07:40
<izhak>
whatwg.org is slowed down.
07:41
<MikeSmith>
yeah
07:41
<MikeSmith>
so maybe it wasn't because of twitter traffic to web-apps-tracker after all
07:42
<MikeSmith>
for what Hixie said previously, I think it might be some crawler hitting the svn server
07:44
<izhak>
the number of crowlers is going to create problems soon
07:46
<izhak>
*crawlers.
08:22
<MikeSmith>
anybody remember when the download attribute was added to the spec?
08:23
MikeSmith
is trying to find the rev number so he can open a reminder vnu bug
08:31
<MikeSmith>
http://html5.org/r/6318
08:36
<zcorpan>
oh, that was more recent than i thought
08:36
zcorpan
was going to guess "about a year ago"
08:37
<MikeSmith>
yeah, seems like at least
08:38
<hsivonen>
https://bugzilla.mozilla.org/show_bug.cgi?id=167303 might be relevant to some spec feature requests
08:38
<MikeSmith>
but it was only last summer
08:38
<hsivonen>
the WONTFIX status that is
08:41
<MikeSmith>
"I am not sure if js is a good language for parsing arrays of bytes"
08:44
<MikeSmith>
god I just looked at the dates on the bug
08:49
<yolin>
is there a html validator in javascript?
08:50
<hsivonen>
yolin: AFAIK, no
08:54
<MikeSmith>
yolin: would be great to have one
08:54
<yolin>
MikeSmith, yeh I was thinking the same, where to start?
08:55
<MikeSmith>
start simple and work up from there
08:55
<MikeSmith>
I think Hixie has some ideas about implementing a checker in JS
08:56
<yolin>
ok, great!
08:56
<jgraham>
I would be careful to start from a position of ignorance. Otherwise I would feel overwhelmed and give up
08:56
<jgraham>
;)
08:56
<MikeSmith>
heh
08:56
<MikeSmith>
yeah :)
08:56
<yolin>
:D
08:57
<jgraham>
But seriously, it sounds like a nice idea
08:57
<yolin>
is the current validator at w3c code avaliable?
08:57
<yolin>
its in Java right? could be ported to Javascript?
08:57
<MikeSmith>
if you mean the current-current w3c validator, you don't want to look at that code
08:57
<yolin>
oh...
08:57
<MikeSmith>
yolin: the java code id validator.nu
08:58
<MikeSmith>
the existing W3C validator is perl code and based on SGML parsing and other ugly stuff
08:58
<MikeSmith>
yolin: I really would suggest not looking at existing code
08:58
<MikeSmith>
and not using any off-the-shelf general schema-checking mechanism
08:58
<yolin>
okay
08:59
<MikeSmith>
but doing it all from scratch
08:59
<MikeSmith>
but that's just my opinion
08:59
<MikeSmith>
though, there is no existing general schema checking in JS
08:59
<MikeSmith>
as far as I know
09:00
<yolin>
I would have to implement the HTML5 parsing algo?
09:00
<MikeSmith>
no
09:00
<jgraham>
There is already an implementation of that
09:00
<MikeSmith>
not that at least
09:00
<MikeSmith>
yeah
09:00
<jgraham>
Dunno if it has parse error detection though
09:00
<MikeSmith>
there are a couple of implementations
09:00
<MikeSmith>
I think it does
09:00
<MikeSmith>
well
09:00
<MikeSmith>
it depends on which one you mean
09:01
<jgraham>
I was thinking of the dom.js / David Flanagan one
09:01
<MikeSmith>
yeah, I think that one might
09:01
<MikeSmith>
there is another one that Aria Stewart did for node.js
09:02
<yolin>
dom.js is for dom manipulation in node, will it help for validating html?
