00:02
<GPHemsley>
and f for fetch
00:15
<GPHemsley>
hmm
00:15
<GPHemsley>
it appears that Gecko does not do any sniffing at all—at least, not for GIFs
00:16
<GPHemsley>
hmm
00:16
<GPHemsley>
opera.com just sent me Japanese
00:17
<GPHemsley>
...apparently I had a language cookie set
00:17
<GPHemsley>
weird
00:21
<GPHemsley>
well, this lack of sniffing in no engine but Gecko is gonna make testing/debugging kinda annoying
00:27
<GPHemsley>
on the other hand, http://damowmow.com/temp/ is good for testing what happens when the server doesn't send a Content-Type header
00:29
<GPHemsley>
Hixie: s/write/right/ in your WebVTT comment commit
06:46
<wirepair_>
anyone here familiar with firefox's processing of CSP rules?
08:27
<annevk>
so people, GitHub is sponsoring WHATWG!
08:27
<annevk>
we got bronze for life :-)
08:27
<annevk>
so I guess I'll make a private repo for the info on the twitter accounts in due course
08:28
<annevk>
does that make sense?
08:31
<SimonSapin>
annevk: I might end up implementing Encoding (or not) in rust rather than python
08:31
<annevk>
so are we going to use images.whatwg.org?
08:32
<SimonSapin>
more fun
08:32
<annevk>
SimonSapin: fair enough :)
08:32
<MikeSmith>
annevk: rockin
08:47
<annevk>
MikeSmith: what's the W3C Team archive list?
08:47
<annevk>
MikeSmith: w3t-archive?
08:50
<MikeSmith>
annevk: yeah
08:54
<annevk>
posted a minor rant there
09:26
<othermaciej>
oh man there are some good new memes up
09:37
<annevk>
http://lists.w3.org/Archives/Public/uri/2012Nov/0000.html is interesting
09:52
<karlcow>
annevk: yup
09:52
<smaug____>
is hallvord ever here?
09:53
<SimonSapin>
annevk: uh, what? Is this ignoring your work while doing the same?
09:53
<karlcow>
SimonSapin: I don't think so.
09:55
<annevk>
SimonSapin: if it is, that's fine
09:55
<annevk>
SimonSapin: curious where we'll end up
09:55
<karlcow>
yes me too. Same than annevk.
09:56
<karlcow>
Whatever the outcomes, if in the end there's something which is better, whatever it is.
09:56
<karlcow>
The communities win.
10:00
<SimonSapin>
I just spent two hours writing up a 600-words email for www-style …
10:08
<jgraham>
smaug____: I think I might have seen hallvord here before
10:08
<jgraham>
But not often at least
10:19
<konza>
Hi all, I have subscribed to the mailing list. Can any one explain how can I contribute to the development. By contributing i mean to write code.. Can any one explain
10:21
<annevk>
konza: what are your interests?
10:22
<konza>
annevk, I love coding.. :) Could you please explain your question a bit?
10:24
<annevk>
konza: well WHATWG doesn't do a lot of coding (other than in English), so I was wondering why you joined the mailing list
10:24
<annevk>
like what you want to contribute to
10:25
<konza>
annevk, I thought some open source works are also done by this group
10:26
<annevk>
well yeah, I suppose, some of us work on http://code.google.com/p/html5lib/
10:27
<annevk>
but I suspect you might be better of contributing to Gecko/WebKit/Chromium if you like to work on code related to what we work on
10:27
<Ms2ger>
Let's just make that Gecko :)
10:33
<smaug____>
++Ms2ger
10:34
<MikeSmith>
zcorpan, jgraham : so I can add the symlinks for those websocket tests now
10:35
<zcorpan>
MikeSmith: ok
10:35
<zcorpan>
thanks
10:35
<MikeSmith>
zcorpan: including I will clobber the existing echo test that's there with a symlink to the one in the repo
10:36
<MikeSmith>
which I assume will not break anything
10:36
<MikeSmith>
for you guys
10:36
<MikeSmith>
since your tests are relying on the echo service you have there
10:36
<zcorpan>
the new one presumably works for us. i don't know what the existing one looks like or what uses it currently
10:36
<jgraham>
I think our echo is probably the default
10:37
<MikeSmith>
ah OK
10:37
<MikeSmith>
yeah it's likely the default I guess
10:37
<jgraham>
I just copied all the files we had in our handler directory that seemd to be referenced
10:37
<MikeSmith>
hai
10:41
<zcorpan>
"can i get two weeks?" "NO."
