00:34
<AryehGregor>
Okay, hg's response to a config file format error seems to be a backtrace.
00:35
<AryehGregor>
Seriously.
00:35
<AryehGregor>
How can this possibly be considered acceptable in real software?
01:08
<AryehGregor>
Wow, people had \
01:08
<AryehGregor>
Wow, people had fun filing bugs the day before the deadline.
01:09
<AryehGregor>
Especially Microsoft.
01:19
<variable>
AryehGregor, bugs for what?
01:19
<AryehGregor>
http://w3.org/Bugs/Public
01:19
<AryehGregor>
Spec bugs.
01:19
<variable>
ah
07:50
<grantg>
What's with the data property of an ImageData Object returned from createImageData or getImageData being read-only?
07:50
<grantg>
All the browsers except IE9 allow you to write (and should for raw display-out capabilities).
07:51
<grantg>
This is blocking my js gameboy color from using the canvas object for display out.
07:51
<grantg>
It has to fall back to drawing via BMP images created by data uris
07:51
<grantg>
site: http://grantgalitz.org/gameboy/
07:52
<grantg>
The each value inside the CanvasPixelArray object should be able to be written to
07:53
<grantg>
But IE9 is the browser that disagrees. WTF?
07:53
<annevk>
that's just a bug in IE9 then
07:53
<annevk>
data is marked readonly because you cannot write to data, but you can write to the object it returns
07:53
<grantg>
Yeah, you shouldn't be able to alter the array count / add / delete the indices, and they should have to fall between 0 and 255
07:54
<grantg>
write with restrictions. ;)
07:54
<grantg>
annek: ok
07:54
<grantg>
Then the IE9 team read that wrong. :P
07:55
<annevk>
there's a whole lot of things that work that way
07:55
<annevk>
so that'd be scary
07:55
<grantg>
Since there's "limited writing" to the indices.
07:55
<annevk>
but maybe they don't do it consistently wrong
07:55
<grantg>
IE9 throws a read-only error on setting an indice.
07:56
<grantg>
Even though the value to write is valid and so is the indice.
07:56
<grantg>
wird
07:56
<grantg>
*weird
07:57
<grantg>
Unless my emu has a hidden bug that's only triggering a fault in IE9. :P
07:57
<grantg>
The emu actually works btw.
07:57
<grantg>
At least most roms do.
07:57
<grantg>
Some bugs still. :(
07:58
<grantg>
The JS N64 emu is even worse.
07:58
<grantg>
Which is why it isn't live yet.
07:58
<grantg>
It has tons of issues with WebG.> :(
07:58
<grantg>
*WebGL
08:00
<annevk>
N64? that'd be great
08:00
<grantg>
annevk: ARM emulation for the GBA is coming up shortly (before the N64 emu).
08:00
<grantg>
At least ARM has documentation. :P
08:01
<grantg>
*more than the GBC :P
08:01
<grantg>
yeah
08:01
<grantg>
Almost all the N64 game roms crash
08:01
<grantg>
some exceptions though
08:01
<grantg>
Mario Kart & Diddy kong racing
08:02
<grantg>
Though WebGL matrices are way off on them.
08:02
<grantg>
And a buggy Firefox 4 beta 7 pre isn't making life any easier.
08:04
<grantg>
Though the documentation on the reality chip for the n64 is sometimes hard to come by, just like the gameboy's cpu docs.
08:04
<grantg>
Damn Nintendo changing the Zilog Z80 CPU around.
08:06
grantg
almost has kirby 64 running on the js n64 emu at least.
08:08
<grantg>
Though I need to work on the GBC emu some more
08:08
<grantg>
Pokemon Crystal badly fucks up after the intro
08:08
<grantg>
And the clock cycle timing isn't right yet.
08:09
<grantg>
Though Pokemon Yellow seems to run perfectly. :P
08:09
<grantg>
color and all
08:09
<grantg>
except sound
08:09
<grantg>
Sound is fucked up on my part.
08:09
<grantg>
Still not ready on sound.
08:10
<grantg>
annevk: Works in IE8 at 1fps. ;)
08:10
<grantg>
(IE 8, not 9)
08:10
<grantg>
IE9 works at 5 fps
08:12
<grantg>
annevk: When is chrome going to fix the memory leak with assigning long arrays onto a canvas context?
