| 01:44 | <TabAtkins> | Argh, I have too many CDs and games to buy right now. >_< Luckily my birthday's coming up, so I should have some cash to spend. |
| 01:47 | <TabAtkins> | http://i.imgur.com/G7WyP.gif |
| 01:56 | <cardona507> | TabAtkins: when is your bday? how old are you gonna be? |
| 01:56 | <TabAtkins> | mid november, and 25. |
| 01:57 | <cardona507> | interesting - for some reason I thought you were older than that |
| 01:58 | <TabAtkins> | It is clearly the beard. |
| 01:59 | <cardona507> | or the wealth of hacker knowledge |
| 02:01 | <TabAtkins> | ^_^ |
| 09:54 | <zcorpan_> | annevk: could you point to navigator.resolveURL? |
| 09:55 | <annevk> | what do you mean? |
| 09:55 | <zcorpan_> | i mean i can't find it in the spec |
| 09:55 | <annevk> | oh, it's location, sorry |
| 09:55 | <annevk> | http://www.whatwg.org/specs/web-apps/current-work/complete.html#dom-location-resolveurl |
| 09:56 | <zcorpan_> | ah |
| 09:57 | <zcorpan_> | WorkerLocation doesn't have it though |
| 09:57 | <annevk> | oh, oversight? |
| 09:58 | <zcorpan_> | dunno |
| 09:58 | <annevk> | as I said, I've not been involved |
| 09:59 | <annevk> | ]http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2008-November/017275.html |
| 10:01 | <annevk> | ah, it's for passing URLs to workers |
| 10:18 | <annevk> | AryehGregor, you could just move http://www.w3.org/Bugs/Public/show_bug.cgi?id=10267 to the WebIDL component of the WebApps WG |
| 10:28 | <zcorpan_> | annevk: setAttributeNS(xmlns_ns, 'xmlns', 'foo') should be allowed, also setAttributeNS(null, 'xmlnslol', 'foo') should be allowed |
| 10:29 | <annevk> | are they not? (but why?) |
| 10:29 | <annevk> | oh, I guess setAttribute should only ban "xmlns" and starts with "xmlns:", not starts with "xmlns" |
| 10:31 | <zcorpan_> | yeah, and they should be allowed if the namespace argument is right, iirc |
| 10:32 | <annevk> | for compat you mean |
| 10:32 | <annevk> | ? |
| 10:33 | <annevk> | manipulating namespace attributes is an extremely bad idea |
| 10:33 | <zcorpan_> | it's what browsers do (iirc), dunno if it's required for compat |
| 10:33 | <annevk> | oh sure |
| 10:33 | <annevk> | but it doesn't make sense |
| 11:10 | <annevk> | using the correct dashes is tricky |
| 11:13 | <MikeSmith> | I'm curious what editing apps others use for editing javascript |
| 11:13 | <annevk> | I use TextWrangler at the moment, but then I do not code anything serious in JavaScript... |
| 11:14 | <annevk> | or at all, really :) |
| 11:15 | <MikeSmith> | I'm using vim but I guess I would be nice to have something that docs better completion |
| 11:16 | <MikeSmith> | for Java I have this "eclim" thing set up that hooks vim into eclipse |
| 11:16 | <MikeSmith> | I suppose I could try that for js files and see see what I get |
| 11:16 | <annevk> | I wonder if there've been studies done to how much time that actually buys you |
| 11:16 | <MikeSmith> | *does better completion |
| 11:17 | <annevk> | it seems to me that figuring things out is the main problem; writing it down not so much |
| 11:18 | <annevk> | but maybe if you're not familiar with the APIs it helps |
| 11:18 | <annevk> | if the autocompletion comes with a few lines of documentation |
| 11:18 | <MikeSmith> | I think for the Java case, it buys a lot -- because of relatively high overhead (for lack of a better word) of programming in Java |
| 11:18 | <MikeSmith> | java is really unforgiving |
| 11:20 | <MikeSmith_> | and I find at least that using eclipse makes it a lot easier to catch certain kinds of problems |
| 11:20 | <MikeSmith_> | for java I mean |
| 11:20 | <MikeSmith_> | js is a lot different language of course |
