| 03:22 | <GPHemsley> | Hixie: To confirm, I should nix sniffing archives specifically, but keep the rest? |
| 03:23 | <GPHemsley> | Or comment out the rest? |
| 03:54 | <GPHemsley> | Interesting... |
| 03:54 | <GPHemsley> | apparently tagging a <tr> with class=XXX adds an extra cell to the row |
| 03:56 | <GPHemsley> | ...which seems to stem from having content :before |
| 05:10 | <navin> | hi |
| 05:12 | <navin> | hello |
| 05:56 | <kennyluck> | Was there ever a thread on whatwg talking about changing <img> to a non-void element? |
| 05:56 | <kennyluck> | I am being asked by some folks about this idea. I am pretty sure it's a crazy idea, but it would be nice to share some old discussions of course. |
| 06:06 | <Aniruddh> | Hi |
| 06:24 | <Aniruddh> | Welcome to #whatwg dbaron |
| 06:24 | <kennyluck> | orz |
| 06:25 | <dbaron> | Aniruddh, it's better not to welcome individual people (especially when you don't know them); many people have IRC clients configured to join many channels at once, and asking for their attention is unnecessary. |
| 06:26 | <Aniruddh> | Welcome to #whatwg Somatt_wrk |
| 06:29 | <Aniruddh> | Welcome to #whatwg ehsan |
| 06:30 | <Aniruddh> | Welcome to #whatwg tomasf |
| 06:30 | <gavin> | Aniruddh: are you a bot? |
| 06:33 | <Aniruddh> | Welcome to #whatwg yoshiaki |
| 06:33 | <yoshiaki> | thank you, Aniruddh |
| 06:46 | <Aniruddh> | Welcome to #whatwg mattgifford |
| 06:47 | <Aniruddh> | Welcome to #whatwg Transfusion |
| 06:48 | <Aniruddh> | Welcome to #whatwg maikmerten |
| 06:48 | <Aniruddh> | Welcome to #whatwg Mr-ex777 |
| 06:50 | <Aniruddh> | Welcome to #whatwg capella |
| 06:51 | capella | mozillians be there |
| 06:52 | <Aniruddh> | Welcome to #whatwg jayne |
| 06:53 | <Aniruddh> | Welcome to #whatwg mattgifford |
| 06:53 | <Aniruddh> | Welcome to #whatwg yoshiaki |
| 06:55 | <MikeSmith> | Hixie: if you have a second can you maybe please kick Aniruddh |
| 06:55 | <MikeSmith> | seems like it must be some bot |
| 06:58 | <hsivonen> | twisting polyglot enough to lose the endorsement of TAG members? https://www.w3.org/Bugs/Public/show_bug.cgi?id=19925#c7 |
| 06:59 | <Aniruddh> | Hi |
| 07:00 | <Aniruddh> | Sorry, I had bot mode activated |
| 07:00 | <Aniruddh> | I was afk |
| 07:05 | <MikeSmith> | hsivonen: clearly the solution here is, we need two specs: one "HTML/XHTML Compatibility Authoring Guide" and one "Robust Profile for Documents which Compensates for Imperfect Parsers" https://www.w3.org/Bugs/Public/show_bug.cgi?id=19925#c9 |
| 07:08 | <annevk> | zcorpan: looks updated to me |
| 07:09 | <annevk> | zcorpan: I think GitHub had some issues yesterday, might be a hickup on their end |
| 07:22 | <MikeSmith> | btw I added collection of statistics on POST input type (textarea vs file input vs entity-body) to the validator statistics code |
| 07:25 | <MikeSmith> | hsivonen: looking at the stats for validator.nu vs html5.validator.nu it seems like they're being used quite differently |
| 07:26 | <MikeSmith> | e.g., GET-based input for validator.nu is only 6% |
| 07:27 | <MikeSmith> | also 89% JSON output |
| 07:27 | <MikeSmith> | which would seem to be from some third-party service |
| 07:27 | <MikeSmith> | or services |
| 08:01 | <hsivonen> | MikeSmith: I suspect the main user of validator.nu is Roger Johansson’s Firefox extension, but that’s just a guess |
| 08:02 | <MikeSmith> | OK |
| 10:12 | <matjas> | ok, http://mathias.html5.org/specs/javascript/ 301s to http://javascript.spec.whatwg.org/ |
| 10:27 | <annevk> | matjas: btw, if you want you can use the new xref repository |
