07:40
<jacobolus>
is http://www.w3.org/TR/2009/WD-SVG-Transforms-20090320/ actually implemented anywhere?
08:16
<hsivonen>
fi.msn.com is still in IE7 Standards mode in IE9...
08:25
<zcorpan_>
http://www.sitepoint.com/forums/showthread.php?p=4699335 shows it was a right move to make the "EN" part be ignored in the doctype...
08:33
<annevk>
hmm, Gecko still has globalStorage?
08:46
<zcorpan_>
seems so
08:47
<zcorpan_>
http://google.com/codesearch?q=globalStorage++lang:javascript&ct=rr&cs_r=lang:javascript
09:19
<annevk>
http://www.w3.org/2010/web-notifications/ -- regressed a bit MikeSmith! :)
09:23
<MikeSmith>
yeah, unfortunately
09:23
<MikeSmith>
I was told it was "hard to read"
09:44
<wirepair>
MikeSmith are you about?
09:44
<MikeSmith>
wirepair: yeah, here
09:44
<wirepair>
where do i submit typos to?
09:44
<wirepair>
or where should i, for the w3c html5 spec
09:45
<annevk>
w3.org/Bugs/Public/
09:45
<wirepair>
thanks
09:45
<annevk>
it's easiest from the WHATWG copy
09:45
<annevk>
fwiw
09:46
<wirepair>
o
09:46
<wirepair>
hehe
09:46
<MikeSmith>
wirepair: yeah, either of the ways annevk mentioned is preferred -- because they both cause a new bug to be entered into our bugzilla
09:46
<wirepair>
would i still submit to w3.org?
09:46
<wirepair>
or..?
09:47
<MikeSmith>
that submits it for you
09:47
<MikeSmith>
the embedded comment thing in the whatwg version of the spec does a call to bugzilla to create a new bug
09:48
<MikeSmith>
if you log in before you use the comment form, it'll also include your e-mail address in the comment
09:48
<wirepair>
severity = trivial yeah?
09:50
<wirepair>
alright submitted; "only" was repeated twice in the HTMLAllCollections under the namedItem algorithm
09:53
<MikeSmith>
ok
09:54
<MikeSmith>
so, am I doing something wrong or does Firefox not yet support <img src="foo.svg">
09:54
<MikeSmith>
that is, specifying an SVG file in the URL for an img element
09:55
<MikeSmith>
it works in Minefield but in FF4.0b6 it doesn't display
09:55
<wirepair>
is the server returning the proper content-type?
09:55
<wirepair>
i know chrome will fail to load
09:55
<MikeSmith>
in this case I am testing it locally
09:56
<wirepair>
ah.
09:57
<wirepair>
is there any place that i could look up the reasoning behind certain decisions for html 5? i'm particularly interested in the point of having document.open have a replace flag
09:57
<wirepair>
s/have/having
10:00
<MikeSmith>
somebody was working on putting together a Rationale page at the whatwg wiki
10:00
<wirepair>
cool, looking now
10:01
<wirepair>
not listed hehe ;/
10:01
<MikeSmith>
http://wiki.whatwg.org/wiki/Rationale
10:01
<MikeSmith>
the work of variable-san
10:01
<wirepair>
guess i could pour over the whatwg mailing list too
10:01
<MikeSmith>
wirepair: so if there's nothing there, the places to look are in the whatwg mailing-list archives and in commit messages
10:02
<MikeSmith>
or, just ask
10:02
<MikeSmith>
ask Hixie or ask others here if they recall the discussions about it
10:03
<wirepair>
will do
10:03
<wirepair>
thanks
10:03
<jgraham>
MikeSmith: I believe Gecko landed support for SVG in <img> very recently
10:04
<MikeSmith>
jgraham: ah, OK
10:04
<MikeSmith>
thanks
10:06
<hsivonen>
MikeSmith: SVG as <img> is pretty recent. Could well be post-beta6
10:06
<MikeSmith>
hsivonen: ok
10:06
<hsivonen>
oh. jgraham said so already
10:06
<hsivonen>
http://www.w3.org/2010/web-notifications/ doesn't finish loading in Opera
10:07
<hsivonen>
hmm. now it does. weird
10:07
<MikeSmith>
yeah -- I do find it's working in my Minefield, that doesn't seem to work as expected if I set different width and height for the image (but I can deal with that later)
10:08
<MikeSmith>
hsivonen: dunno why that would happe -- I'm not doing anything fancy on that page
10:08
<hsivonen>
probably some random oddity
10:08
<MikeSmith>
could be something dumb in my CSS too
10:12
<MikeSmith>
jgraham: so, I have the following ready for testing
