mozilla :: #servo

 
27 Jun 2017
00:04heycamI have terrible trouble remembering when to use and_then instead of map
00:04heycammy brain just gets confused with what all the Options are doing
00:04bzheh
00:10bzemilio: ping
00:10cbrewsterajeffrey: so this case with iframes may be tricky with Servo... Maybe when an iframe is unbound, if the nested BC is same-origin we make sure the BC is removed from the document
00:10cbrewsternot sure if thats the _correct_ solution
00:11ajeffreycbrewster: yes, this may be tricky,
00:11cbrewsterajeffrey: We may not need to worry about the cross-origin case as you couldnt even append anything to the iframe's contentDocument anyways
00:12ajeffreyAFAICT every time a node is removed from the document we have to go via the constellation,
00:12ajeffreyto see which browsing contexts to discard.
00:12ajeffreyGulp.
00:12cbrewsterwhy every node? and not just iframes
00:13ajeffreybecause the node might have iframe descendants.
00:13cbrewsterwouldn't unbind_from_tree still be called on the iframe?
00:13ajeffreyah
00:13ajeffreyyou might be right
00:14* ajeffrey goes and looks
00:14cbrewsterif not we would be leaking pipelines
00:15ajeffreyOK, so when an iframe gets removed, we call unbind_from_tree,
00:16ajeffreywhich blocks on the constellation, and discards BCs correctly.
00:16cbrewsterajeffrey: ahh we block there anyways
00:17ajeffreyso why is the iframe still considered part of a document when it gets reattached?
00:18cbrewsterajeffrey: we check if document has a BC not if the BC is discarded.
00:18cbrewsterajeffrey: maybe we need to set BC to None on the Document
00:18emiliobz: pong
00:19ajeffreycbrewster: the important check is self.upcast::<Node>().is_in_doc_with_browsing_context() in bind_to_tree().
00:20ajeffreythat should be false!
00:20cbrewsterajeffrey: right, so either we check if the BC is *discarded* in is_in_doc_with_browsing_context or make sure document.BC gets set to None when the BC is discarded
00:20ajeffreycbrewster: ah, sounds likely.
00:21cbrewsterAFAICT document.BC would still be Some even after the BC is discarded
00:21ajeffreygot to dash now, be back in a moment.
00:21bzemilio: Again, commented in bug, but cloned() works on Option<&T>
00:21bzemilio: so I would need as_ref().cloned() instead of map(|s| s.clone())
00:21emiliobz: so we have an Option<Arc<ComputedValues>>, why do we need the clone() then?
00:21bzemilio: Can do that if you prefer
00:21cbrewsterajeffrey: sounds good, Im hopping on a plane so Ill be away for a few hours
00:21bzemilio: _That_ is a good question
00:21bzemilio: As in, &quot;why do we need to addref&quot;?
00:22emiliobz: right, we already have a strong reference, seems like we&#39;re just addreffing a temporary to drop the previous value, which is kinda dumb
00:22bzCould be
00:22bzWant me to take out the clones?
00:23bzThey do seem fishy
00:23bzwell, hmmm
00:24emiliobz: oh, ok, so the pseudos.get(..) does yield an Option<&Arc>, so it can be replaced with cloned(). The compute_pseudo_element_style_with_rulenode clone is not needed at all
00:24bzright
00:24bzOK, I will fix this
00:24bzBecause I agree that this is stupid.
00:24* bz should copy/paste less
00:24bzThe fact that clone() goes from &Arc to Arc is exciting too
00:25* bz needs to work in his mental model for what types are where
00:25emiliobz: nor the lazily_compute_pseudo_element_style&#39;s clone fwiw, so I think that looks good with those two removed.