| 10:27 | <annevk> | matjas: https://github.com/whatwg/xref |
| 10:27 | <annevk> | matjas: specification-data is no longer maintained |
| 10:28 | <annevk> | matjas: see the other specs for how to make Anolis use it |
| 10:28 | <darobin> | ah, that's good to know for my integration project too |
| 10:28 | <annevk> | darobin: ah yeah, was meaning to ping you about that at some point |
| 10:28 | <annevk> | darobin: and maybe figure out if I can patch Anolis to make it all work somehow |
| 10:28 | <darobin> | annevk: no worries, I'm rather far behind on actually putting that together tbh |
| 10:29 | <darobin> | I'd initially planned to do that during TPAC, but TTWF folks seemed to want a WebIDL.js update more |
| 10:31 | <matjas> | annevk: thanks, will do |
| 10:31 | <matjas> | annevk: btw how/when does javascript.spec.whatwg.org get updated? are you using GitHub hooks for that, or what? |
| 10:32 | <annevk> | matjas: yes, see the admin panel |
| 10:32 | <annevk> | matjas: a shell script fetches index.html from GitHub |
| 10:32 | <annevk> | matjas: note that using whatwg/xref requires updating Anolis |
| 10:33 | <matjas> | lol @ cgi-script URL |
| 12:16 | <zcorpan> | http://resources.whatwg.org/logo-xhr.svg |
| 12:19 | <beverloo> | haha |
| 12:19 | <beverloo> | not everone may get that reference :) |
| 12:20 | <zcorpan> | that's fine |
| 12:21 | <zcorpan> | wonder why firefox doesn't render all lines |
| 12:22 | <kennyluck> | Is there any presentation in http://wiki.whatwg.org/wiki/Presentations that talks about why folks should participate in the Standards community? |
| 12:23 | <zcorpan> | hmm seems that page hasn't been updated lately :-( |
| 12:25 | <zcorpan> | kennyluck: quick scan suggests no |
| 12:25 | <zcorpan> | kennyluck: what's the audience? |
| 12:25 | <kennyluck> | zcorpan, hmm OK thanks. |
| 12:26 | <kennyluck> | I am going to do a video presentation on Sunday to kids so I am brainstorming on how to persuade them to join these kids of effort. |
| 12:26 | <kennyluck> | You know, kids like to create apps. |
| 12:27 | <kennyluck> | Perhaps I would say saying in this line "by contributing to standards, you are for sure not doing any make work (meaning that Standards never fails). " |
| 12:27 | <kennyluck> | s/kids of effort/kinds of effort/ |
| 12:28 | <zcorpan> | "If you want to be in the Time Wasting Business, choose Standards!" |
| 12:28 | <kennyluck> | *shrug* |
| 12:30 | <zcorpan> | maybe point out that it's easier to contribute than they might think, and that it's not a magic castle |
| 12:31 | <zcorpan> | i mean, it's easier to get involved than they might think. getting what they want done get done might be harder than they think |
| 12:32 | <kennyluck> | Yeah, that makes sense. |
| 12:33 | <jgraham> | It's not a magic castle? |
| 12:33 | jgraham | cries |
| 13:08 | <zcorpan> | seems i hit https://bugzilla.mozilla.org/show_bug.cgi?id=629897 |
| 13:17 | <kennyluck> | zcorpan, yeah, I think the rule here is that you can't have a comma before a command. |
| 13:17 | <zcorpan> | opera and webkit ignore the comma |
| 13:19 | <zcorpan> | it seems like a silly restriction. d="" allows you to omit commands, commas, zeros, spaces, and do all sorts of things, but is unforgiving when it comes to an *extra* comma |
| 13:25 | <zcorpan> | is "const" specified yet? |
| 13:32 | <MikeSmith> | zcorpan: just replied to that bug |
| 13:33 | <MikeSmith> | I guess to fix this I get to become the maintainer for the SVG 1.1 schema |
| 13:33 | <MikeSmith> | since the SVG WG refuses to take any responsibility for it |
| 13:35 | <MikeSmith> | ah great |
