09:23
<zcorpan>
annevk-cloud: "If url's relative flag is set, set encoding override to utf-8." http://url.spec.whatwg.org/#query-state
09:23
<zcorpan>
annevk-cloud: isn't that backwards? or am i misreading it?
09:24
<zcorpan>
annevk-cloud: i read that encoding override gets set to utf-8 for urls with relative schemes
09:28
<zcorpan>
annevk-cloud: also, should ws/wss not always use utf-8? not that they're useful in <a href> but still
09:31
<SteveF>
arunranga: FWIW moves afoot to have editors drafts at /TR http://lists.w3.org/Archives/Public/spec-prod/2014JanMar/thread.html
10:49
<annevk>
Is WHATWG down?
10:50
<annevk>
Hmm just slow
11:04
<annevk>
Was there no Gecko bug for https://www.w3.org/Bugs/Public/show_bug.cgi?id=24005 ?
11:09
<zcorpan_>
annevk: added tests https://critic.hoppipolla.co.uk/showcommit?first=c94f486e&last=70f343af&review=437
11:10
<wefo>
Hmm.
11:10
<wefo>
Can somebody please confirm with me that line 1 loads the font "whenever it feels like it", and line 2 loads the font "right now, before loading anything else"? http://pastebin.com/raw.php?i=7m8LnWeK
11:10
<annevk>
zcorpan_: I can do the utf-8 thing for ws/wss
11:11
<wefo>
That is, somebody who knows this stuff.
11:11
<annevk>
zcorpan_: your bug report seems accurate
11:11
<zcorpan_>
annevk: ok
11:12
<annevk>
wefo: the second line uses an invalid data URL; it won't get you anywhere
11:12
<wefo>
Yeah, I just noticed.
11:12
<wefo>
Because it breaks on my computer.
11:12
<wefo>
annevk: I was told to use the latter.
11:12
<annevk>
wefo: there's no guarantees on when a data URL is interpreted and applied to a page with respect to fonts at the moment
11:12
<wefo>
Because it supposedly works "block-mode".
11:12
<wefo>
So then that guy was lying.
11:13
<wefo>
I will have to do the insane hack after all.
11:13
<annevk>
wefo: well if you include the entire font inline it's reasonable to assume it'll apply faster
11:13
<wefo>
Include the entire font inline?
11:13
<annevk>
wefo: since the font will download together with the style sheet, but you need to know how to use data URLs
11:13
<wefo>
Oh... is that the direct data stuff in base64 or something?
11:14
<wefo>
I was going to use this insane hack to determine if the user indeed has the font loaded: http://www.kirupa.com/html5/detect_whether_font_is_installed.htm
11:15
<wefo>
Frankly, it's kind of ingenious.
11:15
<wefo>
I like it when things like that "work", even if they are not ideal.
11:23
<jgraham_>
annevk: I nominate you to do that code review btw
11:24
<annevk>
jgraham: zcorpan_: done
11:25
<annevk>
Oh I see there is a bunch more...
11:25
<jgraham>
Heh
11:25
<annevk>
jgraham: so this move: https://critic.hoppipolla.co.uk/57951ecc?review=437
11:25
<annevk>
jgraham: that should just be fine no?
11:26
<annevk>
jgraham: you already reviewed a bunch
11:27
<jgraham>
annevk: I reviewed all the trivial bits
11:27
<jgraham>
Basically all the hard stuff is the .js file
11:28
<annevk>
Can you give me a review link for that file?
11:29
<annevk>
I have the feeling I keep looking at individual commits
11:30
<jgraham>
https://critic.hoppipolla.co.uk/showcommit?review=437&filter=pending
11:31
<zcorpan_>
annevk: you can drag-and-drop in the commit list to get a squashed review
11:31
<jgraham>
Yeah, or use the filter links just above it
11:34
<annevk>
Ah great
11:34
<annevk>
So this looks okay to me... I take it you have run it all?
11:35
<annevk>
My only nit is using a string rather than array...
11:36
<annevk>
I guess you didn't want to type all the quotes or something?
11:37
<zcorpan_>
yeah
11:39
<zcorpan_>
i've run it, yes. but some browser bugs cause some tests to time out, so it's not awesome
11:39
<zcorpan_>
also search for XXX
11:41
<jgraham>
zcorpan_: If some tests time out that's probably OK as long as the other tests still run
11:41
<zcorpan_>
ok, yeah they do
11:42
<jgraham>
I can live with that. We can always improve it in followup commits if there's something that causes problems
11:53
jgraham
assumes annevk didn't *really* care about the array thing
11:54
<jgraham>
zcorpan_: Worth squashing before pushing or should I just merge?
