00:07
AryehGregor
eventually remembers that "object" is JavaScript-speak for "associative array"
00:13
<othermaciej>
AryehGregor: I'm not sure if that's a totally correct way to think about it - objects also have object-like aspects to them, besides the fact that their set of properties is open-ended
00:14
<AryehGregor>
Yes, true. But in this context it just means "associative array", pretty much.
00:14
<othermaciej>
the buzzword "Web OS" makes me want to stab myself in the eye
00:14
<AryehGregor>
I'm not used to languages with a generic Object type that you can stick random attributes on.
00:14
AryehGregor
takes sharp objects away from othermaciej
00:15
<tantek>
IMHO "Web OS" is a far more palatable (and useful) buzzword than "Object Oriented"
00:15
<othermaciej>
AryehGregor: isn't basically the same thing true in Python and Ruby?
00:16
<othermaciej>
tantek: I like "Web Platform" or "Open Web Platform" or "Standards-Based Web Platform"
00:16
<AryehGregor>
Python has a dedicated dict type that you'd use for this kind of thing. (If you didn't use named parameters, that is.)
00:16
<othermaciej>
I don't think "Web OS" means anything
00:16
<tantek>
othermaciej - good points. "platform" is more correct
00:16
<othermaciej>
AryehGregor: I understand that it has a dict type, but as I understand it Python objects are also dictionaries under the covers
00:16
<AryehGregor>
I think you could do something like foo = Object(); foo.bar = "baz" and so forth, but I don't think that's possible in any way that's sensible for passing to an argument.
00:16
<tantek>
the "Web OS" has nothing to do with the typical definition of an "OS"
00:17
<tantek>
of course the term "Web App" is just as bad, for the same reasons
00:17
<tantek>
and belies "OS"
00:17
<AryehGregor>
My problem is I'm not used to object literals, so when you say "pass an object to a function" I think about setting a variable to the object and assigning a bunch of stuff on different lines before passing it.
00:17
<tantek>
if there are "Web Apps" then surely they must run on a "Web OS"
00:17
<AryehGregor>
Is Chrome OS a "Web OS"? :)
00:18
<othermaciej>
AryehGregor: it's a "Cloud OS"
00:18
<othermaciej>
at least, that's what I read on TechCrunch
00:18
<othermaciej>
tantek: the way I understand it, applications run on a platform but that platform does not have to be an OS
00:19
<othermaciej>
for example, Java is a platform (though that's not necessarily the nicest comparison example)
00:19
<othermaciej>
Python plus its common extensions is arguably a platform of sorts
00:19
<othermaciej>
Many people think of "LAMP" as a platform, and while the acronym includes an OS, that's the least important bit from the point of view of the apps
00:20
<AryehGregor>
I'd say the web server is less important than the OS.
00:20
<AryehGregor>
For LAMP.
00:20
<tantek>
othermaciej - another good point. there are Java apps, but certainly the notion of a "Java OS" has failed pretty miserably
00:20
<AryehGregor>
L vs. W affects whether you, e.g., can shell out to Unix utilities and handy things like that.
00:21
<AryehGregor>
A vs. I vs. L vs. N vs. whatever doesn't really make a noticeable difference to the app author.
00:21
<othermaciej>
the Web server being Apache lets you implement parts of your app in the form of Apache config rules
00:21
<othermaciej>
(how much people do that, I dunno)
00:21
<AryehGregor>
Nobody I've heard of relies on that very heavily.
00:21
<AryehGregor>
Generally a few rewrites at most.
00:22
<othermaciej>
Linux vs. FreeBSD vs Mac OS X probably doesn't matter hugely in terms of shell capabilities
00:22
<othermaciej>
agreed Windows is a different world
00:22
<othermaciej>
AryehGregor: sure, but if there are even a few rewrites, you need to "port" that part of your web service to move to a different web server
00:22
<AryehGregor>
Usually the rewrites are nonessential, though.