00:25emiliobz: yeah, that&#39;s problematic sometimes. There&#39;s people that advocate for using Arc::clone(foo) explicitly
00:26emiliobz: but so far we haven&#39;t used it (maybe we should)
00:27bzRight
00:27bzOK, that&#39;s all I have
00:27bzI will post upated patches and re-request reviews as needed once I address heycam&#39;s comments
00:28emiliobz: btw, I fixed up the aPseudoElement arg in bug 1376077 already :P
00:28firebothttps://bugzil.la/1376077 FIXED, emilio+bugs@crisal.io ProbePseudoElementStyle&#39;s aPseudoElement argument doesn&#39;t make much sense
00:29bzAh, so I should rebase, eh?
00:29* emilio chose a pretty bad commit message though, reading back
00:29bzExcept then all your interdiffs in mozreview will suck
00:29bzOh, this doesn&#39;t affect my things, good.
00:29emiliobz: yeah, well... no need for it, the rebase is minor, so...
00:53bholleyhiro: emilio yt?
00:53emiliobholley: yes, though not for long I think
00:54bholleyemilio: see my comment re web animations
00:55emiliobholley: I&#39;m pretty sure you can
00:55emiliobholley: (you just pass a &quot;::before&quot; or &quot;::after&quot;) as an argument to the API IIRC
00:56emiliobholley: _but_
00:57emiliobholley: I think container->MayHaveAnimations() || SpecifiesTransitions() || SpecifiesAnimations() may do the trick, depending on how the initial style interacts with web animations (which I&#39;m not sure about)
00:57bzbholley: heya
00:57bz(not a ping, just a hello)
00:58birtlesemilio: let me know if I can help
00:59bholleybirtles: any insight on the above?
00:59birtlesbholley: what&#39;s the context?
00:59emiliobirtles: (This is about bug 1376245)
00:59firebothttps://bugzil.la/1376245 NEW, bobbyholley@gmail.com stylo: Attach styles explicitly to generated content
00:59bholleybz: hi!
00:59bholleywe could presumably all chat about this f2f at the reception in a few minutes, but I was hoping to get off a try push before then
00:59emiliobirtles: bholley wants to optimize the common &quot;we&#39;ve already computed the ::before/::after style for the pseudo&quot; to avoid the traversal machinery
01:00emiliobirtles: but the pseudo-element styles the element stores don&#39;t have animations computed (which is to avoid a lot of wasted work when restyling pseudos for animations)
01:01birtlesemilio: yeah, might be better to chat f2f in a minute -- but yeah, you can get a hold of a pseudo using a currently chrome-only CSSPseudoElement interface and then trigger animations on them from the web animations API
01:02emiliobholley: birtles: ok, sure, see you now then :)
01:02bholleyhm, ok. Let&#39;s just head to the reception and chat then
01:02* bholley -> reception
01:02* bholley is pingable on bholley_mobile
01:03bzbholley: See you there
02:40cbrewsterIs there a way to convert a HandleObject to a HandleValue without using rooted!(..)?
03:55wizwizwizwizis there a rust font rendering library which is fast and can output subpixel accurate font renderings?
04:08wizwizwizwizi guess i can just render at a higher res and downsample :-P
04:50cbrewsterwizwizwizwiz: pcwalton has been working on pathfinder https://github.com/pcwalton/pathfinder I believe he is working on a 2.0 that&#39;s not been released yet.
05:46heycam&quot;error: can&#39;t perform LTO when using multiple codegen units&quot;
05:46heycamhaven&#39;t seen that before
05:56stshinefinally got a sane debugger set in vscode :/
06:30bholleybirtles: yt?
06:30bholleyheycam: yeah, that means you can&#39;t do an opt build with multiple codegen units (are you setting that in your RUSTFLAGS?)
06:30bholleyhiro: yt?
06:30heycambholley: I was setting that, yeah
06:31* heycam remembers an email saying it improved compilation times or something
06:31bholleyheycam: yeah, but only works on debug builds
06:31bholleyheycam: it messes up opt builds
06:31bholleyemilio: yt?
06:31heycambholley: ah
06:34heycambholley: so when you&#39;re doing some testing on wikipedia, are you testing something other than page load? I was just trying to profile for bug 1372061 and I found hardly any Servo_* functions in my profile.
