mozilla :: #devtools

16 Mar 2017
01:17asdffdsaIs there a version of firefox for android that has a functioning "about:debugging" where one can "Load Temporary Add-Ons"?
08:48bwintonasdffdsa: That all works for me in Developer Edition.
09:03soleochameau: fantastic pastries! \o/
09:52solemikeratcliffe: https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Adding_a_new_Telemetry_probe
09:57solenchevobbe: do you want to connect to hear me ramble about confusing developers? in 3min
09:57solenchevobbe: if not that's also fine teehehhehe
09:57nchevobbesole: yep, I was planning to attend
09:57gasolin_asdffdsa, you can check support state on bug 1305413
09:57firebothttps://bugzil.la/1305413 INVALID, sleroux@mozilla.com DB migration for highlights changes not running when updating from version 16 to 17
09:58solerickychien: gasolin I'm going to talk about docs and not confusing developers, if you want to join us in the devtools vidyo room :)
09:58gasolin_asdffdsa, sorry, its bug 1305143
09:58firebothttps://bugzil.la/1305143 NEW, nobody@mozilla.org Support installTemporaryAddon via remote debugger on Android
09:59solegasolin_: ^
10:01rickychiensole: sound great
10:02gasolin_sole, sorry I'm on the bus, so I'll miss the present:'(
10:11rickychienhmm... voice from mic is weird. It's hard to understand what's talking for me. lol
10:13julienwrickychien, the sound is extra weird here since the start
10:13julienwexcept sometimes
10:14julienwrickychien, do you manage to follow with the text ?
10:16rickychienjulienw: yes I'm trying
10:18julienwit's better when she speaks to somebody, than when she just gives the talk, right ?
10:19nchevobbehttp://searchfox.org/mozilla-central/source/devtools/client/responsive.html/docs/browser-swap.md
10:19nchevobbeSearchfox handles markdown files pretty well !
10:37nchevobbesole:
10:40solenchevobbe: thanks! hope it wasn't TOO RAMBLY
10:41nchevobbesole: no it's good to point those things out. I remember having trouble when I set up the dev environment
10:41nchevobbeall those differents docs to read, which tell different things and all
10:41soleYes!
10:42nchevobbedefinetely not easy
10:42solenchevobbe: and when it says "before you send a patch, rebase" (but no mention of how to rebase)
10:42nchevobbe:)
10:43nchevobbesole: I also think there's one thing that really helped me understand what was going on in the team, hence reducing confusion, was attending the weekly meeting
10:44nchevobbesole: but I only started when patrick invited me in, I saw it was a thing, but I couldn't tell if I was welcome there, if it was truely public, ..
10:45nchevobbesole: that's a side note to your talk, but it fills in the bigger picture of people having an understanding of why the team are doing some things, not doing others, ...
11:02solenchevobbe: absolutely :o
11:02soleOtherwise it's "oh, they're doing things..."
13:06nbprickychien: julienw: pbro: http://blogs.atlassian.com/2013/05/alternatives-to-git-submodule-git-subtree/
13:07julienwI know this already :)
13:07nbpThis is currently the best alternative I found to handle dependencies both within a single repo, and contribute upstream.
13:08julienwyeah I agree it seems quite good; did you use it yourself ?
13:08nbpjulienw: no, I did not had the opportunity yet.
13:08julienwok
13:09nbpjulienw: but I did use "svn -M" before, which does not have direct equivalent in git, but git subtree is the closest
13:09julienwgit submodules are better if we don't forget to use --recursive when cloning -- this hsould have been the default
13:10nbpjulienw: git submodules are painful to merge.
13:10julienwmm yeah I think this reminds me of something :D
13:16nchevobbei really like this page
13:20julienwnchevobbe, is the sound better ? I think I put the mic a little too far this time :)
13:20nchevobbei get some bits, some not, but overall I can follow
13:21julienwwhen he turns the head completely to the right we hear nothing grrr
13:43nchevobbe
13:52julienwnchevobbe, don't know if pbro told it, we disconnected because we'll have some work on whiteboard and it's painful on vidyo anyway
13:52nchevobbesure, np
13:53nchevobbeis there anything more planned on vidyo this afternoon ?
13:54julienwidk
13:57jryanssole: thanks for all your work on contribution guides! :D
14:15asdffdsabwinton: desktop firefox developer edition or android firefox developer edition?
14:16bwintonasdffdsa: Yeah, sorry about that, I mis-read your message. But gasolin_ had a better answer, I think.
