16:27
<krijnh>
Have no fear, the (smoke producing) Pentium II IRC logging service is here! \o/
16:28
<gsnedders>
http://google-web-toolkit.googlecode.com/svn/trunk/user/src/com/google/gwt/core/client/impl/StackTraceCreator.java — oh wow.
16:28
<gsnedders>
That's just scary.
16:29
<annevk>
krijnh is talking about http://plixi.com/p/61729229 I believe o_O
16:29
<krijnh>
Yeah
16:29
<gsnedders>
Tries to read Error.message in Opera to get the stack, which was (well, still is) where the stack was put in Futhark with a non-default preference option.
16:29
<gsnedders>
Won't work at all in Carakan…
16:29
<MikeSmith>
damn, krijnh is back
16:30
<MikeSmith>
now I have to be more careful about what I say
16:30
<jgraham>
david_carlisle: You are the resident expert on all things X, right :)
16:30
<david_carlisle>
but you hate us
16:30
<annevk>
so much hate
16:31
<krijnh>
I thought we only hate blind people?
16:31
<annevk>
whenever something bad happens, I always think, oh well, there's also XML
16:31
<jgraham>
If I have some XPath with a default namespace (XPath 2 has this right?) can I mix explicit prefixes and the default in the same expression?
16:31
<david_carlisle>
yes
16:32
<annevk>
krijnh, hmm yeah, can't spread this hate thing thin
16:32
<david_carlisle>
the default-xpath-namespace just changes the interpretation of unprefixed names, has no effect on prefixed ones
16:32
<annevk>
dilemma
16:32
<jgraham>
So if the XHTML ns is the default then //html:div/span should find all span children of div elements both in the XHTMLNS
16:32
<jgraham>
david_carlisle: Perfect, thanks
16:33
<david_carlisle>
xpath today
16:33
<david_carlisle>
xslt tomorrow
16:33
<karlcow>
MikeSmith: as usual, you are █████████████████████ it's ok.
16:33
<david_carlisle>
whole world using xml by the weekend
16:33
<jgraham>
david_carlisle: Actually I quite like XPath
16:33
<jgraham>
It hasn't made me want to use XSLT yet theough :p
16:33
<karlcow>
damn, krijnh sanitizer
16:33
<jgraham>
*though
16:35
<david_carlisle>
you'd probably prefer xquery anyway (which is basically a cut down xslt, without the xml syntax)
16:41
<annevk>
Safari accepts a different set of encodings compared to Chrome
16:42
<annevk>
Chrome is way more restricted
16:55
<annevk>
yay
16:56
<annevk>
something else I figured out
16:56
<annevk>
IE treats aliases of iso-8859-1 as windows-1252, but not iso-8859-1 itself, unless it is loaded directly
16:57
<annevk>
also, Firefox and IE are leading encoding proliferation
16:57
<annevk>
Firefox foremost I guess
16:58
<annevk>
I wonder if they are interested in fixing that
16:59
Philip`
notes that reducing encoding support can be a security risk
16:59
jgraham
wishes for a way to tell emacs to check in the current buffer to a vcs after every save (but not to do it for all buffers)
17:00
<jgraham>
hober: In case you were wondering, that was your queue to say "I've got some code that does just that" :)
17:01
<jgraham>
er "cue"
17:02
<jgraham>
silly language
17:03
<annevk>
Philip`, example?
17:03
<karlcow>
I wonder if someone wrote a nested tables counter for Web pages…
17:09
<Philip`>
annevk: Any not-sufficiently-ASCII-like encoding (EBCDIC, ISO-2022-*, etc) - if a server emits it but a browser doesn't support it and uses a default decoder instead, you can get XSS
17:12
<annevk>
ah yes
17:12
<annevk>
I am only testing 7- and 8-bit so far
17:12
<annevk>
nobody does 7-bit
17:30
<oojacoboo>
latest browser stats on my site are surprisingly good
17:31
<oojacoboo>
IE only 24% with IE6 and IE7 only getting 5%
17:31
<oojacoboo>
75% webkit/gecko
17:31
<oojacoboo>
for an industry and still thinks they live in the 50s
17:38
<karlcow>
http://www.rfc-editor.org/rfc/rfc6067.txt
17:43
<hober>
hmm. http://wiki.whatwg.org/index.php?title=CanvasContexts&diff=5608&oldid=prev
17:45
<hober>
jgraham: get in the habit of hitting C-x v v after you C-x C-s :)
17:46
<jgraham>
hober: Not really good enough since I don't want to be promted for a commit message
17:46
<jgraham>
And since the whole deal is that I am likely to forget
17:47
<jgraham>
(the use case is reducing buggy sites where you remove code and see if the bug still happens. It would be great to have every intermediate step atomatically saved so if you over-reduce you can back up a few steps and try again)
17:48
<jgraham>
*automatically
18:01
<hober>
you could do something like (add-hook 'after-save-hook (lambda () (when (something-p) (vc-checkin (buffer-file-name) nil "Commit message"))))
18:01
<hober>
untested of course
18:01
<hober>
and the impl of `something-p' is left to the reader :)
18:04
<hober>
ahh, an explanation of the weird wiki edit I noted earlier: http://lists.w3.org/Archives/Public/public-html/2010Dec/0075.html
19:19
<zcorpan>
"“Accept” and “Accept-Language” are fairly important" - http://calendar.perfplanet.com/2010/bloated-request-response-headers/
19:22
<zcorpan>
hmm, maybe we should kill the Referer header for same-origin requests?