| 11:22 | <MikeSmith_> | but I've gotten used to checking stuff using the js interpreter consoles like in firebug and web inspector and dragonfly |
| 11:22 | <MikeSmith_> | and node.js shell |
| 11:22 | <MikeSmith_> | and having completion on, e.g., DOM methods |
| 11:23 | <MikeSmith> | so having it for editing would seem like a plus |
| 11:30 | <Moo^> | MikeSmith: Aptana Studio (Eclipse distribution for web developers) comes with decent Javascript editor which can do some static analysis |
| 11:30 | <MikeSmith> | Moo^: thanks |
| 11:30 | <MikeSmith> | I seem to remember reading about that somewhere else as well |
| 11:31 | <Moo^> | MikeSmith: and I assume you are used to Eclipse so it is a perfect fit for you |
| 11:32 | <MikeSmith> | well, Eclipse itself is not a perfect fit for me |
| 11:33 | <MikeSmith> | I started using Eclipse just for java editing |
| 11:33 | <MikeSmith> | and I don't actually use Eclipse directly |
| 11:33 | <MikeSmith> | I use vim |
| 11:33 | <MikeSmith> | and I essentially run eclipse headless, with it hooked into vim through this eclim thing |
| 11:34 | <MikeSmith> | anyway, I was also just trying DashCode |
| 11:34 | <MikeSmith> | by accident |
| 11:35 | <MikeSmith> | because that seems to be what my OSX environment is set up to use by default for javascript |
| 11:36 | <MikeSmith> | and it actually seems to have pretty extensive built-in completion |
| 11:43 | <annevk> | http://iscrockford.partofhtml5.com/ haha |
| 11:48 | <JerichoM> | hey lhnz |
| 11:48 | <MikeSmith> | Crockford _parts_ HTML5 -- like Moses parting the Red Sea |
| 11:48 | <annevk> | that site has terrible markup btw |
| 11:49 | <annevk> | http://areunicorns.partofhtml5.com/ doesn't even validate |
| 11:49 | <JerichoM> | did you enjoy port scanning me? |
| 11:50 | <MikeSmith> | http://isvalidation.partofhtml5.com would be a good one |
| 11:53 | <MikeSmith> | annevk: btw, I was using Berjon's respec thing recently and noticed something that is a serious misfeature in it |
| 11:53 | <MikeSmith> | and I think maybe something like it is in anolis and the css3 postprocessor thing |
| 11:54 | <MikeSmith> | which is, it tries to turn instances of "[[foo]]" into bibliographic citations |
| 11:55 | <MikeSmith> | the whole thing of using macros like that -- or whatever the hell people call them -- is just fricking dumb |
| 11:55 | <MikeSmith> | I mean to do it in an HTML file |
| 11:55 | <MikeSmith> | this is what we have markup for |
| 11:56 | <annevk> | ms2ger added a <span data-anolis-ref> feature to Anolis (or some such; see Web DOM Core source file) |
| 11:56 | <MikeSmith> | good |
| 11:57 | <MikeSmith> | because the problem with that asshatted [[foo]] stuff is, what happens if the source that you're using refspec or whatever to generate output from actually already uses doubled square brackets to mean something? |
| 11:57 | <MikeSmith> | which is not a hypothetical problem |
| 11:58 | <MikeSmith> | case in point is the ES5 spec |
| 11:58 | <MikeSmith> | which has 1243 instances of terms in double square brackets |
| 12:01 | <JerichoM> | lol, lhnz... lol a PSH |
| 12:23 | <annevk> | mijn.ing.nl really should use autofocus="" |
| 12:23 | <annevk> | it always screws up my focus after I'm almost done having me overwrite my username with part of my password -- so annoying |
| 12:39 | <david_carlisle> | <MikeSmith> this is what we have markup for |
| 12:40 | <david_carlisle> | you should use xmlspec and xslt, using XML makes you a happier person |
| 12:40 | <MikeSmith> | heh |
| 12:41 | <MikeSmith> | david_carlisle: don't tell me you're actually using xmlspec? |
| 12:41 | <MikeSmith> | if so, I think you must be the the last one using it :) |