00:22
<AryehGregor>
I'm not aware of any app that requires them.
00:23
<AryehGregor>
You just fall back to ugly URLs if the admin can't be bothered to take the extra step on installation.
00:23
<AryehGregor>
Of course, the same is normally true for use of shell utilities, but that can be a somewhat bigger deal.
00:24
<AryehGregor>
Certain features of MediaWiki will just fail if diff3 isn't available, for instance.
00:24
<AryehGregor>
Usually also not a huge deal, but I'd say slightly more. :)
00:24
<AryehGregor>
P > M > L > A
00:26
<AryehGregor>
Why are the two parts I dislike the most, the most important and least portable ones? :(
00:31
<Hixie>
Lachy: hard to miss that error in testing
08:07
<annevk3>
morning
08:07
<gsnedders|work>
mornin'
08:30
<annevk3>
hsivonen, Philip`, do you have any data related to encodings?
08:31
<gsnedders|work>
Philip` does
08:32
<annevk3>
ah yeah, just found http://philip.html5.org/data/meta-content-type.txt
08:32
<gsnedders|work>
http://philip.html5.org/data/charsets.html
08:33
<annevk3>
sweet
08:34
<hsivonen>
annevk3: what kind of data?
08:36
<annevk3>
unique encoding labels
08:36
<tantek>
anybody here have a particular opinions about the microdata predefined vocabularies?
08:48
<annevk3>
tantek, I like the idea, but I don't feel strongly either way
08:48
<annevk3>
hsivonen, actually, I guess ARIA role=heading will influence the eventual document outline
08:49
<annevk3>
hsivonen, so I think your critique was correct
08:50
<tantek>
annevk3 - the major issue (with the predefined vocabs) is that they are for the most part a static fork/snapshot of microformats vocabs (hCard and hCalendar), and are very much going to have all the bugs (issues in progress) that a snapshot typically does. and frankly this group/community lacks the direct experience and day-to-day usage to keep up with and properly resolve.
08:51
<tantek>
(lots of outstanding issues on hCard and hCalendar that have been challenging for *me* to resolve, and I invented them and have been living/breathing them for nearly 5 years now)
08:52
<gsnedders|work>
What was that downloadable web cache that Philip`was using?
08:52
<annevk3>
tantek, I thought they were snapshots of vCard and vEvent and such
08:52
<tantek>
anyway - I'm moving forward with producing hCard and hCalendar 1.0.1 specs which incorporate all issue resolutions, and have UA conformance requirements with similar wording to HTML5 - I can pretty much guarantee these will be both be better and better reflect the reality of web author usage than the "predefined microdata" section in HTML5
08:53
<tantek>
and will advocate dropping predefined microdata vocabs once hCard 1.0.1 and hCalendar 1.0.1 are published.
08:54
<tantek>
annevk3 - that's the problem, they were *snapshots*
08:54
<tantek>
and as such, rapidly went out of date
08:59
<hsivonen>
tantek: I like the idea of having non-dotted-token vocabs for common stuff
08:59
<hsivonen>
tantek: I'd be OK with microformats.org managing the non-dotted-token vocabs
09:23
mpt
wonders why the March 2009 mailing list archives are so incomplete
09:24
<gsnedders|work>
Because completeness is overrated
09:24
<mpt>
Oh, because I'm looking at the archives for the wrong mailing list, durrrr
09:36
<tantek>
thanks hsivonen - that's good to know.
09:36
<gsnedders|work>
Am I right in reading the spec to say that <table><tbody><td>foo will throw a parse error upon the td open tag being processed in the tree constructor?
09:36
tantek
returns to fixing bugs in predefined non-dotted-token vocabs.
09:39
<annevk3>
gsnedders|work, seems like it should
09:39
<gsnedders|work>
Oh wait, I'm being stupid. Ignore me.
09:39
<gsnedders|work>
Gah. Too late.
