| 00:00 | <annevk> | ojan: it's a valid concern; I've been trying to think of something simpler, but haven't really been able to come up with anything |
| 00:00 | <annevk> | ojan: it's also kind of neat the way the scope chain thing works, but I guess it's a misfeature for extensibility |
| 00:00 | <zewt> | for example, "select * from users where id > 100" optimizes completely differently on (random database i just tested on) than "id > 1000000" (the former uses a seq scan, the latter an index scan) |
| 00:01 | <AryehGregor> | (Once someone asked in #mysql about some query with GROUP BY for days of the week, which was running really slowly on his data set. I advised him to rewrite it as a UNION of seven copy-pasted queries with different WHERE clauses, and he reported the performance was drastically better.) |
| 00:01 | <ojan> | annevk: i think it's also a misfeature for developers... |
| 00:01 | <annevk> | ojan: okay |
| 00:01 | <zewt> | (since postgres knows that there are tons of users > 100, but very few > 1000000) |
| 00:01 | <annevk> | ojan: I guess you don't like 'with' either? |
| 00:01 | <ojan> | annevk: i think the vast majority of developers are surprised by this when they learn about it and tend to learn about it by hitting a weird unexpected bug trying to access a global |
| 00:01 | <AryehGregor> | zewt, yeah, obviously. It's going to have to do cardinality estimates, and it can't do that without the actual numbers. WHERE foo=6 might be a totally different execution plan from WHERE foo=7, if 6 and 7 have very different frequencies. |
| 00:02 | <ojan> | annevk: i'm more OK with "with" since you do it more explicitly |
| 00:02 | <ojan> | annevk: i don't love it, but it's different in my view |
| 00:02 | <annevk> | fair enough |
| 00:02 | <AryehGregor> | The MySQL optimizer will sometimes figure out it doesn't have to run the query at all -- "impossible where condition noticed after reading const tables" or such. |
| 00:02 | <AryehGregor> | So it knows it has to return zero rows even before you actually run the query. |
| 00:02 | <ojan> | annevk: doesn't really matter either way...we're stuck with it |
| 00:03 | <AryehGregor> | But it depends totally on the values of the parameters. |
| 00:03 | <zewt> | i don't think postgres will do that, but it'll at least reduce it to an index scan, which is probably near-constant-time |
| 00:03 | <zewt> | (when the condition never happens) |
| 00:03 | <annevk> | we could have scope-interface HTMLButtonElement { }; interface HTMLButtonElement {}; maybe |
| 00:03 | <annevk> | ojan: ^^ |
| 00:03 | <ojan> | annevk: yeah, that seems nicer |
| 00:03 | <AryehGregor> | Surely [Scope] interface HTMLButtonElement? |
| 00:04 | <AryehGregor> | Not a whole new definition type? |
| 00:04 | <annevk> | I'm not sure why we have all this weird [] syntax instead of just native support |
| 00:04 | <annevk> | but sure |
| 00:04 | <AryehGregor> | "native support"? |
| 00:04 | <ojan> | annevk: i don't feel strongly about any of this as long as we can add new APIs without compat concerns |
| 00:05 | <zewt> | oh, postgres sometimes does it |
| 00:05 | <AryehGregor> | I'm sure it does. |
| 00:05 | <annevk> | AryehGregor: e.g. sometimes you use readonly attribute and sometimes you use [Replaceable] attribute |
| 00:05 | <AryehGregor> | pgsql is generally much smarter than MySQL. |
| 00:05 | <zewt> | if i say "id = 1 and id = 2", it adds "One-Time Filter: false", which i presume means the same thing |
| 00:05 | <annevk> | pretty weird |
| 00:05 | <annevk> | I'd rather we just use keywords all over |
| 00:05 | <zewt> | it just doesn't figure out "id > 100 and id < 0" is always false (for integers) |
| 00:05 | <AryehGregor> | annevk, well, there are like half a dozen that are magic, yeah. readonly, getter/setter/legacycaller/etc., static. |
| 00:05 | <annevk> | and I actually think we should drop "attribute" |
| 00:05 | <annevk> | we don't say "method" either |
