mozilla :: #devtools

11 Jul 2017
07:38nchevobbeHonza: Did you saw that https://bugzilla.mozilla.org/show_bug.cgi?id=1222633 ?
07:38firebotBug 1222633 FIXED, dd.mozilla@gmail.com Add support for <link rel=preload>
07:39nchevobbeHonza: I don&#39;t know if it&#39;s planned (or done :)), but it would be nice if netmonitor indicates requests using preload
07:39Honzanchevobbe: yes, good point
07:41Honzanchevobbe: bug 1379904 (reported for that)
07:41firebothttps://bugzil.la/1379904 NEW, nobody@mozilla.org The Network panel should indicate preloaded requests
07:42nchevobbeHonza: nice !
09:29pbrojdescottes: ping
09:29jdescottespbro: pong
09:30pbrojdescottes: I&#39;m changing a test right now that tests HTMLTooltip. And I see failures where this.xulPanelWrapper.openPopupAtScreen is undefined. The function doesn&#39;t exist. But the panel wrapper is a XUL <panel> element
09:30pbroany idea what can make the function undefined?
09:30pbrowhat I&#39;m doing is I&#39;m taking the HTML_URI in this test devtools\client\shared\test\browser_html_tooltip_arrow-01.js and moving it to an external .xul file isntead
09:31jdescottesI don&#39;t think I&#39;ve ever seen that!
09:31pbro(sorry TEST_URI)
09:34jdescottescould there be a security issue that would prevent calling certain methods on XUL elements in your document?
09:36pbroI don&#39;t know. I&#39;ll try things
09:41soleHonza: did you try the docs locally to make sure they worked fine? with the image and everything
09:41Honzasole: yes
09:41soleah thanks!
09:41soleSometimes gitbook can be tricky
09:41soleand doesn&#39;t include some files
09:41solethen I&#39;ll r+ :D thank you Honza !
09:41Honzasole: thank you :-)
09:42solereviewed! \o/
09:46pbrojdescottes: found the problem. Loading the XUL file from http://example.com/ as a root URL was causing the problem. Using CHROME_URL_ROOT works. This is basically gTestPath
09:47jdescottespbro: ok glad you found a fix, because I still can&#39;t find anything in the code that would explicitly whitelist some methods on this object :/
09:48pbroit&#39;s got to be due to a security thing where xul can&#39;t be executed in content
09:48pbroI mean, you can load the xml file all you want, but it&#39;s just not going to behave as xul
09:50jdescottesoh right
09:50jdescottesyes makes sense
11:17Honzanchevobbe: I am looking at bug 1308566, there are three remaining reviews for Brian, these are the ones I should do?
11:17firebothttps://bugzil.la/1308566 ASSIGNED, nchevobbe@mozilla.com Use the object inspector instead of variables view
11:45nchevobbeHonza: I just assigned you the 2 patch that you should look at
11:45nchevobbeHonza: since you already did the release actor one, there&#39;s only one left
11:46nchevobbethe one that actually uses the object inspector in the console
11:47Honzanchevobbe: the &quot;Use ObjectInspector in Console.&quot; seems to be assigned to no one ...
11:47nchevobbeHonza: weird
11:47nchevobbehttps://reviewboard.mozilla.org/r/152874/diff/6#index_header
11:48nchevobbei don&#39;t know why mozreview is adding &quot;r? cleared&quot;
11:49Honzanchevobbe: the commit message is still using r=bgrins, perhaps change that?
11:49nchevobbeoh yeah
11:51Honzanchevobbe: also, the ` Update reps bundle to v0.9.` patch is huge, but individual changes have been reviewed on github already no?
11:52zer0Honza: in which bug you&#39;ve added the `extend` in DevToolsUtils?
11:52zer0I can&#39;t find it again
11:52nchevobbeHonza: yes. The only thing there that maybe need to be checked is that I didn&#39;t do anything wrong removing the mochitests
11:52Honzanchevobbe: ok
11:52nchevobbeHonza: oh, this is on its own patch
11:53nchevobbeso yeah, the reps bundle one is really straightforward, everything was already reviewed in GH
11:54Honzazer0: originally, it&#39;s been bug 1370172
11:54firebothttps://bugzil.la/1370172 FIXED, odvarko@gmail.com Stop using sdk/core/heritage in Console panel
11:54zer0yes, but I remember you added the extend to devtoolsutils in a patch
11:55zer0Honza: ^
11:55Honzazer0: yep, looking fo the #
11:58Honzazer0: bug 1378850
11:58firebothttps://bugzil.la/1378850 ASSIGNED, odvarko@gmail.com Stop using sdk/core/heritage in DevTools webconsole hudservice
11:59zer0thanks!