11:54
<annevk>
I'd like to hear some reasoning
11:55
<annevk>
But I'm not going to block on it :-)
11:55
<zcorpan_>
i commented in critic
11:55
<zcorpan_>
jgraham: no opinion
11:56
<jgraham>
Let's just merge
11:57
<jgraham>
Done
11:57
<jgraham>
zcorpan_, annevk: Thanks
11:57
<zcorpan_>
thank you!
12:18
<MikeSmith>
annevk: http://krijnhoetmer.nl/irc-logs/whatwg/20140115#l-803
12:22
<hsivonen>
what make people post to Usenet using DOS encodings?
12:27
<hsivonen>
*makes
12:29
<MikeSmith>
fashion statement
12:30
<icaaq>
hi, here they use the label element http://filamentgroup.com/lab/bulletproof_icon_fonts/ to "label" som text. is that really semantically correct?
12:34
<hsivonen>
"Windows codepage 1252, the codepage commonly used for English and other Western European languages, was based on an American National Standards Institute (ANSI) draft. That draft eventually became ISO 8859-1, but Windows codepage 1252 was implemented before the standard became final, and is not exactly the same as ISO 8859-1."
12:34
<hsivonen>
http://msdn.microsoft.com/en-us/library/ee719641.aspx
12:34
<hsivonen>
so windows-1252 even existed first!
12:34
<Ms2ger>
Sounds like the box model in old-IE
13:10
<annevk>
hsivonen: man, the more I hear about ISO and encodings, ...
13:11
<annevk>
MikeSmith: interesting suggestion :-)
13:29
<annevk>
MikeSmith: that doesn't quite work for suggesting what it is
13:31
<MikeSmith>
annevk: maybe it's not necessary to suggest what it is. In the end it's just an opaque string, right?
13:32
<annevk>
It's an opaque string, most typically used as part of the DNS or as a local network address.
13:32
<annevk>
Maybe something like that could work, if I can find good references for both.
13:45
<zcorpan>
Hixie: http://www.w3.org/html/wg/drafts/html/master/embedded-content-0.html#text-tracks-exposing-in-band-metadata doesn't seem to exist in whatwg html
13:47
<zcorpan>
Hixie: what's the deal?
13:48
<MikeSmith>
annevk: A domain label is either an IN-ADDR.ARPA suffix or an opaque string that represents one octet of an Internet address.
13:49
<MikeSmith>
hmm what does that even mean
13:49
<Ms2ger`>
How about IPv6?
13:50
<annevk>
Ms2ger`: well we know that's not a domain label :-)
13:51
<zcorpan>
a domain label is an extremity of Jesus
13:51
<MikeSmith>
heh
13:51
<annevk>
"The text attribute, on getting, must return UTF-16 text converted from data of the text track cue that the TextTrackCue object represents." more like what's the deal with that text
13:52
<Ms2ger`>
No ip6.arpa?
14:12
<zcorpan>
w3c-test.org doesn't use wpt-serve yet?
14:13
<jgraham>
MikeSmith was working on it
14:14
<MikeSmith>
still "working" on it yeah
14:15
<MikeSmith>
meaning, procrastinating
14:16
<MikeSmith>
if only the CSS WG didn't exist, things would all be much simpler
14:35
<zcorpan>
not sure how i got here but it was a bit funny http://9gag.com/gag/agyOxmx
14:39
<wefo>
zcorpan: "These are Satan. Bears."
14:41
<annevk>
GitHub down?
14:52
<mathiasbynens>
annevk: WFM
15:12
<annevk>
mathiasbynens: ta
16:01
<annevk>
IBM emailed the WHATWG list. Can someone confirm hell froze over?
16:02
<jgraham>
Can't verify, too busy dodging flying pigs
16:12
GPHemsley
is in cold Georgia. Does that count?
16:39
<wefo>
Wait... there is a textAlign property to contexts? So all those hours of measuring the width and doing complex math calculations to get it centered were wasted?
16:39
<wefo>
Is this some kind of bleeding edge feature? I hate never being able to tell. There is no clear documentation whatsoever.
