mozilla :: #fx-team

11 Sep 2017
08:19pastAryx: try _showUpdateNotification() in nsBrowserGlue.js
08:19Aryxpast: thank you
08:19pastyw
10:04Standard8where do bugs relating to Clear Recent History / site storage go?
10:52Gijsmak: I'm looking at bug 1398252 and I'm confused...
10:53firebothttps://bugzil.la/1398252 NEW, nobody@mozilla.org Dragging identity box to the bookmarks toolbar to bookmark current tab makes tab dragging indicator
10:53Gijsmak: the bookmarks drop indicator container has position:absolute, but I don't see any ancestors that have position:relative - how did this ever work? :)
10:53Gijs(I mean, if we ignore the fact that the navigator-toolbox got position:relative as of bug 1397236 )
10:53firebothttps://bugzil.la/1397236 FIXED, jaws@mozilla.com Bookmarks folder in Bookmarks Toolbar unexpectedly closes after adding bookmark from Star button
10:54GijsLike, it clearly did work, and it goes back to working if I remove the position:relative from the #navigator-toolbox
10:54Gijsbut I don't understand how the positioning does work in that case
10:54Gijsis that just a XUL oddity or something?
11:16Gijsjohannh: is bug 1398264 being caught by screenshots?
11:16firebothttps://bugzil.la/1398264 ASSIGNED, gijskruitbosch+bugs@gmail.com Private browsing icon missing from private windows on Nightly
11:16GijsI would kind of expect it to...
11:17johannhno screenshots for private mode :(
11:17johannhfeel free to write a config
11:17johannhor make a bug about it
11:17johannh:)
11:20johannhGijs: ^
11:20Gijsfiled https://bugzilla.mozilla.org/show_bug.cgi?id=1398730
11:20Gijsthanks
11:20firebotBug 1398730 NEW, nobody@mozilla.org Should have a configuration for private windows
11:20johannhGijs: thanks!
11:21GijsIs Ekanan Ketunuti on IRC?
11:21Gijsand/or can anyone reproduce bug 1398264 on Windows?
11:24pastwfm on windows
11:27Gijsok, glad I'm not entirely crazy for failing to repro, at least
11:41jcristauis there a documentation of photon-related whiteboard flags i can find somewhere?
11:43Gijsjcristau: I doubt it, but it's fairly simple...
11:44Gijsjcristau: we have a few different subteams, and bugs relevant to those teams get tagged with photon-{subteam}, so [photon-structure], [photon-visual], [photon-animation] etc.
11:44Gijsjcristau: [triage] means we haven't prioritized the bug yet, which happens tue/wed/thu in subteam meetings.
11:44jcristauGijs: and the reserve-* flags are for backlog type things?
11:45Gijsjcristau: at this point I think almost all the mvp (minimum set of bugs we wanted to fix from the beginning) has been finished up by teams
11:45Gijsjcristau: so everything gets put in the 'reserve' because it wasn't part of the original mvp
11:45jcristauah, ok
11:45Gijsjcristau: I don't think that distinction is too helpful, but Marco Mucci can probably explain why (not) better than me
11:45Gijsjcristau: anyway, otherwise, p3 means 'high importance within the reserve', p4 means lower importance, and bugs get labeled p1 when people pick them up
11:45Gijsjcristau: does that help?
11:46jcristauyes, it does
11:46jcristauthanks
11:47jcristaui'm looking through bugs nominated for tracking for 57, so understanding how the team's triage works helps :)
11:48jcristauso i guess anything marked as P4 i don't need to track
11:50Gijsjcristau: well, it's up to you to disagree with the team's decision, if you think that's necessary
11:51Gijsjcristau: but yeah, it would seem unlikely that people would mark things as p4 if they also thought they needed to track 57
13:21daleharveyAnyone point me to around where the awesomebar popup gets positioned, trying to figure out whats up with https://bugzilla.mozilla.org/show_bug.cgi?id=1393406, would have expected http://searchfox.org/mozilla-central/search?q=PopupAutoCompleteRichResult&case=false&regexp=false&path=*.js to have more results
13:21firebotBug 1393406 ASSIGNED, dharvey@mozilla.com [macOS] [Photon] Regression: The Location Bar Drop Down Result List is 1px too high
13:23makdaleharvey: http://searchfox.org/mozilla-central/rev/00fa5dacedb925022f53d025121f1a919508e7ce/browser/base/content/urlbarBindings.xml#1949 (And from there you go into platform code for popups...)
