04:39
<Zerxer>
now that I'm in the right area.. messing around with the <audio> tag and I can't get ogg files to play nicely in firefox. keeps skipping. perhaps I am just using a bad encoder? suggestions?
05:03
<Hixie>
hard to tell without a sample page :-)
05:04
<Hixie>
(#developers on irc.mozilla.org may be able to help better with mozilla-specific issues, btw)
09:13
<MikeSmith>
annevk: about media queries, is it a conformance error for a media query to use a media feature with a media type that it does not apply to?
09:14
<MikeSmith>
the spec does not seem to explicitly say it is a conformance error
09:15
<MikeSmith>
e.g., @media print and (scan: progressive)
10:03
<annevk>
it's not
10:03
<annevk>
you could warn if you want
10:06
<annevk>
http://bitworking.org/news/2009/12/xrd
10:07
<MikeSmith>
annevk: OK, I'll ask hsivonen if he thinks it's worth warning about
10:07
<annevk>
you could look up an old thread on media queries on www-style
10:07
<annevk>
hsivonen asked about this before
10:11
<MikeSmith>
annevk: OK
10:14
<MikeSmith>
annevk: found the thread
10:15
<annevk>
cool
12:29
<Dashiva>
annevk: And now we have two ladders to paint, twice the fun
12:30
<annevk>
it made me think of Microdata / RDFa mostly
12:30
<annevk>
though maybe there the problem is somewhat more scoped
12:32
<Dashiva>
Microdata and RDFa are which material you build the ladder from. Still out there, but somewhat more relevant than the color.
14:34
<annevk>
anyone knows why Ubuntu 9.10 is no longer offered as installation option under updates?
14:34
<annevk>
pretty sure it was before
14:34
<annevk>
but I didn't do it because I thought there might be trouble
14:36
<annevk>
http://guides.macrumors.com/MacBook_Pro_Buyer's_Guide is most excellent
14:37
<Dashiva>
Would we be having this autobuffer thread at all without safari's bug?
14:38
<annevk>
lots of talk, little practical impact
14:40
<wavded>
Does anyone have any experience with the Firefox 3.6 File API implementation?
14:41
<wavded>
I'm trying to do a XHR POST with file read in with FileReader, but I want to add more params to the post. xhr.sendAsBinary() doesn't seem to allow me to do that...
14:41
<wavded>
Any help would be appreciated
14:42
<annevk>
you prolly need to wait until they do an impl of some XHR2 extensions
14:42
<annevk>
i.e. the FormData object
14:43
<wavded>
The docs said you could use xhr.send() with a FileReader but my attempts don't seem to work..
14:44
<wavded>
Hmm... there docs for XHR2 somewhere?
14:45
<annevk>
http://dev.w3.org/2006/webapi/XMLHttpRequest-2/ has the latest
14:45
<wavded>
thx
16:44
<TabAtkins>
Dashiva: Nope, the entire problem is spawned from Safari's bug causing it to autobuffer everything. If it was already acting like Firefox, then no one would have complained, and we wouldn't have people trying to draw ridiculous distinctions like autobuffer=auto being somehow different than autobuffer not being present at all (let alone autobuffer=no actually having a usefully different...
16:44
<TabAtkins>
...definition).
19:32
<annevk>
hmm, I only get it if I run update-manager -d
19:32
<annevk>
but then sites also advice not to install it...
19:32
annevk
is confused
19:32
<annevk>
(still about Ubuntu 9.100
19:32
<annevk>
)
19:32
<gratz|home>
get what?
19:32
<gratz|home>
oh right
19:33
<gratz|home>
what advises not to install it and why?
19:35
<annevk>
there's a series here: http://www.insidesocal.com/click/2009/11/pulling-the-trigger-on-ubuntu.html
19:35
<annevk>
since I think I have a similar card for video I'll wait
19:35
<annevk>
especially with my MacBook being in repairs and all :)
19:37
<gratz|home>
heh
19:43
<wavded>
annevk - seems as though you are right, I haven't been able to find anyway in FF 3.6 to push file data via XHR along with other POST data. :-(
20:03
<AryehGregor>
annevk, you're on 9.04, not 8.04 or anything, right? Maybe apt-get update will fix it, if it thinks it's a development release for some crazy reason.
20:10
<annevk>
AryehGregor, yeah, I'm on 9.04
20:11
<annevk>
and I found a way to make the upgrade option appear, but I decided to wait
21:27
<foolip>
hmm, which spec defines encodeURIComponent?
21:36
<Philip`>
foolip: ECMAScript
21:50
<foolip>
Philip`: found it, thanks
22:00
<zcorpan>
MikeSmith: yo
22:01
<MikeSmith>
zcorpan: hej
22:01
<zcorpan>
MikeSmith: maybe checking for valuelessness and lessthanness should be in the tokenizer instead of the switch block
22:02
MikeSmith
takes a look
22:02
<zcorpan>
MikeSmith: also, i think case INVALID: is redundant with default:
22:03
<MikeSmith>
about checking for valuelessness and lessthanness, I think I made a conscious choice to do it the current way but can't recall why
22:03
<MikeSmith>
but I will look again
22:04
<MikeSmith>
about INVALID, I guess that's true -- I'll remove it
22:06
<MikeSmith>
ah
22:06
<MikeSmith>
zcorpan: I remember
22:07
<MikeSmith>
the reason I don't do the valuelessness check in the tokenizer is that I was thinking, we don't want to report valuelessness for invalid attributes
22:07
<MikeSmith>
I would think we'd just want to report it only for the case of attributes that we already know are valid
22:07
<zcorpan>
why not?
