdHMvaW1hZ2VzL2pwZy1pbWctcGFydGlhbC1sb2FkLWV4cGVjdGVkLnBuZyBi L0xheW91dFRlc3RzL2h0dHAvdGVzdHMvaW1hZ2VzL2pwZy1pbWctcGFydGlhbC1sb2FkLWV4cGVj dGVkLnBuZwpkZWxldGVkIGZpbGUgbW9kZSAxMDA2NDQKaW5kZXggOGRkMWQyMDI1MDQ1OTE2Mzli OTViMzM2YWY4MjAyZTYyZDUxNjU4NC4uMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAw MDAwMDAwMApHSVQgYmluYXJ5IHBhdGNoCmxpdGVyYWwgMApIY21WP2QwMDAwMQoKbGl0ZXJhbCAy ODczCnpjbWVBU0BOPyhvbEh5YHVWQnEhaWEweX55VXsrdllWMmE+aTAqWik9aF5obEElQFdzJGxI YG9rPG19P29UIXRoSQp6NloxMyFxYSp7UClEJU10MTdrfXxCVTV2OEJYaUBfTD16QnllTWtSTjFf ckx2by1VM2Q2PzVMK0cyfWF9QWowYXsKelFTKUU5ZWE/YTJ0d3cjT21vTSFwYHBXJCppa1lGfExp JT9HaEtCOGU1ZXk4NWpqUmo7IVdhemViOF5wU2wpcjBtCnprMzdzPyRQbnxxMzlCa2hlV019KzY7 UmZHQHoxXj1sbCkraXtWdEItZlBDPm1DV2ExOTlmcFAoMlN5cEFmaU9IJAp6IXI4NTV7dE9IaTte TCN9ZE5neTc3SHAlclVeRXZEKklhT2VLaWBnTG95TFpAJEFOOE8yMldRJW12djRGTyNxe3YKQnxF bUIxCgpkaWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvaHR0cC90ZXN0cy9pbWFnZXMvanBnLWltZy1w YXJ0aWFsLWxvYWQtZXhwZWN0ZWQudHh0IGIvTGF5b3V0VGVzdHMvaHR0cC90ZXN0cy9pbWFnZXMv anBnLWltZy1wYXJ0aWFsLWxvYWQtZXhwZWN0ZWQudHh0CmRlbGV0ZWQgZmlsZSBtb2RlIDEwMDY0 NAppbmRleCA4YjEzNzg5MTc5MWZlOTY5MjdhZDc4ZTY0YjBhYWQ3YmRlZDA4YmRjLi4wMDAwMDAw MDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwCi0tLSBhL0xheW91dFRlc3RzL2h0dHAv dGVzdHMvaW1hZ2VzL2pwZy1pbWctcGFydGlhbC1sb2FkLWV4cGVjdGVkLnR4dAorKysgL2Rldi9u dWxsCkBAIC0xICswLDAgQEAKLQpkaWZmIC0tZ2l0IGEvTGF5b3V0VGVzdHMvaHR0cC90ZXN0cy9p bWFnZXMvanBnLWltZy1wYXJ0aWFsLWxvYWQuaHRtbCBiL0xheW91dFRlc3RzL2h0dHAvdGVzdHMv aW1hZ2VzL2pwZy1pbWctcGFydGlhbC1sb2FkLmh0bWwKZGVsZXRlZCBmaWxlIG1vZGUgMTAwNjQ0 CmluZGV4IDIzOWMxOTdkZDA0YjRkMzExNzQyMjAwOWQwNWI0MzJmMjRjYmI5ODAuLjAwMDAwMDAw MDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAwMDAKLS0tIGEvTGF5b3V0VGVzdHMvaHR0cC90 ZXN0cy9pbWFnZXMvanBnLWltZy1wYXJ0aWFsLWxvYWQuaHRtbAorKysgL2Rldi9udWxsCkBAIC0x LDMwICswLDAgQEAKLTxodG1sPgotPGhlYWQ+Ci08c2NyaXB0PgotaWYgKHdpbmRvdy50ZXN0UnVu bmVyKSB7Ci0gICAgdGVzdFJ1bm5lci5kdW1wQXNUZXh0KHRydWUpOwotICAgIHRlc3RSdW5uZXIu d2FpdFVudGlsRG9uZSgpOwotfQotCi1mdW5jdGlvbiBjaGVja1dpZHRoKCkKLXsKLSAgICB2YXIg