| 00:05 | <AryehGregor> | I don't know why we have two different syntaxes. |
| 00:06 | <AryehGregor> | You'd just have to make sure it's not ambiguous. |
| 00:06 | <AryehGregor> | zewt, oh, really? I'm surprised, that seems easy. |
| 00:06 | <annevk> | () makes it that way |
| 00:06 | <annevk> | () -> method |
| 00:06 | <annevk> | no () -> attribute |
| 00:06 | <ojan> | annevk: makes sense to me |
| 00:06 | <AryehGregor> | MySQL does. |
| 00:07 | <AryehGregor> | zewt, MySQL does figure that out. |
| 00:07 | <AryehGregor> | Impossible WHERE noticed after reading const tables |
| 00:07 | <annevk> | but there's not a lot of people who want to simplify Web IDL these days |
| 00:07 | <annevk> | some kind of stabilizing frenzy |
| 00:07 | <Hixie> | did you guys ever get mikesmith's diff -F idea to work, btw? |
| 00:07 | annevk | rather has continuous improvement |
| 00:08 | <AryehGregor> | Changing WebIDL is annoying because then all the specs have to change. |
| 00:08 | <annevk> | Hixie: yes, but it is not there yet because it removed a feature svn diff had |
| 00:09 | <Hixie> | what command line did you use to get it to work, and what did it remove? |
| 00:09 | <Hixie> | (i'm doing it for something else) |
| 00:09 | <annevk> | Hixie: specifically because you use a custom diff the diff result you get no longer gives you the SVN revision numbers that where actually used (they can be different from input under some conditions) |
| 00:10 | <Hixie> | svn passes you the labels, so that's not a big deal |
| 00:10 | <Hixie> | but in this case it doesn't matter for me |
| 00:10 | <annevk> | oh, if svn does that maybe we can make it work somehow |
| 00:10 | <annevk> | anyway |
| 00:11 | <annevk> | I created diffwrap.sh |
| 00:11 | <annevk> | contains |
| 00:11 | <annevk> | #!/bin/sh |
| 00:11 | <annevk> | diff -u -F '<h[1-6]' ${6} ${7} |
| 00:11 | <annevk> | and then |
| 00:11 | <annevk> | the svn diff command becomes |
| 00:11 | <annevk> | svn diff --diff-cmd ./diffwrap.sh -r |
| 00:11 | <annevk> | and after -r your revision numbers |
| 00:12 | <Hixie> | add "-L {$} -L {5}" to your diff arguments, that should get you the labels right iirc |
| 00:12 | <Hixie> | er |
| 00:12 | <Hixie> | {$4} and {$5} respectively |
| 00:12 | <Hixie> | er |
| 00:12 | <Hixie> | {$3} |
| 00:12 | <Hixie> | not 4 |
| 00:12 | <Hixie> | probably |
| 00:12 | <Hixie> | try echoing all the arguments to find out which you want :-P |
| 00:12 | <Hixie> | anyway |
| 00:12 | <Hixie> | why isn't this working for me then |
| 00:13 | <annevk> | I don't want to know what is passed to svn diff btw |
| 00:13 | <annevk> | I want to know what svn diff ended up getting |
| 00:13 | <annevk> | e.g. if I just pass in revFrom |
| 00:13 | <Hixie> | yes, i understand |
| 00:13 | <annevk> | k |
| 00:13 | <AryehGregor> | Wait, wait, wait, wait. Are ES people using "globalization" as a synonym for "localization"? That's got to be the most confusing thing ever. |
| 00:13 | <annevk> | did you 755 on diffwrap.sh? |
| 00:13 | <Hixie> | so i'm doing svn diff source --diff-cmd 'diff' -x "-u6 -F '<h[1-6]'" |
| 00:13 | <Hixie> | instead of using a wrapper |
| 00:14 | <annevk> | you need the wrapper |
| 00:14 | <Hixie> | but the -F thing isn't working right |
| 00:14 | <zewt> | AryehGregor: i don't know, but those words mean utterly different things |
| 00:14 | <Hixie> | why? |
| 00:14 | <AryehGregor> | zewt, do they? I've never heard "globalization" used with respect to software. |
| 00:14 | <annevk> | Hixie: Mike told me to read http://svnbook.red-bean.com/en/1.2/svn.advanced.externaldifftools.html |
| 00:14 | <AryehGregor> | Only "localization" and "internationalization". |
| 00:14 | <annevk> | Hixie: I didn't and just followed his instructions |
| 00:14 | <annevk> | Hixie: but I assume it's explained there |
| 00:14 | <zewt> | it has no particular meaning in the field, but as a word it has no relation to "localization" that i know |