22:08
<MikeSmith>
because it will cause two error messages to be generated for each invalid attribute, if it lacks a value
22:08
<zcorpan>
yeah
22:09
<MikeSmith>
I guess my rationale was, if the attribute is invalid, it doesn't matter to the user whether it lacks a value
22:09
<zcorpan>
fair enough
22:10
<MikeSmith>
about lessthanness, where is the code checking that now?
22:10
<MikeSmith>
oh
22:10
<MikeSmith>
you mean the literal "<" character
22:11
<zcorpan>
yes :)
22:11
<MikeSmith>
I guess I could move that check
22:12
<MikeSmith>
hmm, but again, that's the same kind of case
22:13
<MikeSmith>
if I move it into the tokenizer, it will mean that you'd get two error messages for an invalid attribute if it has a value with "<"
22:13
<zcorpan>
a < will cause browsers to ignore the whole pi
22:13
<zcorpan>
an unknown pseudoatt won't
22:13
<MikeSmith>
ah, OK
22:13
<zcorpan>
same with valueless
22:14
<MikeSmith>
I see
22:15
<MikeSmith>
that trumps the two-message issue
22:15
<MikeSmith>
I will move those now and push it up to qa-dev when I'm done
22:16
<MikeSmith>
btw, I also made some fixes in media-query handling
22:16
<zcorpan>
did you fix case-sensitiveness?
22:17
<MikeSmith>
for xml-stylesheet pseudo-attribute names? yeah
22:19
<MikeSmith>
and also handling of XSLT indicators that have parameters
22:20
<zcorpan>
yep
22:20
<MikeSmith>
and added a warning if type/subtype is something other than text/xsl or text/css
22:20
<zcorpan>
that's in the patch i'm looking at
22:22
<zcorpan>
MikeSmith: i think there are some things that could be removed in the tokenizer, given that some characters are not allowed in xml
22:22
<MikeSmith>
OK
22:22
<zcorpan>
like case '\u000C': and case '\u0000':
22:22
<zcorpan>
but maybe it's not worth it to fiddle with something that works :)
22:23
<MikeSmith>
nah, I think it's worth getting it right
22:23
<MikeSmith>
so I can remove those
22:26
<zcorpan>
what does silentLineFeed() do?
22:27
<zcorpan>
'\r' also can't appear in pi data
22:29
<MikeSmith>
I am not sure what the purpose of silentLineFeed() is.. I just inherited it from the original copy/paste
22:29
<MikeSmith>
but I think basically it's for handling the CRLF case for HTML parsing as the spec requires
22:29
<MikeSmith>
to drop the CR
22:30
<zcorpan>
likely
22:31
<MikeSmith>
so what should I do for the "\r" case? emit a specific error?
22:31
<zcorpan>
remove, since it can't happen
22:31
<zcorpan>
\r is normalized to \n
22:31
<zcorpan>
by the xml parser
22:50
<zcorpan>
+ // TODO warn about apos (IE) and TRADE (Opera)
22:51
<zcorpan>
hsivonen: &TRADE; works in opera now
22:51
<MikeSmith>
OK, I can remove that comment
22:52
<MikeSmith>
zcorpan: so U+000C is explicitly disallowed in XML? or does it get normalized?
22:52
<zcorpan>
it's disallowed
22:53
<MikeSmith>
OK
22:58
<zcorpan>
MikeSmith: maybe you can remove the stuff in the tokenizer that checks for semicolonless entities
22:58
<MikeSmith>
ah yeah
22:58
<MikeSmith>
hmm
22:58
<MikeSmith>
I guess
22:58
<MikeSmith>
so
22:59
<MikeSmith>
the XML parser will catch those?
22:59
<MikeSmith>
oh, I guess not
22:59
<zcorpan>
the table doesn't contain semicolonless entities
23:00
<zcorpan>
so it's dead code i think
23:04
<MikeSmith>
which table?
23:05
<zcorpan>
the entity table
23:06
<MikeSmith>
OK
23:06
<MikeSmith>
but I should keep the checking for semicolonless character references, right?
23:07
<zcorpan>
yes
23:10
<zcorpan>
did you implement the CharRef Legal Character thing?
23:14
<MikeSmith>
zcorpan: you mean an explicit check for "Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]" ?
23:14
<zcorpan>
yes
23:15
<MikeSmith>
not yet
23:15
<MikeSmith>
but I can add it now
23:15
<zcorpan>
ok
23:17
<MikeSmith>
I guess I can add it in the checkChar method
23:17
<zcorpan>
not handleNcrValue ?
23:18
<MikeSmith>
ah, for character refs
23:19
<zcorpan>
yeah... the xml parser has already checked Char for the pi data
23:20
<MikeSmith>
OK
23:33
<zcorpan>
time to sleep
23:33
<zcorpan>
g'nite