14:16asdffdsaoh ok np
14:33ntimjdescottes: I'm curious to see how about:debugging looks like on android
14:33ntimFirefox for android
14:37jdescottesntim: it's not available afaik
14:37nchevobbejdescottes: ntim : yeah, just tested, I got a "invalid address" message
14:38jryansprobably for the best, i am sure it would look quite silly :S
14:38asdffdsayeah it seems like there's only a feature request for it atm. the developer edition of firefox android is "aurora" right?
14:40jryansclarkbw: should i file the CDN or are you on it>?
14:40clarkbwjryans: can you do that?
14:40jryansclarkbw: yep :)
14:41ntimgood to know, thanks for testing :)
15:03ntimIt would be really cool to have a WebExtension scratchpad where you can test different WE APIs without necessarily having to setup the boilerplate
15:03ntimI might do something like that
15:03ntimI guess I could repurpose devtools prototyper to be a WE scratchpad
15:06bwintonntim: That would be cool
15:09rplbwinton: ntim: this should be already possible (by creating a WebExtension with all the permission, install it from about:debugging#addons and open the scratchpad panel from the Addon Debugger)
15:11ntimrpl: simply writing the manifest is a lot of work :D
15:14rplntim: hehehe, I don't disagree, it is definitely annoying (that's one of the reasons why we are planning to introduce a "web-ext create" command and make it able to bootstrap a base webextension project, optionally using a specified git repo as a template)
15:17ntimrpl: that would be awesome to have!
15:18ntimrpl: I would probably name it web-ext init to be consistent with other tools
15:18rplntim: nevertheless, in the meantime, we can create and share an "all permissions test addon" as a git repo to be cloned and used as it is
15:19ntimrpl: It would be nice if it had a blank HTML page that&#39;s accessible from a browserAction, so we can also test different HTML features (I&#39;m thinking of <iframe mozbrowser> which I&#39;m working on atm)
15:19pbroTrying to list all of the pages we have on wiki.mozilla.org to see how much we need to cleanup/move/deprecate. I&#39;ve listed things here https://docs.google.com/document/d/1XrrvgSZka1uEeYynZRf99FI0ejiCIz_rGyuKHrFx8rI/edit# Feel free to let me know if I forgot any page
15:19rplI&#39;m not against to rename the command to init (especially given that it has not been merged yet, it is still in development and we are still in time to discuss about it and change it)
15:23rplntim: definitely, in the &quot;all permissions test addon&quot; I think that it is more than reasonable to hook a blank page on every extension context that is already available (e.g. also the sidebar, the option ui page, the devtools panel, the proxy API context etc.)
15:28ntimrpl: agreed
15:32solepbro: Nice list, I like it says where things should go to
15:53rindolfHi all! What can be done about https://github.com/devtools-html/debugger.html/issues/2381 ?
16:06solejryans: which work?
16:07jryanssole: well, your effort overall to clean up docs and try to show a better path for contributors, like your presented earlier :)
16:07solejryans: oh ok! I wasn&#39;t sure what exactly
16:07sole... were you referring too (hit return too early)
16:08solejryans: I don&#39;t think you were watching? it was early
16:08jryanssole: yeah, i read the notes after
16:09jryanssole: i added a paragraph about the code of conduct (copied from rust) to https://wiki.mozilla.org/DevTools/GetInvolved, maybe there are other places to mention it as well?
16:09soleyou added a paragraph where?
16:10solejryans: sorry I&#39;m feeling really drained
16:10jryanssole: it&#39;s right above the &quot;Contents&quot; box
16:10soleI see it!
16:10jryanssole: feel free to move of course, i wasn&#39;t sure where it should be really...
16:10soleI thought you&#39;d added that to the doc pbro made
16:10solejryans: https://docs.google.com/document/d/1XrrvgSZka1uEeYynZRf99FI0ejiCIz_rGyuKHrFx8rI/edit#
16:11jryansah, okay!
16:11pbrosole: jryans: out of that list, some pages make more sense in /devtools/docs/
16:11pbroso I filed https://bugzilla.mozilla.org/show_bug.cgi?id=1347977
16:12firebotBug 1347977 NEW, nobody@mozilla.org Move some DevTools technical docs to the tree
16:12jryanspbro: seems reasonable for the tech ones at least!
16:12soleyep yep
16:13pbrojryans: sending you a review request for one of them now. You wrote that doc I think, so let me know what you think
16:13soleThe last thing I thought was to look at the wiki and find tech docs tehre
16:14jryanspbro: did you manually reformat or use a tool?
16:14pbromanually, because it was a short one
16:14jryanspbro: i think http://pandoc.org/ can be used, if you want to automate more of them
16:14pbronice, thank you
16:17* rindolf is desperate
16:17rindolfand frustrated
16:18solealso if you can make pandoc work at all :D
16:18jryansrindolf: jlast may know about this since it seems related to jest? not sure if he&#39;s around.