| 12:41 | <hsivonen> | what does xmlspec solve that HTML+Anolis don't? |
| 12:41 | <MikeSmith> | it seems like nobody's been maintaining that in a long time |
| 12:42 | <MikeSmith> | david_carlisle: Robin Berjon's respec is worth taking a look at |
| 12:42 | <david_carlisle> | Not actually tried anolis but the idea of using html in authoring is just scary. |
| 12:42 | <david_carlisle> | xmlspec is extensively maintained but massively forked |
| 12:42 | <david_carlisle> | mathml has a fork |
| 12:43 | <david_carlisle> | xslt/xquery have one |
| 12:43 | <david_carlisle> | xml itself uses it still I think |
| 12:43 | <david_carlisle> | and so it goes on |
| 12:43 | <MikeSmith> | great thing about respec is it doesn't require you to install anything |
| 12:44 | <MikeSmith> | the transformation is all done inside your browser |
| 12:44 | <MikeSmith> | it gives you a way to save a copy of the generated DOM |
| 12:44 | <MikeSmith> | this reminds me, Google Docs has a better way |
| 12:44 | <david_carlisle> | could respec pull all the data out of unicode.xml to make th mathml operator dictionary on the fly, or similarly typeset all the character tables in the xml entites spec (just from the same markup) or colour an cross link a relanx ng file? |
| 12:45 | <MikeSmith> | if you want to save a copy of a document, it generates it as a download |
| 12:45 | <MikeSmith> | respec should do that same kind of download thing |
| 12:45 | <MikeSmith> | david_carlisle: sure it could |
| 12:45 | <david_carlisle> | "colud" but do it |
| 12:46 | <david_carlisle> | anyway I want to type my specs in emacs (I saw earlier you are a vim user, so you wouldn't understand the love) |
| 12:46 | <MikeSmith> | I used to use emacs a lot too |
| 12:46 | <MikeSmith> | nxml-mode |
| 12:47 | <MikeSmith> | I may start using it again more, once hober finishes his elisp text/html parser |
| 12:47 | <david_carlisle> | yes I saw that... |
| 12:48 | <david_carlisle> | I just finally this year (after 23 years) cracked and stopped using email as my regular mail client |
| 12:48 | <MikeSmith> | "email" is the name of an actual app? |
| 12:48 | <david_carlisle> | sorry emacs not email |
| 12:49 | <david_carlisle> | I can't type:-) |
| 12:49 | <MikeSmith> | ah |
| 12:49 | <MikeSmith> | yeah, I tried that for a while |
| 12:49 | <MikeSmith> | GNUs |
| 12:49 | <MikeSmith> | I think Haakon still uses Emacs for mail |
| 12:49 | <MikeSmith> | Gnus or something |
| 12:50 | <MikeSmith> | david_carlisle, hsivonen : while you're both here, have you guys chatted more about annotation-xml handling? |
| 12:50 | MikeSmith | is not caught up on bug mail |
| 12:51 | <david_carlisle> | well we had a conversation the other day somewhere (here? in mail, can't remeber) |
| 12:52 | <david_carlisle> | I think/hope the outcome is that he's going to look again if the spec change as outlined by hixie in the bug entry could work in FF4 timeframe |
| 12:52 | <MikeSmith> | last I saw was Henri's comment http://www.w3.org/Bugs/Public/show_bug.cgi?id=9887#c27 |
| 12:52 | <MikeSmith> | and Simon's http://www.w3.org/Bugs/Public/show_bug.cgi?id=9887#c28 |
| 12:52 | <MikeSmith> | Simon: "Another option is to make <annotation-xml><div> special." |
| 12:54 | <david_carlisle> | hang on 'll see if I can find the thread... |
| 12:57 | <david_carlisle> | from about here |
| 12:57 | <david_carlisle> | http://krijnhoetmer.nl/irc-logs/whatwg/20100913#l-423 |
| 12:57 | <david_carlisle> | to about here |
| 12:57 | <david_carlisle> | hsivonen> well, I suppose I should go ahead and implement it soonish |
| 13:19 | <MikeSmith> | david_carlisle: what kind of implementation you mean/ |
