08:23
<kennyluck>
Lachy, for bug-ish Selectors API feedback, would you prefer mailing list or bugzilla?
08:24
<Lachy>
if it's very minor, I can fix it now. If it requires more thought, then mail.
08:27
<Lachy>
kennyluck, ^
08:29
<kennyluck>
Lachy, I'll send a mail then.
09:00
<matjas>
paul_irish: +9001 on externalizing the testharness.js docs
09:01
<matjas>
a separate website isn’t really needed, you could just use GitHub Flavored Markdown™ and something like docdown (https://github.com/jdalton/docdown) to auto-generate it from the .js source (e.g. https://github.com/bestiejs/punycode.js/tree/master/docs#readme)
09:02
<matjas>
so, have you guys seen this Local File Writing API proposal? https://github.com/devongovett/standards/issues/3
09:02
<Ms2ger>
What, another copy of the documentation?
09:02
<Ms2ger>
I can't see how those would ever go out of sync
09:49
<jgraham>
paul_irish: I think that's rather beautiful. What markup format is it using? Can we use that inline and then automatically update the external docs from the inline docs?
11:16
<jgraham>
paul_irish: In general, I think it would be a good idea to make a modern-looking site to cover the W3C testing effort, with documentation from the various frameworks, instructions, stc. I think having all the information split across various wikis and repositories makes it very unapproachable and academic
11:42
<matjas>
Ms2ger, jgraham: that’s exactly how docdown works… it automatically generates it from the source as long as you use JSDoc-style comments
11:43
<Ms2ger>
Can we remove the copy in the readme, then?
11:45
<matjas>
Ms2ger: you could simply replace README.md with docdown’s output
11:46
<matjas>
php -f parse.php README
11:50
<Ms2ger>
Do you have something in python? :)
11:52
<matjas>
http://jonathan.tang.name/code/pyjsdoc
11:52
<matjas>
(haven’t used it)
13:05
<fflorent>
AryehGregor: one thing I did not understand with <a onlyreplace>. The @href attribute targets a full HTML document (with <HTML><HEAD>...) or a partial document ?
13:06
<AryehGregor>
fflorent, the idea was that it targets a full HTML document.
13:06
<AryehGregor>
Note that the idea has been abandoned, though.
13:06
<fflorent>
hmm, ok
13:06
<fflorent>
and why not a partial HTML document ?
13:06
<AryehGregor>
That wouldn't fall back gracefully in browsers that didn't support onlyreplace.
13:11
<fflorent>
AryehGregor: ah, ok. We could imagine a fallback that uses javascript. But I don't know what is the proportion of people having it disabled...
13:12
<AryehGregor>
fflorent, the whole point of onlyreplace is that fallback was supposed to be automatic and transparent.
13:12
<AryehGregor>
Also, the idea was that the URL bar would change like regular, so if the user hit refresh or bookmarked/restored, they would get the page you linked to without onlyreplace.
13:13
<AryehGregor>
If you did <a href="foo.html" onlyreplace> and the user clicked the link in an onlyreplace-supporting browser, if the page wasn't a full document, it would work fine at first. But then if they refreshed, it would just load foo.html from scratch and that would break.
13:14
<fflorent>
AryehGregor: oh, yes. I thought about a mechanism that would update the URL using hash bangs instead, or query parameters
13:14
<AryehGregor>
That defeats the intended goal of onlyreplace.
13:15
<fflorent>
to be transparent, yes I see know ...
13:15
<fflorent>
now*
14:43
<Ms2ger>
paul_irish, yt?
17:15
<zewt>
it's a bit depressing that even firefox has the broken "link to download goes to html that redirects to the actual download" crap
17:15
<zewt>
(the download site, I mean)
17:16
<zewt>
if I copy-link/wget one more time and end up with an index.html ...
17:20
<zewt>
heh ff13 still hard crashes if i open a webgl context with a 10000x10000 canvas
17:23
<zewt>
agh this is why I put off update firefox, they always break stupid things
17:23
<zewt>
now pages in my session aren't loading at all until i focus them, and i have to track down how to fix that
17:25
<zewt>
guessing browser.sessionstore.restore_on_demand
17:25
<Ms2ger>
That's a feature
17:26
<zewt>
a misfeature, you mean
17:29
<zewt>
agh
17:29
<paul_irish>
Ms2ger: sir.
17:30
<zewt>
and it's smooth scrolling on the mousewheel now, even though i have that turned off system-wide
17:30
<Ms2ger>
Aha!
17:30
zewt
*more about:config searching*
17:30
Ms2ger
tries to remember which test he was going to complain about
17:30
<zewt>
guess i should look at the changelogs to see what features I'm getting for having to fix these things
17:31
<paul_irish>
Ms2ger: probably the fontfacerule-constructors one?
17:31
<Ms2ger>
No, I approved that one
17:31
<Ms2ger>
The comment in cssom-fontfacerule, I think
17:32
<paul_irish>
mmmhm
17:33
<Ms2ger>
Also, yay tests!
17:33
<paul_irish>
\o/
17:34
Ms2ger
keeps his comments about their quality to himself
17:36
<paul_irish>
:) Yeah I'd like to take a more focused approach on part of that spec
17:37
<paul_irish>
Was also talking to smfr, rniwa and others about upstreaming LayoutTests. Going to write up some documentation on how that can work
17:38
<kennyluck>
Is testing the Web forward happening right now or something?
17:38
<paul_irish>
just finished up yesterday. it was friday night and all saturday
17:40
<kennyluck>
paul_irish, oh ok. Thanks. Is there public video or something (so that I can share with other people)?
17:43
<paul_irish>
there was a livestream of the friday night talks but i dont think anything recorded... i'll ask divya to get http://testthewebforward.org/ updated with speaker decks
18:02
<Ms2ger>
paul_irish, so what do you think about doing one in Europe focused more on the DOM?
18:03
<paul_irish>
++
18:03
<paul_irish>
That would be awesome. I think Adobe would enjoy putting that on
18:04
<paul_irish>
I'll bring it up
18:23
<jgraham>
Yeah, that is totally something that we should organise
18:32
<kennyluck>
Ms2ger, yay, I just made another CSSOM bug into a mentored bug → https://bugzilla.mozilla.org/show_bug.cgi?id=765599
18:32
<Ms2ger>
\o/
18:35
<kennyluck>
paul_irish++, for the (promised) livestream :p