08:12
<grantg>
Just seems like a bad security issue
08:12
<grantg>
As well as data uri'd images.
08:13
<grantg>
google chrome leaks a lot of shit and runs systems of of memory accidentally when running my emu
08:13
<grantg>
Firefox, opera, ie don't have this prob
08:14
<annevk>
grantg, better ask someone from Google :)
08:14
<grantg>
heh
08:14
<grantg>
Lots of rendering glitches in Opera & chrome
08:15
<grantg>
Opera wins overall
08:15
<annevk>
but euh, keep up the good work; I love playing games in the browser
08:15
<grantg>
Try to visit my emu in Opera and you'll be horrified.
08:15
<annevk>
oh, if you find bugs in Opera please file them at bugs.opera.com/wizard
08:15
<grantg>
I have
08:15
<annevk>
cool
08:15
<grantg>
The opera dev team has ignored them for 6 months at least
08:15
<annevk>
oh, let me poke them for you
08:16
<annevk>
can you /msg me your email or the bug IDs?
08:16
<grantg>
Box-shadow is *trashed* c o m p l e t e l y. Did I say completely trashed. YES.
08:16
<grantg>
annevk: grantgalitz⊙gc
08:16
<grantg>
Should be in the db
08:16
<annevk>
should definitely make sure games work in our browser
08:16
<annevk>
great
08:17
<grantg>
And SVG rendering is fucked up when used in css backgrounds in Opera whena doc is scrolling
08:17
<grantg>
Not to mention the opacity bug where the browser's rendering engine incorrectly stacks the virtual buffers wrong.
08:18
<grantg>
So UI elements like scrollbars and buttons appear behind the opacity-given element when in the element.
08:19
<grantg>
Opera gets its logic right, but not the rendering logic right.
08:19
<grantg>
annevk: Do they have my bugs, they should be there.
08:20
<grantg>
If you need me to re-fill out the reports, ok.
08:21
<grantg>
Opera's rendering is buggy on all OSes. :(
08:21
<annevk>
there's some weird way we store bugs from external reporters and I forgot how to search that
08:21
<grantg>
hmm
08:21
<grantg>
Should I double report?
08:22
<annevk>
if you have the bug IDs that'd make it easier for me
08:22
<grantg>
Not the IDs.
08:23
<grantg>
I haven't been to their site in awhile, nor have I copied down any #s.
08:24
grantg
needs to go soon.
08:25
<annevk>
I found one bug
08:26
<grantg>
Heh, found a GBC rom that shows off Mode 7. lol @ crude 3D on the gameboy.
08:26
<grantg>
annevk: It might be listed under other emails too.
08:27
<grantg>
Like one of my half-dozen AOL accounts.
08:27
<grantg>
yes, AOL. :(
08:27
<annevk>
heh, do they have something in common?
08:27
<annevk>
would definitely be easier if you used just one address
08:28
<grantg>
I should just re-report the damn bugs.
08:28
<annevk>
that works too
08:28
<grantg>
It's easy to see them
08:28
<grantg>
At least half-dozen on the gameboy emu
08:29
<grantg>
Browse http://grantgalitz.org/gameboy/ and you'll eventually find some bugs.
08:29
<grantg>
These 'windows' are movable and resizable.
08:29
<grantg>
Moving them glitches out opera.
08:30
<grantg>
Clearly Opera bugs when you move them.
08:31
<annevk>
yeah
08:31
<annevk>
bah
08:31
<grantg>
lol
08:31
<grantg>
Trashed, right?
08:33
<grantg>
Do you need a rom
08:33
<grantg>
Remember, you can load any rom from any address
08:33
<grantg>
or you can load them in locally
08:34
<annevk>
I was wondering about samples
08:34
<annevk>
no presets?
08:34
<grantg>
Either through the File API or through Base64 copy+paste
08:34
<grantg>
?
08:34
<grantg>
You mean providing roms for you?