10:12
<MikeSmith>
http://sideshowbarker.github.com/es5-spec/
10:13
<MikeSmith>
though not ready for widespread announcing yet
10:13
<MikeSmith>
because I need to check with TC39 chair and secretary about copyright and license
10:13
<jgraham>
MikeSmith: So just on a publicly logged irc channel? :)
10:14
<MikeSmith>
yeah, I know
10:14
<MikeSmith>
I can take it down if it turns out it's not in line with the license
10:14
<MikeSmith>
anyway, I think it is OK as far as the license on the ES5 spec itself
10:14
<jgraham>
MikeSmith: Looks nice. I am really interested in the annotation bits
10:15
<MikeSmith>
so I have not created any annotations yet
10:15
<MikeSmith>
no original ones at least
10:15
<jgraham>
I mean I am really intersted in the technology
10:15
<MikeSmith>
but what I have created is annos for the published ES5 errata
10:16
<MikeSmith>
the technology? it's just using XHR to suck in content from static files
10:16
<MikeSmith>
http://github.com/sideshowbarker/es5-spec/tree/gh-pages/erra/
10:16
<MikeSmith>
those are the files for the errata
10:16
<jgraham>
Yeah, I am looking at the source
10:17
<MikeSmith>
example is, if you go to http://sideshowbarker.github.com/es5-spec/#x7.1
10:17
<MikeSmith>
and mouse over the heading
10:17
<MikeSmith>
it will show # T E A
10:17
<MikeSmith>
and if you click on E, it will show the errata for that section
10:18
<MikeSmith>
…except that I somehow broke it in Gecko
10:18
<MikeSmith>
I had it working before in Minefield at least
10:18
<MikeSmith>
but now it just stays stuck at "Loading..."
10:19
<jgraham>
OK, so annotations are per section and have to be commited as static files?
10:19
<MikeSmith>
yes
10:19
<jgraham>
That is a little less interesting :)
10:19
<MikeSmith>
old school
10:19
<MikeSmith>
well, I didn't set out to make it interesting
10:19
<MikeSmith>
just useful
10:19
<jgraham>
Sure
10:20
<MikeSmith>
jgraham: if you have a github account I would be happy to give you write access so you can create some annos
10:20
<jgraham>
I just wondered how much I could <strike>steal</strike> reuse for annotating the HTML5 spec with testcases
10:21
<jgraham>
But per-section isn't fine-grained enough
10:21
<jgraham>
MikeSmith: (BTW the design is nice)
10:21
<MikeSmith>
thanks
10:21
<MikeSmith>
you can put markup within the section to indicate which part it is annotating
10:21
<jgraham>
Although it is a bit annoying that hovering a heading shifts the content down
10:22
<MikeSmith>
that is what the ES5 errata doc does
10:22
<MikeSmith>
it says stuff like, "In the last step of the algorithm, change foo to bar"
10:22
<MikeSmith>
or whatever
10:23
<jgraham>
Right, so how do you deal with that? Also can the default view of the document be with errata applied?
10:24
<MikeSmith>
applying the errata would violate the license
10:24
<MikeSmith>
afaict
10:24
<jgraham>
Sigh
10:24
<MikeSmith>
http://sideshowbarker.github.com/es5-spec/#license
10:24
<MikeSmith>
"this document itself may not be modified in any way"
10:25
<MikeSmith>
I hope they are not going to judge that an HTML version represents a modification of the document
10:25
<MikeSmith>
and the other thing is, they have no license statement at all in the errata document
10:25
<MikeSmith>
that's the main thing I need to get clarification on
12:39
<annevk>
ivan`, http://www.whatwg.org/demos/workers/shared/001/test.html works in Opera
12:40
<ivan`>
indeed, I had that simpler test working too
12:41
<annevk>
all the links seem to work
12:41
<annevk>
if you're not doing things over HTTP stuff might break
12:44
<ivan`>
all the demos in that /shared/ do work for me, thanks
12:45
<ivan`>
I haven't read the whole spec but I don't understand why onconnect's event has an array of ports, instead of just one port
12:46
<annevk>
oh sorry, Multiviewer is a bit down
12:46
<ivan`>
every example I see uses [0], and I assume the examples aren't broken, so it's always an array of one port?
