01:17
<afa>
Chrome has rolled in a preliminary WebForms validity API that has broken my site
01:17
<afa>
i'm going to report it as a bug, but i want to make sure i'm correct before i do
01:17
<afa>
if a radio or checkbox group has an element checked, the whole control should be valid, shouldn't it
01:18
<afa>
where "whole control" means every input with the same name
01:20
<othermaciej>
afa: you're using the "required" attribute?
01:21
<afa>
yeah
01:21
<othermaciej>
I'm checking what the spec says
01:21
<othermaciej>
if it's not an implementation bug then it's clearly a spec bug
01:22
<afa>
thanks
01:23
<othermaciej>
afa: for checkboxes, the spec says it is based on just the element's checkedness
01:23
<othermaciej>
http://dev.w3.org/html5/spec/Overview.html#checkbox-state
01:23
<othermaciej>
So I guess a "required" checkbox has to be checked, regardless of the group
01:24
<othermaciej>
afa: for radiobuttons, it is supposed to be per group
01:24
<othermaciej>
http://dev.w3.org/html5/spec/Overview.html#radio-button-state
01:26
<othermaciej>
afa: according to that spec, if any radiobutton in the group has checkedness of true, validation should pass
01:28
<afa>
sorry for wandering off
01:28
<afa>
that link crashes opera (and my IRC session)
01:28
<afa>
so anything you said after that, i didnt get
01:29
<othermaciej>
afa: eek, sorry
01:30
<afa>
it's been a while since i worked on this app, so i want to make sure my memory is correct
01:31
<afa>
if you POST a checkbox, then multiple identical keys get sent for each checked checkbox with the same name, correct?
01:31
<afa>
so if c1s1q1 had values 0 and 1 checked,
01:31
<othermaciej>
afa: short version, for checkbox, "required" requires each checkbox to be chacked, regardless of "name", for radio buttons, it's per group
01:31
<afa>
c1s1q1=0&c1s1q1=1 would be POSTed?
01:32
<othermaciej>
I think that's right but I don't really know
01:32
<afa>
i know when i first brought this up, you said this may be a spec bug
01:33
<afa>
and i know that the unintuitive parts of HTML are often there for a reason (perhaps legacy cruft, perhaps others)
01:33
<othermaciej>
if it was true for radio buttons, yeah
01:33
<afa>
so i want to discuss this before i comment on the RFC
01:33
<othermaciej>
@required isn't legacy
01:33
<afa>
i know, but i know in other cases, there have been reasons for the unintuitive parts
01:33
<afa>
so i want to make sure this isnt the case :)
01:34
<othermaciej>
I think for radio buttons it works as you expect, for checkboxes, what the spec says sounds reasonable to me
01:35
<afa>
the radio bug is a problem in chrome
01:35
<afa>
the checkbox is a conflict between my JS implementation and the docs
01:35
<afa>
i may comment on that part of the spec just to get feedback
01:36
<afa>
on its face, it's not wholly unreasonable
01:36
<afa>
but it seems to me that a group validation is more flexible and consistent
01:36
<afa>
so any multiple choice widget, whether mutually exclusive or not, has the same API
01:36
<afa>
if you want to require a box to be checked, put it in its own group
02:21
<afa>
i submitted a comment on the checkbox validity API:
02:21
<afa>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9160
02:21
<afa>
i'd love to hear other's input
03:35
<drclue>
It is amazing to see Microsoft's reaction to the serious threat that HTML5 has to the OS centric market place after the acquisition of On2. The tactic of using a hobbled Internet explorer to hold back Internet centric applications is finally about to fail and they know it. That's what this is all about. The Microsoft dinosaur is thrashing around in it's tar pit. Microsoft has a lot of fellow reptiles in this tar pit , so I hope we get HTM
03:36
<drclue>
The reptiles will continue to mass. Microsoft , Adobe, Xerox and anyone else whose lock is about to be picked
03:37
<drclue>
Rupert Murdoch is worried about his lunch too
03:43
<drclue>
Sports fans the clock is ticking. If HTML5 with CODEC makes it out there in the next few months the battle is won. If the process of integration and distribution languishes in vacillations , there are indeed enough dinosaurs out there , that when arrayed can sap even the might of Google and HTML5. I might seem to be the ranting member of the tin hat squad, but multi national businesses are real. Their interests in various media locks ar
03:50
<drclue>
What was a simple quest to improve an Internet standard , has wide ranging consequences , perhaps wider ranging than many have realized.