08:34
<annevk>
yeah, though I guess you'd violate some copyright then
08:34
<grantg>
I'd call that illegal or borderline illegal
08:34
<grantg>
like jsnes
08:35
<grantg>
I can give you any unlicensed rom though
08:35
<grantg>
http://grantgalitz.org/gameboy/romStorage/proxima.gb
08:35
<grantg>
copy+paste that address into the emu
08:35
<grantg>
or dl the rom and load it in from your comp
08:37
<annevk>
you need some kind of controls page
08:37
<grantg>
x is a
08:37
<grantg>
z is b
08:37
<grantg>
shift is select
08:37
<grantg>
enter/return is start
08:37
<grantg>
annevk: yeah
08:38
<grantg>
and the arrow keys are the arrow keys.
08:39
<grantg>
annevk: You can play most games even right before GBA came out
08:39
<grantg>
2001 GBC games work as good as 1998 GBC games
08:40
<grantg>
Like mario tennis
08:40
<grantg>
and Super Mario Bros Deluxe
08:40
<grantg>
and Pokemon Yellow / Red / Blue
08:40
<grantg>
And Microsof pinball
08:40
<grantg>
etc
08:40
<grantg>
*MicroSoft
08:40
<grantg>
and more...
08:43
<annevk>
box-shadow issue is already resolved btw
08:43
<grantg>
When?
08:43
<annevk>
will hopefully be in public builds soon
08:44
<grantg>
Thank god
08:44
<grantg>
It's been there since the 10.50 alphas
08:45
<grantg>
annevk: Notice when resizing the 'window' vertically very slowly the inside div doesn't resize as well?
08:45
<grantg>
Bug in opera. :P
08:45
<grantg>
very very weird position: absolute bug.
08:46
<grantg>
with inline-block
08:46
<grantg>
some weird logic
08:46
<grantg>
Like a check in opera is getting missed.
08:47
<grantg>
annevk: BTW, the site renders as application/xhtml+xml for browsers that send that mime type in the accept header.
08:48
<grantg>
So it will present itself as true XHTML to the browsers that support it.
08:48
<grantg>
:D
08:48
<annevk>
Hixie, http://www.emendapp.com/sites/www.whatwg.org
08:49
<grantg>
None of that broken fake-o-xhtml shit that 99% of the web does.
08:50
<annevk>
so that you can use inline SVG?
08:50
<grantg>
Correction, 99+%
08:50
<grantg>
:(
08:51
<grantg>
annevk: I don't use inline svg
08:51
<grantg>
Just as CSS backgrounds
08:51
<grantg>
As fallback when CSS3 gradients aren't supported
08:51
<zcorpan>
so that ie6-8 users will still be happy when your site is broken but ie9 users and users of other browsers will be unhappy?
08:52
<grantg>
IE8 gets MicroSoft's proprietary gradient filter
08:52
<grantg>
IE 6 & 7 cannot work with the emu
08:52
<grantg>
they suck and will not run it
08:52
<grantg>
brb
09:03
<annevk>
hg pull is crashing on me?
09:03
<annevk>
is this because I pushed to a different repository before or something?
09:03
<annevk>
gah
09:08
<jgraham>
annevk: No, you should be able to push and pull from different repos
09:08
<jgraham>
What error are you getting?
09:09
<annevk>
I think it's due to VPN
09:09
<annevk>
I disabled VPN and now things work
09:09
<annevk>
can I pull and push to two repositories by default btw?
09:10
<annevk>
so that I don't have to keep them in sync manually?
09:10
<annevk>
also
09:10
<annevk>
*summons* Ms2ger
09:10
<jgraham>
I don't think you can make hg pull and hg push always work with two different repositories
09:10
<jgraham>
at the same time
09:12
<annevk>
too bad
09:16
jgraham
wonders if Leif actually uses System 9
09:18
<zcorpan>
what's System 9?
09:18
<jgraham>
The Mac operating system from before OSX
09:19
<zcorpan>
ah
09:44
grantg
loads a compiled rom of unix made for gameboy z80 derived cpus and loads it into the js emu.
09:45
<grantg>
bliss
09:45
<grantg>
unix in js
09:47
grantg
loads http://grantgalitz.org/gameboy/romStorage/Unix%20(PD).gb into http://grantgalitz.org/gameboy/
09:52
grantg_
managed to "aw, Snap!" Google Chrome once more with the emu (WebKit is unstable and leaks a lot in the File API)
09:53
<grantg_>
*unstable in the file api
10:03
<grantg_>
annevk: When is the bug fix release coming out approx?