06:35firebothttps://bugzil.la/1372061 ASSIGNED, cam@mcc.id.au stylo: StyleChildrenIterator::IsNeeded may be hot, and it&#39;s slow.
06:35bholleyheycam: so, the IsNeeded is much hotter on restyle, because then there&#39;s a frame to query
06:35bholleyheycam: so I fire up a scratchpad
06:36heycambholley: that does make sense, yeah. although I am still surprised to find so little time under Servo_* calls.
06:36bholleyheycam: and paste in https://pastebin.mozilla.org/9025872
06:36bholleyheycam: how much time do you see under Servo_TraverseSubtree?
06:36bholleyheycam: on initial styling?
06:36heycambholley: I did not find Servo_TraverseSubtree at all
06:36bholleyheycam: then your profile is wrong :-)
06:36heycamheh
06:36bholleyheycam: or you have stylo disabled
06:36bholleyheycam: so I&#39;d start fixing that
06:37heycamlet me wait for this new build to finish
06:37bholleyheycam: I have been noticing weirdness from the gecko profiler recently, in that starting and stopping it doesn&#39;t always clear the old buffer
06:37heycamhmm
06:37heycamdon&#39;t think I saw that
06:37bholleyheycam: so I leave it disabled until right before I load the page
06:38heycambholley: yeah I was loading the page, enabling the profiler, hitting reload, then capturing
06:38bholleyhm I&#39;d think that should work
06:41KiChjangis everyone in SF btw?
06:48heycamKiChjang: everyone
07:13noxAlmost everyone.
07:16heycam:(
08:51noxSimonSapin: Any clue where it is specified that https://www.w3.org/TR/SVG2/painting.html#StrokeDashing should allow spaces as separators?
08:51crowbotnox: that&#39;s probably not the spec you want. Please read https://github.com/servo/servo/wiki/Relevant-spec-links
08:52SimonSapinnox: &quot; A <dasharray> is a list of comma and/or white space separated lengths or percentages.&quot;
08:52noxI&#39;m blind
08:52SimonSapinwhich probaly makes it wrong to use the # multiplier?
08:52noxYes.
08:52SimonSapinwait, whats even #*
08:53SimonSapinits not in https://drafts.csswg.org/css-values/#component-multipliers
08:53* SimonSapin shrugs
08:54SimonSapinnox: SVG 1.1 has probably gotten more review than SVG2 https://www.w3.org/TR/SVG/painting.html#StrokeDasharrayProperty
08:54crowbotSimonSapin: that&#39;s probably not the spec you want. Please read https://github.com/servo/servo/wiki/Relevant-spec-links
08:55noxSimonSapin: I just noticed our code is ok with &quot;foo, bar, &quot;
08:55noxwith a trailing comma,
08:55noxain&#39;t that a bug?
08:55SimonSapinyes
08:55SimonSapinprobably
08:55SimonSapinwhat code?
08:56noxSimonSapin: The function for handling &quot;commas but with spaces too, because why not&quot;.
08:56noxSimonSapin: parse_space_or_comma_separated
08:56* SimonSapin puts hands over ears and prepares to boil the ocean
09:08noxSimonSapin: Have done some machinery to handle space/comma/comma-with-space in the same way.
09:45noxSimonSapin: Hum https://github.com/servo/servo/blob/master/components/style/properties/helpers/animated_properties.mako.rs#L832
09:45* SimonSapin facepalms
09:49noxSimonSapin: Thankfully, Result<T, E> implements Sum too.
09:50SimonSapinuh
09:50SimonSapinwell, maybe it didnt at the time?
09:52noxSimonSapin: Don&#39;t know.
09:53noxSimonSapin: #[stable(feature = &quot;iter_arith_traits_result&quot;, since=&quot;1.16.0&quot;)]
09:53noxThank god we raised the minimum...
09:53SimonSapinhehe
09:53noxSimonSapin: And the code for this impl is actually complex.