| 13:35 | <MikeSmith> | attribute d { PathData.datatype } |
| 13:35 | <MikeSmith> | PathData.datatype = xsd:string |
| 13:36 | <MikeSmith> | gotta love the precision there |
| 13:36 | <zcorpan> | anything goes? |
| 13:36 | <MikeSmith> | yeah |
| 13:37 | <zcorpan> | excellent |
| 13:37 | <kennyluck> | zcorpan, (re. opera and webkit ignore the comma) That's not quite true either, you can't put a comma but a space after a command in WebKit and Opera, so… *shrug* |
| 13:37 | <MikeSmith> | $ grep "datatype =" * | wc -l |
| 13:37 | <MikeSmith> | 52 |
| 13:37 | <MikeSmith> | $ grep "datatype =" * | grep "xsd:string" | wc -l |
| 13:37 | <MikeSmith> | 45 |
| 13:37 | <zcorpan> | kennyluck: ok |
| 13:38 | <zcorpan> | MikeSmith: so now you have something to work on eh? |
| 13:38 | <MikeSmith> | hah |
| 13:38 | <MikeSmith> | yeah |
| 13:43 | <MikeSmith> | oh god |
| 13:44 | <MikeSmith> | It's not even clear from the spec what the microsyntax of that d attribute is supposed to be |
| 13:44 | <MikeSmith> | http://www.w3.org/TR/SVG11/paths.html#DAttribute |
| 13:44 | <MikeSmith> | "The definition of the outline of a shape. See Path data." |
| 13:44 | <MikeSmith> | ok, so let's take a look at the definition of "Path data" |
| 13:44 | <MikeSmith> | http://www.w3.org/TR/SVG11/paths.html#PathData |
| 13:47 | <MikeSmith> | ah sweet |
| 13:47 | <MikeSmith> | http://www.w3.org/TR/SVGTiny12/paths.html#PathDataBNF |
| 13:48 | <zcorpan> | also http://www.w3.org/TR/SVG11/implnote.html#PathElementImplementationNotes |
| 13:48 | <MikeSmith> | yes, let's define the allowed value for a attribute using BNF |
| 13:54 | <MikeSmith> | zcorpan: so what is it basically? just a comma-separated list of three items? |
| 13:55 | <zcorpan> | it's not necessarily comma-separated, but it's a flat list |
| 13:55 | <MikeSmith> | god almighty |
| 13:55 | <MikeSmith> | and the items in the list can have spaces in them? |
| 13:55 | <zcorpan> | you have a command, then something following it depending on what the command is |
| 13:55 | <zcorpan> | yes |
| 13:55 | <MikeSmith> | wow |
| 13:57 | <zcorpan> | it would be nice to have helpful error messages here, like "hey dude you have a comma here that i didn't expect, remove it" rather than "attribute "d" invalid" |
| 13:57 | <zcorpan> | i'm not sure what that implies in terms of impl |
| 13:57 | <zcorpan> | but i guess more of a parser than a regex |
| 13:58 | <jgraham> | Yeah, I was going to say :) |
| 13:59 | <MikeSmith> | well there's a way to use a regexp with capturing groups that allows further checking of the capturing groups |
| 13:59 | <MikeSmith> | it's what we're using to check datetime values for the time element |
| 13:59 | <zcorpan> | ok |
| 14:00 | <MikeSmith> | I wonder if anybody has written a regexp for svg path data already |
| 14:00 | <MikeSmith> | if so that'd be a start at least |
| 14:00 | jgraham | wonders if java has something more like parsec |
| 14:01 | <hsivonen> | about implied </p> being a feature: https://twitter.com/julien_igel/status/268175514261401601 |
| 14:01 | <MikeSmith> | jgraham: don't know what parsec is but I doubt java has anything more than the matcher stuff we're using |
| 14:02 | <jgraham> | Haskell parser library |
| 14:02 | <jgraham> | Supposed to be good, but I haven't ever used it :) |
| 14:02 | <MikeSmith> | ok |
| 14:03 | <MikeSmith> | "An inline <div> is spelled <span> ever since HTML 4. " |
| 14:03 | <MikeSmith> | heh |
| 14:42 | <MikeSmith> | oh boy |
| 14:42 | <MikeSmith> | zcorpan: http://www.w3.org/Graphics/SVG/2002/01/pathdata.re |
| 14:45 | <hsivonen> | “now you have two problems” |