03:50
<drclue>
It's no longer just the university soda machine or a novelty toaster , but your TV , Fridge , car , telephone , very aspect of your life. That is a lot of money, and attracts the attention of a lot of very powerful people. HTML5 along with it's video component represents a major threat to the alternatives to same in every field mentioned and mant more. We are speaking of trillions of trillions of dollars, all over what seems a tiny litt
03:53
<drclue>
Attacking big firms like Google has status , and properly crafted FUD and FNORDS will allow the leveraging of the dinosaurs to achieve the goals of political maggots everywhere.
03:54
<drclue>
I hope that someone will step up and calm my fears with rational discourse that lets me know that my fears are simply un-founded.
03:54
<drclue>
Will of th e104 people here , even one stand up and make the case that this is not a real issue?
03:56
<drclue>
Not even one of the 104 people here feel it worth saying I'm wrong , that the current attack on Google is really a fear of HTML5
04:00
<miketaylr>
or, it's late friday evening/early saturday morning for everyone
04:00
<drclue>
It would be a false assumption to consider lack of comment agreement , so I ask it the other way,
04:00
<drclue>
does anyone here feel that recent actions by Microsoft, Xerox,Adobe and other firms via litigation , government manipulation
04:00
<drclue>
and other methods might be born of fears that an Internet Applications Platform might threaten their business models?
04:00
<miketaylr>
so party, on.
04:02
<drclue>
Not sure if that comment was towards me or not , but if it were , I would consider same to be in the context of
04:02
<drclue>
of a virtual multi-alarm fire and th idea that we should but party on , not be a valid answer to the fire
04:09
<drclue>
I've been programming since before the first PC , and certainly before Microsoft appeared on the scene. I've experienced
04:09
<drclue>
first hand many pivotal points in technology. One can research my comments throughout the history of the Internet and
04:09
<drclue>
if they have access to RYBBS and FIDO net historical data my pre-Internet expressions when communications were limited
04:09
<drclue>
to at my beginnings 150 bits/sec. I'm not often wrong nor do I often make this loud a noise
04:30
<drclue>
Come on folks , either agree that HTML5 is a threat to the powers that be and that the current responses of Adobe, Microsoft , Xerox and indirectly Rupert Murdoch and other powers are a serious threat , or call me out as the fool. Or at least say something. This is a real issue IMHO and should be seen as a call to action. If you consider it to be a yawn fest , say so.
04:30
<drclue>
HTML5 is a major seed change in the way businesses address applications development. Those changes have a major effect on
04:30
<drclue>
the bottom lines of trillions of dollars in business. Everyone just too focused on a weekend of party hardy? People think that Google my favorite champion is bullet proof? What is it?
04:32
<drclue>
AOL and Compuserve were considered by many pendents at the time to be bullet proof.
04:36
<drclue>
IMHO Microsoft is a POS , but in this case it is not simply Uncle Bill , nor even computer or Internet centric businesses that have skin in the game. This means that folks you have never even heard of that have significant market power are taking aim.
04:40
<drclue>
I have on multiple occasions suggested that Google implement a search engine that harvests sec.gov and similar global representations of corporate family trees to allow folks to see who owns who and to understand in an easy way the behaviors of various corporate and news entities. A single link to a child's mobile of relationships would prove my point
04:42
<drclue>
Having to dig about to know that Rupert Murdoch owns MySpace.com and has significant interests in Hulu.com
04:42
<drclue>
should not be a thing I need to prove by elaborate means
04:43
<drclue>
To show the interactions between Rupert Murdoch and Microsoft , Direct TV , NBC and such is not something I should need to prove
04:44
<drclue>
Simply citing a google link would make this easy
04:46
<drclue>
Should I need to prove
04:46
<drclue>
1.) Microsoft ownes "Ciao" a German search site. A complainant in the EU anti rust case against Google
04:46
<drclue>
2.) "Foundem" member of Microsoft funded ICOMP (Initiative for a Competitive Marketplace)
04:46
<drclue>
Should I need to prove the other links from Microsoft to the each and every issue at hand
04:47
<drclue>
HTML5 and the CODEC VP8 need to resolve themselves in weeks not months
04:50
<drclue>
Should I need to re-prove the findings of governments around the world regarding Microsoft?
