00:54
<briankardell>
Hixie: are you saying that when video was introduced that search engines couldn't see the fallback content?
01:14
<GPHemsley>
It's amazing how many sites are completely broken in Firefox 3.6 nowadays.
01:15
<GPHemsley>
(Many of them are completely blank because the old parser thinks the whole page is a comment.)
02:15
<Hixie>
briankardell: no, i'm saying that by and large authors don't bother to use the fallback feature of <video>
05:37
<MikeSmith>
heycam: congrats on style scoped going out in Nightly
05:37
<heycam>
MikeSmith, thanks!
07:12
<hsivonen>
GPHemsley: I expect the Web to become incompatible with Firefox 3.6 even faster now that Firefox 3.6 has (finally) dropped off http://gs.statcounter.com/#browser_version_partially_combined-ww-daily-20130115-20130115-bar
07:12
<hsivonen>
IE6 too!
07:13
<hsivonen>
of course, the number of items on that chart is arbitrary, but I still expect it to matter to people’s perception when stuff falls out of view
07:15
<hsivonen>
(aside: it’s sad that they bundle all Android stock browser versions together when they are at least as different as [and actually more different than] the Safari versions that are listed as separate)
07:15
<hsivonen>
(bundling Opera Mobile and Opera Mini into one number is worse, of course)
07:15
<hsivonen>
(and separating Safari on iPhone and iPod Touch. seriously, why would you do that?)
07:17
<tantek>
separate iPhone and iPod Touch? Why not separate iPhone, iPhone 2, iPhone 3, iPhone 3G, iPhone3GS, iPhone 4, iPhone 4S, iPhone 5, iPod Touch, iPod 4 Touch, iPod 5 Touch?
07:19
<hsivonen>
in general, it seems that when it comes to mobile, they throw out everything that’s technically interesting and lists brands instead
07:22
<hsivonen>
aargh. Acid2 requires nearest neighbor image scaling? sadness. https://bugzilla.mozilla.org/show_bug.cgi?id=830731#c12
08:42
<hsivonen>
MikeSmith: any chance of moving http://www.w3.org/1999/xhtml/ to http://www.w3.org/1999/xhtml ?
08:43
<hsivonen>
MikeSmith: and upgrading the mention of HTML5 to CR?
08:45
<Ms2ger>
Anyone ever heard of <table cols>?
08:45
<hsivonen>
at least the namespace URL of SVG doesn’t redirect to another URL
08:48
<MikeSmith>
hsivonen: yes chance
08:49
<MikeSmith>
we should
08:49
<hsivonen>
MikeSmith: great
08:49
<SimonSapin>
annevk (if you read logs), re http://dump.testsuite.org/encoding/single-byte-test.html , you might not need PHP or .htaccess if you use xhr.overrideMimeType()
08:50
<hsivonen>
MikeSmith: being able to do ctrl-l, ctrl-c and pasting into code would be most helpful
08:50
<hsivonen>
(yes, I still search for “XHTML namespace” on the Web when writing code)
08:51
<hsivonen>
(on Mac, I had a system-level macro for that, but on Linux I don’t)
08:51
<hsivonen>
(I had a thing hooked into the a11y APIs on Mac so that I could type a few letters and it expanded to the XHTML namespace URL)
09:19
<darobin>
Ms2ger, jgraham: we just had our first pull request from a total third party :) https://github.com/w3c/html-testsuite/pull/18
09:38
<hsivonen>
\o
09:38
<hsivonen>
oops
09:50
<Ms2ger>
darobin, next you'll have me believe github actually was a good idea...
09:54
<hsivonen>
has Chrome removed its character encoding override menu? or am I just failing to locate the menu?
09:54
<hasather>
hsivonen: Tools > Encoding
09:56
<hsivonen>
hasather: I don’t see an Encoding submenu there
09:56
<smaug____>
I have it
09:57
<smaug____>
v24
09:57
<hsivonen>
I tried Chrome 25 and Chromium 23
09:57
<hsivonen>
both on Linux
09:57
<hasather>
hsivonen: I have it in 23 and 26 on Linux
09:57
<smaug____>
I tried Chrome, not Chromium
09:58
<hasather>
I tried both
09:59
<Ms2ger>
Still these, Chrome 25 dev on Linux64
10:00
<hsivonen>
are you running with Unity global menu or without?