12:12nchevobbeHonza: Sorry for spamming you with https://reviewboard.mozilla.org/r/152874/
12:12nchevobbeI tried various things but I can&#39;t get mozreview to not have the r? cleared thing
12:12Honzanchevobbe: np
12:13Honzanchevobbe: It looks like the commit message is still wrong
12:13Honzanchevobbe: I think you need to change it for the first three patches and push to review again
12:14nchevobbeHonza: that&#39;s weird, I thought we could ask different people for review on different patches
12:21Honzanchevobbe: yes, every patch can have its own commit message
14:38nchevobbeHonza: I updated my review to add some comments https://reviewboard.mozilla.org/r/152870/diff/11-12/
14:46Honzanchevobbe: R+ (the css problem wasn&#39;t coming from your patch)
14:46Honzanchevobbe: looks like ready to land! :-)
14:46nchevobbeHonza: nice, thanks !
14:47nchevobbeHonza: I&#39;m going to generate a new reps bundle with the latest Github changes, but it shouldn&#39;t change anything
14:47Honzanchevobbe: ok
15:45clarkbwHere are the GCLI commands: https://mzl.la/2u9ZtlQ
15:59solelooks like a new dashboard style?
15:59soleclarkbw: are there only 4 popular commands in use?
15:59clarkbwsole: well, help + 3 commands
16:00soleI was counting &#39;help&#39; as a command :D
16:00clarkbwsole: i&#39;m assuming people were using help to find one of the other commands :)
16:00jimbI&#39;m not getting any sound whatsoever out of Vidyo...
16:01pbrojimb: we can hear you apparently
16:04pbrojimb: want me to run it?
16:04jimbpbro: I&#39;m very sorry, but opening up my Firefox seems to kill my Vidyo audio.
16:04jimbpbro: Could you?
16:04pbrook good, I&#39;ll do it
16:04jimbpbro: Sorry, and thanks
16:05sole2017: the year of linux... audio? jimb
16:05julienwjimb, if you&#39;re on linux, you should try installing latest vidyo ... that&#39;s not proposed on v.mozilla.com but can be found elsewhere (if I can find it)
16:07julienwI recommend version 3.6.3
16:08julienwI have a .deb here
16:09solegl why are you shaking your head? :D
16:10glsole: oh it&#39;s just what i do when i am typing
16:10soleahhh I thought you were reacting to jlast&#39;s update
16:10glsole: gotta put my body into typing updates
16:14jimbjulienw: Okay, thanks!
16:15zer0gl: Superluminal DevTools
16:15jimbjulienw: I&#39;m on Fedora
16:18solemparkms: so nice work!
16:20mparkmssole: thank you!
16:21jimbsole: I think it&#39;s the Year of JimB Building Firefox At Random Times From Scratch
16:22jimbsole: Also, as an adamant defender of Linux&#39;s unsalvageable honor, I&#39;ll just point out that if I blame Vidyo, nobody will even blink
16:23* jimb demonstrates that Linux has always been at the forefront of excuses
16:23soleVidyo is a disaster
16:26zer0gl: since the bugs you mentioned seems all related to the Grid Highlighter, feel free to add me as mentor/reviewer for such bugs (I&#39;m also working on a refactoring of virtual canvas, so it would be good following the progress of Micah)
16:27glzer0: will do
16:27glzer0: gonna need some ideas from you on how to handle RTL and writing modes
16:28glzer0: I am not quite sure the bounds will do everything for us, but for instance, vertical writing mode might just benefit from applying a transform
16:28glzer0: if you have some thoughts on how to go about it, please add a comment on the bug
16:29glzer0: we might get to working on RTL by next week
16:29zer0zer0: sure, I discussed with Jen during the cssday nl about that.
16:32julienwjimb, ok, then my .deb is no use :) maybe ask on #avops, they&#39;ll have it ?
16:32julienwI mean they&#39;ll have a rpm for you
16:33jimbjulienw: I&#39;ll check on Vidyo&#39;s own site
16:33jimbjulienw: I&#39;m running 3.3.0
16:33glzer0: this is the current WIP for showing stacked lines + their line numbers and negative numbers https://irccloud.mozilla.com/file/JHyP5962/Screenshot%20from%202017-07-09%2015-40-39.png
16:33jimbdated 2014
16:33solerpm dealers
16:33MacMogIm trying to figure out how the Mozilla/Firefox test tools (libraries, runner) work and how easy it would be to use them for other software
16:33julienwjimb, thre is nowhere to download on vidyo&#39;s own site
16:34julienwwoudl be to oeasy
16:34jimbjulienw: ...