04:52
<drclue>
HTML5 is indeed the future of the Internet, but don't be so arrogant as to believe that life fixes itself
04:52
<drclue>
that good by itself defeats evil. Internet standards are a stdrong position
04:53
<drclue>
But a camp fire no matter how well made , left undefended is defeated by but a small amount of even polluted water
04:55
<drclue>
OK , the audience went from 104 to 102 IRC viewers. Stop jacking off at porn and say something, or do you want to pay evil companies for your right to even make a living?
04:59
<drclue>
Silence is your support of behaviors of companies like Microsoft,Adobe, Xerox and others and your desire to be their slaves.
04:59
<drclue>
If your of little concern for your own worth , I'm sure we can fasten your mouth to a hole in a bathroom stall
05:01
<drclue>
Come on kids , this is not a Google or Microsoft sanctioned discussion room.
05:01
<drclue>
"Ian Hicks" will not care if we say some word of disconcerting nature.
05:02
<drclue>
Do you not have an opinion? This does indeed effect your livelihood.
05:03
<drclue>
Tell me I'm full of shit , I dare ya. Do I need to get a Microsoft fanboy in the room to gets some discussion?
05:03
<drclue>
Is everyone that chicken shit or ignorant to how this effects their life to care?
05:10
<drclue>
Fine , than all of you who will in over 100 people not take a position deserve what ya git, but
05:10
<drclue>
I for one will be looking in each and every breath to how the peoples Internet , paid for by them can continue to be fore them
05:10
<drclue>
and will in the event that the evils are set upon their heels cite the 103 people here as those who left it to others to make sure that the Internet is as relates to standards owned and controlled by the people and not the fab boy and moneyed personality of the moment (infrared laser scanners and other tools resumed and first occurring and detected , deflected)
05:12
<drclue>
Two hours wasted here , but perhaps not. I checked to see if folk could drop their pudd and gained that valuable
05:12
<drclue>
answer
05:38
<mpilgrim>
opera 10.5b2 is better behaved with the dynamic web fonts on http://diveintohtml5.org/ but still not perfect
05:39
<mpilgrim>
the initial page load is fine, but navigating to different pages slowly makes the fonts smaller and smaller
05:39
<mpilgrim>
very strange bug
05:39
<mpilgrim>
easily reproducible, i hope
05:40
<mpilgrim>
email me if you need further details
06:11
<TabAtkins>
... What the fuck was drclue *smoking*?
06:12
<drclue>
I have been told via official venues that HTML5 is a religious / activist venue basically around the globe. and that in further questioning that the platform is covered by various terrorist directives
06:12
<othermaciej>
TabAtkins: you woke it up!
06:12
<Hixie>
we have official venues?
06:13
<TabAtkins>
OH GOD I DID IT I'M SO SORRY MACIEJ
06:13
<Hixie>
btw anything important in my e-mail? i haven't had time to look yet and i'm probably not going to until tomorrow, so if there is, let me know :-)
06:14
<TabAtkins>
I blame the CSS in my IRC. I make joins/quits almost invisible so I can normally ignore them.
06:14
<drclue>
Not just some middle east thing , but any venue violating dividends with no data to support some other perception
06:15
<TabAtkins>
On the other hand, now I get to enjoy a nice word salad, I guess.
06:16
<drclue>
Fuck all this , if nobody can challenge or even agree I I'm out of here as I have far more worthy folks to give the word
06:16
<TabAtkins>
Have fun, drclue.
06:16
<TabAtkins>
So, in other news, *finally* got my damned diploma in the mail today.
06:17
<drclue>
Geeze [TabAtkins[ actually says even something, and at last gets some warm thoughts from me
06:18
<othermaciej>
Hixie: we sent the TransReqs for our FPWDs
06:18
<drclue>
Not sure what to think vs mail and stale
06:19
<othermaciej>
Hixie: will send pubreq once those are approved
06:19
<Hixie>
othermaciej: k
06:20
<othermaciej>
TabAtkins: grats re diploma
06:20
<drclue>
If hixie wants to say something , gold , anyone wants to claim Hixie fuck ya
06:20
<othermaciej>
Hixie: also Tim officially said HTML5 is in scope for HTML5
06:21
<othermaciej>
(dunno if you were around for that one)
06:21
<drclue>
Hixie is not particularly god , but of less mind to be a pain than those making noise
06:22
<drclue>
Look , Microsoft is dying unless they can trash us , so what ya got to say?