| 13:20 | <MikeSmith> | MathML spec change? |
| 13:25 | <david_carlisle> | MikeSmith: sorry, that was quoting Henri saying he should try implenting the html5 parser change, |
| 13:25 | <MikeSmith> | ah |
| 13:27 | <david_carlisle> | MikeSmith: mathML spec hopefully won't change as we're just waiting for the OK from w3C to publish, AC votes are all in, etc, so basically we're done |
| 13:27 | <hsivonen> | btw, was there a problem with Simon's <div> idea? |
| 13:33 | <david_carlisle> | hsivonen: Hi, you're there, probably not, both are just masking the real problem that the really isn't any reason for the parser to abort the math element unless it's doing fixup for mis-nested tagging or something. The attribute version was suggested earlier and now has the advantage the the Math WG signed off on it, checked all the examples in the mathml spec would actually work, etc.... |
| 13:33 | <david_carlisle> | ...without looking too deeply I see |
| 13:33 | <david_carlisle> | http://www.w3.org/TR/MathML3/chapter6.html#interf.graphics |
| 13:33 | <david_carlisle> | which would fail? |
| 13:38 | <hsivonen> | david_carlisle: yeah, that would fail |
| 13:39 | <hsivonen> | david_carlisle: though the problem goes away as SVG support becomes ubiquitous and people no longer care about bitmap fallback |
| 13:41 | <david_carlisle> | hsivonen; yes but as I said in the earlier irc thread, fallback is just one possible use of an annotation. |
| 13:41 | <david_carlisle> | especially once math in text/html becomes possible, far more interesting annotations are structured proof hints etc |
| 13:47 | <MikeSmith> | david_carlisle: not to (re)start an argument here, but as far as "masking the real problem", the real problem is arguably that the annotation-xml design does not lead to ease of implementation |
| 13:47 | <david_carlisle> | annotation-xml really should not be complicated, the only thing that the parser has to do with it is get to the end of it and stuff the data somewhere accessible, honestly the fact that it is proving difficult should be seen as a problem with the html5 parser's intricate rules for foreign content, not a problem with annotation-xml. The currently proposed solution is to make the parser have... |
| 13:47 | <david_carlisle> | ...even more intricate rules to hide some of the weirdness generated by the original rules, that wouldn't have been my choice, but it isn't my call, |
| 13:47 | <david_carlisle> | I can live with it |
| 13:48 | <hsivonen> | david_carlisle: the HTML5 parser's fogeign content support is really designed for SVG and form MathML that doesn't use HTML annotations |
| 13:49 | <hsivonen> | david_carlisle: at the time of designing it, annotation-xml was seen as a way to include semantic mathml and as a way to hack in SVG as seen on Jacques Distler's blog |
| 13:49 | <david_carlisle> | its foreign object thing isn't so different is it? |
| 13:50 | <hsivonen> | david_carlisle: it's fundamentally different |
| 13:50 | <hsivonen> | david_carlisle: because it doesn't need to support non-<svg> SVG children |
| 13:50 | <hsivonen> | david_carlisle: but annotation-xml needs to support non-<math> MathML children |
| 13:50 | <hsivonen> | that's the problem |
| 13:51 | <hsivonen> | Jacques Distler's blog is the showcase of how MathML actually gets used, and AFAIK, there's no XHTML inside annotation-xml on his blog |
| 13:51 | <hsivonen> | so... |
| 13:51 | <david_carlisle> | That's a very artificial problem though, caused by the strangeness elsewhere in the html5 parser rules. |
| 13:51 | <hsivonen> | david_carlisle: what strangeness? |