09:53noxhttps://doc.rust-lang.org/src/core/iter/traits.rs.html#636-641
09:53noxResultShunt and whatnot.
09:54SimonSapinnox: if it were not stable Id have used an explicit for loop
09:55noxYeah same.
09:55noxSimonSapin: That&#39;s what the code was previously.
10:20noxbholley_mobile: What are &quot;the memory reordering optimizations introduced in 1.19&quot;?
10:36Aaronepowernox: I assume it&#39;s this https://github.com/rust-lang/rust/pull/40377 Which was released in 1.18. There are no reordering optimisations in 1.19 at the moment.
10:36crowbotPR #40377: Implement optimization fuel and re-enable struct field reordering - https://github.com/rust-lang/rust/pull/40377
11:27noxemilio: Around?
12:10noxSimonSapin: r? ^
12:41neon64I&#39;m trying to build Servo from source, and the`script` crate maxes out my memory usage. Any ideas whats in it so that it is this resource-intensive?
12:44noxneon64: It&#39;s huge and there is not much we can do about it for now. :(
12:45noxneon64: It contains everything that concerns DOM and similar things.
12:48neon64Aha.. nice . Well my computer isn&#39;t that great tbh - only 4gb ram and a cpu from 2010. So maybe I&#39;ll give up and download the binary
12:50noxSimonSapin: Is codegen-units supposed to help in these circumstances?
12:51SimonSapinon the contrary
12:52SimonSapinneon64: try changing `codegen-units = 4` in Cargo.toml to `codegen-units = 1`
12:52SimonSapinand compiling with `./mach build -j1`
12:52SimonSapinbut even so, 4 GB might not be enough to build the script crate
12:53SimonSapinneon64: yes, if you just want to try it out its much easier to download Servo Nightly. If you want to contribute Im sorry theres not much we can do
12:55neon64SimonSapin: yeah at the moment I don&#39;t have any plans of contributing so I&#39;m sure the download will be fine - and I can always switch to my desktop if need be
12:56neon64Thanks
13:40noxIs it me or geckotries are extra slow today?
13:51emilionox: now I am
13:52emilionox: it&#39;s sometimes very slow... I thought it was for the load, but there should be less people in Europe today so...
13:55noxemilio: Feel free to review my PR when you have time. :)
13:55noxQuite happy about it
14:02emilionox: just woke up, but will do asap :)
14:03emilionox: just skimmed over it, and looks quite nice over al :P
14:05emilionox: btw, I flagged you to review #17523. Should be trivial, but let me know if you don&#39;t want to review it and I can forward it to someone else
14:05crowbotPR #17523: style: Cleanup traversal.rs - https://github.com/servo/servo/pull/17523
14:05emilionox: also, you and your stupid 43 are missed :(
14:05nox43 :O
14:06noxOh.
14:06noxI misparsed that LOL
14:06noxI had two thoughs
14:06nox&quot;WOW we are 43 not to come?&quot;
14:06nox&quot;WOW I must have misparsed that, emilio wouldn&#39;t call anyone stupid&quot;
14:07noxemilio: Looking at the PR.
14:08emilionox: lol. Just to clarify, I meant your stupid liquor, which I still don&#39;t know how you can like :-)
14:08noxr+&#39;d.
14:08noxemilio: Yeah I got it afterwards. :D
14:08emilionox: thanks!
14:08noxemilio: If you don&#39;t make any extra effort,
14:09noxemilio: when you read a number in some English prose,
14:09noxdo you read it in English or in Spanish, in your brain?
14:09noxI read any number in French in my mind, and I never say &quot;quarante-trois&quot; to talk about the liquor,
14:09noxso I confused myself. :)
14:10emilionox: huh, Spanish, I think... fair enough :-)
14:16noxemilio: Maybe everyone is doing mach try in SF or something?