10:43
<konza>
annevk, can you please give the full url of gecko
10:43
<annevk>
konza: http://www.mozilla.org/en-US/contribute/ I suppose
10:44
<konza>
annevk, thanks
10:45
<konza>
annevk, pm?
10:45
<annevk>
sure
10:49
<hsivonen>
http://whatcanidoformozilla.org/
10:49
<konza>
thanks hsivonen
10:56
<MikeSmith>
zcorpan, jgraham : OK all those handlers should now be reachable at http://w3c-test.org/ws
10:57
<MikeSmith>
if not lemme know
10:57
<zcorpan>
MikeSmith: did you mean ws://w3c-test.org/ ?
10:57
<MikeSmith>
yeah
10:58
<zcorpan>
ok, excellent
10:58
<MikeSmith>
well ws: instead of http:
10:58
<MikeSmith>
you want them at the root? I thought you had changed your tests to look for them under /ws
11:01
<zcorpan>
MikeSmith: my error console for http://w3c-test.org/webapps/WebSockets/tests/submissions/Opera/interfaces/WebSocket/send/006.html says "WebSocket handshake failure, invalid response code '404'. "
11:01
<zcorpan>
MikeSmith: no the tests have /ws/ already, i didn't know that was changed
11:02
<MikeSmith>
ok will try to take a look at the server logs
11:06
<zcorpan>
q?
11:06
<zcorpan>
oops
11:08
<annevk>
so I walked in the HTML WG meeting to get some AC power
11:09
<annevk>
hsivonen just suggested to put everything "at risk" and undefine all the things when the time comes (kind of)
11:11
<hsivonen>
I asked what's the procedural reason not to
11:12
<hsivonen>
I didn't really suggest it
11:12
<MikeSmith>
http://lists.w3.org/Archives/Public/public-ietf-w3c/2012Sep/thread.html#msg4
11:12
MikeSmith
ah wait that's not it
11:13
<MikeSmith>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=18975
11:13
<MikeSmith>
-> https://www.w3.org/Bugs/Public/show_bug.cgi?id=18975 registerContentHanlder and registerProtocolHandler open huge security and privacy holes (bug raised by Larry Masinter)
11:14
<smaug____>
annevk: sounds like a fun meeting
11:14
<smaug____>
apparently discussions about the process ?
11:14
<annevk>
it's about CR exit criteria
11:14
<annevk>
since most of HTML is not interoperable if you look at the details, they're in kind of a fit
11:14
<MikeSmith>
ah wrong channel
11:15
<annevk>
so they're looking for loopholes
11:15
<annevk>
MikeSmith: sorry for the distraction :)
11:16
jgraham
thinks it is a bad idea to point out that we will not get interoperability on navigation
11:16
<Ms2ger>
Why's that?
11:17
<jgraham>
Or maybe I should just say that we will never get two complete implementations
12:11
<karlcow>
hmmm I had a parse error for Robin's email
12:11
<karlcow>
"or more precisely for a way of specifying isolates more powerful than <bdi> one possible implementation of which could rely on the dir attribute."
12:12
<karlcow>
http://www.w3.org/mid/5093A450.5000107⊙wo
13:12
<karlcow>
https://tech.dropbox.com/2012/10/caching-in-theory-and-practice/
13:19
<annevk>
"it's a waste of time" "any other comments?"
13:19
<annevk>
this meeting is hilarious
14:03
<hsivonen>
Hixie: this xml-stylesheet is x3d all over again
14:03
<hsivonen>
you notes trigger concern
14:28
<annevk>
http://www.rfc-editor.org/errata_search.php?rfc=1123
14:28
<annevk>
wtf
14:28
<annevk>
all the rejected errata refers to each other for being rejected and nothing then addresses the problems
14:29
<annevk>
and how and http://tools.ietf.org/html/rfc1123 be the latest RFC for host names anyway?
14:30
<annevk>
and just look at the patchwork that is http://tools.ietf.org/html/rfc1035
14:30
<annevk>
you can't just go in and get an understanding of how this works
14:31
<annevk>
that's no documentation of the web architecture if you ask me
14:31
<karlcow>
I guess because nobody has been masochistic enough to touch hostnames spec ;)
14:32
<annevk>
pussies
14:32
<karlcow>
macho ;)
14:33
<annevk>
pew pew
14:36
<hober>
Scooby Doo: Special (Process) Victims Unit
14:36
<hober>
*dun dun*
14:37
<darobin>
at the end of the day we'll find out that it was just a dishonest land developer dressing up a Hixie to scare the locals
14:37
<darobin>
into selling their land
14:38
<karlcow>
:D
14:38
<MikeSmith>
heh
14:39
<GPHemsley>
:)
15:20
karlcow
found an oldie http://www.snee.com/xml/a-href.html
15:23
<MikeSmith>
karlcow: interesting
15:24
<karlcow>
It was a link in my today 10 years old blog post. It survived.
