The failure editor/libeditor/tests/test_resizers_resizing_elements.html
Bug 1652557 Comment 6 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: - It randomly fails when resizing <img>, <div> or <table> elements. - The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. - The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: - It randomly fails when resizing <img>, <div> or <table> elements. - The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. - The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. - Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: - It randomly fails when resizing <img>, <div> or <table> elements. - The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. - The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. - The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. - Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: - It randomly fails when resizing <img>, <div> or <table> elements. - The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. - The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. - The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. - Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. - Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. 7) implies this is not a fission-specific problem.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. _7._ implies this is not a fission-specific problem.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission).
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: - In case of failure, `HTMLEditor::OnMouseUp` is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) isn't called.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In multiple runs, I couldn't reproduce the failure when sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86). However, the failure was reproducible when waiting for the next animation frame (`await new Promise(function(resolve) { window.requestAnimationFrame(function() { setTimeout(resolve)})});`) instead.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 7\. implies __this is not a fission-specific problem__. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In multiple runs, I couldn't reproduce the failure when sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86). However, the failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In multiple runs, I couldn't reproduce the failure when sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86). However, the failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. 8) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. d) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content `](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content `](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 13) The __tests uses `enableObjectResizing`, which isn't supported by Chrome__. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content `](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 13) The __tests uses `enableObjectResizing`, which isn't supported by Chrome__ and [isn't spec'ed](https://w3c.github.io/editing/docs/execCommand/). 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content `](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 13) The __tests uses `enableObjectResizing`, which isn't supported by Chrome__ and [isn't spec'ed](https://w3c.github.io/editing/docs/execCommand/). 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 13) The __tests uses `enableObjectResizing`, which isn't supported by Chrome__ and [isn't spec'ed](https://w3c.github.io/editing/docs/execCommand/). 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). There's no focused content [according to the `FocusManager`](https://searchfox.org/mozilla-central/rev/6cc48251bb97600fdf11a5b4c5f621bfc8606d55/editor/libeditor/HTMLEditor.cpp#5879,5891). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 13) The __tests uses `enableObjectResizing`, which isn't supported by Chrome__ and [isn't spec'ed](https://w3c.github.io/editing/docs/execCommand/). 14) With `--enable-xorigin-tests --enable-fission` after synthesizing mouse clicks, unexpected `onblur` and `onfocus` occur. 7\. implies __this is not a fission-specific problem__. 8\. implies it's not a problem of one test not cleaning up after its execution. 14\. indicates, for fission with xorigin iframes enabled, the problem is likely the same as in [bug 1653160](https://bugzilla.mozilla.org/show_bug.cgi?id=1653160). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). There's no focused content [according to the `FocusManager`](https://searchfox.org/mozilla-central/rev/6cc48251bb97600fdf11a5b4c5f621bfc8606d55/editor/libeditor/HTMLEditor.cpp#5879,5891). In a passing run, there is focused content.
Some findings about editor/libeditor/tests/test_resizers_resizing_elements.html: 1) It randomly fails when resizing <img>, <div> or <table> elements. 2) The pref value of "editor.resizing.preserve_ratio" doesn't matter, it sometimes fails when it's true and sometimes when it's false. 3) The test case can be reduced by removing the event listeners for "beforeinput" and "input" and it still fails sometimes. 4) The test still fails when not setting "dom.input_events.beforeinput.enabled" to true. 5) Running the test only with `--enable-xorigin-tests` (without `--enable-fission`) doesn't seem to cause the failures. 6) Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer doesn't help. 7) ~Sleeping for 100 ms (using `setTimeout`) after all `synthesizeMouse...` calls after clicking a resizer without `--enable-xorigin-tests --enable-fission` causes the failures too.~ This wasn't reproducible anymore. I suspect a physical, unintended mouse-move might have caused the failure. 8) Sometimes the first test (first call of `testResizers`) fails. 9) Not [stopping to propagate `mouse*` events](https://searchfox.org/mozilla-central/rev/358cef5d1a87172f23b15e1a705d6f278db4cdad/editor/libeditor/tests/test_resizers_resizing_elements.html#37,39-41) still let's the test fail (only with --enable-xorigin-tests --enable-fission). 10) Sleeping (like in 6)) after [the first `synthesizeMouseAtCenter` call](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/tests/test_resizers_resizing_elements.html#86) leads to sporadic failures. 11) The failure was reproducible when [waiting for 2 animation frames](https://searchfox.org/mozilla-central/rev/fa7f47027917a186fb2052dee104cd06c21dd76f/dom/animation/test/testcommon.js#273-299) after resetting width and height, and after all `synthesize*` calls. 12) When `is(target, document.activeElement, "target is active element.");` added after clicking on the target to be resized, it always fails, even when the remainder of the test passes. 13) The __tests uses `enableObjectResizing`, which isn't supported by Chrome__ and [isn't spec'ed](https://w3c.github.io/editing/docs/execCommand/). 14) With `--enable-xorigin-tests --enable-fission` after synthesizing mouse clicks, unexpected `onblur` and `onfocus` occur. ~7\. implies __this is not a fission-specific problem__.~ 8\. implies it's not a problem of one test not cleaning up after its execution. 14\. indicates, for fission with xorigin iframes enabled, the problem is likely the same as in [bug 1653160](https://bugzilla.mozilla.org/show_bug.cgi?id=1653160). Searching in the dark: a) In case of failure, [`HTMLEditor::OnMouseUp`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#774) is called twice with `mIsResizing == false` immediately before the failure. In all success cases, it's called first with `mIsResizing == true` and then with `mIsResizing == false`, so perhaps the problem is around that variable. b) In the failure case, [`HTMLEditor::OnMouseDown`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp) and[`HTMLEditor::StartResizing`](https://searchfox.org/mozilla-central/rev/2f9eacd9d3d995c937b4251a5557d95d494c9be1/editor/libeditor/HTMLEditorObjectResizer.cpp#658) aren't called. c) In a failing run, [`HTMLEditorEventListener::MouseDown` returns early](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#373,389), because [`IsAcceptableMouseEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditorEventListener.cpp#258,265) returns `false`. It's because [`HTMLEditor::IsAcceptableInputEvent`](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/HTMLEditor.cpp#6140-6141) returns `false`, because [`EditorBase::IsAcceptableInputEvent` returns `false` because there's no focused content](https://searchfox.org/mozilla-central/rev/23dd7c485a6525bb3a974abeaafaf34bfb43d76b/editor/libeditor/EditorBase.cpp#4526,4537). There's no focused content [according to the `FocusManager`](https://searchfox.org/mozilla-central/rev/6cc48251bb97600fdf11a5b4c5f621bfc8606d55/editor/libeditor/HTMLEditor.cpp#5879,5891). In a passing run, there is focused content.