10:03
<grantg_>
This month or next?'
10:03
<grantg_>
Or "when we feel like it"?
10:05
<annevk>
Hopefully soon is the best I can do
10:06
<grantg_>
Yeah, this is the kind of bug where box-shadow should be disabled by default until it's fixed.
10:07
<grantg_>
breaks whenever the effect is altered (on /off) or moved around.
10:08
<grantg_>
I'm going to sleep
10:08
<grantg_>
nn
10:16
<hsivonen>
jgraham: does http://www.w3.org/html/wg/wiki/ChangeProposals/FigureInP#Details look sane to you?
10:22
<zcorpan>
hsivonen: table-row seems like an annoying style since it can't have a border in the separated table border model (which is the default)
10:24
<zcorpan>
hsivonen: you could say figure { border-collapse:collapse } in the ua style sheet but then that would affect descendant tables
10:25
<hsivonen>
zcorpan: would table-cell make more sense?
10:26
<hsivonen>
zcorpan: hmm. actually, it would not
10:26
<jgraham>
hsivonen: Why do you want it to be scoping?
10:26
<hsivonen>
zcorpan: what would you suggest for the default UA style sheet for figure and figcaption
10:26
<zcorpan>
hsivonen: display:block for both
10:26
<hsivonen>
zcorpan: OK
10:27
<hsivonen>
jgraham: to make <pre> as the figured element work
10:27
<hsivonen>
(and <table>)
10:27
<zcorpan>
we'd want the caption text to wrap to the width of the image, which display:table doesn't do
10:28
<zcorpan>
if we try hard enough to get it right, what we have accomplished is a headache for authors who will need more rules in their reset.css to make a default they can understand
10:28
<hsivonen>
zcorpan: ok
10:29
<hsivonen>
zcorpan: I removed the UA style sheet changes
10:29
<hsivonen>
jgraham: it's covered in the "Rationale" section :-)
10:30
<annevk>
allowing <pre> and <table> nested in <p> through <figure> is something I'm going to object to I think
10:30
<hsivonen>
annevk: why?
10:30
<annevk>
because that is the kind of authoring insanity I was hoping we would go away from
10:31
<annevk>
<table> does not work in <p>, but you can hack around that with <figure>, and it'll validate too
10:31
<hsivonen>
annevk: authors want to put tables inside paragraphs already. Why should we get in the way
10:31
<annevk>
because if this is the only way they can do that everyone will just end up getting confused
10:31
<hsivonen>
annevk: table not working in p is a flaw on the format. I would not hold hacking around flaws of the format against authors
10:32
<hsivonen>
s/on/of/
10:32
<annevk>
we should have fixed the flaw when that was still possible
10:32
<annevk>
adding special cases around the flaw instead of just accepting it as an inherent limitation seems misguided
10:33
<hsivonen>
annevk: you know how hard it is to fix a flaw like that. http://ln.hixie.ch/?start=1137799947&count=1
10:34
<hsivonen>
annevk: if we don't fix figure now, we'll get one of those blog posts 7 years from now
10:34
<annevk>
not at all
10:34
<hsivonen>
actually, it's worse than 7 years. Mozilla still hasn't shipped a release with Web-compatible comment parsing
10:34
<annevk>
http://hsivonen.iki.fi/last-html-quirk/ is the flaw
10:35
<annevk>
and we know we cannot fix it
10:37
<hsivonen>
annevk: what could I do to the Change Proposal draft to make you not object to it if it went to a poll?
10:38
<annevk>
I think the content model would have to remain transparent
10:38
<hsivonen>
annevk: my point is that <p><figure>, <p><table> and SGML comments are all things that were working nicely for authors before Hixie pushed a change on theoretical purity grounds
10:38
<hsivonen>
annevk: that's an edit I'm not willing to make at this point :-(
10:39
<annevk>
I'm fine with <p><figure> as long as <p><figure><table> is not allowed. I don't want weird content model rules.