| 00:15 | <Hixie> | annevk: looks like that's out of date. svn has a -x argument now to pass arguments to diff. |
| 00:16 | <annevk> | oh |
| 00:16 | <annevk> | dunno then |
| 00:16 | annevk | -> bed |
| 00:16 | <annevk> | nn |
| 00:17 | <Hixie> | nn |
| 00:18 | <Hixie> | wtf |
| 00:19 | <Hixie> | svn diff source --diff-cmd 'diff' -x "-u6 -F=<" works fine |
| 00:19 | <Hixie> | svn diff source --diff-cmd 'diff' -x "-u6 -F=<h" shows no function |
| 00:19 | <Hixie> | svn diff source --diff-cmd 'diff' -x "-u6 -F<h" shows no function |
| 00:21 | <AryehGregor> | Maybe <h is interpreted as "use 'h' as the input file"? Try more quoting. |
| 00:21 | AryehGregor | doesn't know if this is being passed to a shell somewhere |
| 00:22 | <Hixie> | svn diff source --diff-cmd 'diff' -x "-u6 --show-function-line=<h[1-6]" works |
| 00:25 | <AryehGregor> | Curious. |
| 00:25 | <AryehGregor> | What about: "-u6 --show-function-line='<h'" |
| 00:27 | <Hixie> | dunno, once it started working i moved on :-) |
| 00:28 | <Hixie> | ok the topics i have so far are: ['Canvas','HTML','HTML Syntax and Parsing','Microdata','Offline Web Applications','Server-Sent Events','Video Text Tracks','Video and Audio','Web Storage','Web Workers','WebSocket API'] |
| 00:28 | <Hixie> | any other topics i should add? |
| 00:29 | <AryehGregor> | Do you have some particular reason for picking that granularity? Implementers normally work on extremely specific sets of things, no? |
| 00:30 | <AryehGregor> | To be fair, I can think of some people who are interested specifically in some of those topics. |
| 00:30 | AryehGregor | doesn't need this anyway, just looks at all the revisions |
| 00:30 | <Hixie> | well the video and track ones were requested |
| 00:30 | <Hixie> | and the others map to what i already had annotated for other reasons :-) |
| 00:31 | <AryehGregor> | Good strategy. |
| 00:54 | <Hixie> | ok you can now subscribe to specific topics |
| 01:15 | <Hixie> | i sent a mail to the list describing how |
| 02:02 | <jwalden> | zewt: re animated tab icons, I wonder if that's perhaps a use case for an attribute to put on <img> (well, its <xul:image> brother, but really <img>) to disable image animation; not a good use case, but perhaps there are use cases which would better fit on the actual web |
| 02:07 | <zewt> | jwalden: https://bugzilla.mozilla.org/show_bug.cgi?id=653035 |
| 02:08 | <zewt> | (which was met with deafening crickets) |
| 02:12 | <jwalden> | zewt: I think the place to propose that would be www-style, more likely |
| 02:17 | <zewt> | (no, because all of the use cases are UA-specific) |
| 03:26 | <MikeSmith> | zcorpan: needsinfo I created because under the current regime I don't officially have authority to move other editors' bugs to resolved |
| 03:26 | <MikeSmith> | so I didn't want to move others' bugs to resolved and have somebody call process foul |
| 03:36 | <Hixie> | MikeSmith: want me to go through them and mark then NEEDSINFO? |
| 03:37 | <MikeSmith> | yeah, sure |
| 03:37 | <MikeSmith> | if you have time |
| 03:37 | <MikeSmith> | there's about 40 of them at this point, I think |
| 03:37 | <MikeSmith> | so I seen you started marking up the spec with topic annotations |
| 03:38 | <MikeSmith> | I think that'll be useful |
| 03:38 | <Hixie> | yeah see whatwg mail |
| 03:38 | <MikeSmith> | ah, ok |
| 03:38 | MikeSmith | is at 295 unread in his inbox right now |
| 03:39 | <MikeSmith> | a few hours ago annevk was trying out that diff thing I suggested |
| 03:39 | <MikeSmith> | for web-apps-tracker |
| 03:39 | <MikeSmith> | and got it working |
| 03:39 | <Hixie> | yeah i used that for this too |
| 03:39 | <MikeSmith> | ah cool |
| 03:39 | <MikeSmith> | annevk got it working but ran into a problem caused by a side effect of it |
| 03:40 | <MikeSmith> | so hopefully he can figure out some way to get around that |
| 03:40 | <Hixie> | yeah we spoke about it |
| 03:40 | <MikeSmith> | ah, OK |
| 03:42 | <Hixie> | http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-November/033946.html |