16:34julienw*too
16:34MacMogfarthest Ive gotten is <https://developer.mozilla.org/en-US/docs/Mozilla/QA/Automated_testing>; any suggestions for better documentation?
16:34jimbjulienw: okay, I&#39;ll try #avops
16:34glzer0: we are also gonna fix alignments with the line numbers if the line numbers happen to be cut off by being at the edge of a viewport
16:34zer0gl, for the stack this is what I suggested to Jen, but it could be simplified; for the negative number this WIP won&#39;t scale well with transformation.
16:34julienwjimb, https://vidyoportal.cern.ch/download.html?lang=en
16:35zer0gl: I already mentioned all of that at Jen, maybe she didn&#39;t recall.
16:35glzer0: ya, handling transformation is mystery case
16:35jimbjulienw: I&#39;ll check that out as soon as I can use a browser again :)
16:35julienwjimb, you can also be cc-ing https://bugzilla.mozilla.org/show_bug.cgi?id=840707
16:35firebotBug 840707 NEW, nobody@mozilla.org INC0015618 - Vidyo Desktop - Linux Audio Issues
16:35julienwfor more informatio nas they come in :D
16:35zer0gl: so I would avoid that.
16:36zer0gl: I mean, we should have a WIP that is taking transformations into account.
16:36zer0From the start.
16:36glzer0: yes, overall, we are avoiding handling the transformation case since you can easily toggle on/off the transformation
16:36zer0gl: but if you&#39;re avoiding handling the transformation that you have a biased design.
16:37zer0gl: we made a great effort on both our side and platform side to support them, we shouldn&#39;t waste everything now, IMVHO.
16:38zer0Otherwise we added a nice feature that we&#39;re going to break it.
16:39glzer0: while i agree, i don&#39;t think there is trivial fix for making all the line numbers visible under all transformations
16:39zer0gl, also this design is flaw when you do not have the same number on the top / left corner.
16:40zer0gl: that a more common use case, trasformation aside.
16:40zer0gl: it might not be trivial, but it&#39;s worth the effort to do not break a feature like that. There are at least a couple of approach we can investigate.
16:41glzer0: It&#39;s an iterative process. Currently, we don&#39;t have a solution for the transformation case, but I don&#39;t think that should hold us back from showing negative line numbers
16:42zer0gl: if it breaks how the highlighter is displayed, I think it should.
16:42zer0gl: one of our goal is polish our tools, if the negative number won&#39;t play well with the transformation applied, it would result to the user as an unpolished tool.
16:42zer0And that&#39;s also the main reason why we&#39;ve added transformation on transformed node too.
16:43glzer0: I also understand that, but I don&#39;t think I see what you are seeing with the current line numbers breaking the highlighter
16:44zer0gl: we have currently two issues with this design and the numbers.
16:44zer0gl: one, is related to transformation: if we have skew or scale, it won&#39;t work well.
16:45zer0the second is related to implicit numbers and the top / left number. The current design won&#39;t work with that either.
16:46zer0We should have a design that addresses both, although the second is probably more common and therefore more important.
16:47glzer0: this is also what was suggested regarding the line number placements https://irccloud.mozilla.com/file/QFWEQyyi/IMG_20170708_133005.jpg
16:47zer0gl: if you want, we can schedule a meeting and I can be more exhaustive and shows some examples.
16:47zer0gl: yeah, I know, this is what Jen suggested to me, but I told to her that this as is won&#39;t scale well for transformation.
16:49zer0gl: I&#39;ve to go. Feel free to add me to any discussions / brainstorming about that; or schedule a meeting. I&#39;ll try to have some proof of concept design too.
16:50gregtatumzer0: gcli is not used on the toolbar buttons anymore thanks to ochameau https://bugzilla.mozilla.org/show_bug.cgi?id=1320149
16:50firebotBug 1320149 FIXED, poirot.alex@gmail.com gcli takes between 20 to 40% of toolbox open time
16:50gregtatumhttp://searchfox.org/mozilla-central/source/devtools/client/definitions.js#545
16:50zer0gregtatum: well, it is.
16:50zer0gregtatum: I know that bug very well unfortunately.
16:50zer0:)
16:50gregtatumoh it is?
16:51gregtatumoh sohot
16:51zer0gregtatum: yep. So, it&#39;s not loaded at open time.