16:57
<dglazkov>
good morning, Whatwg!
17:06
<Hixie>
zcorpan: DataCue doesn't make sense. If you don't know the format, how can you expose it? If you do know the format, then use the interface that that format provides.
17:32
<MikeSmith>
CORS is Recommendation
17:32
<MikeSmith>
now all the confused deputy attacks can begin in earnest
17:32
Ms2ger`
confuses MikeSmith
17:33
<MikeSmith>
...
17:33
<Hixie>
isn't CORS obsolete? i thought Fetch replaced it a while back.
17:34
<MikeSmith>
Hixie: well yeah
17:35
<MikeSmith>
it has joined the ranks of the undead
17:35
<Hixie>
wait is that the one anne was telling me about where they decided to just ignore all the open bugs?
17:36
<MikeSmith>
bingo
17:36
<MikeSmith>
they "resolved" them though
17:36
<MikeSmith>
which is technically different from ignoring them
17:37
<Hixie>
iirc they ignored them for 2 years then resolved them as not relevant since nobody had looked at them in two years
17:37
<MikeSmith>
that would be one way to describe it, yes
17:37
<MikeSmith>
the accurate way
17:37
<Hixie>
which is some bug system ninja juju that i have to say is pretty impressive
17:38
<MikeSmith>
hey this is what we WGs for man
17:38
<MikeSmith>
that kind of magic
17:39
<jgraham>
Wait what? MikeSmith is a deputy?
17:39
<Hixie>
i wish there were journalists who cared about this kind of crap enough to report on it
17:39
<jgraham>
So who's the sheriff in this town?
17:40
<jgraham>
Isn't that a trick they learnt from the chrome bugtracker?
17:40
<Hixie>
i mean, this is just as messed up as some political shenanigans, and has similar long-term effects (which is to say, mostly none)
17:40
<jgraham>
Well it's mildly more interesting than who the president of France is sleeping with, but I'm not sure why that's being reported either
17:42
<MikeSmith>
yeah it's like politics but robbed of all the debauchery
17:42
<MikeSmith>
which just makes it depressing instead of scandalous
17:42
<jgraham>
Anyway, if EME has taught you nothing else it should have been that getting press attention only leads to a bunch of people with extremely shallow understanding of the issues getting rilled up in an entirely impotent way
17:42
<MikeSmith>
jgraham: but, freedom!
17:44
<Hixie>
jgraham: at least it'd make me feel like i wasn't alone in thinking it was ridiculous :-)
17:44
<Domenic_>
"Congratulations Anne on today's publication of the CORS Recommendation!" hmmmm
17:44
<MikeSmith>
jgraham seems to not understand about the Wisdom of the Crowds
17:46
<jgraham>
I only think it's ridiculous if you are measuing against The Process. If you think a Process where things go to Rec. irrespective of bugs, and then are worked on further, is better then CORS is doing all the right things. Now W3C probably don't realise that, but perhaps they will cotton on one of these decades
17:46
<jgraham>
(actually closing the bugs is nonsense of course)
17:48
<arunranga>
I think we should add text that refers to CORS as a snapshot subset of Fetch.
17:48
<Hixie>
jgraham: what i find ridiculous is the hypocrisy of claiming to care about a process and then _blatently_ not caring about it
17:50
<MikeSmith>
well they are standing on the shoulders of some giants, as far as that pattern goes
17:51
<MikeSmith>
well maybe not giants
17:52
<Hixie>
i really don't understand why anyone would claim to believe in one set of principles, but act in a way contrary to them. i really don't see the point. why lie about your values? or if it's not a lie, why not do a good job of actually meeting them?
17:52
<Hixie>
i just don't get it
17:52
<jgraham>
Because they don't see it as a moral problem
17:53
<jgraham>
It's a simple path-finding exercise. "I need to achieve X. Constraints Y exist. How can I get to X as fast as possible given Y"
17:54
<Hixie>
ok but then why claim that one is trying to achieve Z?
17:54
<jgraham>
The only confusion is the fact that the *stated* Y and the reality of Y are quite different
17:54
<Hixie>
right, that's what i'm saying
17:54
<Hixie>
why not state the truth?