13:32daleharveymak, thats perfect thanks
14:18mconleyEnn: woooo - bug 1356674! Good to have you back. :)))
14:18firebothttps://bugzil.la/1356674 ASSIGNED, enndeakin@gmail.com 32ms uninterruptible reflow at get_alignmentPosition@chrome://global/content/bindings/popup.xml:158:
14:49mconleydaleharvey: ping
14:51daleharveymconley: pong
14:52mconleydaleharvey: hey - choosing you randomly since I think you've been styling the URL bar stuff lately... I'm not sure how much that extends to the AwesomeBar panel, but am I right when I say that bug 1391391 is likely a matter of changing a box-shadow rule?
14:52firebothttps://bugzil.la/1391391 NEW, nobody@mozilla.org URL bar drop down shows shadow outside window
14:54mconleydaleharvey: bonus question, do you have the cycles to mentor Maya on that one?
14:54mconleyassuming it's not a quagmire?
15:00daleharveymconley: yeh it does seem possible to have the awesomebar shadow only show at the bottom
15:01daleharveyI cant review patches there, but I can answer simple questions hopefully, so sure
15:01mconleydaleharvey: cool. :) Thanks!
15:10daleharveyis it normal that getBoundsWithoutFlushing gets a value of .5? would assume it to be round
15:32Mossopjaws: Are you running the fx-team meeting tomorrow or did you run last week's?
15:32MossopLast time's even
15:32azakia shark running a meeting, that's something i'd like to see. =O
15:36jawsMossop: Felipe is running this one. I ran the last one
15:36Mossopjaws: Ok. It looks like the last meeting's notes never got converted to the wiki?
15:41sfosterazaki: maybe he's the james bond jaws?
15:45azakisfoster: i'd like to see that too..... lol XD
16:52Standard8Anyone know where bugs get filed about local storage / clear user data?
16:53Standard8aha looks like Core/DOM for local storage
18:37Gijsadw: is there a current bug on screenshots-in-the-urlbar not having a light icon when using a dark theme?
18:37adwGijs: it's probably https://bugzilla.mozilla.org/show_bug.cgi?id=1395284 ? maybe?
18:37firebotBug 1395284 ASSIGNED, adw@mozilla.com Wrong color for the Screenshots icon when added to the URL Bar.
18:38adwi can try applying my patch there to see if that fixes it
18:39adwthe svg doesn't specify context-fill-opacity or whatever
18:39Gijsno worries, that sounds right
18:42adwGijs: yeah, that fixes it
18:49mconleygandalf: hey - plate starting to clear a bit. I'm looking at bug 1397365 now, and hoping you can help get me caught up. Which tests are you still fighting?
18:49firebothttps://bugzil.la/1397365 ASSIGNED, gandalf@aviary.pl Set nodefaultsrc on initialBrowser in tabbrowser.xml
18:50gandalfmconley: https://bugzilla.mozilla.org/show_bug.cgi?id=1397365#c17
18:51mconleygandalf: that's the most up-to-date?
18:51mconleybecause I'm reading in later comments that you've made progress since then
18:51gandalfmconley: if you can take a look at the accessibility ones and the browser_referrer_middle_click that would be useful
18:51mconleysure
18:51gandalfI have a solution to 3 out of 4 accessibility ones but I'm not sure if it's the right solution
18:51gandalfand the click ones are just confusing
18:52gandalfin comment 10 I have some dirty-hacks that makes things work (mostly changing the default URL to load from about:blank to about:mozilla) - maybe it'll give you a clue?