12:46
<annevk>
I see now
12:46
<ivan`>
:)
12:46
<annevk>
there can be multiple ports afaik
12:49
<ivan`>
can a SharedWorker know when a port disconnects (because a tab/window was closed)?
12:49
<ivan`>
I found a bunch of mailing list posts about onclose but I don't see anything in the current spec
12:50
<ivan`>
there's a suggestion of using onunload but this doesn't work everywhere
12:50
<hsivonen>
wow. the poster thread just goes on and on
12:51
<annevk>
ivan`, I think best would be emailing the list
12:51
<ivan`>
thanks
13:19
<nessy>
hsivonen: you just gotta love that poster thread ;)
13:27
<hsivonen>
Philip`: my "tomorrow-ish" is now...
13:27
<hsivonen>
Philip`: http://hsivonen.iki.fi/htmlparser-statistics.jar
13:28
<hsivonen>
Philip`: the interface to implement is http://pastebin.mozilla.org/794419
13:28
<hsivonen>
Philip`: there's a setter on the HtmlParser, HtmlBuilder and HtmlDocumentBuilder
13:28
<hsivonen>
Philip`: it's called once per doc
13:29
<hsivonen>
Philip`: I recommend running the JVM with -XX:-DontCompileHugeMethods
13:30
<hsivonen>
Philip`: without that JVM switch, the perf drops to about one tenth compared to the situation with that switch
16:08
<jgraham>
The bug tracker needs a resolution Resolved: 2004 called and they want their bug back
16:10
<hsivonen>
jgraham: indeed.
17:30
<KrooniX>
Hello
17:30
<KrooniX>
What is the state of the target="_blank" in html5?
17:31
<zcorpan_>
it's valid
17:36
<bzed>
p News · Most Recent
17:38
<KrooniX>
is it recommended?
17:40
<bzed>
gah, sorry for that paste :)
17:41
<zcorpan_>
it's recommended over jumping through hoops with script to make a link open in a new window, but you should consider omitting target altogether and let the user decide how the link should be opened
17:43
<hober>
right. it's the best way to do what it does, but you probably shouldn't want to do what it does
17:43
<KrooniX>
so it is in other words, nothing in the specification, which discourages its use?
17:44
<zcorpan_>
right
17:44
<KrooniX>
it's not that I want to use it, I just wanted to use the specification as an argument for not using it.
17:47
<KrooniX>
thanks
20:28
<AryehGregor>
http://samy.pl/evercookie/
20:29
<AryehGregor>
This contains a way to use canvas to store and retrieve data after the user clears their cookies, via browser cache.
20:30
<AryehGregor>
It retrieves a URL that gives a PNG containing the data, and writes it to canvas, which it then uses to retrieve the data, I guess by reading pixels or whatever. If the request contains a cookie, it copies the cookie data into the returned PNG. Otherwise it just returns a 304 so that the client uses the cached copy, which contains the old cookie version.
20:30
<AryehGregor>
Fun.
20:31
<Philip`>
Why not just cache a text file and use XHR to retrieve it?
20:31
<espadrine>
Scary. It even uses web history!
20:32
<AryehGregor>
Philip`, because that wouldn't use cool HTML5 features!
20:33
<AryehGregor>
espadrine, well, that one's well-known, and fixed in Firefox 4. http://hacks.mozilla.org/2010/03/privacy-related-changes-coming-to-css-vistited/
20:33
<AryehGregor>
vistited :(
20:34
<AryehGregor>
So, I wonder whether browsers could adapt to prevent this attack.
20:34
<AryehGregor>
Is there any simple change that would prevent it?
20:34
<AryehGregor>
Of course, if you can run JS, you can just store the cookies server-side and use fingerprinting, which is extremely reliable if you do it right . . .
20:38
<AryehGregor>
https://panopticlick.eff.org/index.php?action=log&js=yes
20:39
<AryehGregor>
One in 147485.63 browsers has my UA string, and I'm the only one tested so far with my exact list of plugins.
20:39
<Philip`>
Seems the more interesting problem is identifying users, not browsers
20:39
<AryehGregor>
Sure, but the latter is a considerable step toward the former.
20:40
<Philip`>
(I saw some work from Microsoft about trying to identify multiple users sharing a computer users using multiple computers, based on analysing login names and times and things)
20:41
<AryehGregor>
Sites wouldn't have access to login names, though.