17:54
<jgraham>
Well see it from the point of view of the W3C
17:55
<Hixie>
i don't understand the point of view of the W3C, that's my problem
17:56
<Domenic_>
It seems like there are pretty obvious gains, in general, from stating a position of one thing and then doing another thing
17:56
<Domenic_>
That shouldn't be hard to understand
17:56
<jgraham>
If they rigrously enforced all the stated Y, it would take forever to get anything done and people would do work elsewhere. But part of Y is that changing the stated Y is very difficult, so instead they make up Y' which is Y with a whole load of concessions to the people actually trying to do X"
17:56
<Hixie>
Domenic_: what are the gains here?
17:57
<Hixie>
jgraham: nothing is "very difficult"
18:07
<jgraham>
I mean the W3C can only really cause limited damage. It can annoy you and cause the rest of us to have to point everyone we know to whatwg specs
18:08
<Hixie>
i'm not making any judgements as to absolute objective harm :-)
18:09
<jgraham>
OK, well I guess it's pretty normal that your personal sense of disappointment isn't correlated with objective harm
18:10
<Hixie>
btw, i would appreciate a review of https://www.w3.org/Bugs/Public/show_bug.cgi?id=23475#c16 from people here
18:11
<Hixie>
it's a proposal for revamping how focus works
18:11
<Hixie>
(new proposal from yesterday with more readable terms and more web compatability)
18:12
<Hixie>
it handles, amongst other things, scrollable regions (long ignored by specs) and <dialog>s (new)
18:24
<Hixie>
on a different note... i wonder where to spec that the 'select' event fires sometimes
18:26
<Hixie>
(in response to user interaction, i mean)
18:28
<annevk>
I have another REC...
18:29
<annevk>
css3-namespace and cors
18:29
<annevk>
And I'm not a big fan of either
18:35
<Hixie>
annevk: http://w3cmemes.tumblr.com/post/73529051245 :-P
18:36
<annevk>
Hixie: no wonder I act out like in http://w3cmemes.tumblr.com/post/73519371870/anne-van-kesteren-ladies-and-gentlemen-csswgs
18:38
<Hixie>
annevk: :-P
18:42
<aklein>
annevk: hi there, still about?
18:42
<annevk>
aklein: if it's quick
18:42
<annevk>
aklein: I'll be in SF starting Monday 4/5PM
18:43
<annevk>
aklein: until the 31st
18:43
<aklein>
annevk: another time, then, I was just curious if you think there's anything else to be done about ShadowRoot.baseURI without isolation
18:43
<aklein>
annevk: ah, cool, we should meet up at some point
18:43
<annevk>
aklein: after you brought up your point I wasn't so sure anymore what the point would be
18:44
<aklein>
annevk: yeah, that was my thought, just wondering if there were things I hadn't thought of
18:44
<annevk>
aklein: in the declarative scenario it might work better
18:44
<aklein>
the good news is that without <element> being specced, folks have to do something to set up their ShadowRoot anyway so they can fix their URLs by hand for now
18:44
<aklein>
sounds like we're on the same page
18:44
<annevk>
aklein: right
18:45
<annevk>
aklein: I'm still interested in defining baseURI though, potentially removing most of the cruft for now then
18:46
<aklein>
annevk: indeed, the current HTML definition of baseURLs is...not good
18:47
<aklein>
especially since Gecko actually fully supports xml:base
18:47
<aklein>
anyway, I'll let you go for now and catch up later
18:48
<Hixie>
wait what?
18:48
<Hixie>
what's wrong with the HTML definition?
18:48
<Hixie>
HTML supports xml:base
18:48
<annevk>
We don't want to support xml:base :-)
18:49
<aklein>
Hixie: neither Blink nor WebKit support xml:base in HTML documents
18:49
<Hixie>
i'm not following the problem here
18:49
<aklein>
or rather, they appear to support it (Node.baseURI is affected by it) but don't (URL completion/loading is unaffected)
18:50
<annevk>
Hixie: it's not really a problem, it's an opportunity to massively simplify base URL handling
18:50
<Hixie>
ah so by "the current HTML definition of baseURLs is...not good" you mean you want to change it, not that it's buggy
18:50
<aklein>
and avoid tree-walking to do URL handling
18:50
<Hixie>
ok
18:50
<Ms2ger`>
That would be nice
18:50
<Hixie>
i'm fine with dropping xml:base, though xml:base does make web components work better
18:50
<aklein>
Hixie: sorry, there are bugs, xml:base is not the bug
18:51
<Hixie>
we'll need something if we don't have xml:base
18:51
<Hixie>
oh
18:51
<Hixie>
what are the bugs?