10:00
<hasather>
hsivonen: without
10:00
<hsivonen>
I see the Encoding submenu in the View menu on Mac
10:00
<hsivonen>
no Tools menu on Mac
10:01
<hasather>
hsivonen: I use the menu to the right of the address bar
10:01
<Ms2ger>
GPHemsley, what's Firefox 3.6? :)
10:02
<hsivonen>
hasather: ooh. *that* Tools
10:02
<hsivonen>
hasather: thanks
10:02
<hsivonen>
I have another different Tools menu in the Unity global menubar
10:02
<hsivonen>
kudos to Google for burying this UI
10:03
<hasather>
hsivonen: Ah, didn't think of that, I'm not touching Unity :)
10:03
<hsivonen>
Firefox with the Firefox button enabled buries it in an illogical place
10:03
<hsivonen>
(Under a Web Developer menu)
10:04
<smaug____>
well, that is pretty similar to chrome's tools
10:04
<hsivonen>
(Web developer had better declare their encoding instead of using the menu!)
10:04
<hsivonen>
*developers
10:04
<smaug____>
true
10:04
<hsivonen>
the reason the menu exists is for *users* to deal with developers
10:05
<smaug____>
obviously no one has thought about the menu entry because no one ever uses it
10:05
<hsivonen>
right.
10:05
<hsivonen>
I want to add some telemetry so that we could get data for killing the menu
10:06
<hsivonen>
(now that I’ve spent time making the menu be less bad)
10:13
<smaug____>
hsivonen: problem is that it can be very useful for someone
10:13
<smaug____>
yet most of the users don't even know it is there
10:16
<darobin>
Ms2ger: well, at this point it's just one submission, but with luck we can get that grow
10:17
<darobin>
richt: just the man I was looking for! is there an Opera build with the Discovery API in it?
10:18
<darobin>
I'm looking through labs but the site is somewhat broken
10:18
<richt>
yep there is. you want the link?
10:19
<richt>
darobin, I sent you it on a private channel.
10:20
<darobin>
richt: thanks man!
10:21
<darobin>
ah, the site seems to work again, too — cool
10:33
<jgraham>
darobin: Nice :)
10:34
<Ms2ger>
Hi jgraham
10:34
<Ms2ger>
Do you want to do me a little favour? :)
10:35
<Ms2ger>
Apparently Presto does something with <ul align=center>, which it shouldn't
10:35
<Ms2ger>
Want to file a bug for me? :)
10:35
<darobin>
Ms2ger: do you have a recommended hotel to stay at near the Mozilla office in MV?
10:36
<Ms2ger>
Not personally, but I can ask around
10:36
<darobin>
Ms2ger: it's not a big deal, I can just pick something not too far from the offices
10:37
darobin
hates the valley
10:37
<Ms2ger>
holiday inn express, I'm told
10:37
<Ms2ger>
I haven't been to California this century, myself
10:38
<Ms2ger>
Apparently on walking distance
10:38
<Ms2ger>
darobin, ^
10:39
<darobin>
Ms2ger: excellent, thanks!
10:39
<smaug____>
yeah, 15mins walk or something
10:39
<Ms2ger>
Np
10:39
<darobin>
anything that involves not having to drive sounds like heaven
10:49
<smaug____>
darobin: oh, is that hotel closed
10:49
<smaug____>
google maps at least says so
10:49
<smaug____>
closed or moved
10:49
<darobin>
smaug____: ah, damn, I'll check that when I book (which won't be this second)
10:51
<smaug____>
Hotel Avante is a bit farther away, 25mins walk
10:55
<darobin>
smaug____: thanks, I'll check that out if the other one has indeed vanished
11:35
<MikeSmith>
darobin: plh will have a car when he's there I'm pretty sure
11:35
<MikeSmith>
can ride with him
11:36
<darobin>
MikeSmith: ah, good point
11:40
<hsivonen>
darobin: at this time of year, you’d need heating in Hotel Avante and the heater is noisy
11:41
<darobin>
hsivonen: is it noisier than me snoring?
11:41
<darobin>
if plh has a car, I might just go to whatever hotel he's staying at
11:41
<hsivonen>
darobin: I haven’t heard you snoring, but probably yes
11:41
<darobin>
that would be very loud
11:41
<darobin>
as annevk knows :)
11:42
<hsivonen>
darobin: the heater has a fan that blows the warm air into the room
11:42
<darobin>
yeah, I'm familiar with that setup
12:42
<MikeSmith>
hsivonen: so moving of http://www.w3.org/1999/xhtml/ is hindered by the fact that we are still using CVS to maintain the source
12:43
<MikeSmith>
and there's no way to delete directories in CVS
12:43
<MikeSmith>
well there is
12:43
<MikeSmith>
but not from the client side
12:44
<MikeSmith>
and actually even if I removed that dir on the cvs server I don't think it will get removed on the Web server
12:45
<MikeSmith>
so.. I'll need to ask the system team for some help with it I guess
12:54
<hsivonen>
MikeSmith: ah CVS. the fitting companion to Namespaces. :-(
12:54
<darobin>
we're removing that page? why not turn it into something more useful?