06:40
<drclue>
So really , I need to on my own shoulders Internet freedom from a word
08:01
<GPHemsley>
I feel like quoting TabAtkins, but I'd probably just wake up the dragon again
08:01
<GPHemsley>
(Not that it matters, since I'm going to sleep now, myself.)
08:45
<annevk>
Hixie, maybe we should just push for another port again?
08:45
<annevk>
with web sockets
11:19
<Philip`>
JonathanNeal: Your while loops look wrong
11:19
<Philip`>
while (--i) only loops i-1 times
11:19
<Philip`>
You should write normal for loops rather than trying to be clever with micro-optimisations that will make no practical difference :-)
11:45
jgraham
would like to complain about the snow
11:45
<Dashiva>
Why?
11:46
<Dashiva>
Snow is vastly superior to ice and sludge
11:46
<jgraham>
Because yesterday it was raining and washing the snow away. Today it is snowing again
11:46
<Dashiva>
You should complain about the rain then :)
11:46
<Dashiva>
It's too early for spring yet
11:46
<jgraham>
So there will be more ice and sludge in the future compared to it raning again and wasking more snow away
11:49
<Philip`>
You should complain about the ground - if there wasn't any then the snow couldn't build up
12:04
<jgraham>
I expect such a complaint would be dismissed as groundless
12:04
<jgraham>
(sorry)
13:18
<Philip`>
Someone mentioned http://arstechnica.com/open-source/news/2010/02/ogg-theora-vs-h264-head-to-head-comparisons.ars here recently
13:18
<Philip`>
A relevant comment from http://x264dev.multimedia.cx/?p=292#comment-2768 : "We talked about that on IRC, both with x264 and Theora people: all considered it one of the worst articles they had ever seen."
13:19
<Philip`>
So probably best not to pay too much attention to it
13:37
<Philip`>
Hmm
13:37
<Philip`>
Chrome thinks my test page full of "ABCD" is written in Danish and offers to translate it
14:08
<Dashiva>
>it should explicitly say alt="" MUST NOT be blank unless it MAY be left blank
14:09
<Dashiva>
Isn't that implied? Or am I misunderstanding how the terms work
14:12
<Philip`>
Implied requirements are dangerous
14:13
<Philip`>
It seems generally better to be explicit
14:19
Philip`
doesn't like "x MAY y if z" because he's never quite sure if the intention was "x MUST NOT y if not z", or "x MUST NOT (not y) if z", or something else
14:20
<Philip`>
although in the case of <area> it already says it must be text that's somehow equivalent to the image area, so I don't see why it'd need to be stated again
14:33
<Dashiva>
For additional fun, write requirements on the form "x MAY not y"
14:34
<Philip`>
And "x MUST only y if z"
14:35
<Philip`>
and "x MUST y only if z"
14:36
<Dashiva>
Maybe all normative criteria should be written using some kind of formal logic notation
14:40
<Philip`>
RFC2119 is fine as long as you use the words in simple sentences ("if z, then y MUST x", "if not z, then y MUST NOT x", etc)
14:40
<Philip`>
and I think HTML5 generally does that quite consistently now
14:55
<Philip`>
Oddly, Opera kerns text with strokeText on Linux but not with fillText
14:55
<Philip`>
and with neither on Windows
15:06
<annevk>
hmm, Ubuntu says my harddrive is getting sucky
15:07
<annevk>
and video playback/downloading does seem to fail way more often
15:07
<annevk>
grmbl
15:07
annevk
wonders if it's easy to replace a hard drive
15:07
<gsnedders>
In your ThinkPad?
15:10
<annevk>
yes
15:11
<annevk>
this sucks a bit, I was planning on seeing some of the new Damages episodes today
15:11
<gsnedders>
annevk: ThinkPads all make it fairly easy to change HD, AFAIK
15:13
<annevk>
hmm, guess I have to burn some Ubuntu installer disk as well
15:13
<annevk>
meh
15:15
<annevk>
would be nice if I could just put a SSD in
15:25
<annevk>
hmm, those are not cheap
15:29
<Philip`>
Is there a reliable way to make sure a @font-face is loaded before trying to draw with it?