11:04
<Creap>
in http://www.whatwg.org/specs/web-apps/current-work/multipage/tabular-data.html#attr-th-scope that example, couldn't you say that Legs and Tails are also th's, for their corresponding row?
11:07
<annevk3>
in the actual table they are <th>
11:07
<annevk3>
just not in the example markup
11:07
annevk3
files a bug
11:08
<takkaria>
annevk3: I sent feedback on removing device-* from media queries, don't know when it'll arrive on list
11:09
<annevk3>
http://lists.w3.org/Archives/Public/www-style/2009Aug/0585.html
11:09
<takkaria>
oh, I was looking at the July archives, oops
11:09
<annevk3>
Creap, http://www.w3.org/Bugs/Public/show_bug.cgi?id=7440
11:10
<Creap>
thanks
12:39
<jgraham>
Is it defined somewhere what happens if I do Object.prototype.__defineGetter__("document", function() {return "foo"})?
12:41
<Dashiva>
Not in a single place
12:42
<jcranmer>
the proton streams get crossed
12:42
<Dashiva>
But what is the alternate behavior you're wondering about, jgraham?
12:42
<jgraham>
Dashiva: OK, well afaik __defineGetter__ isn't specified anywhere. But assuming it wasw, is there any wat to tell from a spec what will happen if I do the above and access window.document?
12:43
<jgraham>
i.e. whether I would get "foo" or a [HTMLDocument]
12:44
<Dashiva>
Oh, I see what you mean
12:45
<Dashiva>
window doesn't have a regular prototype chain
12:46
<Dashiva>
And in any case, window's own properties would be checked before reaching Object
12:46
<jgraham>
Hmm, true
12:51
<Dashiva>
Now, if you tried doing that on e.g. Document with documentElement, I think it would depend on whether documentElement is ReadOnly or not
12:57
<Philip`>
gsnedders|work: http://www.dotnetdotcom.org/ ?
12:57
<gsnedders|work>
Philip`: Ah, yeah, thx
12:57
<jgraham>
Dashiva: That was what I was thinking. So I started investigating…
12:59
<Dashiva>
Now, DOM 3 Core says documentElement is ReadOnly
15:52
<eighty4>
gsnedders|work: pong?
15:52
<gsnedders|work>
eighty4: pong
15:52
<eighty4>
*ping
15:52
<eighty4>
gsnedders|work: valla 18.00?
15:53
<gsnedders|work>
eighty4: yeah
15:54
<eighty4>
good good
15:55
<eighty4>
but don't expect any wonders, I'm way to tired for that.
15:56
<gsnedders|work>
:)
16:34
<eighty4>
gsnedders|work: still there?
16:34
<gsnedders|work>
eighty4: yes
16:34
<eighty4>
you're so not going to make it in time :)
16:34
<gsnedders|work>
eighty4: I'm so getting a lift :)
16:34
<eighty4>
pff
16:34
<gsnedders|work>
eighty4: You going from Ryd?
16:34
<eighty4>
yeah
16:34
<gsnedders|work>
That helps :)
16:35
<eighty4>
?
16:35
<gsnedders|work>
Well, it is closer :)
16:35
<eighty4>
right
16:35
<eighty4>
but you should see my bike
16:35
<eighty4>
or well... you'll see my bike
16:36
eighty4
s so hating that opera rounds % and that ie don't support display:table-row
16:36
<TabAtkins>
God, I know. Makes it difficult as hell at times.
16:36
<gsnedders|work>
IE8 supports display:table-row, no?
16:37
<TabAtkins>
Yes, but that's not something you can depend on for a widely-deployed page.
16:38
TabAtkins
really wishes for a display:table-row-start value that would make a table-cell, but in a new row.
16:38
<TabAtkins>
There's no reason I should have to wrap this stuff in a container just to make it multi-line.
16:39
<gsnedders|work>
Manipulating tables through DOM is horrible
16:41
<TabAtkins>
is there something special about tables that makes it bad, or is just dealing with the DOM?
