mozilla :: #devtools

17 May 2017
09:47salty-horsehi. does the "simulate touch events" feature in the Responsive Design Mode allow simulating multi-touch pinch?
10:53soleSo it looks like someone took the pitfall document and translated bits into Chinese (?)-compare vs
10:53soleHas this been someone in Taipei?
10:53soleI used google translate and it mostly seems like a translation
10:55solejdescottes: ?
10:55jdescottessole: aucune ide
10:55solejdescottes: misterio!
10:56solejdescottes: trs intrigant
10:56jdescottessole: muy muy intrigante
10:57solejdescottes: I'd say super intrigante
10:57soleAlthough academics hate it when people use super instead of muy
10:57sole"you're destroying the language, we have better superlatives etc etc"
10:58jdescottesSuper Intrigante
10:58jdescottesthe famous spanish super hero
10:59jdescottestranslates your etherpads in chinese, behind your back. Super Intrigante!
10:59soleAll episodes are SO THRILLING and full of mystery
11:01jdescottessomeone needs to get started on a theme song, right now
11:03soleI don't have any music making devices on me right now! Sorry 8-)
11:05soleUnless you count a small laser gun novelty toy and a pacman stuffed toy (don't ask)
11:05sole"things that end up on people's desks"
11:55HonzaHi Patrick, looks good
12:11smaugRandom devtools profiling shows that in parent process there is some heavy Promise usage when opening web console.
12:11smaugPromises are synchronous from browser point of view
12:11smaugcould devtools perhaps use idle callbacks?
12:11smaugto reduce jank
12:12* smaug files a bug
12:20smaugBug 1365574
12:20firebot NEW, When opening devtools there is some heavy Promise usage in the parent process causing jank
12:27julienwsmaug, or setTimeout ?
12:28julienwbut good call here
12:28solesmaug: do you mean requestIdleCallback() ?
12:28smaugsetTimeout might be ok too
12:28smaugsole: yup
12:42julienwI don't really like idle callbacks for this, because when we open the console we don't want to wait for a idle moment before doing stuff ? But definitely we want to let the browser do stuff :)
12:42ochameauwe are still using task.jsm, would we expect some win by switching to async/await?
12:43ochameauand also still using promise.jsm
12:48julienwasync/await uses normal promises, ochameau
12:57ochameausure but using task.jsm and promise.jsm may add yet another layer of slowness of top of that
12:58ochameauI think /browser/ recently mass switched to async/await and got some win
13:58soleochameau: I have heard that too, or I might have read it somewhere. Didn't we mention some mass auto-rewrite of code? with a script or something
13:58solesomething along the lines of "expect a patch that rewrites your code"
13:59tromeyis this about Task.async?
14:00tromeythere was a patch but they only ran the script on some subset of the tree
14:03ochameausole: that is bug 1353542 you heard about and it looks like it wasn't run against /devtools. I tried opening the revisions but it crashed as it was too big.
14:03firebot FIXED, Switch to async/await from Task.jsm/yield
14:12soletromey: oh yes!
14:13tromeyI think it would be good to do this to our code
14:19soletromey: is there a bug on file? Possibly we can file it?
14:20fflorentHi there. It's been a month I didn't get update from DevEdition. I think I received some news in the mailinglist announcing that the channel will soon be updated. Any news on it?
14:20tromeyI don't know of a bug
14:20fflorent(platform: linux64)
14:21stoyanfflorent: I guess you'll get once next beta is released
14:21stoyanSince dev-ed will be migrated to beta
14:22fflorentstoyan: the beta channel has newer version than DevEdition
14:23fflorent54.0b7 for beta VS 54.0a2 for DevEdition
14:24stoyanfflorent: got your point but doubt anyone in this channel can be of help
14:25stoyanfflorent: From the next merge (April 18th), users running 54 Aurora will remain on the Aurora channel but updates will be turned off. In case of critical security issues, we might push new updates to these aurora channel users. Aurora channel users will be migrated to Beta channel in April 17. For this to happen, we need to make sure that the Developer Edition features are working the same way on the Beta update channel (theme, profile, etc).
14:25stoyanfrom the link
14:26fflorentAh right, so I have to wait a little bit again. Thanks
14:27stoyanfflorent: have fun while waiting (;
14:28fflorentstoyan: thanks ^^
14:31firebotBug 1365607 NEW, Switch devtools to async/await from Task.jsm/yield
14:31soletromey: brilliant!
14:31ochameaupbro: bgrins go faster meeting?
14:31bgrinsochameau: what room?
14:32ochameaubgrins: honza one
14:32pbroochameau: can't do it this week. Rust training conflicts
14:46tromeyjryans: around?
14:46jryanstromey: hello!
14:46tromeyjryans: is there any standard for source maps other than that Google doc?
14:47jryanstromey: afaik, that's all there is. fitzgen would know for sure.
14:47tromeyjryans: ehsan asked me to send an intent-to-ship for my platform change and I am trying to fill out the form properly
14:47tromeyI did some searching and also came to that conclusion
14:47jryanstromey: there is also a mailing list where elaborate dreams of the future are discussed
14:48tromeyI guess I should join that
14:50jryanstromey: perhaps, no posts in months i don't think, but still...