20:42
<Philip`>
Some sites (e.g. lots of Microsoft ones) do
20:43
<AryehGregor>
You mean they have access to the name that you use to log into that same site?
20:43
<AryehGregor>
But that's already a unique identifier, why do they care if they have that? To identify duplicate accounts?
20:44
<Philip`>
It's a unique identifier of login names, it's not a unique identifier of users
20:44
<Philip`>
It's interesting when multiple users are logging in with the same account (e.g. one might have stolen the other's password), or when one user is logging in with many accounts (e.g. they're sending spam)
22:20
<MikeSmith>
Opera seems to be ignoring my z-index setting
22:20
<MikeSmith>
http://sideshowbarker.github.com/es5-spec/#x7.1
22:20
<MikeSmith>
mouse over the head and click on the Ⓔ
22:20
<annevk>
we think it's wrong
22:20
<MikeSmith>
oh
22:21
<MikeSmith>
how do I work around it?
22:21
<annevk>
sorry, I was just joking
22:21
<annevk>
time to read a book :)
22:22
<MikeSmith>
heh
22:23
<zcorpan_>
MikeSmith: what's the issue?
22:24
<MikeSmith>
in the box that pops up, look at the left side
22:24
<MikeSmith>
at the little curley thing that leads from the box to the globe
22:24
<MikeSmith>
that curley thing is an SVg image
22:25
<MikeSmith>
that i put z-index=10 on
22:25
<zcorpan_>
it doesn't overlap the border of the other box?
22:25
<MikeSmith>
it does overlap
22:25
<MikeSmith>
but it has a higher z-index
22:25
<zcorpan_>
i mean in opera
22:25
<AryehGregor>
Did Opera remove its UI for form validation? data:text/html,<!doctype html><form action=http://google.com><input required> <input type=submit></form>
22:25
<AryehGregor>
Clicking sends me to Google.
22:26
<MikeSmith>
zcorpan_ the border shows in Opera, but not in other browsers
22:26
<zcorpan_>
AryehGregor: you need name=something for the input to be validated
22:26
<AryehGregor>
Oh.
22:26
<AryehGregor>
Then that's a Firefox bug. :)
22:26
<zcorpan_>
MikeSmith: right
22:26
<MikeSmith>
maybe I need to set an explicit z-index on the box as well
22:27
<zcorpan_>
AryehGregor: ...but the spec changed so that name is no longer needed
22:27
<AryehGregor>
Heh, oh well.
22:28
<zcorpan_>
MikeSmith: is the problem that opera doesn't apply z-index on svg elements?
22:28
<jgraham>
zcorpan_: Do we have a bug report on that?
22:28
<zcorpan_>
MikeSmith: if so maybe you could set z-index on a parent
22:28
<zcorpan_>
jgraham: dunno
22:29
<MikeSmith>
zcorpan_, I don't know -- this is the first time I've tried to do much with SVG
22:29
<jgraham>
zcorpan_: Oh. Well I guess I can try my luck with Jira
22:29
<zcorpan_>
jgraham: or ask emil
22:30
<jgraham>
zcorpan_: The point was more to tell emil
22:30
<jgraham>
Although asking might have the right effect :)
22:33
<zcorpan_>
hmm. there haven't been any emails on hybi since 14th
22:33
<zcorpan_>
what's wrong?
22:35
<jgraham>
Ian hasn't posted anything since his draft on the 1st (I think that is the right date)
22:35
<jgraham>
I assume that everyone realised that discussing v2 features was pointless
22:35
<jgraham>
Since we don't have a v1 yet
22:35
<jgraham>
other than -76
22:36
<zcorpan_>
i wonder if fette is making any progress on -02
22:36
<zcorpan_>
maybe i should ask on the list
22:36
<jgraham>
That might be a good idea
22:37
<jgraham>
Really we needed v1 done a month ago
22:37
<jgraham>
Even that might have been too late
22:39
<AryehGregor>
Where's that usability study that concluded it was best to report input validity errors onchange rather than onsubmit, and not to report them on inputs the user hasn't changed?
22:41
<AryehGregor>
Found it: http://www.alistapart.com/articles/inline-validation-in-web-forms/
22:41
<AryehGregor>
<3 Google
22:43
<zcorpan_>
:invalid:dirty:not(:focus)
22:46
<zcorpan_>
or maybe just :invalid:dirty where :dirty means that the element has received and lost focus at least once
22:46
<AryehGregor>
Tab and I talked about :dirty once.