| 14:46 | <darobin> | hsivonen: it gets a lot more interesting if you use that regex in an XML Schema type :) |
| 14:48 | <MikeSmith> | do we have Dean to thank for that? |
| 14:48 | <MikeSmith> | http://tech.groups.yahoo.com/group/svg-developers/message/10650 |
| 14:49 | <darobin> | MikeSmith: yeah, though IIRC there are later versions by me lying around |
| 14:49 | <MikeSmith> | ok |
| 14:50 | <darobin> | honestly though I really don't think that you want a regex here |
| 14:51 | <zewt> | hsivonen: only two? |
| 14:51 | <darobin> | this is substantially more complicated than datetime |
| 14:51 | <darobin> | or rather, it's substantially longer so providing good error reporting on a regex is going to be painful at best |
| 14:52 | <darobin> | it's not impossible that the current SVG RNG is based on the one I made a loooong time ago, if so I can probably help with tweaking |
| 14:52 | <darobin> | (assuming I recall how that stuff works) |
| 14:53 | <darobin> | you'll find that most of the attribute values are indeed strings because there's rarely much useful stuff you can do within the constraints of XSD simple types |
| 14:53 | <shepazu> | darobin: we'd love if you could help us with an RNG for SVG 2 |
| 14:53 | <darobin> | (which is what RNG uses, or rather, what's implemented) |
| 14:54 | <darobin> | shepazu: that might be possible at some point :) |
| 14:59 | <MikeSmith> | what's the current home for anolis? |
| 14:59 | <MikeSmith> | Ms2ger: your repo? |
| 15:01 | <Ms2ger> | Yeah |
| 15:03 | <MikeSmith> | ok |
| 15:03 | <MikeSmith> | thanks |
| 15:16 | <MikeSmith> | zcorpan: I find http://xn--dahlstrm-t4a.net/svg/tools/scour/svg_regex.py from Erik |
| 15:16 | <MikeSmith> | which might be port-able |
| 15:17 | <MikeSmith> | and which seems to do at least some actual error reporting |
| 15:18 | <zcorpan> | yeah |
| 15:19 | <zcorpan> | wonder if it's up to date |
| 15:19 | <MikeSmith> | dunno but if it's not I'd guess Erik might could be convinced to update it |
| 15:21 | <matjas> | ping jgraham |
| 15:22 | <jgraham> | matjas: If this is about escapes, I don't think I know the answer :) |
| 15:22 | <matjas> | jgraham: ah, i was just wondering if you got the message or not |
| 15:22 | <matjas> | jgraham: fair enough. any idea where this requirement originates from? |
| 15:22 | <jgraham> | Yeah, sorry, I have been a bit behind on replying |
| 15:22 | <matjas> | no worries |
| 15:23 | <jgraham> | I have no memory of it at all. Did I actually add it to the original wiki page or was it someone else? |
| 15:23 | <matjas> | it was you :) |
| 15:23 | <jgraham> | Oh |
| 15:23 | <jgraham> | Maybe I had been hyponotized |
| 15:24 | <matjas> | http://wiki.whatwg.org/index.php?title=Web_ECMAScript&diff=3938&oldid=3798 |
| 15:24 | <TabAtkins> | jarek, annevk: The IRI without url() behavior is totally bogus. WebKit has several serialization bugs where the value obviously won't round-trip. |
| 15:25 | <matjas> | jgraham: ok, i’ll remove it then. thanks! |
| 15:26 | <TabAtkins> | darobin: Pathdata sounds like a relatively trivial regex to write, actually. There's nothing optional, and the structure is simple. |
| 15:26 | <TabAtkins> | I could understand error-reporting being annoying, though. |
| 15:27 | <zcorpan> | TabAtkins: "There's nothing optional" doesn't sound right |
| 15:28 | <TabAtkins> | All of the path commands have a very strict syntax. |
| 15:28 | <darobin> | TabAtkins: I didn't mean that a regex matching it is hard (it isn't) but one that provides good error reporting is less fun |
| 15:29 | <TabAtkins> | I don't even know how you get "good error reporting" out of a regex. |