aW1hZ2UgPSBkb2N1bWVudC5nZXRFbGVtZW50QnlJZCgiaW1hZ2UiKTsKLSAgICBpZiAoaW1hZ2Uu d2lkdGggPT0gMjU2ICYmIHdpbmRvdy50ZXN0UnVubmVyKQotICAgICAgICB0ZXN0UnVubmVyLm5v dGlmeURvbmUoKTsKLQotICAgIHNldFRpbWVvdXQoY2hlY2tXaWR0aCwgMSk7Ci19Ci0KLWZ1bmN0 aW9uIHJ1blRlc3QoKQotewotICAgIHZhciBpbWFnZSA9IGRvY3VtZW50LmdldEVsZW1lbnRCeUlk KCJpbWFnZSIpOwotICAgIGltYWdlLnNyYyA9ICJodHRwOi8vMTI3LjAuMC4xOjgwMDAvcmVzb3Vy Y2VzL2xvYWQtYW5kLXN0YWxsLnBocD9uYW1lPS4uLy4uLy4uL2Zhc3QvaW1hZ2VzL3Jlc291cmNl cy9ncmVlbi0yNTZ4MjU2LmpwZyZtaW1lVHlwZT1pbWFnZSUyRmpwZWcmc3RhbGxBdD0xMDI1JnN0 YWxsRm9yPTYwIjsKLQotICAgIHNldFRpbWVvdXQoY2hlY2tXaWR0aCwgMSk7Ci19Ci08L3Njcmlw dD4KLTwvaGVhZD4KLTxib2R5IG9ubG9hZD0ic2V0VGltZW91dCgncnVuVGVzdCgpJywgMSk7IiBz dHlsZT0iYmFja2dyb3VuZC1jb2xvcjogeWVsbG93OyI+Ci08aW1nIGlkPSJpbWFnZSIgc3R5bGU9 ImJhY2tncm91bmQtY29sb3I6IGJsdWU7Ij4KLTwvYm9keT4KLTwvaHRtbD4KZGlmZiAtLWdpdCBh L0xheW91dFRlc3RzL3BsYXRmb3JtL21hYy9UZXN0RXhwZWN0YXRpb25zIGIvTGF5b3V0VGVzdHMv cGxhdGZvcm0vbWFjL1Rlc3RFeHBlY3RhdGlvbnMKaW5kZXggMzk2MWY4NzMyZjVkN2NkOGZhMWJj YzgzYjk4ZDk3NjdhOTdhZmVlOC4uNGE1MzUyMzc4ZTJiZTY5NWNiYmU4MTM2ZmM5MjE0NTE5MjQ5 ZTQxZCAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0vbWFjL1Rlc3RFeHBlY3RhdGlv bnMKKysrIGIvTGF5b3V0VGVzdHMvcGxhdGZvcm0vbWFjL1Rlc3RFeHBlY3RhdGlvbnMKQEAgLTEy NzQsOSArMTI3NCw2IEBAIHdlYmtpdC5vcmcvYi85MzI0NyBbIERlYnVnIF0gZmFzdC9saXN0cy9s aXN0LW1hcmtlci1yZW1vdmUtY3Jhc2guaHRtbCBbIENyYXNoIF0KICMgKHIxMjQ0ODQpIGluc3Bl Y3Rvci9kZXZpY2Utb3JpZW50YXRpb24tc3VjY2Vzcy5odG1sIGZhaWxpbmcgb24gTWFjIHBvcnRz CiB3ZWJraXQub3JnL2IvOTM1NTIgaW5zcGVjdG9yL2RldmljZS1vcmllbnRhdGlvbi1zdWNjZXNz Lmh0bWwgWyBGYWlsdXJlIF0KIAotIyAocjEyNTE4NSkgaHR0cC90ZXN0cy9pbWFnZXMvanBnLWlt Zy1wYXJ0aWFsLWxvYWQuaHRtbCB0aW1lb3V0IG9uIExpb24KLXdlYmtpdC5vcmcvYi85MzYzNiBo dHRwL3Rlc3RzL2ltYWdlcy9qcGctaW1nLXBhcnRpYWwtbG9hZC5odG1sIFsgRmFpbHVyZSBdCi0K ICMgKHIxMjU3OTApOiBOZXdseSBhZGRlZCB0ZXN0IGluc3BlY3Rvci90aW1lbGluZS90aW1lbGlu ZS1kZWNvZGUtcmVzaXplLmh0bWwgaXMgZmFpbGluZyBvbiBtYWMKIHdlYmtpdC5vcmcvYi85NDM3 NCBpbnNwZWN0b3IvdGltZWxpbmUvdGltZWxpbmUtZGVjb2RlLXJlc2l6ZS5odG1sIFsgRmFpbHVy ZSBdCiAKZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL3BsYXRmb3JtL3dpbi9UZXN0RXhwZWN0YXRp