14:18emilionox: its 7am here, so unless they worked at night... Idk
14:39SimonSapinnox: 4+ years of Rust, and I dont understand this lifetime error
14:40noxSimonSapin: Which error?
14:40SimonSapinnox: Im messing with cssparser to return e.g. &Token instead of Token
14:41SimonSapinah, but if I stop looking at the error message and think about what Im doing it makes sense :)
15:35noxemilio: I broke filter, will look at it.
15:49SimonSapinnox: any idea how to make this compile? https://play.rust-lang.org/?gist=b295372607da5c342ee31b9ad5a687db
15:50SimonSapinI think it would Just Work with non-lexical borrows, but Id rather find something that works in todays rust :)
16:19emilioSimonSapin: wow, that error...
16:19noxSimonSapin: No luck.
16:20SimonSapinnox: thanks for looking
16:20SimonSapinI may have a work-around
16:21noxemilio: Fixed the geckotry, PR should be ready to be reviewed.
16:21noxNext step is killing OneOrMoreSeparated.
16:21emilioSimonSapin: just curious, how does the workaround look like?
16:23SimonSapinemilio: using break instead of return inside the loop (so that the borrow doesnt need to last for the entire function, since it isnt returned), then outside the loop call self.current_token()
16:23SimonSapinthis is in cssparser::Parser
16:23emilioSimonSapin: I see
16:24emilioSimonSapin: are you storing something like a already_tokenized: Vec<Token> or something of that sort?
16:24SimonSapinemilio: current_token: Option<Token>
16:24SimonSapinNone at EOF
16:43emilionox: why a Vec<T> instead of a Box<[T]>?
16:44noxemilio: Because that&#39;s what everything uses in the mako files and filter was the exception. I plan to bring back Box again all at once with every other property, but that means changing many call sites and that felt out of scope.
16:44emilionox: sounds fine
16:45noxemilio: My plan with sequences is to use type Sequence<Filter, Space, None> for the filter property.
16:45noxemilio: And make non-empty sequences have Impossible instead of None.
16:49canaltinovanox: hey, can you r? #17434 when you have time?
16:49crowbotPR #17434: stylo: Implement -moz-prefixed gradients - https://github.com/servo/servo/pull/17434
16:50emilionox: r=me
16:50canaltinovaSimonSapin: hey, and can you r? #17517 when you have time? :)
16:50crowbotPR #17517: stylo: Implement font-language-override descriptor for @font-face rule - https://github.com/servo/servo/pull/17517
16:50noxemilio: Cool.
16:50noxcanaltinova: Oops sorry forgot about that.
16:51canaltinovanox: no worries :)
16:52noxemilio: Good job noticing the wrongly split commits, will fix.
16:53SimonSapincanaltinova: r+
16:53canaltinovaSimonSapin: thanks!
16:58mbrubeckOoh, parallel sort in rayon https://github.com/nikomatsakis/rayon/pull/379
16:58crowbotPR #379: Add parallel sorts - https://github.com/nikomatsakis/rayon/pull/379
16:58mbrubeckI wonder if Servo does any sorting that&#39;s big enough to benefit
17:00stjepang1mbrubeck: &quot;big enough&quot; is &quot;length >= 10_000, more or less&quot;
17:02mbrubeckstjepang1: Good to know
17:11bz_sleep!seen jryans
17:11firebotjryans was last seen 3 days and 22 hours ago, saying &#39;jaws: thanks!&#39; in #fx-team.
17:13jrmuizelgw: ping
17:18SimonSapinjack: so hows the room for that meeting? Its in ~40 minutes, right?
17:19gwjrmuizel: pong
17:21jackSimonSapin: it&#39;s a huge room. we can definitely hook a laptop to vidyo, but am concerned about being able to hear
17:21jackwon&#39;t know if we&#39;ll stay here until after sean is done talking, but ealrier between the intro and the q&a it was *loud* in here
17:21jackET doubled in size since last work week it feels like :)
17:22GankroManishearth: coming back to nsCSSShadowArray -- gfx code can reasonably assume it won&#39;t be mutated -- can it mutate it to update the values without having to copy?