22:46
<AryehGregor>
I don't think we ever had a real proposal going, but it would be really useful for this.
22:47
<AryehGregor>
(Does :dirty get set if the element is changed by script? Can script control the dirty bit directly, perhaps?)
22:54
<AryehGregor>
Hmm, I see.
22:54
<AryehGregor>
I wonder if the meaning of this is specced if you look closely: data:text/html,<!doctype html><form><input required oninvalid="this.setCustomValidity(''); if (!this.validity.valid) this.setCustomValidity('abcd')"><input type=submit></form>?
22:54
<AryehGregor>
It seems to work as I expect in Opera but not Firefox.
22:54
<AryehGregor>
If you try submitting, then enter something in the first field, then submitting, it should submit, right?
22:55
<AryehGregor>
Because it will have no errors left.
22:55
<zcorpan_>
did you include name=foo when testing opera?
22:55
<AryehGregor>
Yes.
22:56
<AryehGregor>
data:text/html,<!doctype html><form><input name=x required oninvalid="this.setCustomValidity(''); if (!this.validity.valid) this.setCustomValidity('abcd')"> <input type=submit></form>
22:58
<zcorpan_>
i've never learned the forms api properly. what does setCustomValidity do?
22:59
<AryehGregor>
It sets a custom validity message. An element with a nonempty custom validity message is invalid, and should display the given message.
23:00
<AryehGregor>
In some undefined fashion.
23:00
<AryehGregor>
I think Firefox displays it much better than Opera.
23:03
<zcorpan_>
so when it's submitted the first time, the script sets setCustomValidity to 'abcd'. then if you type in the field, is setCustomValidity reset?
23:04
<AryehGregor>
The custom validity thing is totally custom, only setCustomValidity() ever changes it.
23:04
<AryehGregor>
So my idea was that when you try submitting, first it clears the custom invalidity, then if it's invalid, it sets the message to override the UA default message.
23:05
<AryehGregor>
With tweaks so that it actually works in Firefox, it works really well, because Firefox just replaces the whole error message with the provided one.
23:05
<AryehGregor>
In Opera it's kind of terrible, but then so is everything about Opera's form validation UI.
23:06
<AryehGregor>
I submitted a bug months ago that pointed out that <input type=password pattern=....*> to require passwords at least 3 characters long will print out the password in cleartext if there's an error.
23:06
<AryehGregor>
Seemingly still not fixed in 10.60.
23:06
<AryehGregor>
DSK-262266.
23:06
<AryehGregor>
I reported it in 10.00.
23:08
<AryehGregor>
I guess if Firefox 4's UI works well, we can enable it somehow for everyone except WebKit and Opera. But that will be a pain, and require JavaScript. Oh well.
23:10
<miketaylr>
so only for FF4?
23:11
<AryehGregor>
Well, we could blacklist only specific WebKit versions.
23:11
<AryehGregor>
Then future WebKit or IE could Just Work.
23:11
<AryehGregor>
Or Unknown Browser X, whatever.
23:14
<zcorpan_>
MikeSmith: the es5 doc is awesome. one annoyance though is that the text moves when you hover a heading -- maybe use position:absolute or something on the annotation links?
23:15
<MikeSmith>
zcorpan_: yeah, I need to fix that somehow
23:15
<MikeSmith>
but note that it only happens in Opera
23:15
<MikeSmith>
I think I need to adjust the line-height of the headings
23:16
<zcorpan_>
the text jumps in firefox too
23:16
<MikeSmith>
the x height of those circled letters is apparently bigger than normal glyphs or something
23:16
<MikeSmith>
ah, OK
23:16
<MikeSmith>
hmm, doesn't jump in minefield at least
23:16
<MikeSmith>
but anyway I will fix it one way or the other
23:17
<MikeSmith>
more serious problem I am having right now is that I can't get text into those anno boxes in Gecko, because readystate on my XHR never changes away from 1
23:17
<MikeSmith>
can't figure out why
23:19
<jamesr>
was 'clear' supported at one point for canvas 2d's globalCompositeOperation?
23:19
<jamesr>
it appears that gecko and webkit currently accept it as a valid value
23:26
<Philip`>
jamesr: Not in the spec
23:26
<jamesr>
Philip`: was there any particular reason not to add it?
