Closed
Bug 1402707
Opened 4 years ago
Closed 4 years ago
Rewrite browser_library*middleclick.js to use Places' async APIs
Categories
(Firefox :: Bookmarks & History, enhancement, P2)
Firefox
Bookmarks & History
Tracking
()
RESOLVED
FIXED
Firefox 58
| Tracking | Status | |
|---|---|---|
| firefox58 | --- | fixed |
People
(Reporter: standard8, Assigned: standard8)
References
Details
(Whiteboard: [fxsearch])
Attachments
(2 files)
As part of the ongoing rewrite of Places' unit tests to use the async APIs, we need to do the same for the browser_library*middleclick.js tests. We can also remove a lot of tab listener code for these tests if we make a slight alteration to BrowserTestUtils.waitForNewTab that allows it to wait for tab opens that aren't the first one after it is called. Hence splitting this out to a separate bug from bug 1094864 due to the wider impact.
| Comment hidden (mozreview-request) |
| Comment hidden (mozreview-request) |
| Assignee | ||
Comment 3•4 years ago
|
||
Somehow this has decided to leak places.xul on shutdown. I'll take a look at that on Monday.
| Assignee | ||
Comment 4•4 years ago
|
||
Comment on attachment 8911593 [details] Bug 1402707 - Change BrowserTestUtils.waitForNewTab to allow waiting for tabs to open a new URL when the tab might not be the next one opened. Thinking about it, some tests might rely on waitForNewTab to detect if it isn't the next tab for test failure purposes. I think I should make it an option.
Attachment #8911593 -
Flags: review?(mak77)
| Assignee | ||
Updated•4 years ago
|
Attachment #8911594 -
Flags: review?(mak77)
| Comment hidden (mozreview-request) |
| Comment hidden (mozreview-request) |
| Comment hidden (mozreview-request) |
| Assignee | ||
Comment 8•4 years ago
|
||
It looks like there's still memory leaks here. I'll have to have another think about what that could be.
| Assignee | ||
Updated•4 years ago
|
Attachment #8911593 -
Flags: review?(mak77)
Attachment #8911594 -
Flags: review?(mak77)
| Assignee | ||
Comment 9•4 years ago
|
||
For some reason the leak is still happening even if I null out gLibrary. I can't see anything obvious here at the moment, so I'm going to put it on the back burner for a few days, whilst I look at other items.
Blocks: 1068032
Priority: P1 → P2
| Comment hidden (mozreview-request) |
| Comment hidden (mozreview-request) |
| Assignee | ||
Comment 12•4 years ago
|
||
I think I've now fixed the leak issues. I've triggered a try build for the new patch.
| Assignee | ||
Comment 13•4 years ago
|
||
Try runs are now green :-)
Comment 14•4 years ago
|
||
| mozreview-review | ||
Comment on attachment 8911593 [details] Bug 1402707 - Change BrowserTestUtils.waitForNewTab to allow waiting for tabs to open a new URL when the tab might not be the next one opened. https://reviewboard.mozilla.org/r/183014/#review202288
Attachment #8911593 -
Flags: review?(mak77) → review+
Comment 15•4 years ago
|
||
| mozreview-review | ||
Comment on attachment 8911594 [details] Bug 1402707 - Rewrite browser_library_*_middleclick.js to use Places' async APIs. https://reviewboard.mozilla.org/r/183016/#review202308 ::: browser/components/places/tests/browser/browser_library_left_pane_middleclick.js:17 (Diff revision 4) > -var gTests = []; > -var gCurrentTest = null; > > -// Listener for TabOpen and tabs progress. > -var gTabsListener = { > - _loadedURIs: [], > +add_task(async function test_setup() { > + // Temporary disable history, so we won't record pages navigation. > + Services.prefs.setBoolPref(ENABLE_HISTORY_PREF, false); use await SpecialPowers.pushPrefEnv? ::: browser/components/places/tests/browser/browser_library_middleclick.js:21 (Diff revision 4) > - is(gBrowser.tabs.length, gCurrentTest.URIs.length + 1, > - "We have opened " + gCurrentTest.URIs.length + " new tab(s)"); > - } > > - var tab = aEvent.target; > - is(tab.ownerGlobal, window, > + // Temporary disable history, so we won't record pages navigation. > + Services.prefs.setBoolPref(ENABLE_HISTORY_PREF, false); ditto ::: browser/components/places/tests/browser/head.js:45 (Diff revision 4) > > +function promiseFocus(win) { > + return new Promise(resolve => { > + waitForFocus(resolve, win); > + }); > +} This is the 4th head.js wit this util, what about a follow-up to move it to TestUtils.waitForFocus()?
Attachment #8911594 -
Flags: review?(mak77) → review+
| Assignee | ||
Comment 16•4 years ago
|
||
| mozreview-review-reply | ||
Comment on attachment 8911594 [details] Bug 1402707 - Rewrite browser_library_*_middleclick.js to use Places' async APIs. https://reviewboard.mozilla.org/r/183016/#review202308 > This is the 4th head.js wit this util, what about a follow-up to move it to TestUtils.waitForFocus()? I've just discovered there's SimpleTest.promiseFocus(), so I'm switching it to use that. afaict there's only one other head.js that defines the function (in downloads).
| Comment hidden (mozreview-request) |
| Comment hidden (mozreview-request) |
Comment 19•4 years ago
|
||
Pushed by mbanner@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/1cadcb3b1963 Change BrowserTestUtils.waitForNewTab to allow waiting for tabs to open a new URL when the tab might not be the next one opened. r=mak https://hg.mozilla.org/integration/autoland/rev/d069236cdcef Rewrite browser_library_*_middleclick.js to use Places' async APIs. r=mak
Comment 20•4 years ago
|
||
| bugherder | ||
https://hg.mozilla.org/mozilla-central/rev/1cadcb3b1963 https://hg.mozilla.org/mozilla-central/rev/d069236cdcef
Status: NEW → RESOLVED
Closed: 4 years ago
status-firefox58:
--- → fixed
Resolution: --- → FIXED
Target Milestone: --- → Firefox 58
You need to log in
before you can comment on or make changes to this bug.
Description
•