18:51
<aklein>
Hixie: I've been meaning to file one about <img>
18:51
<Hixie>
Ms2ger`: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=24238
18:51
<aklein>
it reloads when affected by a base URL change
18:51
<Ms2ger`>
Oh dear
18:51
<Hixie>
aklein: it does? wow
18:52
<aklein>
Hixie: in Webkit/Blink/Gecko at least. really need to try some IEs.
18:52
<Hixie>
aklein: i intentionally made the spec not reload images in that case
18:52
<Hixie>
aklein: it would cause all kinds of hassle with pushState(), e.g.
18:52
<Ms2ger`>
Hixie, no strong opinion here
18:52
<Hixie>
Ms2ger`: should i wontfix then?
18:52
<Ms2ger`>
Hixie, if you prefer not-resolving, that's fine with me
18:52
<aklein>
Hixie: only sometimes
18:52
<aklein>
Hixie: when <img> switches documents it reloads
18:52
Ms2ger`
isn't sure ping will happen anyway
18:52
<aklein>
not necessarily on pushState
18:52
<Hixie>
aklein: oh when it switches documents, ok. that's easy to believe.
18:53
<Hixie>
aklein: that's not when changing base URL, that's when changing documents.
18:53
<aklein>
Hixie: speaking of web components, agreed that it would be nice to have something. that's what I was talking to annevk about above, see https://www.w3.org/Bugs/Public/show_bug.cgi?id=20976#c24
18:53
<aklein>
is there some changing documents hook in HTML I've missed?
18:54
<aklein>
Hixie: err, https://www.w3.org/Bugs/Public/show_bug.cgi?id=20976#c20 is probably a better starting point, that was esprehn's last suggestion
18:55
<Hixie>
aklein: the html spec doesn't currently reload images crossing documents, no. please do file that http://whatwg.org/newbug
18:57
<Hixie>
aklein: the scripts in these comments are expected to be running where? in the context of a component?
18:58
<Hixie>
aklein: i guess web components' documents aren't "live" documents, so they don't run script?
18:58
<Hixie>
aklein: what's the global scope of a script in a component?
19:03
<aklein>
Hixie: when you say "component" I think you may be combining two concepts, "imports" and "custom elements"
19:03
<aklein>
I realize that those were historically more intertwined :)
19:03
<Hixie>
i've no idea what either of those are, but ok
19:03
<Hixie>
i just mean whatever it is that you use to create proprietary widgets using DOM and JS
19:03
<aklein>
I'd say that in your terms, the script in a component runs in the global scope of the hosting page
19:03
<Hixie>
wow
19:03
<Hixie>
ok
19:03
<Hixie>
yeah, that'll give you all kinds of issues
19:04
<aklein>
indeed
19:04
<Hixie>
why not the xbl approach of running the scripts in the binding document?
19:04
<aklein>
I'd have to go look up xbl to answer that question well, dglazkov might have a better answer
19:04
<aklein>
the short answer is there is no other global at present
19:04
<aklein>
that is, none of the current web components specs involve additional global scopes
19:04
<Hixie>
right, but why not?
19:05
<aklein>
though they do involve additional documents
19:05
<Hixie>
seems like you'd want a global scope per resource defining widgets
19:06
<aklein>
Hixie: there's a public-webapps thread about this somewhere
19:06
<aklein>
http://lists.w3.org/Archives/Public/public-webapps/2013OctDec/0483.html
19:06
<aklein>
not sure there are actual answers there
19:06
<aklein>
but there's discussion
19:07
<Hixie>
well that e-mail pretty much summarises my concern, yes :-)
19:08
<Hixie>
looks like dglazkov's answer is that you can have a new scope using modules, or something
19:08
<Hixie>
so that it works even within just one doc
19:08
<Hixie>
which makes sense i guess
19:08
<Hixie>
but in that case, the problem you had earlier is easy
19:09
<Hixie>
you just make the modules be where you put the base URL scopes
19:09
<Hixie>
(presumably modules hook into the script settings object mechanism)
19:20
<Hixie>
man, i hate it when headless computers start acting up and need rebooting
19:20
<Hixie>
such a pain
21:03
<TabAtkins>
annevk-cloud: What are your thoughts on fetch()? Do you have sufficiently organized thoughts to write something down I can put into a spec? Maybe talk about this next week when you're here for ServiceWorker stuff?