10:40
<hsivonen>
annevk: transparent content models are weird rules for normal people (where normal people means people who don't hang out here)
10:40
<annevk>
Having certain elements be allowed as descendants sometimes and not work at all when you remove an element inbetween is just too magical
10:40
<hsivonen>
annevk: I claim that people think more in terms of parent-child than in terms of ancestor-descendant when it comes to content models
10:41
<jgraham>
I think transparent content models are always weird
10:42
<annevk>
I think people would find it very weird if they can suddenly nest <table> inside <p> through <figure> but not directly. If they'd even notice.
10:42
<annevk>
I think having <figure> on the same level as <p> and <pre> etc. therefore makes more sense, though I'd be ok with making it inline and not </p>-implying.
10:44
<annevk>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=10692#c9 -- I wonder how often I will have to keep explaining this to people
10:48
<zcorpan>
annevk: the suggestion in that bug is about the coersion rules, and browsers don't implement the coersion rules, so the html[xmlns] thing is a non-issue
10:49
<annevk>
that's just what the title says
10:49
<annevk>
he's actually asking for something else
10:50
<annevk>
e.g. "I don't necessarily care how this is done as long as the Javascript that is executed on the document, intended to find "xmlns:*" mappings, doesn't have to have two code paths depending on if the document is in HTML5-mode or XHTML5-mode."
10:52
<zcorpan>
might be worthwhile to point out in the bug then that changing the coersion rules does not affect javascript/dom in the browser
10:52
<annevk>
feel free
11:05
<jgraham>
Isn't Manu asking for the attribute to be put in the DOM in *both* the XMLNS NS and in no namespace?
11:06
<jgraham>
It sounds pretty insane, but like it would not break with a simple html[xmlns] selector
11:07
<annevk>
oh
11:07
<annevk>
that would be insane
11:09
<jgraham>
Well yeah, but you need a different example it would break, or some other argument that the insanity isn't worth the perceived value
11:10
<jgraham>
(to me it is obvious that adding that kind of complexity to support the usecase "people want to use features that were designed to not work in HTML" isn't a good idea)
11:11
<jgraham>
Especially given the number of people involved, and there are alterntaives that are designed to work in HTML
11:19
<annevk>
thanks for pointing that out
12:42
<gsnedders>
Do reftests in Gecko work by getting a hash of the entire viewport, or does it depend upon window size?
12:52
<smaug____>
gsnedders: AFAIK reftest uses canvas
12:52
<smaug____>
it paints two pages to canvas and compares them
12:53
<jgraham>
How big is the canvas?
12:55
<smaug____>
jgraham: seems like it takes the canvas size from the element which contains the web page
12:55
<smaug____>
so the size of iframe or so
12:55
<smaug____>
(I'm just reading the code...)
12:57
<gsnedders>
smaug____: okay
12:59
<smaug____>
The browser element (== iframe) has style="min-width: 800px; min-height: 1000px; max-width: 800px; max-height: 1000px"
12:59
<smaug____>
gsnedders: ^
13:00
<jgraham>
Aren't there tests that don't work in an iframe?
13:00
<jgraham>
I'm thinking of viewport0related CSS stuff
13:02
<smaug____>
jgraham: well, it is actually <xul:browser type="content-primary"> so the contents of it is the top level content page
13:02
<jgraham>
Ah
14:00
hsivonen
wonders if all the FOs raised in the first months of the HTML WG's existence have become moot and no longer tracked: http://dev.w3.org/html5/status/formal-objection-status.html
14:46
<zcorpan>
hsivonen: i thought that all FOs would be tracked and presented to the Director when the time comes, but i don't know how it works exactly
14:46
<hsivonen>
zcorpan: that's what I thought, too
14:52
<zcorpan>
hsivonen: i think Hixie meant css whitespace as in whitespace the actual style sheet, not what the white-space property thinks is white space in the document
14:52
<jgraham>
Oh more bugspam
14:53
<hsivonen>
zcorpan: oh. ok.
14:53
<hsivonen>
zcorpan: this stuff is confusing!
14:55
<zcorpan>
maybe it's possible to change css's definition of whitespace to exclude the form feed too, but i don't think the csswg are willing to do that (they didn't want to add u+000b as whitespace to match html5 back when html5 had that, which was in part why u+000b was removed from html5)
14:56
<zcorpan>
(the other reason was that u+000b wasn't consistently implemented as whitespace in browsers anyway)
15:07
<hsivonen>
zcorpan: consistently implemented on the parser layer, yes. Not consistently implemented throughout the engine.
