160857 2016-08-15 12:19:08 -0700 Promise API does not fully process microtask checkpoint: spec violation 2016-08-24 17:20:37 -0700 1 1 1 Unclassified WebKit JavaScriptCore WebKit Nightly Build Unspecified Unspecified NEW https://bugs.webkit.org/show_bug.cgi?id=140043 P2 Normal --- 1 a webkit-unassigned d ggaren sam youennf oldest_to_newest 1220109 0 286076 a 2016-08-15 12:19:08 -0700 Created attachment 286076 This file will demonstrate the bug Promises microtask queue does not fully process microtasks queued while microtask is running. According to spec, microtask queue should loop until it is empty. Why is this bug important? Promises are often used by developers for scheduling. When they are, developers assume Promise.resolve() means "execute this function right after this script returns". This bug violates this promise. FF/Edge/Chrome get this right. It'd be great if all browsers worked the same. Example: Promise.resolved() .then(() => { console.log("Promise 1"); return Promise.resolve(); }) .then(() => { console.log("Promise 2"); } window.setTimeout(function() { console.log('timeout'); }, 0); Might also be related to Bug #140043. I've attached a test case. 1220560 1 a 2016-08-16 13:47:01 -0700 Noticed that my example was not clear. The example should print: Promise 1 Promise 2 timeout In Safari, it prints Promise 1 timeout Promise 2 286076 2016-08-15 12:19:08 -0700 2016-08-15 12:19:08 -0700 This file will demonstrate the bug PromiseEventLoop.html text/html 3204 a PCFkb2N0eXBlIGh0bWw+CjxoZWFkPgoJPHRpdGxlPndlYmtpdCBwcm9taXNlcyBidWc8L3RpdGxl PgoJPHN0eWxlPiAjYXNzZXJ0cyB7IGNvbG9yOiByZWQ7IH0gPC9zdHlsZT4KPC9oZWFkPgo8Ym9k eT4KCTxoMj5Qcm9taXNlcyBhbmQgZXZlbnQgbG9vcCBpbnRlZ3JhdGlvbi48L2gyPgoJPHA+QWNj b3JkaW5nIHRvIHNwZWM6PC9wPgoJPG9sPgoJCTxsaT5Qcm9taXNlcyBhcmUgZXhlY3V0ZWQgYXMg bWljcm90YXNrcyBieSAibWljcm90YXNrIGNoZWNrcG9pbnQiLiA8YSBocmVmPSJodHRwczovL3Rj MzkuZ2l0aHViLmlvL2VjbWEyNjIvI3NlYy1wcm9taXNlLW9iamVjdHMiPkVDTUFTY3JpcHQgUHJv