15:25
<karlcow>
It's kind of encouraging some of the things survive
15:25
<karlcow>
That was also the time where designers were experimenting with small icons http://www.antipixel.com/blog/archives/2002/10/22/steal_these_buttons.html
15:31
<hsivonen>
wow. I didn't realize it has been ten years alrady
15:32
<hsivonen>
also, when did you last see a mention of amphetadesk?
15:32
<karlcow>
:)
15:40
<annevk>
Hixie: it would be nice if http://dev.w3.org/html5/webvtt/ had Bugzilla links
15:40
<annevk>
Hixie: one instance of "WebVTT comments" is not linked
15:40
<annevk>
Hixie: the same for "ASCII digits"
15:41
<annevk>
Hixie: should also use the Encoding Standard
16:09
karlcow
recommends annevk to check out the source code of Arena beta-3, the file HTParse.h
16:11
<karlcow>
"Canonicalization of URIs is a difficult job" — https://gist.github.com/2569ea81038d5e98a7b2
16:11
<karlcow>
I put the file on gist
16:37
<annevk>
my parser handles those scenarios, not entirely in the same way, but then that browser is old :)
16:40
<annevk>
GPHemsley: you might want to look at MIME type usage in http://xhr.spec.whatwg.org/ too
16:40
<annevk>
GPHemsley: and I guess I might use some of the terminology you introduced, such as XML type
16:41
<annevk>
gotta go
17:07
<GPHemsley>
annevk: I don't see anything related to sniffing, per se, in XHR. But if there's something you want me to add to mimesniff to make your XHR life easier, let me know.
17:16
<smaug____>
hsivonen: ping
18:47
<Hixie>
hsivonen: what do you mean? (re xml-stylesheet and "you notes trigger concern")
18:53
<GPHemsley>
Hixie: There are points in HTML were MIME sniffing is explicitly encouraged/required, correct?
18:54
<GPHemsley>
Hixie: And, if so, can you provide me with a list of where that occurs (or point me to one)?
19:36
<Hixie>
GPHemsley: encouraged of whom? implementors?
19:36
<Hixie>
GPHemsley: mentions of MIME Sniff can be found by searching for [MIMESNIFF] then clicking the <dfn> links that those are next to
20:58
<Ms2ger>
Anne might be interested in http://docs.sqlalchemy.org/en/rel_0_7/core/engines.html#sqlite
21:04
<Hixie>
i'm amused at the meme with the caption "It seems that the people who most enjoy complaining about W3C Process are those who most enjoy Process discussions", but it seems incorrect
21:05
<Hixie>
since i enjoy complaining about w3c process, and i hate the process discussions so much i don't even turn up to tpac
21:08
<say2joe>
lol
21:10
<say2joe>
wish i lived in France
22:37
<GPHemsley>
Hixie: Well, the lists of mentions are quite long. I was hoping for you to narrow it down a little, since you're more familiar with exactly what the spec says. But yes, where implementors are encouraged to proactively sniff.
22:38
<Hixie>
the lists arne't long, it's like three areas
22:38
<Hixie>
ignore the <dfn>Content-Type metadata</dfn>, that's just references to Content-Type, not sniffing
22:40
<GPHemsley>
well, the lists only work on the one-page version, which my computer seems to hate
22:40
<GPHemsley>
which makes me afraid to go fishing for information directly
22:41
<GPHemsley>
oh, OK, I see
22:41
<GPHemsley>
I was working from memory when I asked
22:41
<GPHemsley>
and it was indeed only the content-type link that was long
22:55
<GPHemsley>
Hixie: Am I right in my understanding that the type SHOULD always be sniffed for resources called by <img> and <object>?
22:59
<GPHemsley>
Hixie: And the same for <link> (e.g. for a shortcut icon)?
23:00
<GPHemsley>
Hixie: Also, do you define precisely what an "image" is?
23:05
<GPHemsley>
Hixie: And your only use of the text-or-binary algorithm (in <object>) is just to ensure against binary files tagged as 'text/plain'?