12:55
<MikeSmith>
darobin: was just trying to move it to be without the slash
12:55
<MikeSmith>
i.e., to make the actual Namespace URI
12:56
<darobin>
it's not as if millions of URLs pointing to it were going to suddenly cry out in terror, and be suddenly silenced
12:56
<darobin>
oh dear me, that's a bad mistake...
12:56
<MikeSmith>
hsivonen: actually it's even better than CVS and Namespaces
12:56
<MikeSmith>
hsivonen: because we make heavy use of conneg
12:57
<MikeSmith>
so it's the best all worlds
12:57
<darobin>
I wonder if that couldn't save us here
12:58
<darobin>
well, not conneg specifically
12:58
<darobin>
but maybe some modrewrite magic would work
12:58
<MikeSmith>
CVS rubbing elbows with conneg to serve of deferenceable documents for Namespace URIs
12:58
<darobin>
redirect xhtml/ to xhtml, and ensure that xhtml serves (through internal redirection) whatever else
12:58
<darobin>
you should throw some PHP in there
12:58
<MikeSmith>
darobin: I tried a little of that already but don't work
12:58
<darobin>
damn
12:59
<MikeSmith>
could be I did it wrong though
13:05
<Ms2ger>
A bit of CVS, a bit of PHP, a bit of Namespaces and a bit of conneg
13:05
<Ms2ger>
Sounds like the W3C
13:05
<Ms2ger>
Either that or the Divine Comedy
13:05
<MikeSmith>
we aim to please
13:06
<MikeSmith>
and we don't disappoint
13:06
<MikeSmith>
as far as providing pure entertainment value for your dollar
13:08
<hsivonen>
MikeSmith: how come conneg needs a slash?
13:09
<MikeSmith>
it doesn't
13:09
<MikeSmith>
that part is not not currently using conneg
13:10
<MikeSmith>
but if I were to put a file in there named xhtml.html it would show up as http://www.w3.org/1999/xhtml
13:10
<hsivonen>
I see
13:12
<MikeSmith>
Ms2ger: it's more of The Good Soldier Schweik than the Divine Comedy
13:14
Ms2ger
waves at darobin
13:20
<darobin>
Ms2ger?
13:21
<Ms2ger>
Oh, nothing, just a PR :)
13:21
<hsivonen>
so do we still think that letting some ampersands followed by ASCII letters not be errors is a good idea?
13:22
<Ms2ger>
I don't, if I counted the negatives correctly :)
13:24
<hsivonen>
I don’t either. So do I r+ a patch to match the spec while waiting for Hixie to change the spec?
13:28
<Ms2ger>
The validator is more strict than the spec right now? I'd r-
13:30
<hsivonen>
Ms2ger: both the validator and View Source in Firefox
13:48
<hsivonen>
stuff I learned today: don't view a patch in gmail and save as. Use the download link in the Gmail UI instead.
13:49
<hsivonen>
save as makes the patch fail to apply
15:45
<Ms2ger>
darobin: I'm told: '"holiday inn express Mountain view" is now called "Hotel Strata"'
15:52
<zewt>
hsivonen: is it expanding tabs or something?
15:52
<Ms2ger>
zewt, introducing CRs, possibly
15:52
<zewt>
(lately i've stopped using hard tabs in code completely; call them a long failed experiment)
15:52
<zewt>
(also it's easier to convince people to not use tabs than to fix broken ts=4 defaults in their editors)
16:07
<hsivonen>
zewt: I did not diagnose
16:09
Ms2ger
looks for an Opera person
16:11
<hasather>
Ms2ger: can I help?
16:12
<Ms2ger>
hasather, https://bugzilla.mozilla.org/show_bug.cgi?id=831232#c2 ; consider yourself informed :)
16:13
<hasather>
Ms2ger: alright :)
16:31
<dglazkov>
good morning, Whatwg!
16:35
<nimbu>
hey dglazkov you should do some PR with @stshank
16:35
<nimbu>
w.r.t web components
16:39
dglazkov
does PR
16:39
<dglazkov>
how do I do PR, btw?
16:39
<dglazkov>
and who is @stshank
16:41
<nimbu>
dglazkov: he is a tech journalist afaik
16:41
<nimbu>
@stshank on twitter
16:41
<nimbu>
dglazkov: you do PR by being your smooth suave self!
16:41
nimbu
may not know how to spell suave
16:41
<dglazkov>
yay shampoo!
16:42
<nimbu>
dglazkov: https://twitter.com/stshank/status/291511013612867586
16:42
<nimbu>
ahahah
16:44
<dglazkov>
twittered
16:52
<nimbu>
very nice dglazkov
17:13
<annevk>
So soon we can write cursor:progreß?