19:23
<zcorpan>
or will that Break The Web?
19:25
<zcorpan>
sadly we won't be able to remove the User-Agent header
19:25
<zcorpan>
(at least we haven't added it for websockets!)
19:25
<AryehGregor>
Just use SPDY, then most of this stuff is only sent once, and is compressed to boot.
19:26
<zcorpan>
i haven't read up on SPDY yet
19:27
<Peter->
zcorpan: many anti-hotlink scripts rely on that
19:27
<Peter->
although they usually accept no referer too
19:27
<zcorpan>
Peter-: if they accept no referer then it should be fine
19:40
<Hixie>
so... i'm confused
19:40
<Hixie>
is elliot's draft normative or not?
19:41
<hober>
he says it's normative, but he's removed all normative language from statements that *don't* just say what's already said in another spec
19:48
<Philip`>
Launchpad seems to fail if you disable Referer
19:48
<Philip`>
so you'll break that part of the web, at least
19:49
<zcorpan>
Philip`: ok
19:49
<Hixie>
hober: that sounds like a huge recipe for confusion
19:53
<hober>
Hixie: indeed
19:53
<hober>
it would have made more sense for him to do the opposite: drop RFC2119 terminology when restating reqs from html&xml specs
19:53
<hober>
and only use RFC2119 terms for *additional* restrictions
19:56
<zcorpan>
what would make sense is if it didn't have additional restrictions, just non-normatively describe the intersection of the restrictions in the text/html and xhtml5 restrictions
19:57
<hober>
I'm assuming that the spec is intended to define a useful/interesting subset of that intersection
19:57
<zcorpan>
still, i don't see the usefullness of it, so i don't care
19:57
<hober>
that's where the additional restrictions would come in
19:57
<hober>
but yeah, if it's just intended to describe the actual intersection, no additional normative reqs are necessary
19:58
<zcorpan>
i don't know what the intention is. maybe the intention is just to waste people's time
20:21
<zcorpan>
twitter says @johnfoliot is "similar to me"
20:27
<karlcow>
zcorpan: the moustache maybe
20:28
<zcorpan>
i don't have one!
22:06
<annevk>
hmm, updating html5-diff
22:06
<annevk>
I guess I can do that
22:06
<annevk>
even though technically I'm on vacation
22:19
<Hixie>
so http://www.whatwg.org/specs/web-apps/current-work/complete.html#playing-the-media-resource says that play() loads the resource "if necessary"
22:19
<Hixie>
which some people have interpreted as meaning "if the src='' attribute has changed"
22:20
<Hixie>
but i really just meant if the element is in the empty state
22:20
<Hixie>
anyone have a better way of phrasing this that is both accurate and helpful?
22:34
<zcorpan>
Hixie: "loading the media resource if it hasn't already started loading, and beginning playback."
22:35
<Hixie>
what's "the media resource"?
22:35
<zcorpan>
see the spec! :)
22:35
<Hixie>
well there's two here, right
22:35
<Hixie>
silvia interpreted as meaning the one that was specified in src=""
22:35
<Hixie>
even if the one previously specified had started playing already
22:37
<zcorpan>
i don't follow. if you change src="", it gets loaded regardless of play()
22:37
<Hixie>
<source src>, sorry
22:37
<bfrog>
is websocket going to be fixed and out in another year or so do you guys think?
22:37
<bfrog>
or is it going to be a never ending draft
22:37
<annevk>
I wish I'd know
22:37
<Hixie>
bfrog: dunno, ianfette in #chromium might know
22:37
<Hixie>
or ask the hybi chairs what their timetable is
22:37
<zcorpan>
Hixie: prepend "current"
22:38
<zcorpan>
hmm maybe that'll confuse things
22:38
<annevk>
WebSockets went out of WHATWG and does not appear to be going anywhere :/
22:38
<Hixie>
zcorpan: yeah i think that someone who misinterpreted "as necessary" would misinterpret "current media resource" in the same way
22:39
<bfrog>
whats the alternative to websocket anyways, there doesn't seem to be anything particularly good that can replace it easily
22:39
<zcorpan>
Hixie: maybe the text about loading isn't helpful
22:40
<annevk>
bfrog, there's none
22:40
<annevk>
bfrog, apart from prolly using Flash or something crazy like that
22:40
<bfrog>
right
22:41
<Hixie>
zcorpan: yeah, maybe i should just remove that
23:05
<hober>
This seems like as good a place as any to mention that I'm looking for work. If you're looking to hire a full-stack web developer with some HTML5 familiarity, toss me a line. :)
23:15
<zcorpan>
hober: don't fancy browser qa?
23:17
<hober>
zcorpan: browser qa sounds like the sort of thing that would drive somebody to write their own HTML spec
23:17
<zcorpan>
eventually, yeah, maybe
23:22
<Hixie>
it has happened, can't be denied
23:23
<Dashiva>
But who will test the testers
23:24
<krijnh>
-_-