mozilla :: #fx-team

17 Jul 2017
08:59pastisn't there a bug on file for centering the location bar?
08:59past(for photon)
09:05mikedeboerpast: centering? I think that work is revolving around the flexible spaces impl
09:05mikedeboer...which we have bugs for
09:06pastok, is that photon or visuals?
09:06paster, structure or visuals?
09:07pastI've been staring at the spreadsheet and can't find anything related
09:07johannhit's structure
09:07johannhwell there's bug 1366813
09:07firebot FIXED, Add customizable flexible space item in customize mode
09:08johannhbut it's already resolved without actually applying that flexible space in the toolbar
09:08johannhI can't see a bug for adding flexible space on both sides of the urlbar, there probably should be one
09:09johannhhow does the flex space even work? is it supposed to be resizable?
09:10pastI doubt it, but the spec I'm looking at doesn't mention anything
09:12johannhhm, we see a ton of these in mozscreenshots recently
09:13johannhwhere the tab bar is not scrolled consistently
09:13johannhit's because of something we changed in Photon
09:14johannhlazyirc: how do I programatically scroll the tab strip all the way to the left/right
09:19pastit looks like bug 1363485 is the missing piece
09:19firebot ASSIGNED, Move the Back, Forward, Reload and Stop buttons outside of the location bar container to the start o
09:21johannhright, we can't even put a separator between the buttons and the url bar
09:23pastbut it doesn't look like we have a bug for implementing the default experience (hoe button on the left, spacers on both sides, etc.)
09:24pastmaybe I missed it
09:29johannhI don't see it either
09:29johannhthere's also the question of migration
10:56johannhI can't use mozreview, it keeps crashing :(
10:57nhnt11johannh: are you trying to r+/- a patch?
10:57johannhoh maybe stylo
10:57johannhnhnt11: yup
10:57nhnt11johannh: yeah, I had similar problems
10:57nhnt11johannh: wait for the little header section to fully load
10:57nhnt11it works
10:57johannhnhnt11: do you have stylo enabled
10:57nhnt11I was going to report it
10:57nhnt11but I forgot
10:57johannhnhnt11: cool, please do
12:55makargh, this mozreview crash
14:25globnhnt11: got a bug number for that?
14:26nhnt11glob: had to leave the cafe I was at so I didn't file it
14:27* nhnt11 will file it after dinner if no one has done it by then
14:33makglob: nhnt11 there's a bug for my crash already. Wonder if it's same signature as yours
14:33firebotBug 1379617 NEW, stylo: Crash in nsIFrame::SetNextSibling
14:37globmak, nhnt11: i've had reports of people hitting bug 1381475 with mozreview too
14:37firebot NEW, Stylo: Crash in mozalloc_abort | abort | geckoservo::glue::Servo_ResolveStyle
14:42makI didn't crash there, the only other crashes I have are in js world
14:44Gijsanyone know how to get an assertion stack trace from the content process on windows?
14:44GijsI have a MOZ_ASSERT() but nothing shows up in my terminal, all I get is a content process crash.
14:45Gijs(with no stack, that so far I have had 0 luck catching in msvc)
14:45timdreamjaws: ping, did evanxd talked to you about ?
14:45firebotBug 1365133 ASSIGNED, Updated Preferences re-org according to proposal from May 11th
14:45jawstimdream: no, looking now
14:46jawscomment 236?
14:46johannhGijs: haven't tried it on the content process, but visual studio lets you debug on crash, right?
14:46Gijsyeah, doesn't work
14:46timdreamjaws: he is supposed to comment at the end of day but he didn't.
14:46Gijsas in, I'm attached, nothing happens
14:46timdreamjaws: yah
14:47jawstimdream: okay reading now
14:47johannhdisable e10s? :)
14:47Gijsjohannh: assert only happens in e10s
14:47timdreamjaws: another workaround found was adding a timeout *before* opening the tab.
14:48timdreamjaws: which suggests nothing in the about:preferences itself is at fault? not sure...
14:48johannhok, sorry, my vast pool of ideas is exhausted then :)
14:49Gijsjohannh: any idea how to get pids for the content processes?
14:51jawstimdream: yeah i see the patch from evan's try push,
14:51johannhGijs: unfortunately not
14:52johannhGijs: so, logging doesn't work for you either?
14:52Gijsjohannh: not sure what the question is
14:53Gijsjohannh: I'm in C++ land, I want a C++ stack
14:53Gijsa printf() wouldn't really help
14:53timdreamjaws: I don't think the one you pointed out works? I asked him to try before leaving office.