18:53gandalfI suspect that some of that are just tests that don't wait for the right event, but others seem to be related to contentscripts not being injected
18:53mconleygandalf: well, was going to start with the middle_click one, but would you prefer I look at the a11y stuff first?
18:53gandalfyeah!
18:53mconleysure, okay
18:53gandalfI think it'll be easier
18:54gandalfI'm working on browser_ext_commands_onCommand.js
18:54gandalfthanks!
18:56mconleyheh, don't thank me yet - thank me if I find something that helps. :p
18:56gandalfoh, right, retracting the conditional courtesy. r=me
18:56mconley:)
18:57mconley(was joking, the thanks is appreciated all the same)
18:57gandalf:
18:57gandalf:)
19:08azakimconley: thank you.
19:08azaki(i dunno what's going on, i just saw you say not to thank you yet, and decided to be non-conformist)
19:09azaki=)
19:10azakimaybe i'll kick it up to 11 and send you flowers, too.
19:11azakithen the establishment will know i mean business.
19:37gandalfyeah! Hey, @mconley - . Deal with it.
20:11mconley:D
20:47dmoseDoes anyone know how to get the URI associated with a RemotePage:Init callback?
20:48dmoseI've tried port.target.browser.currentURI, but it sometimes returns about:blank, presumably because it's too early
20:48dmoseMossop: ^
20:49dmosereading RemotePageManager.jsm has not yielded enlightenment
20:55gandalfMossop: nsBrowserGlue has a welcoming number of 3000 lines of code. Do you have any recommendations where should I hook L10nRegistry into it? :)
20:56Mossopgandalf: _beforeUIStartup probably makes sense
20:57dmoseMossop: any thoughts on the above RemotePageManager question?
20:58gandalfthnx!
20:58dmosei.e. is there some browser API that will show me what page is loading assuming aboutURI has fired?
20:58dmoses/i.e./e.g./
20:58dmoseer loadURI, i mean
20:59Mossopdmose: I don't know that you can get the uri at that point. You could add it to the message being sent from content
21:00dmoseMossop: in other words, have RPM itself use its own knowledge to add it to the message?
21:00Mossopdmose: Right
21:01dmoseMossop: ok, thanks
21:01dmoseMossop: one thing that occurs to me is creating two RPM objects, one for about:home and one for about:newtab
21:02dmosei don't have a sense of how painful that extra bloat would be
21:02MossopSHouldn't cost much on the RPM side. Not sure about your side though
21:22mconleygandalf: soooo, no luck reproducing these accessibility failures locally when I revert the change you made to accessible/tests/mochitest/browser.js
21:22gandalfyou're saying that you can't reproduce the failures with just `nodefaultsrc="true"` set on initialBrowser?
21:23mconleygandalf: I'm looking strictly at the accessibility failures you mentioned
21:23gandalfyeah
21:23mconleyI applied your patch, reverted the change you made to accessible/tests/mochitest/browser.js, and then ran all of the tests that were failing
21:23mconleyand they all pass for me on OS X.
21:23mconleyWhich OS are you testing on?
21:23gandalflinux, but I do have a mac next to me
21:23gandalfI can test it there
21:23gandalfgive me a sec
21:25gandalfmconley: https://pastebin.mozilla.org/9032146
21:25gandalfno failure?
21:28mconleytesting
21:29mconleygandalf: mm, when I just apply that one change, yes, I can reproduce the failure.
21:30mconleygandalf: so, just so I'm clear, with the latest patch in the bug, you _still_ get the accessibility test failures on Linux?
21:30gandalfok, when I apply the whole patch this test is still failing
21:30mconleyon OS X too?
21:30gandalfjust the test_scroll.xul
21:30mconleyOkay, so I'm just going to focus on test_scroll.xul.