17:22Gankroor will that clobber some cached CSS state
17:25SimonSapinjack: laptop in a big room, I wonder if its even worth trying to call
17:29larsbergthis room also feels a lot smaller than hawaii
17:29larsbergand that room felt small and loud
17:30larsbergthe best part of this meeting so far is edunham and aneesh making crazy configurations out of Manishearth&#39;s lazer-cut Rust logos
17:30jrmuizelgw: we&#39;re in Mission I it would be nice if you could come
17:32gwjrmuizel: I&#39;m in the ET all hands right now, and then we have a Servo team meeting after that. I was planning to head over after lunch though - will you be in the same room all day?
17:39jrmuizelgw: yep
17:39canaltinovajeremychen: hey, I saw that `font-variant-alternates` has wrong parsing algorithm in servo side and in stylo-failures.md, that failure references bug 1355721 but I don&#39;t see the strong relation. Are you working on correcting this behavior currently? If not I can work on it
17:39firebothttps://bugzil.la/1355721 ASSIGNED, jeremychen@mozilla.com stylo: Implement @font-feature-values rule support
17:39gwjrmuizel: ok, I&#39;ll head over straight after lunch.
17:39jrmuizelgw: great
17:41jeremychencanaltinova: no I&#39;m not working on that at the moment. Feel free to take it! :)
17:42canaltinovajeremychen: cool, thanks!
17:47SimonSapinugh, matching borrowed tokens without non-lexical borrows is a nightmare
17:52SimonSapintheres just no way we can port all of servos css parsing to this new api :(
17:53SimonSapinjack: please ping the Rust team on https://github.com/rust-lang/rfcs/issues/811 :)
17:53crowbotIssue #811: Non-lexical borrow scopes and better treatment of nested method calls - https://github.com/rust-lang/rfcs/issues/811
18:05canaltinovabholley: hey, are you guys at stylo home room?
18:06canaltinovabholley: there is servo meeting or something in emerging tech room but I guess I should go to stylo room :)
18:06bholleycanaltinova: yeah
18:06bholleycanaltinova: we&#39;re all over in parc55
18:07bholleycanaltinova: 4th floor, lombard
18:07jackSimonSapin: nox: join my room
18:07bholleycanaltinova: are you in the servo meeting?
18:07bholleyjack: you guys doing a team meeting?
18:07canaltinovabholley: yeah, but it hasn&#39;t started yet
18:07canaltinovabholley: oh it started
18:07noxjack: Do I need mic?
18:07jackbholley: yes
18:08bholleyjack: ok - you guys want to come over to lombard when you&#39;re done?
18:08avadacatavrahi SimonSapin :) saw you joined the meeting
18:09SimonSapinHi
18:09SimonSapinnox: mute
18:09jackbholley: you want just stylo folks right?
18:09SimonSapinthanks :)
18:09bholleyjack: yeah
18:09bholleyjack: (also lmk if you want me to come join your meeting, happy to if that&#39;d be helpful)
18:09jycoops, I am in the graphics room! don&#39;t want to distract them by leaving, but will try to come to the stylo room later
18:15jdmnox: SimonSapin: https://public.etherpad-mozilla.org/p/servo-all-hands
18:25stshineIf you want more contributions to webrender you need to add more comments to it :/
18:30KiChjangwelcome to SF!
18:30noxtill++
18:33tillnox: :)
18:37noxBREAD ISSUES
18:38noxHI JDM
18:41KiChjangdid jdm go missing on the first day again?
18:41jdmnevar
18:42KiChjanglies, you were nowhere to be found in hawaii
18:42est31organ donor lol
18:43noxgw looks like he is listening, but he is secretly thinking about webrender version 47.