16:43
<gsnedders|work>
Tables are just especially bad
16:44
<eighty4>
and now my server is playing with me :(
16:44
<eighty4>
http://play.broken8.se/box/index.php ffs
16:51
<eighty4>
there we go. I so want that to work in ie
16:52
<TabAtkins>
Possibly solution: pop some jQuery in a conditional comment that fakes it. This requires saying "screw you" to IE<8 users who have scripting turned off, but screw them.
16:54
<cardona507>
thats cool eighty4 - is it javascript? or php?
16:55
<cardona507>
ic now
17:30
<Philip`>
"... unless maybe the user has a terabyte of free disk space" - pfah, who *doesn't* have a terabyte of disk space these days?
17:31
<AryehGregor>
I have less than 200 GB free right now. :(
17:35
<TabAtkins>
i have 40gb free.
17:36
<TabAtkins>
this laptop's old, and i regularly fill it with movies.
17:36
<AryehGregor>
One of the servers I'm root on has over a terabyte free.
17:36
<AryehGregor>
But that's not really usual on home desktops yet. :)
17:37
<TabAtkins>
nod. 500gb is fairly common for new computers, or higher-end slightly older computers.
17:37
<TabAtkins>
but 1tb or more isn't common yet unless you've added the space yourself.
17:38
<AryehGregor>
500 GB total is common, 500 GB free isn't so common. Not after a while's use, anyway.
17:39
<AryehGregor>
I have over 500 GB total here.
17:39
<TabAtkins>
ok, yeah
17:39
<AryehGregor>
. . . actually, right now it's about exactly 500 GB total, since I'm using my second disk for RAID 1.
17:40
<TabAtkins>
next birthday: new computer, so I'm not relying on a 5-year old laptop for all my personal computing.
17:44
<Philip`>
1TB disks cost about 20% more than 500GB disks when I last checked
17:44
<Philip`>
(They were external disks, though, so I suppose they have higher fixed costs)
17:44
<TabAtkins>
Yeah, but assuming equal quality, you're reducing your mean time to failure by consolidating
17:45
<TabAtkins>
Obviously, assuming you're RAID0ing the 500gb disks
17:49
<AryehGregor>
If you care about MTBF, why are you using RAID 0?
17:49
<AryehGregor>
Er, MTTF, whatever.
17:49
<smoofles>
oy
17:50
<smoofles>
where can one ask about features for html 5, or rather just bounce an idea – is this channel ok?
17:50
<TabAtkins>
Clearly you wouldn't, is the point. ^_^ But if you're doing a better RAID, you're losing some disks.
17:50
<TabAtkins>
smoofles: yup
17:51
<smoofles>
cools, I was wondering if it would make sense to have the video tag support an alt attribute
17:51
<TabAtkins>
So for a simple RAID1, you're paying a 50% premium. More expensive than the 1tb disk, but more reliable too.
17:51
<smoofles>
that would hold the url for an alternatively-encoded video file
17:51
<AryehGregor>
smoofles, that's what <source> is for, surely?
17:51
<smoofles>
oh
17:51
<TabAtkins>
For that case, use <source> children elements rather than the @src attribute.
17:52
<smoofles>
aah, didn't know about the <source> one
17:52
<TabAtkins>
no problem
17:52
<TabAtkins>
Works for <audio> too.
17:52
<smoofles>
coolio
17:52
<smoofles>
I read over the source elements stuff
17:53
<smoofles>
good, that basically settles the codec discussion for me, thanks :)
17:53
<TabAtkins>
Yup. Still annoying that you have to provide two files, but at least it works.
17:55
<smoofles>
yeah, well, on a lot of the sites people provide more versions anyway, everything from swf, mov, avi, mp4, etc
17:55
<TabAtkins>
Yeah.
17:55
<TabAtkins>
Someday, though. Someday...
17:58
<smoofles>
I guess 10 years from now it will be just a couple of thousand jpg files and an mp3, sort of like Apple's "http streaming" idea, since bandwidth will increase that much – problem solved :P
17:58
<TabAtkins>
haha, that'll work.