| 03:43 | <Hixie> | annevk: if you want me to include the affected topics in my html5.org ping, let me know |
| 03:43 | <Hixie> | i should probably find a way to put them in the checkin comment, too |
| 03:43 | MikeSmith | finishes reading scrollback about the diff stuff |
| 03:44 | <MikeSmith> | didn't know about -x |
| 03:44 | <MikeSmith> | AryehGregor: yeah, that use of "globalization" is misguided |
| 03:45 | <MikeSmith> | that is what TC39 is calling their locales API |
| 03:45 | <MikeSmith> | which I guess is probably the context for why you mentioned it |
| 03:46 | <MikeSmith> | it's kind of like when people started calling themselves "serial entrepreneurs" |
| 03:46 | <MikeSmith> | and now serial whatevers |
| 03:47 | <MikeSmith> | when not just a few years ago the only context for using "serial" was in serial murderer or serial some-other-criminal-type |
| 04:03 | <zewt> | MikeSmith: i tend to assume "attention grab" when people don't use the same word for something that the rest of living english-speaking humanity uses |
| 04:04 | <MikeSmith> | hmm |
| 04:04 | <MikeSmith> | yeah |
| 04:04 | <MikeSmith> | there is that too |
| 04:04 | <MikeSmith> | but I don't think they are consciously doing it for that reason |
| 04:04 | <MikeSmith> | they are just naively saying, Hey, there's a word that sounds good |
| 04:05 | <zewt> | if people are implementing a localization framework who don't even know the word "localization", i'd be very worried |
| 04:05 | <MikeSmith> | without taking time to consider how the word it normally used in practice and what negative connotations it might have |
| 04:05 | <MikeSmith> | no, they know the word localization |
| 04:05 | <MikeSmith> | I think it's that what they are producing is not strictly about localization |
| 04:06 | <MikeSmith> | it's more about what normally called "locales" support |
| 04:06 | <MikeSmith> | I guess |
| 04:06 | <MikeSmith> | and it seems like until recently most everybody had been calling the JS thing "The Locales API" |
| 04:24 | <zewt> | locales is a standard part of localization |
| 04:24 | <zewt> | not to state too much the obvious :) |
| 05:46 | <heycam> | annevk, yeah I would say hold off on cosmetic changes like dropping "attribute" so as to avoid churning lots of specs for little gain. |
| 05:47 | <heycam> | AryehGregor, thanks for filing this bunch of bugs recently, btw. probably can't get to them until next week or later due to https://twitter.com/#!/heycam/status/138489725102211073 |
| 06:05 | <Hixie> | heycam++ on not changing webidl in backwards-incompatible ways |
| 06:05 | <Hixie> | i have real problems to work on and don't need make-work |
| 07:26 | <matjas> | zcorpan told me this the other day: “a-zA-Z1-9 and = after an entity without semicolon inside an attribute makes the entity get unconsumed and emitted as text” |
| 07:26 | <matjas> | where can i find that in the spec? |
| 07:29 | <MikeSmith> | matjas: you looked through http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#consume-a-character-reference already? |
| 07:30 | <MikeSmith> | the middle of the "Anything else" part |
| 07:31 | <matjas> | “If the character reference is being consumed as part of an attribute, and the last character matched is not a U+003B SEMICOLON character (;), and the next character is either a U+003D EQUALS SIGN character (=) or in the range U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9), U+0041 LATIN CAPITAL LETTER A to U+005A LATIN CAPITAL LETTER Z, or U+0061 LATIN SMALL LETTER A to U+007A LATIN SMALL LETTER |
| 07:31 | <matjas> | Z, then, for historical reasons, all the characters that were matched after the U+0026 AMPERSAND character (&) must be unconsumed, and nothing is returned.” |
| 07:31 | <matjas> | thanks, MikeSmith! |
| 07:31 | <MikeSmith> | cheers |
| 07:40 | <MikeSmith> | getting support for command element and API into webkit not looking so good |