21:42
<Hixie>
so... why does inserting an audio element into another document do something to whether the audio is playing or not?
21:45
<jgraham>
A better question might be "why do we have a model where moving things between documents isn't just a deep clone?"
21:45
<jgraham>
To which the answer is probably "legacy"
21:45
<jgraham>
So I think my answer to your question is "see /topic"
21:47
<Hixie>
why would moving things between docs be anything interesting at all?
21:47
<Hixie>
i don't understand what's going on here
21:48
<Hixie>
however, firefox's inability to play my test audio isn't helping
21:48
<Hixie>
what formats does firefox support?
21:49
<Hixie>
i've tried ogg, wave, and mp3
21:50
<Hixie>
oh ffs, it's mime types
21:51
<Hixie>
i hate mime types so much
21:51
<jgraham>
Hixie: Moving elements between documents is quite problematic in general because of the prototype chain
21:51
<Hixie>
so pointless
21:51
<Hixie>
jgraham: yeah the problem there is that we have this crazy idea of mutable prototypes and per-page prototypes
21:51
<Hixie>
ok, with the mime type set, wave works. ok.
21:53
<aklein>
Hixie: thanks for doing the experimentation, apologies for my laziness
21:53
<aklein>
the other one I've been wrestling with (and which needs more testing) is urls in CSS
21:54
<Hixie>
that one is conveniently not my problem :-)
21:54
<Hixie>
but yeah
21:55
<Hixie>
audio was easy to test because i can hear it
21:55
<Hixie>
now to test <img>...
21:55
<Hixie>
maybe i need a counter cgi that returns a different number each time or something
21:56
<aklein>
I would have posted a live dom viewer example for that if it had been easy...I just did it locally
21:58
<Hixie>
hehe
22:28
<Hixie>
aklein: firefox isn't doing anything when the img is inserted into another doc either: http://software.hixie.ch/utilities/js/live-dom-viewer/?saved=2758
22:30
<Hixie>
but if the resolved url is different, then it does
22:30
<Hixie>
interesting
22:34
<Hixie>
anyone got IE around to test with the three tests in https://www.w3.org/Bugs/Public/show_bug.cgi?id=24312 ?
22:39
<JonathanNeal>
Hixie: still need someone to look?
22:43
<Hixie>
JonathanNeal: yup
22:48
<Hixie>
is it me, or is textContent wrong for elements? http://dom.spec.whatwg.org/#dom-node-textcontent
22:48
<Hixie>
it says to replace the nodes, but i thought if the only child was a text node, it was preserved
22:52
<TabAtkins>
We were just talking about this on blink-dev, and we think we want to change the spec to preserve the node in that case.
22:53
<Hixie>
k well i'll assume Top People are on it then
22:53
<JonathanNeal>
Hixie: unsure of how to follow the test, but I have IE up.
22:53
<Hixie>
JonathanNeal: ok, let me walk you through it
22:54
<Hixie>
JonathanNeal: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2755 - what audio plays?
22:54
<Hixie>
JonathanNeal: and what does it say in the log?
22:54
<JonathanNeal>
It actually says "Invalid Source" to me in IE11
22:55
<Hixie>
ugh
22:55
<Hixie>
ok, let's ignore that.
22:55
<Hixie>
ok, next test. uh
22:55
<Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2758
22:56
<Hixie>
there should be two numbers in the log
22:56
<Hixie>
what are they?
22:57
<JonathanNeal>
checking
22:57
<JonathanNeal>
the two numbers are 27 and 0
22:58
<Hixie>
0!
22:58
<Hixie>
interesting
22:58
<Hixie>
and in http://software.hixie.ch/utilities/js/live-dom-viewer/saved/2759 ?
22:59
<Hixie>
(should be two numbers again)
23:03
<JonathanNeal>
just a moment
23:03
<Hixie>
no rush
23:03
<JonathanNeal>
17, 0
23:05
<Hixie>
JonathanNeal: fascinating. thanks.
23:09
<JonathanNeal>
what does it mean?
23:24
<Hixie>
JonathanNeal: means that IE is further away from the other browsers than is useful to include IE in our sample set :-)
23:49
<Hixie>
man
23:49
<Hixie>
i could really do with dreamhost giving me more visibility into what processes are taking up all the memory when it decides that my machine is out of RAM and should be rebooted