21:31gandalfthe other three are fixed by the change to accessible/tests/mochitest/browser.js aURL
21:31mconleypretty sure that was passing, but now I'll just run it by itself.
21:31mconleyright, but if you revert that change, do they fail on OS X?
21:31gandalfI'm not confident if my fix to accessible/tests/mochitest/browser.js is the appropriate one either :)
21:31mconleyif you revert the change to accessible/tests/mochitest/browser.js
21:31gandalfif I revert that change they fail on linux. I'm about to test on mac
21:31mconleydo the tests that you listed fail on OS X?
21:31mconleyokay.
21:33kitcambridgehi - is there a way to check if a JS xpcom component actually implements a method? i'm seeing exceptions like: "JavaScript component does not have a method named: "onBeginUpdateBatch"'JavaScript component does not have a method named: "onBeginUpdateBatch"' when calling method: [nsINavBookmarkObserver::onBeginUpdateBatch]" nsresult: "0x80570030
21:33kitcambridge(NS_ERROR_XPC_JSOBJECT_HAS_NO_FUNCTION_NAMED)
21:34Mossopkitcambridge: Not unless you can unwrap the component.
21:34kitcambridgei could make the method an empty function, and copy-paste everywhere, but i'm wondering if there's an easier way
21:34kitcambridgeMossop: ah, ok, in this case i can't. copy-paste it is. :-P thanks!
21:36gandalfmconley: did you reproduce?
21:36gandalf(I'm rebuilding on mac and it's taking its sweet time)
21:37mconleygandalf: nothing has changed - if I just use nodefaultsrc="true" in tabbrowser.xml, I can reproduce test failures. If I use your full patch, I cannot reproduce the accessibility test failures on OS X.
21:38gandalfok, can you switch to browser/base/content/test/referrer/browser_referrer_middle_click.js ? I'll investigate the accessibility on mac
21:38mconleyOkay.
21:58gandalfok, build complete
22:13gandalfmconley: yes, I'm getting test_scroll.xul timeout after applying the current patch from the bug on mac.
22:13mconleyvery odd
22:14gandalfI am not, however getting the test_scroll_caret.xul to hang with my whole patch
22:14gandalfare you on today's m-c?
22:15gandalfcause I'm not. I'm on Friday's. Will update to today
22:15mconleygandalf: I am, yes - I pulled this morning and applied your patch on top.
22:16gandalfok, updating the mac build
22:16gandalfdid you reproduce the browser_referred_middle_click ?
22:18mconleygandalf: yep
22:45mconleygandalf: this fixes the referrer test:
22:45mconleyhttps://irccloud.mozilla.com/pastebin/srQ1sphZ/
22:46mconleythe test harness was waiting for browser-delayed-startup-finished before loading a tab, but because _handleURIToLoad is asynchronous and kicked off from delayedStartup, it was overriding the tab that was being opened with about:blank
22:46gandalfmconley: soo, I rebuilt m-c to tonights, and the test_scroll.xul still times out on mac, with the patch applied. :/
22:46gandalfmconley: thanks! I'll test how many tests it fixes :)
22:47gandalfmconley: if you have time for one more, can you try browser/components/extensions/test/browser/browser_ext_windows_allowScriptsToClose.js ?
22:47mconleygandalf: not sure my patch here is the best way, btw. This seems like it'll be a recurring problem
22:48mconleywhere browser-delayed-startup-finished is no longer the point where it's safe to load stuff in the default case - we have to wait for _handleURIToLoad to do it's thing as well.
22:48mconleygandalf: I'm afraid I'm all out of time for today - got spread over a number of different bugs. Maybe tomorrow!
22:49gandalfok! thank you
22:49gandalfI'll work with that and see how far I can get by myself. Will update the bug :)
22:49mconleyokie dokie
22:51gandalfmconley: yep! all three referrer tests fixed. :D
12 Sep 2017
No messages
   
Last message: 8 days and 3 hours ago