16:51gregtatumshoot
16:51gregtatumyou&#39;re right
16:51zer0gregtatum: but when you click on the button, with some witchcraft, we&#39;ll end up to call the gcli.
16:52gregtatumzer0: http://searchfox.org/mozilla-central/source/devtools/client/shared/developer-toolbar.js#47-58
16:52glzer0: Sure, let&#39;s schedule a meeting. My final note is that we won&#39;t be able to get everything perfect the first iteration, but I am more of the position of pushing forward what devrel needs and going fast.
16:53glzer0: of course, i would welcome alternative design of the line number placements if we can get it to address all the scenarios
16:53zer0gl: I&#39;m not saying has to be perfect, I&#39;m saying it shouldn&#39;t broke anything visually and leave a wrong impression to the user. I agree we can always improve it.
16:54zer0gregtatum: I worked on that in the last work week in paris. :)
16:55gregtatumzer0: ah nice, I was quite surprised to find the GCLI backing those buttons when i was working on the toolbar
16:55zer0gregtatum: unfortunately there was a regression about the buttons&#39; state
16:55zer0gregtatum: yeah. I think we should probably have a totally different approach about them.
16:56gregtatumzer0: agreed, there is a lot of unneeded complexity
16:56zer0It&#39;s a bit hard to follow the flow.
17:05gregtatumzer0: telemetry only fires from GCLI, not from the toolbar, I just tried it with a console.log on the telemetry
17:06zer0which command did you tried?
17:06zer0gregtatum: ^
17:06jimbjulienw: The RPM from CERN seems to have worked, thanks!
17:06* jimb casually downloads RPMs from a different continent
17:06soleCERN!
17:06solethis is so good
17:06zer0gregtatum: the code involved it&#39;s the same so it&#39;s odd if it doesn&#39;t fire.
17:07jimbsole: Yeah! This version of Vidyo has extra Higgs bosons, so it&#39;s MASSIVE
17:07gregtatumzer0: screenshot and measure
17:07soleahaha
17:07jimbsole: I will forever remember this meeting as the one in which you proposed human sacrifice and brain juicing as a regular devtools workweek activity
17:07zer0gregtatum: they don&#39;t have telemetry that&#39;s why.
17:07jimb&quot;seems reasonable&quot;
17:07gregtatumzer0: it&#39;s on Inputter.prototype._handleReturn
17:08zer0gregtatum: try paintflashing
17:08solejimb: No I said I didn&#39;t want to!
17:08gregtatumzer0: the telemetry is on the keyboard handler, so we&#39;re good
17:08jimbsole: Oh, sorry, IRC bug, I couldn&#39;t see what you just typed
17:08zer0zer0: ah, okay so it&#39;s not the telemetry of the commands
17:08zer0gregtatum: ^
17:09solejdescottes: ochameau ping?
17:09zer0gregtatum: we might then want to remove the telemetry from paintflashing
17:09gregtatumzer0: correct http://searchfox.org/mozilla-central/source/devtools/shared/gcli/source/lib/gcli/mozui/inputter.js#567
17:10zer0gregtatum: we might then add something for the button click, so we have the telemetry for both.
17:12zer0but I&#39;ve no idea if it&#39;s a useful data in that case, but we might compare what&#39;s the ratio of usage between the two, also considering that some buttons are not visible by default but they can be invoked by gcli anyway.
17:23glzer0: this is current WIP patch for the negative line numbers and line number stacking. https://raw.githubusercontent.com/TigleyM/firefox-devtools-patches/6e483f3d1f5bc17f77249697956a3436e8dff378/WIP/Bug1369942.patch
17:24firebothttps://bugzil.la/1369942 ASSIGNED, tigleym@gmail.com Display Negative Line Numbers in CSS Grid Inspector
17:24glzer0: you will notice that it doesn&#39;t work on retina displays
17:24glzer0: but still just a WIP
17:24glzer0: i think this will help you find breaking examples with transforms
17:25julienwjimb, good ! for me it fixed all issues I used to have on linux
17:28solepbro: jdescottes ochameau we join the meeting right? in ReleaseCoordination
17:29jdescottessole: yes, we will join (in a meeting atm)
17:29solejdescottes: thanks! no prob
17:29zer0gl: thanks!
17:36soleI feel like we&#39;re crashing somebody&#39;s meeting 8-)
17:36ochameausole: did we missed anything?