18:10
<jorlow>
oops.....one of the test shell unit tests is going to go red from 24612
18:10
<jorlow>
i'll have a fix in soon
18:10
<AryehGregor>
. . . is that meant for #chromium?
18:11
<jorlow>
:-) sorry
18:11
AryehGregor
used to have #chromium where he now has #whatwg on his screen, so was confused for a second there
18:24
<zcorpan_>
annevk2: html5-differences could point out in the abstract that it also lists changes between published html5 drafts
19:08
<zcorpan_>
annevk2: "The address element is now scoped by the new concept of sectioning." - could be changed to say that it's scoped to nearest ancestor article element (or body)
19:10
<Lachy>
is anyone here familiar with SQL?
19:11
<mpilgrim>
Lachy: i am
19:11
<mpilgrim>
though it's been a while
19:11
<AryehGregor>
Lachy, you mean like the formal standard, or some SQL-based RDBMS in practice?
19:11
<Lachy>
I'm need to write a test for webdatabase that inserts multiple rows, but I don't know how to construct an INSERT statement that can actually insert more than 1 row
19:11
<Lachy>
AryehGregor, SQLite specifically
19:11
<AryehGregor>
I'm not sure that's possible in a RDBMS-independent way.
19:11
<AryehGregor>
In MySQL you can just do INSERT INTO foo (x,y,z) VALUES (1, 2, 3), (4, 5, 6), ...;
19:12
<Lachy>
I don't care if it's RDBMS-independent. I just need to test our implementation that uses SQLite
19:12
<AryehGregor>
Couldn't you just do multiple INSERT statements?
19:13
<AryehGregor>
It looks like in SQLite, the only way to do it is: INSERT INTO foo (x, y, z) SELECT * FROM (SELECT 1, 2, 3 UNION ALL SELECT 4, 5, 6 UNION ALL ...);
19:13
<AryehGregor>
At least according to a quick Google search.
19:13
<AryehGregor>
The usual way is probably just multiple INSERTs.
19:14
<AryehGregor>
The "S" part of "SQL" is kind of a misnomer. :D
19:14
<Lachy>
no, because I'm trying to test the resultSet.rowsAffected API, which means I need to do it within a single executeSql statement, rather than several
19:15
<AryehGregor>
Then you should be able to use a subquery. INSERT INTO foo (x, y, z) SELECT * FROM (SELECT 1, 2, 3 UNION ALL SELECT 4, 5, 6 UNION ALL ...);
19:15
<AryehGregor>
Alternatively, if it's not actually possible, you don't have to test it. :)
19:15
<Lachy>
I will try that soon
19:18
<eighty4>
TabAtkins: yeah I know. I could just rewrite it to do float:left for ie<8 since those handles % more or less correctly imo
19:18
zcorpan_
blinked at "ie<8" without actually reading the whole line
19:19
zcorpan_
misread it as "ie<3"
19:19
<eighty4>
zcorpan_: definitely not :)
19:20
<eighty4>
is it wrong that I'm more concerned about getting stuff to work in opera then in ie6 and 7?
19:20
<zcorpan_>
nope ;)
19:20
<eighty4>
good
19:22
<zcorpan_>
maybe you'd get a different answer from someone who doesn't work at opera
19:22
<eighty4>
;)
19:22
<eighty4>
it's not like I have any visitors anyway :D
19:23
<eighty4>
but you really should fix how opera handles %
19:23
<zcorpan_>
i'm not sure what we do wrong with %
19:26
<eighty4>
you're rounding it up :)
19:27
<eighty4>
or if it was down
19:27
<eighty4>
12.456% gets 13 or if it was 12
19:28
gsnedders
rounds everything down
19:29
<gsnedders>
zcorpan_: We always round down in CSS
19:29
<gsnedders>
zcorpan_: Also, foolip kept trying to find you.
