that's probably what's confusing the mutation events ordering

so, the behavior indicates that this script tag's script gets run *after* the document itself has loaded, and not during document parsing

that's the only way i can make sense of the behavior

KiChjang: any chance the script element is marked async?

or defer?

jdm: here's what i'm looking at: github.com/servo/servo/blob/master/…ationObserver-document.html#L79-L91

now it doesn't seem to me that it's been marked as async or defer

but it *might* be the case that our code thinks otherwise

doubtful

KiChjang: so what are the symptoms that you observe?

lemme grab my PR

lemme try and remember what i saw...

KiChjang: what are the text node additions that log that you refer to?

it would be this: github.com/servo/servo/blob/master/…s/MutationObserver-document.html#L8

i believe the test harness appends new text node elements to this div

ah, right

KiChjang: what are the actual failures that are reported?

so, insertionStep2 expects an array of 2 elements, but it sees only 1

insertionTest2, sorry

i've looked into the mutation observer's sequence -- the only element in the array is not the event of the insertion of either inserted_script nor inserted_element

s/sequence/record

it's the text node under <div id="log">

KiChjang: I think it might make sense to try to make a reduced testcase that demonstrates this behaviour that you can easily run in servo and firefox and compare the output

standups: more XR integration work

standups: github.com/Manishearth/elsa (for gandalf)

standups: Spent the day working out how to make script_thread something that can execute futures.

standups: rewriting a surprising number of referrer-policy tests.

standups: meetings, writing, etc...

ugh why is everything terrible

the googlevr build isn't working anymore

(maybe its all android builds?)

oh no

Task "Servo daily: decision task. On failure, ping: SimonSapin, nox, emilio" complete with status 'failed'. Inspect: tools.taskcluster.net/tasks/JRT3kT8xT6CKmO9oFceDMw

r? Manishearth servo/media #176 and servo/servo #22522 please :)

Manishearth: please, let me know if you don't feel confident enough reviewing the Player changes and I'll ask vjaquez to do the review after he returns from PTO

Hi, I need a help to tackle assigned issue! while repeating test, it started to spit 'assert_not_equals: startTime got disallowed value 0' for all the tests. is it kind of cache problem? then how can i resolve this?

haedaal: you will want to figure out what in the test is causing that failure to appear, and trace that back to the code changes that you made

haedaal: it's possible that you have now exposed the startTime property whereas previous the test was getting undefined, now it's getting 0

jdm: swap you a review of servo/rust-mozjs #450 for one of servo/servo #22491?

can do!

jdm: ta

Jordan Le'Oran, Tristen Gaspadarek, Larissa Maestro, Caitlin Rose, Orono Noguchi, Timothy Shann, Mark Turner, Christopher Young, Matias S. Koedt, Marek Skiba, Sean Doherty, Kyle Gritchen

Zach Cox, Stephen & Leslie Braun, Mark Wetzstein, Brad Kaiserman, David Bornstein, Micah Hauptman, Stephen Engel, Dustin Milligan, Brock Williams, Jon Coplon, Ben Schnetzer, Dave Cowen, Nina Emkin

Ben Painter, Trevor Stoddart, Joe Ramirez, Gabriel Hart, Aaron Tap, David & Rene Pacheco, Lennon Kloser, Andrew Dunlap, Max maulitz, Glenn Schloss, Justina Adorno, Lanisha Javon Gholston

jdm: Manishearth larsberg mbrubeck SimonSapin ^

ajeffrey: bleah, where do I find the documentation for KeyboardProperties?

jdm: creator.magicleap.com/learn/referen…1____lr1__0_1_1ui_1_1_keyboard.html

thanks

silly me for trying to find it in the local packages I had installed

ajeffrey: there's a "speechToTextEnabled" boolean o.O

jdm: yup, might come in handy :)

there's also back, forward and clear keys that I should hook up at some point.

jdm: not sure how exokit is supporting key-down and key-up, since the lumin api only gives key-pressed creator.magicleap.com/learn/referen…r1__0_1_1ui_1_1_keyboard_event.html

ajeffrey: ah, the C API exposes different things than the lumin RT API

jdm: argh

very curious

jdm: not sure what to do about this, I'd rather not mix-n-match C and Lumin API calls :(

yeah, I'm fine with what's there right now

jdm: also, there may be different APIs for the physical and virtual keyboards.

I'll ask on the ML forum.

maybe I'll try adding hyper logging

jdm: Could it be that the responses are compressed ?

I still haven't taken time to fix the decoder thing...

eijebong: I mean, we still receive an incorrect response body in that case, right?

jdm: No, we might receive nothing if the first packet is a 1 byte sized one IIRC

hmm

maybe I'll try to fix that problem today and see if anything improves

I'll try to find motivation this weekend if you don't do it

The switch kind of killed my motivation to do code stuff o:)

standups: reviews

ugh ./mach create-wpt --reftest just created a 115MB diff by changing a bunch of stuff in all MANIFEST.json files :\

:o

anyway... it's friday and it's dinner time, so I'll investigate on monday :)

should be an easy fix in run_create

dispatch the update-manifest command first instead of adding the --manifest-update command to test-wpt

cool, thanks for the pointer

jdm: Yeah it'd need some work to have multiple decompression algorithms but it would at least fix our problems...

standups: Wrote a task executor for script_thread which allows async Rust functions to be used synchronously in JS. github.com/asajeffrey/servo/tree/script-async-experiments

ajeffrey++

Upside: allows other pipelines in the same script thread to get on with things.

Downside: might not be entirely spec-compliant, sigh. servo/servo #22501

jdm: HAHAHA

So sorry for you

Hmm, LuminOS now has WebView UI items? creator.magicleap.com/learn/guides/lumin-os-0-94-0-release-notes

jdm: Any more thoughts on #22491?

hi, who's managing servo's infra?

CI and bors in particular :)

pietroalbini: jdm and Manishearth do most of the infra mgmt at the moment.

jdm and Manishearth, just to make sure you're aware of this, travis-ci is going to migrate all the public repositories to travis-ci.com on january 31st

yep, I've been keeping up with it

ajeffrey: I'm fine with landing it

ajeffrey: barring the fact that almost nothing can land right now

jdm: ta

jdm, do you have an homu branch that works with the checks api?

pietroalbini: no, I guess we need to schedule time to make that happen

jdm, yeah, we (rust infra) learned of this today by accident, it's going to be fun

I'm probably going to look into adding the support to our fork soon btw

oh, rust has its own fork?

is it github.com/rust-ops/homu ?

jdm, yep

jdm, it shouldn't have diverged *too* much afaik

yeah, there's nothing particularly surprising there

jdm: btw, blair tells me that webvr is going to go away completely

Manishearth: oh, I thought we had talked about that

I mean, I suspect blair has some unrealistic expectations

oh

but certainly nobody is working on the webvr standard

well

and google really wants to remove its implementation

I asked for what the cross interactions between the two specs are

but keeps putting that off because of existing content

and blair said there aren't any bc webxr replaces webvr

that is the goal of webxr, yes

when it comes to cross interactions the only way they won't matter is if webvr doesn't exist at all

bc otherwise they can interfere

theoretically, yeah

for my current impl I'm reusing stuff and ignoring that, but eventually we have some choices to make

that sounds like some big news in the standards world

then again i'd rather have standards die quick and fast

lest we get stuck in backwards-compat hell

"💣 Failed to start rebuilding: Unknown error"

(followed by a retry)

standups: Off for winter break.