bWlzZXMgRW5xdWV1ZUpvYjwvYT4gKyA8YSBocmVmPSJodHRwczovL2h0bWwuc3BlYy53aGF0d2cu b3JnLyNlbnF1ZXVlam9iKHF1ZXVlbmFtZSwtam9iLC1hcmd1bWVudHMpOnF1ZXVlLWEtbWljcm90 YXNrIj5FbnF1ZXVlSm9iIGluIEhUTUwgc3BlYzwvYT48L2xpPgoJCTxsaT5NaWNyb3Rhc2sgY2hl Y2twb2ludCBydW5zIGluIGEgbG9vcCB1bnRpbCBhbGwgbWljcm90YXNrcyBoYXZlIGJlZW4gZGVs aXZlcmVkLiA8YSAgaHJlZj0iaHR0cHM6Ly9odG1sLnNwZWMud2hhdHdnLm9yZy8jcGVyZm9ybS1h LW1pY3JvdGFzay1jaGVja3BvaW50Ij5IVE1MIHNwZWM6IHBlcmZvcm0gYSBtaWNyb3Rhc2sgY2hl Y2twb2ludDwvYT48L2xpPgoJCTxsaT5NaWNyb3Rhc2sgY2hlY2twb2ludCBpcyBwZXJmb3JtZWQg aW1tZWRpYXRlbHkgYWZ0ZXIgY2FsbGluZyBzY3JpcHQuIDxhIGhyZWY9Imh0dHBzOi8vaHRtbC5z cGVjLndoYXR3Zy5vcmcvI2NsZWFuLXVwLWFmdGVyLXJ1bm5pbmctc2NyaXB0PiI+SFRNTCBzcGVj OiBjbGVhbiB1cCBhZnRlciBydW5uaW5nIHNjcmlwdDwvYT48L2xpPgoJPC9vbD4KPGRpdiBpZD0i YXNzZXJ0cyI+PC9kaXY+Cgo8c2NyaXB0PgovKgpUZXN0MSYyOgpodHRwczovL2h0bWwuc3BlYy53 aGF0d2cub3JnLyNjbGVhbi11cC1hZnRlci1ydW5uaW5nLXNjcmlwdApJZiB0aGUgSmF2YVNjcmlw dCBleGVjdXRpb24gY29udGV4dCBzdGFjayBpcyBub3cgZW1wdHksIHBlcmZvcm0gYSBtaWNyb3Rh c2sgY2hlY2twb2ludC4KClRlc3RzIHdoZXRoZXIgbWljcm90YXNrIGNoZWNrcG9pbnQgaXMgcGVy Zm9ybWVkIGltbWVkaWF0ZWx5IGFmdGVyIGNhbGxpbmcgc2NyaXB0LgoKVGVzdDM6Cmh0dHBzOi8v aHRtbC5zcGVjLndoYXR3Zy5vcmcvI3BlcmZvcm0tYS1taWNyb3Rhc2stY2hlY2twb2ludAoicGVy Zm9ybSBhIG1pY3JvdGFzayBjaGVja3BvaW50IiBydW5zIGluIGEgbG9vcCB1bnRpbCBhbGwgbWlj cm90YXNrcyBoYXZlIGJlZW4gZGVsaXZlcmVkLgoqLwoKdmFyIGNoZWNrcG9pbnRzID0gW107Cgpm dW5jdGlvbiBhc3NlcnRUcnVlKHN0YXRlLCBtZXNzYWdlKSB7CglpZiAoIXN0YXRlKSB7CgkJbGV0 IGVsID0gZG9jdW1lbnQuY3JlYXRlRWxlbWVudCgnZGl2Jyk7CiAgICBlbC5pbm5lclRleHQgPSBt ZXNzYWdlOwoJCWRvY3VtZW50LnF1ZXJ5U2VsZWN0b3IoJyNhc3NlcnRzJykuYXBwZW5kQ2hpbGQo ZWwpOwogIH0KfQoKZnVuY3Rpb24gdmFsaWRhdGVSZXN1bHRzKCkgewoJbGV0IHAxID0gY2hlY2tw b2ludHMuaW5kZXhPZigncHJvbWlzZSAxJyk7CiAgbGV0IHAxYyA9IGNoZWNrcG9pbnRzLmluZGV4 T2YoJ3Byb21pc2UgMSBjaGlsZCcpOwogIGxldCBwMiA9IGNoZWNrcG9pbnRzLmluZGV4T2YoJ3By b21pc2UgMicpOwogIGFzc2VydFRydWUocDEgPT09IDAsICJ0aW1lb3V0L3Njcm9sbCBzaG91bGQg