14:53johannhGijs: right, that's what I'm asking, I usually printf like hell
14:54jawstimdream: yeah that one works, that is from his try push with 50 runs, it has the same timeout in it
14:54timdreamjaws: Anyway that was trying to mimic another test which didn't fail at all
14:54johannhbut if it's really a stack you need, maybe #developers or #content knows more
14:54timdreamjaws: ?
14:55timdreamjaws: timeout == didn't work?
14:55jawstimdream: the timeout did work for him
14:55timdreamjaws: on you mean timeout in the code
14:56jawstimdream: yeah sorry lol i see the confusion now :)
14:57jawstimdream: i think this is the right idea, but we can't check in a patch with the setTimeout call in in. we will need to either way for the database to tell us it has been inserted or read the database continuously until we can see that it has been inserted
14:57timdreamjaws: sigh. I ask him not to try mixed approach. What you pointed added both the timeout and split add_task()
14:57jawstimdream: then we should also clean up the database after each test, so we don't have the following test relying on the previous test
14:58timdreamjaws: I doubt the problem is in the database. If it is I should fail at earlier line, right?
14:58* jaws looks
14:59timdreamThe item shouldn't even show up. We shouldn't fail only after selection.
15:00jawsoh because it's failing on ` let list = await waitForCondition(() => win.document.getAnonymousElementByAttribute(ourItem, "class", "actionsMenu"));`
15:00timdreamThe fact "ourItem" exists suggests the database is right.
15:01timdreamjaws: I assume we are waiting for a XBL binding there.
15:01timdreamjaws: I assume we are waiting for a XBL binding there.
15:02jawsyeah, that is what that await is already doing, and it is never getting applied?
15:02jawsbut in this try push we don't have this issue?
15:03timdreamjaws: that's the mystery. When it failed the screenshot shows the binding but the DOM didn't.
15:03timdreamjaws: can you point me the try URL so I don't need to scroll through his 100 try pushes?
15:04timdreamThe one you are talking about
15:06timdreamjaws: ok. That one fixed the problem with setTimeout(). Which like you said we can't check-in.
15:07timdreamjaws: what's weird is a setTimeout() before Preferences opens fixes "things". Not sure if the "things" === XBL binding.
15:09jawstimdream: we weren't seeing a failure with "browser_applications_selection.js" though, right? and that test runs before this test
15:10jawsand that test doesn't insert anything in to the database
15:10timdreamjaws: no we don't.
15:10jawstimdream: how about we try changing browser_change_app_handler.js to use the same mimetypes that browser_applications_selection.js uses (and stop inserting a fake one)
15:11timdreamjaws: So you still think DB is at fault?
15:11jawstimdream: not necessarily that DB is at fault, but if there are other things that might need to be set up and that the browser_change_app_handler.js is not doing
15:11timdreamjaws: ok we could try tomorrow
15:12jawstimdream: okay cool
15:13timdreamjaws: ttyl thanks for helping
15:14jawstimdream: you're welcome, i hope we can get this finished tomorrow
15:15timdreamjaws: I don't know not a lot of leads here....
15:15jawsyeah it's frustrating
16:04jawsGijs: do you know of a way to draw 'outside' of a stack?
16:05jawsGijs: i've got -moz-stack-sizing:ignore; on a child of the stack, but i can't draw outside of the stack
16:15Gijsjaws: nope
16:15jawsokay well at least i wasn't missing anything
16:16jawsgetting this pocket button positioned correctly with different font sizes is becoming a major PITA
16:25jawsoh i might have something
17:01ddurstelan: do you have a minute (ok, 3) to talk about trello use?
17:53bgrinsmconley: can i clear out the bi weekly meeting minutes doc? wanted to add an update for tomorrow's meeting but it looks like it has info from last week
17:54mconleybgrins: shoot - did we not dump the meeting notes?
17:54* mconley looks
17:55mconleybgrins: looks like we didn't. Let me copy the doc and you can clear
17:55bgrinsok, thanks!
17:55mconleybgrins: okay, you're good to go
18:21elanddurst: I have some time now
18:43elanddurst: I'm sending Thomas Elin to the rescue, he will be able to talk to you about trello
18:43ddurstjust saw that
18:43ddurstfor purposes of quantum, my email probably covers everything. I'll sync up with trello as instructed
19:04sfosterpaolo: hey, any idea when you'll be looking at that download/animations review? I have a minor color change to add but I can hold off if you are mid-review
19:08adw_6a68: if i want to submit a patch to the screenshots extension, should i post it to bugzilla/mozreview, or is there a github?