23:26
<Philip`>
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-June/012063.html has the information I'm aware of
23:26
<Philip`>
jamesr: The default state for features is to not add them, and there wasn't any particular reason to add it
23:27
<jamesr>
but it's observable
23:27
<jamesr>
i.e. if bullet point 2 is true, then they should be added
23:27
<Philip`>
It's an observable spec-compliance bug in browsers that implement it :-)
23:28
<Philip`>
Is bullet point 2 true?
23:28
<jamesr>
dunno
23:28
<Philip`>
As far as I'm aware, Opera hasn't had to add it for compatibility
23:28
<Philip`>
which suggests it's not used much on the web
23:28
<zcorpan_>
MikeSmith: seems to wfm
23:29
<MikeSmith>
zcorpan_: really? in Firefox or in Minefield?
23:29
<MikeSmith>
it doesn't stay stuck at "loading..."?
23:29
<zcorpan_>
MikeSmith: minefield... although only when i have the web console open...
23:29
<zcorpan_>
which is helpful for debugging it...
23:29
<MikeSmith>
heh
23:30
<MikeSmith>
I'm testing in Minefield with console open but not seeing what I expect to see
23:30
MikeSmith
tries FF4 beta
23:32
<jgraham>
(usually you get "only works with console open" because you left a console.log in. Which you probably know already
23:32
<jgraham>
)
23:33
<zcorpan_>
we need to standardize console.log
23:34
<miketaylr>
a standardized console API would be great
23:36
<MikeSmith>
I think Web Inspector already supports the Firebug console API
23:36
<MikeSmith>
Dragonfly does too?
23:37
<miketaylr>
possibly, but the webkit has added some newer stuff that firebug doesn't support
23:38
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/644
23:39
<zcorpan_>
in particular i think browsers should expose console even when the console isn't open, and we need to agree on which members to have on console
23:40
<zcorpan_>
otherwise scripts around the web will break in some browsers because the author left in a console.something
23:40
<zcorpan_>
s/will//
23:40
miketaylr
has done that before >_>
23:42
<zcorpan_>
can someone file a bug on html5? i really need to go to bed now :)
23:42
<zcorpan_>
nn
23:45
<othermaciej>
MikeSmith: does w3c bugzilla use Bcc for Cc'd bugs?
23:45
<MikeSmith>
othermaciej: no, it uses a normal Cc afaict
23:46
<othermaciej>
MikeSmith: I'm getting new bugmail for all HTML5 bugs but I don't see public-html in the recipient list on the email and I don't think I am on the other lists, so I am unsure of what is happening
23:47
<MikeSmith>
othermaciej: because you're subscribed to the issue-tracking list, I think
23:47
<MikeSmith>
I'll turn off those
23:47
<othermaciej>
ah, I guess I just forgot about subscribing and have been bitbucketing the emails
23:48
<othermaciej>
I do want to be getting them, though via public-html so my mail filter catches them
23:48
<MikeSmith>
you aren't manually subscribed to that list -- its subscriber list is populated out of the user DB
23:48
<othermaciej>
I see
23:49
<othermaciej>
so everyone in the WG is on the public-html-issue-tracking list?
23:49
<othermaciej>
er, public-html-wg-issue-tracking
23:50
<MikeSmith>
othermaciej: no, just some people.. I can't remember when we set up the group in the DB for that, but you are in it
23:50
<MikeSmith>
we have maybe 20 people or so in that group
23:50
<othermaciej>
I see
23:50
<MikeSmith>
othermaciej: so, I can change right now to having notifications for new bugs go to public-html, as you proposed. Do you want me to do that now?
23:50
<othermaciej>
looking at the archives it looks like it gets about the same set of mail that I suggested public-html should be getting
23:50
<othermaciej>
MikeSmith: I would appreciate it - I think the other chairs agree that we should at least try it out that way
23:50
<MikeSmith>
PhilipJ seemed to be OK with new bugs notifications going to public-html
23:51
<MikeSmith>
OK
23:51
<MikeSmith>
I will flip the switch on it now
23:51
<othermaciej>
thanks!
23:51
<othermaciej>
I guess we can probably obsolete the separate issue tracking list then
23:51
<othermaciej>
I'm off to a meeting, ttyl
23:53
<jamesr>
Philip`: there are a _lot_ of failures in this suite in WebKit and Gecko ToT :(
23:57
<jamesr>
Philip`: where is the rounding/wrapping behavior for canvaspixelarray defined? it's tested by 2d.imageData.object.round/wrap