| 13:52 | <hsivonen> | the concept is not strange at all: 1) the parser starts in an HTML scope. 2) <svg> and <math> switch to SVG or MathML scope 3) certain elements open a nested HTML scope |
| 13:53 | <hsivonen> | the problem is that annotation-xml doesn't know whether it wants to be a nested HTML scope or not |
| 13:54 | <david_carlisle> | the fact that /> doesn't do what one might expect (you may have needed to special case script and other limited things for security) but if /> worked as an empty element on general elements (and the html parser didn't re-arrange html elements out of same) but... |
| 13:54 | <hsivonen> | david_carlisle: that would still leave the namespace problem |
| 13:54 | <david_carlisle> | ... I'm really not asking that the entire html5 parser design be ripped up now, so we are where we are |
| 13:54 | <david_carlisle> | namespaces: who cares;-) |
| 13:54 | <hsivonen> | david_carlisle: if we are doing hypotheticals, we wouldn't have this problem it Namespaces didn't exist |
| 13:56 | <david_carlisle> | Well personally I;d have been happy enough if the rule was that the html parser didn't do namespaces, and so if you wanted to namespace you dom fragments you'd have tto post process them in script. At least that way you would have the correct tree structure to work with |
| 13:57 | <hsivonen> | david_carlisle: that would have required drastic changes to e.g. Gecko's MathML implementation |
| 13:57 | <hsivonen> | david_carlisle: chameleon namespaces are bad |
| 13:57 | <david_carlisle> | Browsers come and go, documents last for thousands of years |
| 13:58 | <hsivonen> | david_carlisle: right, so if you want to support existing XML MathML content, you have to support the non-HTML namespace |
| 13:58 | <hsivonen> | if you also want to support mathml local names in the HTML namespace, you've got chameleon namespaces, which are bad |
| 13:59 | Philip` | notes that IE6 seems likely to last for thousands of years |
| 13:59 | <david_carlisle> | But no, you could have still had the rule that <math> flipped in the mathml namespace, and in practice the actual rules would be more complicated that can be specced in a couple of lines of irc grumbe, but... I just wish the world was different |
| 14:00 | <david_carlisle> | and /> should be empty elements even now, so there. |
| 14:11 | <david_carlisle> | hsivonen: isn't it the case that if svg's foreign object element has any non-svg non xhtml xml then because of /> the tree will be incorrectly parsed as it will be parsed as html scope? |
| 14:12 | <hsivonen> | david_carlisle: the only other possibility is <math>, and then you are in a math scope, so all is good |
| 14:12 | <hsivonen> | and /> on <svg/> and <math/> themselves is honored |
| 14:15 | <david_carlisle> | That isn't what the svg spec says, it says you can put any well formed xml in there. What you mean is the only other possibility that works is math, what you are saying is that if there is other cml guarded by a switch something like |
| 14:15 | <david_carlisle> | http://www.w3.org/TR/SVG/extend.html#ForeignObjectElement |
| 14:15 | <david_carlisle> | then the html parser will do something strange most likely, but such other html just isn't supported. |
| 14:15 | <david_carlisle> | this is not much different than annotation-xml, perhaps they have just complained less |
| 14:16 | <david_carlisle> | sorry, "such other xml" |
| 14:20 | <david_carlisle> | the problem is less for foreignobject as the expectation there is that the extended markup is rendered so in a browser context limiting to those vocabularies that can be rendered is perhaps reasonable, but that could be some home-brew xml css styled, for annotation-xml as the expectation isn't to render it it is less acceptable to restrict to the browsers native rendering vocabularies |