15:29
<Philip`>
Firefox and Opera seem to block onload as long as the font is used in a (non-display:none) element within the page itself, so that's okay
15:30
<Philip`>
Chromium doesn't, so it randomly works only half the time
15:30
<Philip`>
and if I do something like draw with the font, then wait a bit to make sure it's really loaded and draw again, Chromium crashes
15:30
<Philip`>
so that's not so good
15:31
<Philip`>
and I'm not sure how else to do it
15:42
<Philip`>
Oh, seems to work better if I actually use characters from the font in the page
15:57
<gsnedders>
Philip`: Detect the character width?
16:01
<Philip`>
gsnedders: That would only help detect whether the font has been loaded, and wouldn't cause it to become loaded
16:02
<gsnedders>
Philip`: position: absolute; visibility: hidden; top: 0; left: 0; I guess would cause it to load
16:02
<gsnedders>
Philip`: Without interfering with anything
16:07
<Philip`>
That seems to work
16:07
<Philip`>
(I was doing opacity:0 which also seems to work)
16:08
<Philip`>
with a 0.5s delay before drawing to make sure it's really loaded
16:08
<Philip`>
because WebKit doesn't delay onload, it seems
16:20
<Philip`>
Opera doesn't even get textAlign correct :-(
18:20
<Philip`>
Current canvas text-drawing test scores:
18:21
<Philip`>
Opera is trailing with 52%
18:21
<Philip`>
Chromium is a bit better at 56%
18:21
<Philip`>
Firefox is racing into the lead with a staggering 61%
18:21
<Dashiva>
Awesome performance by Firefox
18:22
Philip`
gets bored and decides he has enough tests for now
18:22
<Dashiva>
Reminds me of a survey done here in Norway recently, which asked which dialect was the sexiest, and proclaimed that the northern areas were the winner.
18:22
<Dashiva>
Now, the northern area dialect had got a whooping 12% of the votes
18:23
<Dashiva>
And the least popular one had 7%
18:28
<annevk>
Philip`, are you uploading the new version of the test suite?
18:33
<Philip`>
annevk: I will do once I've fixed some other broken tests
20:54
<Philip`>
http://philip.html5.org/tests/canvas/suite/tests/results.html
20:54
Philip`
can't be bothered to run them all on Windows too
20:55
<Philip`>
It's possible some tests are buggy - if anyone notices any, please let me know
20:55
<Philip`>
(Also there's some things I haven't tested at all yet, like <video>)
20:56
<Philip`>
annevk: ^
20:57
<Philip`>
(Also, actual tests at http://philip.html5.org/tests/canvas/suite/tests/)
20:58
<Philip`>
(Maybe the results page should link back there...)
21:03
<JonathanNeal>
Hello all!
21:05
<Philip`>
JonathanNeal: Hello, and http://krijnhoetmer.nl/irc-logs/whatwg/20100227#l-282
21:06
<JonathanNeal>
Philip`, I don't quite follow.
21:06
<annevk>
Philip`, oh, so Opera has the highest score rather than the lowest
21:07
<annevk>
seems like a lot of trivial bugs we ought to fix
21:07
<JonathanNeal>
Oh I see it.
21:13
<Philip`>
annevk: Highest if you count all of the tests, and weight them all equally
21:15
<Philip`>
(and if I don't have serious bugs in my test runner)
21:16
<Philip`>
Most are trivial and boring, though there's some like http://philip.html5.org/tests/canvas/suite/tests/index.2d.text.draw.baseline.html where it might be nice if top/bottom/middle worked properly
21:18
<Dashiva>
Philip`: Maybe this is the kind of stuff Larry was talking about as overspecifying
21:18
<Philip`>
Non-existence of createImageData is maybe the biggest thing
21:20
<Philip`>
Dashiva: Maybe, but they're almost all things that could come up in practice
21:22
<Philip`>
I'd probably call it "sensible, unambiguous, testable level of specification" rather than "overspecification"
21:24
<Dashiva>
I agree personally, but minor bugs that don't get fixed even over time may suggest otherwise
21:33
<Philip`>
Dashiva: It may suggest that expecting implementations to have 100% conformance is unrealistic and not a good use of anybody's time (and so the Rec criteria are broken and harmful), but problems are much more likely to get fixed if they're clearly specified and tested than if they're not
21:33
<Philip`>
so I don't think it suggests that it's not good to specify all the behaviour