bm90IGZpcmUgYmVmb3JlIFByb21pc2UudGhlbi4gU3BlYyAoMSkiKTsKICBhc3NlcnRUcnVlKChw MiAtIHAxKSAgPT09IDEsICJQcm9taXNlIDIgc2hvdWxkIGZpcmUgaW1tZWRpYXRlbHkgYWZ0ZXIg cHJvbWlzZSAxLiBTcGVjICgyKSIpOwogIGFzc2VydFRydWUoKHAxYyAtIHAyKSAgPT09IDEsICJQ cm9taXNlIDEgY2hpbGQgc2hvdWxkIGZpcmUgaW1tZWRpYXRlbHkgYWZ0ZXIgcHJvbWlzZSAyLiBT cGVjICgyKSIpOwogIGlmIChkb2N1bWVudC5xdWVyeVNlbGVjdG9yKCcjYXNzZXJ0cycpLmNoaWxk Tm9kZXMubGVuZ3RoID09PSAwKSB7CiAgCWxldCBlbCA9IGRvY3VtZW50LmNyZWF0ZUVsZW1lbnQo J2gxJyk7CiAgICBlbC5pbm5lclRleHQgPSAiVGVzdHMgcGFzc2VkIjsKICAgIGRvY3VtZW50LmJv ZHkuYXBwZW5kQ2hpbGQoZWwpOwogIH0KICBlbHNlIHsKICAJbGV0IGVsID0gZG9jdW1lbnQuY3Jl YXRlRWxlbWVudCgnaDInKTsKICAgIGVsLmlubmVyVGV4dCA9ICJUZXN0cyBmYWlsZWQuIgogICAg ZG9jdW1lbnQuYm9keS5hcHBlbmRDaGlsZChlbCk7CiAgICBlbCA9IGRvY3VtZW50LmNyZWF0ZUVs ZW1lbnQoJ3ByZScpOwogICAgZWwuaW5uZXJUZXh0ID0gIkV2ZW50cyBvYnNlcnZlZFxuIiAgKyBj aGVja3BvaW50cy5qb2luKCcsICcpOwogICAgZWwuaW5uZXJUZXh0ICs9ICJcblxuRXZlbnRzIGV4 cGVjdGVkXG4iICsgCiAgICAicHJvbWlzZSAxLCBwcm9taXNlIDIsIHByb21pc2UgMSBjaGlsZCwg dGltZW91dCwgc2Nyb2xsIgogICAgZG9jdW1lbnQuYm9keS5hcHBlbmRDaGlsZChlbCk7CiAgfQp9 CgovLyBURVNUCgovLyBSZXNvbHZlIHByb21pc2UgMQpQcm9taXNlLnJlc29sdmUoKQogIC50aGVu KGZ1bmN0aW9uKCkgewogICAgY2hlY2twb2ludHMucHVzaCgicHJvbWlzZSAxIik7CiAgICByZXR1 cm4gUHJvbWlzZS5yZXNvbHZlKCk7CiAgfSkKICAudGhlbihmdW5jdGlvbigpIHsKICAJLy8gUmVz b2x2ZSBwcm9taXNlIDEgY2hpbGQgaW5zaWRlIHByb21pc2UgMSBjYWxsYmFjawogICAgY2hlY2tw b2ludHMucHVzaCgicHJvbWlzZSAxIGNoaWxkIik7CiAgICB2YWxpZGF0ZVJlc3VsdHMoKTsKICB9 KTsKCi8vIFJlc29sdmUgcHJvbWlzZSAyClByb21pc2UucmVzb2x2ZSgpCgkudGhlbihmdW5jdGlv bigpIHsKICAgIGNoZWNrcG9pbnRzLnB1c2goInByb21pc2UgMiIpOwogICB9KTsKCi8vIFNldCB1 cCBldmVudHMgdGhhdCBtdXN0IGJlIGV4ZWN1dGVkIGFmdGVyIFByb21pc2UuCndpbmRvdy5zZXRU aW1lb3V0KGZ1bmN0aW9uKCkgewogICAgY2hlY2twb2ludHMucHVzaCgndGltZW91dCcpOwp9LCAw KTsKd2luZG93LmFkZEV2ZW50TGlzdGVuZXIoJ3Njcm9sbCcsIGZ1bmN0aW9uKCkgewogICAgY2hl Y2twb2ludHMucHVzaCgnc2Nyb2xsJyk7Cn0pOwp3aW5kb3cuc2Nyb2xsQnkoMCwxMCk7Cjwvc2Ny aXB0Pgo8L2JvZHk+