| 14:57 | <MikeSmith> | for anybody who might care: about my ramblings earlier on trying to find ways to get decent completion when editing javascript, I installed the "Eclipse IDE for JavaScript Web Developers" flavor of eclipse |
| 14:59 | <MikeSmith> | and I find it has "libraries" that enable completion for standard DOM interfaces |
| 14:59 | <MikeSmith> | in addition to ones for built-in Javascript objects |
| 15:00 | <MikeSmith> | and it works as expected in vim through eclim |
| 15:00 | <MikeSmith> | do everything is fine and dandy |
| 15:45 | <jgraham> | MikeSmith: You tried emacs + js2-mode for javascript? It rocks |
| 15:46 | <MikeSmith> | not tried it yet |
| 15:46 | <MikeSmith> | oh |
| 15:47 | <MikeSmith> | Steve Yegge |
| 15:47 | <MikeSmith> | cools |
| 15:47 | <MikeSmith> | *cool |
| 15:48 | <jgraham> | I'm not sure how much autocomplete it has; I mainly use it for the awesome highlighting, indentation and useful warning messages |
| 15:50 | <jgraham> | (in general my emacs-fu with autocomplete is limited to meta-/) |
| 21:23 | gsnedders | should try and work out how readonly parent/top/self actually are… |
| 22:13 | <annevk> | oh god, Paul Duffin just said lack of versioning is the problem |
| 22:13 | <annevk> | haha |
| 22:13 | <annevk> | guess I'll leave that thread to someone else |
| 22:14 | <annevk> | can't really explain everything all over again whenever someone does not do any research |
| 22:14 | gsnedders | wonders whether it's worthwhile pointing to XHTML 1.0 as an example of using versioning to introduce backwards-incompatible changes |
| 22:39 | <hsivonen> | you should write a 386 post about version once and then drop the URL everywhere |
| 22:39 | <hsivonen> | *versioning |
| 22:41 | <annevk> | I guess, but it's hard |
| 22:41 | <annevk> | but it should be done, yes, at some point |
| 22:41 | <annevk> | because clearly this does not scale |
| 22:42 | <annevk> | but summarizing x years of experience on a post-it is tricky :) |
| 22:44 | <gsnedders> | annevk: Also: don't lead me into reading more Murakami. I already have a full-time university course and a part-time job. Your blog is dangerous. :P |
| 22:45 | <annevk> | heh, you can block the URL :) |
| 22:46 | <jgraham> | After Dark is rather wonderful though |
| 22:47 | <jgraham> | Who is Paul Duffin and where is he bein misinformed about versioning? |
| 22:50 | <annevk> | www-style |
| 22:50 | <annevk> | see the namespace thread |
| 23:12 | <MikeSmith> | \(^\t<h[1-6].\+\)\@<!\([1-9][0-9]\?\.\)\+[1-9][0-9]\? |
| 23:13 | <aho> | that's a very good question |
| 23:13 | <aho> | :> |
| 23:36 | <MikeSmith> | aho: I didn't know til today that vim can do negative lookahead |
| 23:37 | <MikeSmith> | or whatever it's called |
| 23:37 | <gsnedders> | negative lookahead assertions |
| 23:37 | <MikeSmith> | @<! |
| 23:37 | <gsnedders> | (which, as it happens, aren't JIT'd in any browser) |
| 23:37 | <MikeSmith> | oh really? |
| 23:37 | <MikeSmith> | in regular expressions? |
| 23:38 | <gsnedders> | MikeSmith: AFAIK yes |
| 23:38 | <MikeSmith> | so yet another great reason to avoid using regular expressions unless you really can't avoid it |
| 23:38 | <gsnedders> | Most regexp are JIT'd |
| 23:39 | <MikeSmith> | ok |
| 23:40 | <gsnedders> | backreferences, assertions prevent stuff being JIT'd. I think there are other things too, but I can't remember off the top of my head. |
| 23:55 | MikeSmith | is listening to "Punta Banda" by Bostich & Fussible on WFMU on Transpacific Sound Paradise from WFMU - Freeform Radio the Way it Oughta Be |