19:29
<zcorpan_>
gsnedders: ok
19:30
<zcorpan_>
i guess we have a bug already?
19:31
virtuelv
misread ie<8 as ie is less than boobies
19:31
<TabAtkins>
Firefox I know tries to round intelligently, so frex 8 floating boxes set to 12.5% width always take up exactly 1 line.
19:31
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/219
19:31
<TabAtkins>
I don't know if they use heuristics to tell when they should round up/down or what.
19:31
<gsnedders>
zcorpan_: Dunno. What we do is fine per spec, because rounding is undefined.
19:32
<Hixie>
you shouldn't round
19:32
<virtuelv>
Hixie: I assume you should do sub-pixel positioning, then?
19:32
<Hixie>
yes
19:33
<Hixie>
possibly with render-time snap-to-grid
19:33
<Hixie>
but there shouldn't be any cumulative rounding error
19:33
<gsnedders>
We fail a lot of the Mozilla rounding reftests because of rounding differences
19:33
<Hixie>
i filed a bug on opera needing to do subpixel layout about 3 decades ago
19:33
<gsnedders>
Though as far as I could tell, our behaviour is mostly fine per spec, just different.
19:34
<zcorpan_>
3 decades?
19:34
gsnedders
tries to remember if Hixie has even been alive for three decades
19:36
<zcorpan_>
it would be slightly better if we rounded down just one pixel
19:36
<zcorpan_>
but now we round 0.9% to 0%
19:36
<gsnedders>
Yeah, right
19:36
<eighty4>
I should file that "bug" I found.... still not sure it's actually a bug
19:36
gsnedders
finds a site compat bug on the Flickr Pro "place order" page
19:36
<eighty4>
0.9% could be 100px
19:36
gsnedders
guesses about the cause
19:37
<TabAtkins>
zcorpan_: Really? That's, um, quite bad.
19:37
<zcorpan_>
indeed
19:37
<gsnedders>
TabAtkins: YA RLY
19:37
<TabAtkins>
Even on standard resolutions that's 10px or more. >_<
19:37
<TabAtkins>
gsnedders: ORLY?
19:40
zcorpan_
finds CORE-1006
19:40
<mpilgrim>
"CW: I still plan on being involved but probably less than in the past." <-- I honestly don't see how that's possible
19:41
gsnedders
laughs at the number of duplicates
19:41
<mpilgrim>
(from http://lists.w3.org/Archives/Public/public-html/2009Aug/1331.html )
19:44
<gsnedders>
zcorpan_: You going to be in tomorrow?
19:45
<zcorpan_>
gsnedders: yes
19:48
<gsnedders>
Then I'll speak to you about this then
20:07
<Hixie>
jgraham: the spec gen script is hanging when doing the w3c version
20:08
<Hixie>
which i think means it's hanging doing the annotations
20:08
<eighty4>
would this be considered a bug? http://play.broken8.se/opera_bug/ watch it in ff browsers and then opera. Try to remove overflow: hidden on #content.
20:09
<Hixie>
ok that could possibly be because my server went down
20:09
<Hixie>
wtf
20:14
<gsnedders>
eighty4: Oh, that issue again… WebKit does what Opera does, and I think that is right
20:15
<zcorpan_>
Hixie: maybe we should allow <object> to have an open-ended set of attributes, like <embed>
20:15
<zcorpan_>
Hixie: as a shorthand for <param ...>
20:15
<eighty4>
gsnedders: yeah, I never did anything about it last time
20:15
<gsnedders>
eighty4: overflow:hidden is odd though
20:16
<eighty4>
gsnedders: but why would the display property change how margin and float works?
20:16
<gsnedders>
It shouldn't,.
20:17
<eighty4>
yet it does :)
20:18
<gsnedders>
I didn't say it did not :)
20:18
<eighty4>
i.e. a bug?
