00:33
<Philip`>
Hmm, I need to go on holiday instead of continuing the alt discussion - what a dreadful shame :-(
00:39
<Philip`>
http://www.bbc.co.uk/iplayer/ - <meta name="DCTERMS.modified" content="2076-02-12T12:00:00Z" /> - yay for metadata
00:43
<Philip`>
(And DCTERMS.modified doesn't even exist, according to the declared schema)
00:50
eseidel
wonders if anyone is interested in a theoretical discussion about <canvas> patterns? :)
00:50
<eseidel>
http://www.whatwg.org/specs/web-apps/current-work/#createpattern
00:51
<eseidel>
Safari's current behavior is slightly differnet from the spec, in that if an image is still loading off the network when createPattern is called with an HTMLImageElement parameter, the pattern will draw the image when it's finally loaded
00:51
<eseidel>
it will even draw it in incremental states :)
00:51
eseidel
should really email the list
01:00
<Philip`>
eseidel: That sounds unpleasantly unpredictable
01:00
<eseidel>
Philip`: yeah, I'm gonna change iet to make a copy
01:00
<eseidel>
Philip`, Hixie: but i sent a note to the whatwg list all the same
01:00
<eseidel>
this will now force web authors to listen for image load evnets
01:00
<eseidel>
which they might be doing anyway
01:00
<Philip`>
I think the spec just says it should throw an exception on createPattern if it's not complete
01:01
<eseidel>
If the image argument is an HTMLImageElement object whose complete attribute is false, then the implementation must raise an INVALID_STATE_ERR exception.
01:01
<eseidel>
oh
01:01
<eseidel>
it does
01:01
<eseidel>
nm then
01:01
<Philip`>
"If the image argument is an HTMLImageElement object whose complete attribute is false, then the implementation must raise an INVALID_STATE_ERR exception."
01:04
<Philip`>
http://philip.html5.org/tests/canvas/suite/tests/results.html#2d.pattern.image.incomplete possibly isn't entirely helpful since it assumes new Image().complete == false (since the spec says so), which is untrue in reality
08:43
<heycam>
hsivonen, yt?
09:30
<hsivonen>
heycam: I am now.
10:25
<heycam>
hsivonen, was gonna ask: how come http://s.validator.nu/xhtml10/text.rnc removes xml:space from <pre>?
10:26
heycam
afk again
10:31
<hsivonen>
heycam: hmm. must be a bug
10:33
<hsivonen>
heycam: It's now a common attribute, isn't it?
10:34
<hsivonen>
hmm. it's not a common attribute. sigh
10:50
<jacobolus>
annevk: i'm a bit late here, but if you want a compact camera w/ a small lens, there is no way 10 Mpx will actually be useful (the extra data will be some combination of noise and blur)
10:55
<mcarter>
do any of the opera SSE implementations have a way to set the retry interval, like the current spec does?
11:01
<annevk>
mcarter, no
11:01
<annevk>
jacobolus, ok, thanks
11:01
<annevk>
jacobolus, I will most likely buy this model: http://www.dpreview.com/reviews/specs/Panasonic/panasonic_dmctz5.asp
11:02
<mcarter>
annevk, does it (the application/x-dom-event-stream SSE) auto-reconnect when a response completes?
11:03
<annevk>
mcarter, it should, yes
11:03
<mcarter>
annevk, i don't suppose you recall if it does so immediately or not?
11:03
<jacobolus>
mcarter: I don't think so
11:03
<jacobolus>
mcarter: try it! :)
11:04
<annevk>
mcarter, what jacobolus said
11:08
<mcarter>
so it seems to reconnect automatically after a 5 second delay
11:09
<jacobolus>
really? did it do that in 8.x? huh.
11:10
<mcarter>
annevk, do you recall if there's a way to attach a trigger for when the stream completes, if i wanted to, for instance, remove the stream when the response completes?
11:13
<hsivonen>
heycam: hmm. My local copy works. For some reason, the deployment copy isn't pulling stuff from svn correctly or something, so you see a partial & broken "fix"
11:13
<hsivonen>
heycam: I'll investigate
11:23
<hsivonen>
heycam: fix deployed
11:23
<annevk>
mcarter, maybe use a special event that removes the stream?
11:24
<mcarter>
annevk, yeah, that works. But in the case of a network breakage it falls apart
11:24
<annevk>
mcarter, as in, the page registers a special listener for the "removestream" event and when it gets that from the server it kills the connection
11:24
<annevk>
mcarter, true
11:24
<hsivonen>
heycam: thanks
13:37
<hsivonen>
we should have a standard form for explaining why Web spec feature suggestions don't work like the Slashdot spam form (http://tech.slashdot.org/comments.pl?sid=623883&cid=24314727)
15:05
<heycam>
hsivonen, thanks
16:54
<takkaria>
I wonder if IE's authoratitive=true MIME header will disable sniffing on the charset too
22:44
<Hixie>
hsivonen: that form is pretty close already
23:10
<jcranmer>
how does the W3C moderation system work? First post is sent to moderator, and, if accepted, all subsequent posts go through?
23:58
<Lachy>
jcranmer, no. For their first post, they're sent a confirmation email asking them to manually agree for it and all subsequent mails to be archived. They probably have some sort of spam filter in place too. They do have list moderators, but most things are automated