15:09
<zcorpan>
hsivonen: iirc there was something with ie not always considering u+000b to be whitespace in the parser, but i forget the details
15:20
<hsivonen>
is this pre-LC component barred from accepting new bugs?
15:28
<hsivonen>
it would have been nice to disable bugmail during the mass update
15:31
<MikeSmith>
hsivonen: true
15:31
MikeSmith
does that now
15:32
<MikeSmith>
problem is that when I do disable it, I usually forget to turn it back on
16:11
<f1lt3r>
anyone know where I could find: "Robin Berjon"? (re: <device>)
16:58
<TabAtkins>
f1lt3r: Blog? Twitter? Email?
17:00
<Philip`>
Street address?
17:29
<mathrick>
hi guys
17:29
<mathrick>
may I ask how exactly the whole versioning mechanism came to be in WebSQL?
17:29
<mathrick>
because as far as I can see it's 1) completely broken 2) totally unsuitable for the single stating purpose, ie. ensuring that a known and supported schema version is being used
17:30
<TabAtkins>
gsnedders: As long as your initial guess is "sufficiently close", yeah, it'll terminate.
17:31
<mathrick>
http://pastebin.com/MUJTBXza
17:31
<mathrick>
my understanding of these snippets is that you can never successfully change the schema version unless you happen to guess the existing version exactly
17:32
<mathrick>
you can also never *read* the version, making it basically impossible to check if you know how to operate on the data safely
17:57
<AryehGregor>
Amusing: http://code.google.com/p/chromium/issues/detail?id=57816
18:24
<mathrick>
also, http://www.w3.org/TR/webdatabase talks about read/write mode a lot, but no method takes a parameter called "mode"
18:24
<mathrick>
in particular, transaction() / readTransaction() don't, and it's implied they should
18:24
<mathrick>
http://www.w3.org/TR/webdatabase/#dom-database-transaction
18:44
<gsnedders>
TabAtkins: Yeah, that's what I found looking online. So I concluded my working had gone wrong. :)
18:45
<gsnedders>
(I later found my mistake, and just made it to his English Language lecture on time)
18:45
<gsnedders>
s/his/my/
18:45
gsnedders
blatantly uses actions way too much
18:52
<gsnedders>
Hixie: http://barkavenueblog.blogspot.com/2009/04/tokyo-cat-cafe-big-business.html (via wilhelm)
20:01
<mikekelly>
hi fans
20:01
<mikekelly>
I came to ask about the PUT/DELETE removal from http://dev.w3.org/html5/spec/Overview.html#attr-fs-method
20:01
<mikekelly>
any word on that?
21:13
<othermaciej>
are any of the people who commented on the link relation thread planning to write up a proposal of the wiki-based approach? (AryehGregor? hober? TabAtkins?) I haven't seen a volunteer yet and if no one does it then the IANA approach will win by default, which would be unfortunate given the number of people advocating the microformats wiki instead
21:13
<TabAtkins>
othermaciej: What's the deadline?
21:13
<othermaciej>
October 27
21:13
<TabAtkins>
Yeah, I'll do it if no one else picks it up.
21:18
<hober>
I'll help with it, but I don't want to drive. :)
21:19
<hober>
tantek would be a good person to have help on that
21:21
<TabAtkins>
Yeah, that's what I was thinking too.
21:36
<mathrick>
okay, guys, WTF
21:36
<mathrick>
why does executeSql work async?
21:36
<mathrick>
is the fact the transaction callback is run asynchronously already not enough?
21:37
<mathrick>
I'm surprised that ResultSets can be accessed without involving callbacks somewhere
21:47
<Hixie>
mathrick: anything that might take more than zero time is async, otherwise the browser might lock up while the call is happening
21:48
<mathrick>
Hixie: then to be honest rows.length should also be async, as as the spec notes it might be expensive
21:48
<mikekelly>
hi Hixie
21:48
<mathrick>
besides, there are already *Sync() variants, so the lockup is still possible
21:48
<mikekelly>
Hixie, did you get back to anyone on that missing PUT/DELETE thing?
21:49
<mathrick>
Hixie: and unrelated to async, I still think that the whole changeVersion() interface is horribly broken
21:49
<mathrick>
why does it not pass in the current DB version?