18:44* wafflespeanut envies at whoever&#39;s typing in etherpad
18:45KiChjangwafflespeanut, i think that&#39;s till
18:45jdmnope, me
18:45wafflespeanutwow
18:46tillyeah, I certainly couldn&#39;t do that
18:49Manishearthnox: document.write 4eva
18:49noxZankoku na tenshi no yo-- oops wrong eva.
18:50nox&quot;Ship vehicle&quot; is a pleonasm.
18:50Manishearthnox: also i had bread yesterday and there was nothing you could do about it
18:50frewsxcvmore like document.wrong
18:50noxYou had fake break Manishearth.
18:51avadacatavrafrewsxcv: you here?
18:51frewsxcvi&#39;m sitting in nyc
18:51frewsxcvbut i&#39;m here in this channel :)
18:54noxPLS SEND HALP
18:54frewsxcvhow&#39;s stylo nox
18:55noxavadacatavra: Say hi to azita!
18:57noxfrewsxcv: Pretty fine,
18:57noxfrewsxcv: I&#39;m making code smaller.
18:58jdmok, meeting over, notes done
18:58noxjack: Am gone.
18:58KiChjangwow... jdm&#39;s typing speed is ungoly
18:58KiChjangungodly*
18:58jdmI felt slow
18:58jdmtoo many typos
18:58jdmjntrnr&#39;s name is really frustrating to type quickly
18:58KiChjangjust type john next time
18:59jdmor mr t
18:59noxjdm: Just write only the vowels,
18:59noxthat will show him
18:59jdmha
19:00KiChjangoaaue
19:00noxDoctor Jntrnr and Mister Oaaue.
19:01wafflespeanut> zhiting: ?????
19:01wafflespeanutreally?
19:01jdmI totally missed what zhitingz and simran said
19:01jdmthe room was loud
19:02wafflespeanutheh
19:22marmistrzHi! I have trouble building rust, precisely: servo/osmesa-src
19:23marmistrzThe build log is here: https://pastebin.com/czfPKWmL
19:23marmistrzI&#39;m using Arch Linux. I have system-wide llvm installed, version 4.0.1-1
19:26marmistrzI have a lot of warnings about statically linked LLVM but the Arch Linux LLVM doesn&#39;t look like statically linked: https://www.archlinux.org/packages/extra/x86_64/llvm-libs/
19:26marmistrzDo you have any ideas?
19:30jdmmarmistrz: we have a known issue with using osmesa with the most recent llvm
19:30jdm(I think 4.0+)
19:30* jdm looks in the logs for the solution
19:31jdmmarmistrz: http://logs.glob.uno/?c=mozilla%23servo&s=22+Jun+2017&e=22+Jun+2017&h=osmesa#c704204
19:33marmistrzJust remove the osmesa-src from dependencies, right?
19:37zhitingzjdm: I&#39;m saying I&#39;m working on magicdom as the intern project. Sorry, I should speak louder.
19:40jdmmarmistrz: I think so
19:40jdmzhitingz: omg I did not know this
19:41jdminteresting
19:51marmistrzThis did the trick!
20:00gwjack: do you know where those empty rooms are that azita mentioned?
20:04jack4th floor
20:04jackRoom 10
20:04jackWas the manager one this afternoon
20:04jackBut apparently lots
20:04jackIf them are empty
20:07est31yes there is a pr &quot;update osmesa&quot; that apparently will fix it one day
20:07est31argh
20:07est31I forgot to scroll down
20:08est31disregard pls
20:11Manishearthbholley: there?
20:13SimonSapinnox: progress, I guess? https://imgur.com/a/KIOe9
20:21avadacatavrajack: are people going to those rooms?