20:20
<gsnedders>
Yeah
20:20
<gsnedders>
W
20:20
<gsnedders>
*Which behaviour is right is debatable, I think
20:20
<zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/220 another opera bug i stumbled upon yesterday
20:21
<gsnedders>
We have quite a few bugs with button, though
20:21
<zcorpan_>
seems display:list-item on parent/ancestor triggers it
20:23
<Hixie>
zcorpan_: why?
20:24
<Hixie>
jgraham: regen script is indeed hanging on the w3c copy for some reason
20:24
<zcorpan_>
Hixie: because browsers support it and it's a lot shorter in markup
20:25
<zcorpan_>
Hixie: authors have to choose between short markup and no fallback (embed) and verbose markup with fallback (object)
20:25
<Hixie>
zcorpan_: *shrug*
20:25
<Hixie>
zcorpan_: they shouldn't do either
20:26
<zcorpan_>
because plugins are evil?
20:28
<Hixie>
well i dunno about "evil", but certainly they're not to be encouraged
20:28
<Hixie>
what with being media-specific, platform-specific, even browser-specific
20:28
<Hixie>
and proprietary
20:31
<jgraham>
Hixie: Oh
20:31
<jgraham>
I will have a look
20:35
<annevk3>
zii.com looks interesting
20:37
<zcorpan_>
Hixie: "The id attribute represents its element's unique identifier (ID)." - not sure why <a name> is much different
20:38
<TabAtkins>
annevk3: Huh, interesting. An android phone with multi-touch, cool.
20:38
<TabAtkins>
And a free phone with purchase of the SDK.
20:38
<Hixie>
zcorpan_: woah, that's bogus.
20:39
zcorpan_
sees that lang="" uses the word "specifies" instead of "represents"
20:40
<annevk3>
Hixie, represents is used all the time for various things
20:40
<annevk3>
Hixie, e.g. also for the cookie attribute
20:40
<Hixie>
file bugs
20:41
<annevk3>
well I don't really care
20:41
<Hixie>
(especially if it's the hyperlinked "represents")
20:41
<Hixie>
(cos that makes no sense)
20:41
<annevk3>
it's not linked
20:41
<Hixie>
ok well then it's basically a meaningless statement
20:41
<Hixie>
just fluff
20:41
<Hixie>
and certainly not a definition of any kind
20:42
zcorpan_
filed a bug
20:44
<zcorpan_>
maybe you should replace all non-normative sentences in the spec with "Fluff fluff fluff..."
20:44
<Dashiva>
zcorpan_: Normative spec view
20:45
<gsnedders>
zcorpan_: Or, "This is filler text."
20:46
<zcorpan_>
Hixie: "An error occured while submitting your comment. Please let ian⊙hc know."
20:48
zcorpan_
tries again
21:01
<sebastiandeutsch>
Don't know if I'm right here. Can I ask a technical canvas question? Is it possible to instanciate a canvas, then draw on it and the blit it (drawImage) on another canvas?
21:06
<TabAtkins>
sebastiandeutsch: You *should* be able to pull the data from the first canvas with getImageData, then blit it over with putImageData on the second.
21:07
<TabAtkins>
Haven't used <canvas> directly, but the spec seems to say that's how you'd do it.
21:08
<jgraham>
Hixie: Clearly the annotations script is broken but I'm not quite sure why
21:08
<TabAtkins>
Goddammit, people have no patience. Dude left after *2* minutes of me reading the spec for him.
21:13
<annevk3>
he prolly figured out drawImage worked fine
21:16
<jgraham>
Hmm maybe it is just that w3c.org is being really slow
21:16
<jgraham>
s/w3c/w3/
21:17
<jgraham>
I guess I should make a cache of the data somehow and use the cached values if the requests take too long
21:17
<jgraham>
But not this evening
21:51
<Philip`>
sebastiandeutsch: I guess you'll never read this but you can just use drawImage to draw one canvas onto another
21:56
TabAtkins
is having fun rolling his own security system (industry-standard encryption, but homebrew system for employing it), but wonders if it's sufficient.