17:14
<tantek>
oh no
17:14
<annevk>
This whole lets use Unicode case folding for language identifiers madness got to stop.
17:16
<tantek>
yeah
17:23
<espadrine>
annevk: reading through the URL parsing algorithm, it seems that URLs like "http:news.ycombinator.com" (without the double slash) yield to a bogus state in which we keep adding to the buffer and don't end the algorithm properly (even though there is a parse error).
17:23
<espadrine>
annevk: Is there any reason a scheme-relative URL must start with //? Browsers seem to accept it when there is no double slash, and I recall that TBL considered it unfortunate since it doesn't serve a purpose.
17:23
<espadrine>
annevk: Also, what does it mean to have both a parse error and to continue the algorithm?
17:24
<annevk>
It depends on the base URL what such a URL will do.
17:25
<annevk>
Parse errors indicate issues with the syntax, but not all such issues result in total failure.
17:25
<annevk>
Most in fact, do not, and are happily "ignored".
17:26
<annevk>
espadrine: So in parsing your URL, you hit step 2 in http://url.spec.whatwg.org/#scheme-state
17:27
<annevk>
espadrine: Then it either goes to relative or authority state or authority first slash state
17:28
<espadrine>
in this case, it ends up in the "authority state", and "appends c to buffer" until buffer is empty. Should we assume that then, the algorithm stops? What is the behaviour when the pointer goes beyond the end of the buffer?
17:29
<espadrine>
in any case, I believe it makes the algorithm fail where browsers currently succeed
17:31
<annevk>
it appends to buffer until you hit step 2
17:31
<annevk>
the EOF code point
17:32
<annevk>
at which point it reprocesses the string
17:34
<espadrine>
ah, right, my bad. So it does work.
17:36
<annevk>
Hopefully :-) I very much appreciate you double checking!
17:36
<espadrine>
:)
18:00
<volkmar>
Hixie: I think valueAsNumber and valueAsDate need some love
18:00
<volkmar>
every time I look at them, I feel like some things are missing
18:01
<Hixie>
send feedback or file a bug, listing use cases :-)
18:01
<volkmar>
Hixie: just a quick question before doing that
18:02
<volkmar>
Hixie: on setting valueAsNumber
18:02
<volkmar>
if the algorithm to convert to a string doesn't work
18:02
<volkmar>
what should happen?
18:02
<volkmar>
my code does set the value to the empty string
18:02
<volkmar>
but I can't find if I decided that or if it is written somewhere
18:04
<Hixie>
doesn't the spec answer?
18:04
<Hixie>
(i don't know off hand, would have to just read the spec to answer)
18:05
<volkmar>
Hixie: no the spec say to run the algorithm and set the value to the returning string
18:06
<volkmar>
but the algorithm, for time for example, is something like "set a valid time string that represents the input milliseconds..."
18:06
<volkmar>
there is no error handling
18:07
<volkmar>
and for time, it's quite easy to imagine a number that isn't working (like anything greater or equal to 24*60*60*1000 or a negative number)
18:12
<Hixie>
hm, let me see...
18:12
<Hixie>
i'm confused
18:13
<Hixie>
why would the algorithm not work?
18:13
<Hixie>
what value can you set valueAsNumber to that would fail?
18:14
<volkmar>
Hixie: what .valueAsNumber = -1; should do?
18:14
<volkmar>
it type='time'
18:15
<volkmar>
iow, what would .value return after that?
18:16
<volkmar>
Hixie: actually, the |create a Date object| concepts solve the issue of float numbers
18:16
<Hixie>
"23:59:59", no?
18:17
<volkmar>
Hixie: why?
18:18
<volkmar>
I mean, it's a way of seeing it
18:18
<Hixie>
valueAsNumber applies, so it doesn't throw InvalidStateError. valueAsDate applies, so you use the first branch. create a Date object with value -1, which is one second before 1970-01-01. Convert that to a string, which returns a valid time string for that time, which is 23:59:59. Set the value to that string.
18:19
<Hixie>
seems pretty unambiguous in the spec
18:19
<Hixie>
(s/for that time/for that time in UTC/)
18:20
<volkmar>
ok, I see
18:20
<volkmar>
this branch to Date is just odd
19:29
<Ms2ger>
Quoting Feynman on www-style? Would upvote.
19:31
<jgraham>
"Let's work in a lap dancing club"?
19:33
<Ms2ger>
jgraham, btw, did you file a bug for <ul align=center> or should I?
19:35
<jgraham>
I didn't
22:12
<Hixie>
anyone have test cases for the kind of thing discussed here?: http://www.w3.org/mid/DF0A84C4-AEAB-4716-B23F-FB3BA48BDE3C⊙gc