21:49
<mathrick>
you have no way to enumerate DBs, and no way to learn the current version
21:50
<mathrick>
so you have to try each possible combination of versions until one works
21:52
<Hixie>
mathrick: yeah, the .rows being sync and the changeVersion() API sucking are known bugs. That spec is dead though so I doubt anything will happen to fix those problems.
21:53
<mathrick>
*sigh*
21:55
<mathrick>
Hixie: so we have WebSQL, which is a huge pain to use, and broken in many places, and will never be supported by Mozilla, and IndexedDB, which is similarly broken but doesn't give a useful query language
21:55
<TabAtkins>
For the broken parts, comment on it!
21:55
<mathrick>
proposed specs should not have glaring bugs apparent on the first reading ever :(
21:55
<TabAtkins>
it = IndexedDB.
21:55
<mathrick>
TabAtkins: the draft at W3C doesn't have a nice comment box like the whatwg documents do
21:55
<mathrick>
but in a nutshell, it's race conditions waiting to happen
21:56
<TabAtkins>
No, but it has a mailing list.
21:56
<mathrick>
"Every method for making asynchronous requests returns an IDBRequest object that communicates back to the requesting application through events." <-- here be race conditions
21:57
<Hixie>
mathrick: yes. Not much I can do about either though. If you want Web SQL DB resurrected, convince Mozilla and Microsoft to implement it (I'll spec the SQL profile if necessary). If you want IndexedDB to not suck, send comments on it, it's a work in progress.
21:57
<mathrick>
http://www.w3.org/TR/IndexedDB/#idl-def-IDBRequest <-- No method for providing error handlers upon creation, and no method to read if it's finished with an error or not
21:57
<Hixie>
IRC is not a good place to send feedback
21:57
<TabAtkins>
mathrick: I'm not an editor of that spec, nor is Hixie. Email the list.
21:57
<Hixie>
it's especially not a good place for IndexedDB since none of the IndexedDB editors are here :-)
21:59
<mathrick>
yeah, I'm just being disheartened by the piles of brokenness I encounter every time I try to use the newer APIs
22:15
<jgraham>
http://reprog.wordpress.com/2010/09/05/bibliographic-data-part-3-has-anyone-anywhere-ever-read-the-whole-of-the-rda-specification/ Never let anyone tell you that HTML5 is too long, ever again
22:17
<paul_irish>
oh my.
22:17
<karlcow>
"As best I can tell, the current total is about 1640 pages — so if you printed it out on standard 80 gsm paper, it would be 17 cm thick (nearly 7 inches) and weigh something over 8 kg (18 pounds)."
22:18
<karlcow>
it is still less than the changelog of Irak article on wikipedia ;) http://booktwo.org/notebook/wikipedia-historiography/
22:18
<AryehGregor>
Is there really no way to tell Bugzilla not to notify everyone when you make mass changes like renaming a component?
22:19
<jgraham>
Well it is a bit unfair to compare full history against a single snapshot
22:20
<Hixie>
complete.html is about 500-600 pages, but that doesn't include things like SVG, CSS, MathML, HTTP, etc, which really should be part of the same spec IMHO, if we're going to spec The Platform
22:21
<Hixie>
what are the 1640 pages speccing?
22:21
<Hixie>
the blog post wasn't clear about that
22:21
<jgraham>
http://www.rdatoolkit.org/home
22:22
<karlcow>
http://www.rdatoolkit.org/constituencyreview
22:24
<karlcow>
from what I can see it is also using a large font. This kind of stats would be better counting words more than pages. Still it is a lot
22:24
<karlcow>
they use extensive prose to describe each feature
22:25
<karlcow>
example: 2.3.1.1
22:25
<karlcow>
A title is a word, phrase, character, or group of characters that
22:25
<karlcow>
names a resource or a work contained in it.
22:25
<karlcow>
More than one title may appear in the resource itself (e.g., on a title page, title frame, etc.; as a caption title, running title, etc.; on a cover, spine, etc.; on a title bar, etc.), on a jacket, sleeve, container, etc., or in material accompanying the resource.
22:26
<karlcow>
they have also a *lot* of examples
23:51
<Hixie>
Philip`: yt?