00:28
<smaug____>
jsbell: Entry webkitGetAsEntry(); that webidl doesn't allow returning null, but the prose says null can be returned
00:28
<smaug____>
so Entry -> Entry?
01:23
<Domenic>
callback interfaces are icky
01:23
<Domenic>
but if this is a legacy API then what can you do
08:27
<hsivonen>
annevk: earlier I was looking for CC0 test cases for the Encoding Standard. Now there are some: https://github.com/hsivonen/encoding_rs/commit/186ab87b68a6cd1eb3fa4895ce6dc87c66f11cf8
12:53
<aleray_>
hi, using html5lib, when I run `html5lib.parseFragment(myhtml, treebuilder="lxml")` I get a list of Element, rather than an ElementTree. Is it normal?
13:08
<MikeSmith>
gsnedders: ⬆
13:41
<nox>
aleray_: I think so.
13:42
<nox>
aleray_: https://html.spec.whatwg.org/multipage/syntax.html#parsing-html-fragments
13:42
<nox>
Cf. step 14.
14:07
<aleray_>
nox, thanks
14:28
<aleray_>
nox, lxml html5parser seems todo it "by hand": https://github.com/lxml/lxml/blob/master/src/lxml/html/html5parser.py#L95
15:42
<JakeA>
Domenic: https://html.spec.whatwg.org/multipage/webappapis.html#dom-createimagebitmap - the only step that's in parallel here is resolving the promise. Does this mean the promise is kinda pointless and the work will be on the main thread?
15:43
<JakeA>
Eg steps 10 & 11
17:19
<Domenic>
JakeA: hah, yes, looks like a bug.
17:20
<JakeA>
Domenic: cheers, will write it up
17:20
<Domenic>
I wonder where implementations actually go to be "in parallel"
17:20
<Domenic>
seems like there could be race conditions depending...
17:20
<Domenic>
or maybe there's a copy of the incoming data made at the top of the algorithm?
17:20
<gsnedders>
aleray_: yeah, there's no way to give a fragment as a tree, because you essentially have multiple top-level eleemnts
17:22
<JakeA>
Domenic: pretty early. I think the intention is for decoding & raster to happen in another thread if the input is blob
17:23
<Domenic>
I'll be curious what junov says...
17:23
<Domenic>
Is the idea that the promise-returning only matters for the blob input?
17:24
<Domenic>
so for img/video/etc. cases we should just collapse them into "Return a new promise resolved with ..."?
17:26
<JakeA>
Domenic: I'd love it if we could do SVG rasterizing in another thread. That means it should be done for <img> too, since the resize operation means the SVG needs to be re-rastered.
18:18
<Domenic>
Can someone check to see if Gecko uses the replace argument to window.open? smaug____?
18:19
<smaug____>
Domenic: you mean if that is implemented?
18:19
<Domenic>
smaug____: yeah
18:19
<smaug____>
(in a meeting but will check after that)
18:19
<Domenic>
thanks!
18:32
<Domenic>
smaug____: nevermind, I was able to test.
18:50
<smaug____>
Domenic: back
18:51
<smaug____>
the outcome was?
18:51
<smaug____>
aha, I see the github email
18:53
<smaug____>
Domenic: did you test document.open case
18:53
<jsbell>
smaug____: Entry? change made
18:53
<Domenic>
smaug____: ah right good point
18:53
<smaug____>
jsbell: thanks
18:53
<smaug____>
Domenic: since IIRC Gecko supports replace in document.open
18:54
<Domenic>
smaug____: Chrome's bindings for document.open() literally do dynamic dispatch to window.open() O_o
18:54
<jsbell>
almost missed that; did I miss anything else? (also, bugs welcome!)
18:54
<Domenic>
to be fair that does seem to be what the spec implies
18:54
<smaug____>
we do something similar... reading some code
18:54
<Domenic>
smaug____: yeah looks like document.open()-as-window.open() does not support replace either http://software.hixie.ch/utilities/js/live-dom-viewer/?saved=4321
18:55
<smaug____>
yes, http://searchfox.org/mozilla-central/source/dom/html/nsHTMLDocument.cpp#1396
18:55
<Domenic>
Hmm that XXXbz means maybe we should ping bz
18:55
<smaug____>
Domenic: and see bz' comment about ignoring replace there
18:58
<smaug____>
Domenic: well, bz just added the comment in somewhat unrelated patch: https://hg.mozilla.org/mozilla-central/rev/f870d9cd5af1#l1.519
18:59
<smaug____>
he split the Open() method to two, one dealing with non window-opening case and other for window opening
18:59
<Domenic>
oh i see
19:02
<smaug____>
Domenic: and the non-window opening case certainly does care about replace param, but that is about different replace param. http://searchfox.org/mozilla-central/rev/5a3efff7dee9d3233f532126afdbcb82ffcc1ff4/content/html/document/src/nsHTMLDocument.cpp#1672-1677 http://searchfox.org/mozilla-central/source/dom/webidl/HTMLDocument.webidl#36,38
19:03
<Domenic>
yeah exactly
19:03
<Domenic>
nice and confusing
19:03
<Domenic>
i mostly just don't want to remove this thing if someone was planning on implementing it soon
19:04
smaug____
doesn't even know what replace does in window.open
19:05
<Domenic>
we broke the spec without realizing it
19:05
<Domenic>
but it was meant to cause window.open() to navigate the named window with replacement instead of normally
19:05
<Domenic>
so instead of fixing the spec i figure we should just remove it
19:56
<aleray_>
gsnedders, Hi
19:57
<aleray_>
thanks, I though there war a createParent option (orsomething like this) to address that
19:57
<aleray_>
*was
19:59
<aleray_>
In https://html5lib.readthedocs.io/en/latest/html5lib.html?highlight=parsefragment#html5lib.__init__.HTMLParser.parseFragment
19:59
<aleray_>
the "container" option
21:30
<jsbell>
TabAtkins: updated bikeshed, getting "Unknown Status 'ED' used." Any guesses what's bogus in my metadata?
21:31
<jsbell>
https://github.com/w3c/FileAPI/blob/gh-pages/index.bs for reference
21:32
<TabAtkins>
jsbell: Nothing's bogus, it's a semi-breaking update that'll flush out some minor fixes in some places. The W3C statuses are now limited to groups that I know are W3C groups.
21:33
jsbell
assumes https://github.com/tabatkins/bikeshed/commit/f02935307493078ab95b09236b27d2f471aebb44 ... ah
21:33
<jsbell>
TabAtkins: Group: webplatform in this case
21:33
<TabAtkins>
Yes, not known yet. Gimme a sec.
21:34
<jsbell>
Cool, found megaGroups = { ... }
21:34
<TabAtkins>
I need to make it friendlier, so if there's no name clash I go ahead and let you use it for an unknown group, but throw a warning asking you to let me know about it.
21:35
<TabAtkins>
Okay, update pushed.
21:35
<jsbell>
yay, works
21:36
<TabAtkins>
You'll also start getting priv-sec nagging. ^_^
21:36
<jsbell>
yep, ignoring it in FileAPI (not my spec...)