17:36soleNo, I thought they&#39;d be like WHO ARE YOU but they just went ahead
17:37soleochameau: ^
17:37solePerhaps pbro should say hi
17:37sole(when done)
17:37soleit&#39;s an interesting insight nevertheless
17:38pbrosole: so, Erin invited me and Bryan to a meeting before this one, to ask GoFaster questions
17:38pbronot sure why, maybe to make sure the Nightly meeting stayed short and smooth
17:42solepbro: is this the nightly meeting?
17:42pbrosole: yes
17:42solepbro: what did you get us into? :P
17:43pbrosorry ... Thomas said a couple weeks ago that it&#39;d be good for us to show up at this one
17:43pbrothis is about assessing the risks on nightly
17:43soleyeah from what I see!
17:43soleI wonder if there are cards for devtools here
17:43* sole looks at trello
17:44soleoh yeah
17:48pbrointeresting though ... people talking about our stuff
17:48pbrothere&#39;s a life after we land stuff in nightly ? oh wow :)
17:52pbrosole jdescottes ochameau : hope you disconnected
17:53ochameauyes, it looks that a super long meeting!
17:53soleI was being nosy and still there :P
18:46bzIs there a tracker for new debugger frontend breakage?
18:52zer0I notice just now the new Nightly&#39;s reload / stop icon! It&#39;s rad!
18:53bzAnd more to the point, should I be reporting a github issue for https://bugzilla.mozilla.org/show_bug.cgi?id=1380085 ?
18:53firebotBug 1380085 NEW, nobody@mozilla.org New debugger frontend breaks serviceworker debugging
18:53bzOr is that bug enough?
18:54* bz is not clear on which bug trackers who&#39;s ignoring.
19:00gregtatumjlast: yulia ^
19:05yuliabz: thanks for the report and for bringing it up here! we usually use github but this should be enough :)
19:06bzyulia: Great, thanks!
19:10jlastthanks bz
19:11bzOr https://pastebin.mozilla.org/9026873 if you prefer the json
19:11bzer, wrong window
19:12bzjlast: you&#39;re werlcom
19:12jlastadding it to the issue
19:12jlasthttps://pastebin.mozilla.org/9026873
19:12jlasthttps://github.com/devtools-html/debugger.html/issues/3313
19:13jlastbz - do you remember what the status of react + input behavior is?
19:13jlastsorry if that&#39;s vague
19:16bzjlast: that pastebin was unrelated to my bug report
19:16jlastahh - alright. wasn&#39;t sure
22:33victoriaHi everyone! For those of you who havent met me yet, Im the new DevTools designer. Ive mentioned this to some of you already, but heres my centralized doc for DevTools UX tasks and ideas. Hoping to continuously document things here so it doesnt get lost in emails/chats/meetings. You can also check out my Trello (linked in the doc) to see what
22:33victoriaIm actively doing on a daily basis. https://docs.google.com/document/d/148L0fEXC96xH-pQdpgcSd7fVJhdpQzcm9MK1xxvRdUo/
23:08ochameauvictoria: hi!
23:09ochameauI just requested some new builds with the add-on shim - https://treeherder.mozilla.org/#/jobs?repo=try&revision=d29bcf2d801c
23:09ochameauvictoria: I&#39;ll ping you again tomorrow to make you review that by testing it by yourself with these builds
23:10ochameauI&#39;ll ensure that these builds are correct tomorrow before pinging you
23:11victoriaochameau: Awesome, thanks! Will stay tuned
23:11ochameauI only partially implemented 3 screens. the page when a key shortcut is pressed, the similar one when you cliked on the system menu and the &quot;welcome page&quot;.
23:12victoriaSounds good. I received the graphics needed for that welcome page and am making a new mockup for it. There&#39;s code at the dev edition staging site that can be borrowed https://bedrock-demo-craigcook.us-west.moz.works/en-US/firefox/developer/
23:15ochameauvictoria: cool! for now I completely ignored the bottom part as you told me it was still in progress. But that shouldn&#39;t be hard to incorporate all that.
23:16victoriaSounds good!
23:18victoriaBtw, I thought about the idea you mentioned of having a Go back button on Install Complete page. It seems like it&#39;s unnecessary if the installation page always opens in new tab (a &quot;Close Tab&quot; button may make sense).
23:20ochameauyes I stripped all my code that was trying to complete the original shortcut. It may just be simplier and better as it was clearly going against the &quot;welcome page&quot;.
23:22victoriaAh okay! I think since most people who need DevTools will get it automatically installed, it&#39;s okay to have a bit of an extra step or two for those who are installing it for the first time (at least on a new computer or whatever)
12 Jul 2017
No messages
   
Last message: 74 days and 18 hours ago