09:03
<MikeSmith>
it has a separate html parser that can be used standalone
09:04
<MikeSmith>
https://github.com/andreasgal/dom.js/blob/master/src/impl/HTMLParser.js
09:04
<yolin>
oh great
09:05
<MikeSmith>
hmm
09:05
<MikeSmith>
it doesn't look like it reports parse errors though
09:06
<jgraham>
In all honesty, if you want to write a validator, adding parse error detection to the parser isn't the most difficult part :)
09:06
<MikeSmith>
heh
09:06
<hsivonen>
yolin: you could try compiling the Validator.nu code to JS using GWT
09:06
<hsivonen>
yolin: you'd need to rework parts that touch java.io
09:06
<hsivonen>
at least
09:07
<hsivonen>
yolin: or you could use GWT only for the non-schema parts and use a JS-based RELAX NG engine for the schema parts
09:07
<yolin>
lemme check validator.nu
09:08
<jgraham>
FWIW I think Hixie's idea was that you implement one class per tag and let that class be responsible for running whatever checks are relevant to that tag. I guess tehre was more to it than that, but that's al I remember at the moment
09:09
<yolin>
yes, thats seems like a good idea to handle the complexity
09:13
<MikeSmith>
jgraham: yeah, that's the way I remember it too
09:13
<MikeSmith>
so it would all be custom code specifically for checking HTML
09:13
<MikeSmith>
rather than a generic mechanism
09:14
<MikeSmith>
and one big advantage is that you could fine-tune the error messages
09:15
<MikeSmith>
hsivonen: so I think it would be useful to streamline and consolidate that presets in the validator
09:15
<MikeSmith>
e.g., ARIA and SVG and MathML are all now actually part of HTML
09:15
<hsivonen>
MikeSmith: yeah, makes sense to flatten those instead of offering them as options
09:15
<MikeSmith>
so I think we don't really need separate options for those
09:16
<MikeSmith>
OK
09:16
<MikeSmith>
will do that too, then
09:16
<hsivonen>
though the UI should probably still say which version of MathML and SVG it is using as a reference
09:16
<MikeSmith>
hsivonen: also do we actually want to keep the separate SVG 1.1 checker?
09:17
<MikeSmith>
I wonder if anybody actually uses that or wants it
09:17
<hsivonen>
also, let's get rid of "(experimental)". this stuff is now real in practice
09:17
<MikeSmith>
yeah
09:17
<MikeSmith>
that too
09:17
<shepazu>
I would prefer to keep an SVG 1.1 checker… I don't think that most SVGs are (or even will) be in an HTML document
09:18
<MikeSmith>
shepazu: OK
09:18
<MikeSmith>
there's a problem with the schema, though
09:18
<hsivonen>
MikeSmith: I think we should keep the schema that allows <svg> as the root element, yeah
09:18
<MikeSmith>
hai
09:18
<MikeSmith>
shepazu: I'm trying to recall, but remember I e-mailed the SVG WG a while back?
09:18
<hsivonen>
XHTML Basic can probably be zapped
09:19
<MikeSmith>
OK
09:19
<MikeSmith>
will do that
09:19
<hsivonen>
MikeSmith: thank you
09:19
<shepazu>
MikeSmith: emailed about what?
09:19
<MikeSmith>
shepazu: to ask that the SVG 1.1 relaxng schema be updated to fix ... something
09:19
<MikeSmith>
I would need to check my mail to see
09:19
<shepazu>
oh, yeah
09:20
<shepazu>
we'd need to make a new SVG RNG
09:20
<MikeSmith>
yeah
09:20
<shepazu>
berjon was going to do that for us
09:20
<shepazu>
but...
09:20
<MikeSmith>
the response from heycam and chris at the was, that SVG 1.1 is not an official product of the SVG WG and they are not maintaining it
09:21
<shepazu>
yeah
09:21
<MikeSmith>
which kind of bummed me out to hear that
09:21
<MikeSmith>
kind of put a damper on my enthusiasm for pursuing the fix
09:21
<MikeSmith>
whatever it was
09:22
<MikeSmith>
anyway, i'll look back at it at some point
09:22
<shepazu>
wish we had the resources to make an RNG
09:22
<shepazu>
I'll bring it up again
09:22
<MikeSmith>
I think I may have already made the change in the validator.nu copy of the schema
09:22
<MikeSmith>
can't remember
09:22
<MikeSmith>
but I think maybe all I was asking was for that change to be ported back upstream
09:22
<MikeSmith>
ah yeah yeah yeah
09:23
<MikeSmith>
I remember now
09:23
<MikeSmith>
it's because you dudes made a change to the spec, like, just a few months ago!!