16:18jlasthi
16:18rindolfjryans: ok, thanks
16:18rindolfjlast: hi
16:18solejlast was busy trying to fix a terrible thing in debugger.html
16:18solesomething docker related!
16:18rindolfjlast: yarn run test-all is still failing
16:18jlasthaha - yeah, some of us are updating the docker mc image
16:19jlastyep, we&#39;ll get to that after the docker issue on CI
16:19rindolfjlast: yarn run webpack seems to make the situation a little better but not perfect
16:20jlastthanks. sorry for the inconvenience
16:20rindolfjlast: you&#39;re welcome
16:20jlastfeel free to push your work with `--no-verify` and we can review it
16:20rindolfjlast: ah
16:20rindolfjlast: well, I didn&#39;t patch anything yet
16:33ntimpbro: I can&#39;t seem to access the doc
16:33ntimbut maybe I&#39;m not supposed to access it :)
16:35gasolin_rpl, ntim, &quot;all permissions test addon&quot; seems like a good candidate for testpilot
16:36rindolfMaybe I&#39;ll learn a new programming language while I wait.
16:38ntimgasolin_: why would it be ? it&#39;s supposed to be helping developers
16:41gasolin_ntim, I assume testpilot can be used to provide new addons to measure some assumptions
16:42gasolin_ntim, but maybe you are right testpilot is a more user facing place
16:46pbrontim: that&#39;s weird, the doc is in our shared drive, so it should by default be shared. I&#39;ll see if I can fix that
16:47pbrontim: https://docs.google.com/document/d/1XrrvgSZka1uEeYynZRf99FI0ejiCIz_rGyuKHrFx8rI/edit?usp=sharing
16:47ntimpbro: thanks!
19:55mykjryans: you probably already know this, but qbrt is the reason i&#39;ve been asking about uplifting positron devtools patches to firefox :-)
19:55jryansmyk: yes, that was my guess once i saw the post :)
19:56mykjryans: qbrt also does some wacky things to reuse devtools in firefox; namely: expand browser/omni.ja and then move browser/ to a different place
19:57mykjryans: (which breaks using --jsdebugger, unfortunately)
19:57jryansmyk: ah, do you have a link to where this happens?
19:57mykjryans: would love to find a different solution for that problem, which is essentially a variant of bug 1244163
19:57firebothttps://bugzil.la/1244163 NEW, nobody@mozilla.org DevTools inaccessible to XUL apps using &quot;firefox -app&quot;
19:58mykjryans: here&#39;s where qbrt expands the omni.ja: https://github.com/mozilla/qbrt/blob/master/bin/postinstall.js#L205-L214
19:59mykjryans: and here&#39;s where it moves the directory: https://github.com/mozilla/qbrt/blob/master/bin/postinstall.js#L240-L254
20:00mykjryans: it separately moves the prefs files, which need to be in a specific place: https://github.com/mozilla/qbrt/blob/master/bin/postinstall.js#L225-L239
20:01mykjryans: once the files are in the right place, qbrt&#39;s devtools.manifest file registers them: https://github.com/mozilla/qbrt/blob/master/devtools.manifest
20:04jryansmyk: well, devtools had a work week this week, and we seemed to reach consensus to head towards moving out of m-c to a separate repo deployed as an add-on. maybe this would make it easier for you?
20:06mykjryans: i bet it would!
20:06mykjryans: is there a bug/issue/other way to track that work?
20:08jryansmyk: no bugs yet, but here are meeting notes about the effort https://docs.google.com/document/d/1mTX9wskT7nr-x5Kp0bqS-Cs5RHQEmX9PCCkr3e9FoWU/edit which mentions ochameau&#39;s prototyping repo
20:09jryansmyk: the readme is maybe of interest https://github.com/ochameau/ff-dt/blob/master/README.md
20:11mykjryans: is that the upstreamiest repo, or is there an upstream to it?
20:13jryansmyk: that&#39;s it for now, but it&#39;s not really ready for use yet, he&#39;s force pushing things, etc. it&#39;s just a testing ground. we would make an official one with a better name once we&#39;re ready to switch to this.
20:14mykjryans: right, sure; just wondering if that&#39;s the one i should watch for now; sounds like it is!
20:15jryansmyk: right! we only discussed it as a team yesterday, so still many steps to take until we&#39;re ready for real.
20:16mykjryans: sure, but useful to know; thanks for the reference! in the meantime, if you have any suggestions for improvements to qbrt&#39;s use of devtools, i&#39;m game to hear them!
20:17jryansmyk: okay! i&#39;ll check it out in a bit
23:48ejpbruelquit
17 Mar 2017
No messages
   
Last message: 41 days and 11 hours ago