20:33cpetersonStylo release plan: https://docs.google.com/document/d/1UH91n525xfmbHB49mTLMfE-3iuko_qc20YjS86xRVAU/edit#heading=h.3g0rdgxt9jay
20:33ferjmhttps://docs.google.com/document/d/1UH91n525xfmbHB49mTLMfE-3iuko_qc20YjS86xRVAU/edit
20:38Aryxcanaltinova: hi, there are stylo failures for the push: https://treeherder.mozilla.org/#/jobs?repo=autoland&revision=348133f59b33924a5a2093c5f15ed5f293ed4fc4&filter-resultStatus=testfailed&filter-resultStatus=busted&filter-resultStatus=exception&filter-resultStatus=retry&filter-resultStatus=usercancel&filter-resultStatus=runnable
20:40canaltinovaAryx: pushed the updates
20:40Aryxthank you
20:44bholleywafflespeanut: ping
20:45shinglyuhttps://treeherder.mozilla.org/#/jobs?repo=try&revision=0ac549771fc9d7c16447619f4df190664adc74e6
20:48avadacatavrahttps://public.etherpad-mozilla.org/p/servo-food
20:50shinglyuhttps://treeherder.mozilla.org/#/jobs?repo=try&revision=78ec0d5395330203b52db66c9d350daf04b70b32
20:50shinglyu^ two week old wpt result
20:50shinglyuI&#39;ll push a new one now
20:50gwjack: thanks - we found a room over in parc
20:52jackgw: who is we? :) i thought you were going to hang out with QR :)
20:52gwjack: yea - QR :)
20:52jackthey have a permanent room i thought
20:53gwpcwalton: we moved to balboa btw
20:53gwpcwalton: still in parc
20:53pcwaltongw: ok :)
20:56wafflespeanuthey canaltinova, pinged you in #17268
20:56crowbotPR #17268: Add support for subgrids - https://github.com/servo/servo/pull/17268
20:58canaltinovawafflespeanut: responded :)
20:58wafflespeanutcanaltinova: thanks :)
20:59gwjack: they have a permanent room but we were being too loud, and it&#39;s a bit cramped in there :)
21:10shinglyutill: Are you free tomorrow morning?
21:10shinglyutill: Astely from Taipei is also interested in the embedding topic
21:19KiChjangnox: why did you do document.write?
21:20noxNeeded for gdocs somehow.
21:41noxjyc: I changed Separator in #17530.
21:41crowbotPR #17530: Improve sequence values in style - https://github.com/servo/servo/pull/17530
21:42jychi nox! thanks for the heads up!
21:49luke_nukem[m]-Mfitzgen: Heya. How are you intending to distribute the mozjs rust bindings?
21:50fitzgenluke_nukem[m]-M: probably via `git = &quot;<url>&quot;` dependencies in cargo, at least initially
21:50fitzgenluke_nukem[m]-M: maybe at some point in the future actually publishing on crates.io
21:51luke_nukem[m]-MHmm okay. Was wondering due to the mozjs_sys dependency
21:57avadacatavrajdm: ping
22:04jdmavadacatavra: sorry, tracking down a bug with someone else at the moment; I&#39;ll be late
22:04avadacatavrajdm: np I&#39;m in the home room
22:07avadacatavraNvm going to registration room
22:13jdmheh
22:14avadacatavrathey have fruit here
22:18avadacatavratill: if you want to talk xows, jdm and i will be meeting somewhere when he&#39;s done with something
22:22TYLinemilio: the :hover bug I just mentioned https://bugzilla.mozilla.org/show_bug.cgi?id=1366163#c2
22:22firebotBug 1366163 NEW, nobody@mozilla.org stylo: rendering of <video controls> and <audio controls> is wrong
22:22shinglyuhttps://treeherder.mozilla.org/#/jobs?repo=try&revision=07e4d0c6178d8ef4289299b1e0e9a57d1154ce3d
22:23shinglyu^ Latest reftest, mochitest, wpt test result
22:23shinglyuhmmm wpt is all green
22:28luke_nukem[m]-Mfitzgen: A serious question - when it comes time to distribute the bindings via crates.io, what mozjs version will be targetted? The reason I ask is because many distributions (linux) rely on older versions - in particular GJS (Gnome JS bindings) uses 38 currently, but is being updated to 52 which is still unreleased.
 
27 Jun 2017
   
Last message: 13 seconds ago