09:23
<MikeSmith>
this "second edition" trickery or whatever
09:24
<MikeSmith>
where you try to pretend it's not, um, a living standard that occasionally might need updating
09:24
<MikeSmith>
made a change that loosened some constraint
09:25
<shepazu>
aren't all specs "living standards"?
09:27
<MikeSmith>
well, the good ones are at least
09:27
<MikeSmith>
if by "living" it means acknowledging that they are going to continue to be updated
09:28
<MikeSmith>
and not updated just by addition of errata or subsequent "editions"
09:31
<MikeSmith>
shepazu: anyway, since we are going to be continuing to have that SVG 1.1 standalone-checking option, it would be good to have somebody take ownership of that schema and make sure it actually checks the constraints in the current spec
09:31
<MikeSmith>
actually this also affects the HTML checker
09:31
<MikeSmith>
because that same schema is used to check SVG in HTML
09:32
<shepazu>
yeah, but nobody on the SVG WG knows how to make an RNG (as far as I know)
09:32
<MikeSmith>
well
09:32
<MikeSmith>
geez
09:32
<MikeSmith>
when I started working with RNG I didn't know either
09:33
<MikeSmith>
and it ain't rocket science man
09:33
<MikeSmith>
anybody who has ever edited a DTD can learn it easily
09:33
<MikeSmith>
the compact syntax at least
09:33
<MikeSmith>
not the XML syntax
09:34
jgraham
keeps reading RNG as "Random Number Generator"
09:34
<shepazu>
if you want to help us, MikeSmith, that would make it easier
09:35
<shepazu>
but right now, I'm going to sleep
09:36
<MikeSmith>
hai
09:36
<MikeSmith>
I can definitely help you
09:36
<MikeSmith>
get your beauty sleep :)
09:37
<zcorpan>
MikeSmith: hey if you could do my manual testing today i could leave early and go to the gym or so
09:38
<zcorpan>
or doesn't it work like that? :(
09:40
<MikeSmith>
heh
09:40
<MikeSmith>
manual testing should be something given as penace, during confession
09:40
<MikeSmith>
that would be a genuine incentive to not sin further
09:42
<jgraham>
It kind of does work like that
09:42
<jgraham>
But with the normal assumption that we are all sinners
12:41
<david_carlisle>
hsivonen_: "though the UI should probably still say which version of MathML and SVG it is using as a reference", if you belived the whatwh living standards mantra you'd use the latest and greatest version of each (and I wouldn't have had to knock out a mathml2/3 hybrid: for you:-)
12:42
<david_carlisle>
s/whatwh/whatwg/
13:35
<jgraham>
david_carlisle: (that sort of requires that the people writing the spec also believe in the living standard mantra and don't introduce big backwards-incompatible changes or complexity that prevents people from adopting the new version)
13:35
<jgraham>
(not sure if that applies to MathML 2 -> 3 or not)
14:10
<davidb_>
hi all!
14:29
<david_carlisle>
jgraham: MathML3 -> 2 hasn't got backward compat issues, but Mike and Henri wanted to be able to keep validator closer to mathMl2 as that's closer to what browsers implement today (apart from IE/mathPlayer which does mathML3) (same reason people other than browser makers ask for versioned html, actually)
15:40
<Ms2ger>
"SVG 1.1 is not an official product of the SVG WG"
15:40
<Ms2ger>
Say what?
15:42
<shepazu>
Ms2ger: MikeSmith meant to say that the SVG 1.1 *RNG* was not produced by the SVG WG, and that we don't maintain it
15:42
<shepazu>
the RelaxNG schema
15:42
<shepazu>
not the spec
15:42
<Ms2ger>
That makes slightly more sense
15:44
<Ms2ger>
Totally not what he said, though
17:36
<jgraham>
I love it when irccloud dies
17:36
<jcranmer>
looks ike it
17:36
<Ms2ger>
jgraham, so, my review comments? :)
17:37
<jgraham>
Ms2ger: I fixed them didn't I?
17:37
<jgraham>
Dunno how one says on github "this commit fixes these issues"
17:38
<Ms2ger>
Not in HG
17:38
<jgraham>
Hmm OK
18:00
<jgraham>
Ms2ger: Pushed
18:21
<Ms2ger>
jgraham, thanks