b25zIGIvTGF5b3V0VGVzdHMvcGxhdGZvcm0vd2luL1Rlc3RFeHBlY3RhdGlvbnMKaW5kZXggMmU5 NDcyZmMxZDFkOGRhMTNhOTg3OGEwNzA3ZmZlODk2NjNkNzlhOS4uNmRlZWFkMDY1MTc3M2U2Mzhk ZGMyMGM4MTAwZWQxMWVkMDc3ZDQwNCAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvcGxhdGZvcm0v d2luL1Rlc3RFeHBlY3RhdGlvbnMKKysrIGIvTGF5b3V0VGVzdHMvcGxhdGZvcm0vd2luL1Rlc3RF eHBlY3RhdGlvbnMKQEAgLTIyODYsNyArMjI4Niw2IEBAIHRhYmxlcy9tb3ppbGxhX2V4cGVjdGVk X2ZhaWx1cmVzL21hcnZpbi90YWJsZV9vdmVyZmxvd19oaWRkZW5fdHIuaHRtbAogaHR0cC90ZXN0 cy9mdWxsc2NyZWVuL2Z1bGxzY3JlZW5lbGVtZW50LWRpZmZlcmVudC1vcmlnaW4uaHRtbAogaHR0 cC90ZXN0cy9mdWxsc2NyZWVuL2Z1bGxzY3JlZW5lbGVtZW50LXNhbWUtb3JpZ2luLmh0bWwKIGh0 dHAvdGVzdHMvaGlzdG9yeS9iYWNrLXdpdGgtZnJhZ21lbnQtY2hhbmdlLnBocAotaHR0cC90ZXN0 cy9pbWFnZXMvanBnLWltZy1wYXJ0aWFsLWxvYWQuaHRtbAogaHR0cC90ZXN0cy9pbnNwZWN0b3Iv Y29uc29sZS1jcm9zcy1vcmlnaW4taWZyYW1lLWxvZ2dpbmcuaHRtbAogaHR0cC90ZXN0cy9pbnNw ZWN0b3IvaW5zcGVjdC1lbGVtZW50Lmh0bWwKIGh0dHAvdGVzdHMvaW5zcGVjdG9yL3Jlc291cmNl LWhhci1jb252ZXJzaW9uLmh0bWwKZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL3BsYXRmb3JtL3dp bmNhaXJvL1Rlc3RFeHBlY3RhdGlvbnMgYi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS93aW5jYWlyby9U ZXN0RXhwZWN0YXRpb25zCmluZGV4IDQxZDg0YWNlMzI5MGNmYjllMjAwNDczYzg1Yzg0MzQyYmFl ODBlYjUuLjcyOGJjNDYzODBiYmEyZTVlMjJkZDhhYzAxNWQxZDU2OGE0YWY0ZjYgMTAwNjQ0Ci0t LSBhL0xheW91dFRlc3RzL3BsYXRmb3JtL3dpbmNhaXJvL1Rlc3RFeHBlY3RhdGlvbnMKKysrIGIv TGF5b3V0VGVzdHMvcGxhdGZvcm0vd2luY2Fpcm8vVGVzdEV4cGVjdGF0aW9ucwpAQCAtMjc4Myw3 ICsyNzgzLDYgQEAgdGFibGVzL21vemlsbGFfZXhwZWN0ZWRfZmFpbHVyZXMvbWFydmluL3RhYmxl X292ZXJmbG93X2hpZGRlbl90ci5odG1sCiBodHRwL3Rlc3RzL2Z1bGxzY3JlZW4vZnVsbHNjcmVl bmVsZW1lbnQtZGlmZmVyZW50LW9yaWdpbi5odG1sCiBodHRwL3Rlc3RzL2Z1bGxzY3JlZW4vZnVs bHNjcmVlbmVsZW1lbnQtc2FtZS1vcmlnaW4uaHRtbAogaHR0cC90ZXN0cy9oaXN0b3J5L2JhY2st d2l0aC1mcmFnbWVudC1jaGFuZ2UucGhwCi1odHRwL3Rlc3RzL2ltYWdlcy9qcGctaW1nLXBhcnRp YWwtbG9hZC5odG1sCiBodHRwL3Rlc3RzL2luc3BlY3Rvci9jb25zb2xlLWNyb3NzLW9yaWdpbi1p ZnJhbWUtbG9nZ2luZy5odG1sCiBodHRwL3Rlc3RzL2luc3BlY3Rvci9pbnNwZWN0LWVsZW1lbnQu aHRtbAogaHR0cC90ZXN0cy9pbnNwZWN0b3IvcmVzb3VyY2UtaGFyLWNvbnZlcnNpb24uaHRtbAo=