14:51tromeythanks, already joined :)
14:54jryansyeah, that contains some of the dreams :)
14:55jryanstromey: also various things on fitzgen's blog, but maybe you've seen them before
14:55tromeyoh no
15:12bgrinsochameau: jdescottes: was thinking about it more, and we certainly will need to be shipped with m-c source, so that jsdebugger works properly when running / testing (and so that browser devs can open BT without waiting for a system addon to be downloaded)
15:13jdescottesbgrins: the addon should automatically install if --jsdebugger is passed
15:14bgrinsah nice
15:14bgrinsis there a patch that i can use to test this?
15:15bgrinsand also, what if I open the BT at runtime but not via command line flag?
15:15jdescottesbgrins: see Bug 1361080
15:15firebot NEW, [meta][devtools-addon] Create a UI shim to install/bootstrap devtools addon
15:15jdescottes(in development)
15:45jdescottesbgrins: for the BT, the devtools menu item should trigger the UI shim that will guide the user to install the addon. We also support the "main" key shortcuts (console, inspector, debugger, F12), which will also trigger the UI shim and propose to install the addon
15:57ochameaubgrins: that WIP patch would do the trick, but still downloading the add-on would be painful. I agree that it would be handy to have it in m-c.
15:58ochameauI trust jdescottes to find a way to get that
15:59jdescottesa lot depends on that, so much pressure ...
16:11fitzgentromey: the google doc is all there is -- if youre looking for something informal, theres a good intro post somewhere... let me find it
16:14fitzgentromey: anything on my blog is mroe about what could have been -- a more DWARF like source map extension to encode scopes and how to recover them and their bindings' values
16:14fitzgennothing that actually exists in practice
16:52jryansochameau: i am really starting to doubt everything... you can load privileged chrome documents in the child process, right?
16:56ochameaujryans: I'm not sure as I don't know about any example yet
16:57ochameauabout:home is unprivileged
16:57ochameaubut I can't think of a reason why you couldn't
16:59jryansochameau: okay, i swear i had browser toolbox running in the child as a prototype, so surely that means it is possible...
17:40tromeythanks fitzgen
17:41fitzgentromey: in gneeral I'd love for someone to take source maps, pull a wasm on them (isomorphic binary format that's more compact), add scopes and things, and properly standardize
17:41fitzgentromey: I never got to, unfortunately
17:41fitzgentromey: (hint hint)
17:42tromeyit seems fun but I don't know if it's in the cards for me to do
17:42tromeyI do plan to dig through the existing bugs to see what maybe needs to be done
17:42tromeythere are even some open PRs
17:43fitzgenas far as standardization and stuff, the trouble was just getting anyone interested :-/
17:45tromeyI found a thread where the answer was no
17:46tromeyanyway that's how I read your response :-)
17:53fitzgentromey: not necessarily a "no" -- those questions should definitely have answers, and they *should* very compelling answers, but it was important IMO to enumerate them
17:53fitzgenbecause if we can't then thats a bad sign
17:53tromeythat makes sense
17:54fitzgenanyways, when that thread went on, I had hope of the source-map-rfc repo taking off
17:54fitzgenwhich it didn't
17:54tromeybtw about error.stack, at least one person asked me if we could source-map it
17:54tromeybut I think the answer is no
17:54fitzgenno other devtools teams from other browsers or compilers targeting JS ever responded to pings for comment or anything
17:54tromeysince afaik it's just a string by the time we get it
17:54tromeymaybe a standard would force their hand somehow
17:54tromeyI dunno
17:55fitzgentromey: I don't think we want to let web JS know if we devtools are open and if we are activel source mapping or what
17:55fitzgentromey: thats a huge can of worms
17:55tromeythe standard or the stack thing?
17:55fitzgentromey: the stack thing
17:56fitzgentromey: I don't think it is a good idea for web JS to be able to observe whether these things are active or not
17:56tromeyI agree
17:56fitzgentromey: and then theres the whole question of "what if the request for the source map is still in flight? do we block Error().stack calls?"
17:57tromeywe definitely would not want that
17:57tromeyI was wondering if we could somehow do it in the console
17:57tromeybut it's not really on my roadmap, just idle thinking
17:58fitzgenagree that doing all this stuff in the devtools is the way to go
18:02Sky_SymbolHello, I'm writing tests for the netmonitor panel. What's the best way to interupt a test while it's running, so I can look arround?
18:05stylizit@Sky_Symbol you can put a "debugger;" statement in your code and then run your test with the --jsdebugger flag
18:06stylizitSky_Symbol: ^
18:07Sky_Symbolstylizit: ok, let's try
18:09stylizitSky_Symbol: Oh and don't forget to open the Browser toolbox in Developer > Browser toolbox
18:10Sky_Symbolgreat, thank you :-)
18:11stylizitSky_Symbol: No worries!
18:28tromeyis it possible for an uninitialized 'let' binding to ever evaluate as truthy
18:29tromeylike a conditionally initialized one
18:29tromeylet x; if (...) { x = something } ... if (x) { ...}
18:29tromeyI would have thought this should throw
18:30tromeybut my current theory for my windows 7 pgo failures is that I'm seeing this behavior
18:32tromeyno, it doesn't throw, and I guess I was thinking of the function hoisting weirdness
18 May 2017
No messages
Last message: 5 days and 9 hours ago