19:09RyanVM is the canonical repo
19:11adw_6a68: one more q: if i'm in bootstrap.js, what do i call to do whatever happens when the toolbar icon is clicked?
19:14adwit looks like bootstrap.js starts the webextension with this: webExtension.startup(startupReason) ?
19:14adwbut i have no idea what happens next
19:17adwlooks like browser.browserAction.onClicked in background/startBackground.js is what i want, but how can i call that from bootstrap.js?
19:21sfosterGijs: I dont fully understand the implications of those changes in the patch on bug 1336839. The await on the removeTab seems useful but AFAICT the load/newTab part should be functionally equivalent either way? I can stamp it if you like but if you were hoping for deep insight from me it will either need to wait or I can pass the review along to someone
19:21sfosterelse :)
19:21firebot NEW, Intermittent browser/components/customizableui/test/browser_880164_customization_context_menus.js |
19:25adwoh maybe i can send a message via api.browser.runtime?
19:28adw"TypeError: api.browser.runtime.sendMessage is not a function" :-(
19:52_6a68adw: hey! to send a message from bootstrap to webextension, I think you'll have to add a runtime.onConnect listener on both sides of the connection
19:52_6a68right now, we only use onMessage, which is one-way from webextension to bootstrap:
19:53_6a68you could wire up onConnect in the same spot (or I guess, replace onMessage with onConnect)
19:53adw_6a68: do you think that's the right way to do it?
19:54_6a68adw: yeah, based on the embedded webextension docs
19:54adwi mean, do you think messaging the extension from bootstrap to handle the clicking on the page action is the right way to do this?
19:55adwgenuinely don't know
19:56_6a68yeah, I think that's the best option, unless it's possible to create menu items from the webextension side (I'm not sure if the pageAction API will be updated for the photon release)
19:57adwok, thanks
19:58_6a68sure! we have a #screenshots channel for that project, though you're welcome to just ping me in here too
20:03adwok cool, that works
21:08jawsdo we have something like waitForEvent outside of test utils?
21:35Gijsjaws: not that I'm aware of...
21:35Gijsjaws: you mean you want a promise for an event?
21:36andymdo we have meeting notes for tomorrow's desktop meeting? wanted to type in some notes
21:37jawsGijs: yeah it would be better when writing non-test code if we could use things like waitForEvent
21:37jawsinstead we write nested callbacks
21:37andymthanks Mossop
21:43Mossopmconley: Are you tomorrow's meeting leader?
21:48* Mossop wonders why there are three thumbnails hidden beneath the scrollable area in about:newtab
21:58adwmixedpuppy: am i right that you own browser/extensions/pocket or at least can review changes to it?
21:58mixedpuppywell, pocket should really own that, but I can review
22:03adwmixedpuppy: ok. where should i submit patches? bugzilla/mozreview? github?
22:04mixedpuppyadw: bugzilla -> firefox::pocket and reviewboard is fine
22:16mconleyMossop: ooh, am I?
22:16* mconley checks
22:16mconleyWell well, so I am
22:16mconleythanks for the reminder!
22:18Gijsjaws: maybe move BTU's waitForEvent or promiseEvent or whatever to BrowserUtils?
22:18Gijsjaws: bonus points if you return a proxy on the promise that has a cancel() method
22:19Gijsjaws: IME we often end up with cases where we add a listener, and want to remove it on first use or when some other condition is met
22:20GijsMossop: re: about:newtab: I think this has "always" (at least, for a considerable amount of time) been the case but people notice it more on nightly after the temporary switch to a smaller about:newtab page
22:20GijsMossop: see also
22:21firebotBug 1380583 NEW, Newtab thumbnails got from 4x2 to 4x3
22:24MossopGijs: Ah interesting. It's trying to display a 5x3 grid in a space that can only handle 4 columns so it becomes 4 rows and the scrollbar only lets me see 3 rows
22:25GijsMossop: it would be interesting to know when that broke... but I think we've been shipping it for a while :s
22:26GijsMossop: that is, maybe "broke" is subjective here, but I expect our tests check what happens in a clean profile (and with window sizes that might not trigger this?) and so if we do consider it a bug, maybe it's a good place for more testing.
18 Jul 2017
No messages
Last message: 7 days and 8 minutes ago