| 15:30 | <darobin> | you haven't done enough Perl :) |
| 15:30 | <TabAtkins> | This is true. |
| 15:32 | <MikeSmith> | TabAtkins: https://bitbucket.org/validator/syntax/src/default/relaxng/datatype/java/src/org/whattf/datatype/AbstractDatetime.java?at=default#cl-191 |
| 15:32 | <Ms2ger> | You've done more than enough perl if you've never done perl |
| 15:33 | darobin | thought he heard some strange hissing sound. Weird. |
| 15:35 | <Ms2ger> | That must've been me spitting at the sound of "perl" |
| 15:36 | <darobin> | Ms2ger: I'd troll you back, but frankly with Python it just seems unfair |
| 15:41 | <MikeSmith> | hmm Java doesn't have anything like Python "yield" |
| 15:41 | <MikeSmith> | as far as I know |
| 15:42 | <MikeSmith> | (I mention because Erik's pathdata parser uses it) |
| 15:45 | <MikeSmith> | maybe the pathdata parsing code from Batik might be usable for validation |
| 15:46 | <darobin> | MikeSmith: googling around brings up some Java implementation of yield |
| 15:46 | <darobin> | (not sure if any of those is any good) |
| 15:46 | <MikeSmith> | ok |
| 15:47 | <darobin> | e.g. http://jimblackler.net/blog/?p=61 |
| 15:47 | <MikeSmith> | thanks |
| 15:47 | <jgraham> | MikeSmith: Surely in java you wouldn't use that, but would use a class to store state |
| 15:48 | jgraham | doesn't really know how iteration in java typically works |
| 15:49 | <darobin> | jgraham: you cast a subclass of the iterator interface through using a pattern factory |
| 15:54 | <jgraham> | Can't tell if joking or java |
| 15:54 | <MikeSmith> | hah |
| 16:02 | <MikeSmith> | so http://svn.apache.org/repos/asf/xmlgraphics/batik/branches/svg11/sources/org/apache/batik/parser/PathParser.java looks like it might actually be useable as a starting point at least |
| 16:05 | <darobin> | MikeSmith: if you need help pinging heycam is prolly the right thing to do there |
| 16:05 | <MikeSmith> | ok |
| 16:41 | <dglazkov> | good morning, Whatwg! |
| 17:17 | <heycam> | MikeSmith, there's some info here on using Batik's parser classes: http://xmlgraphics.apache.org/batik/using/parsers.html |
| 17:26 | <[tm]> | heycam: thanks |
| 17:28 | <[tm]> | hoping i can repurpose some of that code without needing to depend on a bunch of other batik code |
| 17:29 | <[tm]> | the stuff in the abstract superclass |
| 17:31 | <heycam> | I can't remember how tightly it is entwined with the rest of it, but hopefully not too much |
| 17:33 | <[tm]> | anyway in trying to prioritize my choice of what to work on as far as my contributions to the validator code, a simple rule I've tried to follow is, if zcorpan files a bug, i try to fix it |
| 17:34 | <[tm]> | that rule has worked out pretty well so far |
| 17:34 | <heycam> | sounds like a fine strategy |
| 17:34 | <[tm]> | yeah |
| 17:35 | <jgraham> | That is also how Opera's bugtracker assigns priority |
| 17:35 | <jgraham> | (not really) |
| 17:35 | <jgraham> | (but maybe it should) |
| 18:31 | <GPHemsley> | annevk: Can you figure out what's going on with the encoding here? http://atlc.sourceforge.net/bmp.html#_toc381201084 |
| 18:35 | <GPHemsley> | (The tagged encoding is clearly not correct.) |
| 18:46 | GPHemsley | wonders who it helps to have file format specs use C code instead of describing the bytes. |
| 18:46 | <jsbell> | hackers? |
| 19:19 | <Hixie> | GPHemsley: implementors, they can copy the data straight out |
| 20:29 | <Yuhong> | Just installed IE10 for Win7 Release Preview released today! |
| 20:29 | <Yuhong> | Results for http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1905 in IE10 mode: false, true |