| 07:40 | <MikeSmith> | https://lists.webkit.org/pipermail/webkit-dev/2011-November/018662.html |
| 07:41 | <MikeSmith> | at least not as currently spec'ed |
| 08:32 | <annevk> | there's a benefit to cleaning up IDL syntax too, but mkay |
| 08:33 | <annevk> | Hixie: CANVAS topic should include the element |
| 11:58 | <annevk> | smaug____: do you know the CORS code a little bit? |
| 11:58 | <annevk> | smaug____: I wonder how Mozilla implemented the Content-Type stuff |
| 13:21 | <annevk> | non-living standards strike again: http://www.ietf.org/mail-archive/web/websec/current/msg00846.html |
| 13:28 | <annevk> | okay event delegation is neat |
| 13:42 | <jgraham> | annevk: What in particular? |
| 13:47 | <annevk> | jgraham: ? |
| 13:47 | <bga_> | lol xslt! http://www.meta-language.net/sample.html |
| 13:48 | <AryehGregor> | heycam|away, no rush. It's not blocking my test writing; I've just got comments in the code pointing out what I should look at when the bugs are resolved. |
| 13:52 | <jgraham> | annevk: What do you mean by "event delegation". Or what part of it is neat? jQuery seems to have an API for attaching an event listener to a bunch of elements which was called "delegate" but afaict "delegation" is just having one event handler that deals with events coming from multiple nodes in the subtree of the node to which the handler is attached |
| 13:52 | <jgraham> | But possibly I don't understand the terminology here |
| 13:53 | <annevk> | see the API proposal on www-dom |
| 13:56 | <jgraham> | Oh, so the important point of the API proposal is matching the nodes on which you actually care about the event is baked in? |
| 14:17 | <annevk> | jgraham: euhm, the important point is that can catch events on objects inserted at any point in time, without having to explicitly insert listeners on those objects and without having to do a lot of filtering yourself |
| 14:25 | <jgraham> | annevk: Right, the "without having to do a lot of filtering yourself" is what I said about matching the right nodes being baked in |
| 14:33 | <annevk> | jgraham: ah, "baked in" threw me of |
| 15:42 | <AryehGregor> | Hixie, HTML has [OverrideBuiltins] partial interface Document { . . . }, which WebIDL doesn't currently allow. What's the intent? That the properties in that partial interface should override builtins, but the ones DOM4 defines for Document should not? Is this how implementations behave? |
| 16:01 | <hsivonen> | http://occupyhtml.org/ some people still like plug-ins |
| 16:03 | <bga_> | plugins is ok |
| 16:09 | <annevk> | MikeSmith: are you there? |
| 16:16 | <jarek__> | why {term} is defined in CSS2.1 as: |
| 16:16 | <jarek__> | [ NUMBER S* | PERCENTAGE S* | LENGTH S* | EMS S* | EXS S* | ANGLE S* | TIME S* | FREQ S* ] |
| 16:16 | <jarek__> | | STRING S* | IDENT S* | URI S* | hexcolor | function |
| 16:16 | <jarek__> | and not as: |
| 16:17 | <jarek__> | [ NUMBER S* | PERCENTAGE S* | LENGTH S* | EMS S* | EXS S* | ANGLE S* | TIME S* | FREQ S* | STRING S* | IDENT S* | URI S* | hexcolor | function] |
| 16:17 | <jarek__> | how is that different? |
| 16:17 | <jarek__> | I mean, why the first part is between [] brackets? |
| 16:22 | <divya> | because those are length units |
| 16:22 | <divya> | the rest are not |
| 16:24 | jarek__ | searches for definition of length unit |
| 16:27 | <divya> | http://dev.w3.org/csswg/css3-values/ |
| 16:27 | <jarek__> | in CSS2.1 padding is defined like this: |
| 16:27 | <jarek__> | ‘padding-top’ <length> | <percentage> |
| 16:28 | <jarek__> | so this grouping does not make sense :/ |
| 16:31 | <jarek__> | divya: btw, I have read your article on CSS vocabulary last week, great writing |
| 16:31 | <divya> | thanks jarek__! |
| 16:38 | <AryehGregor> | What's the difference in WebIDL between an implements statement, and using a partial interface? |
| 16:38 | <AryehGregor> | Just that you can save some typing if you want to add the same members to multiple interfaces? |
| 16:40 | <annevk> | is implements automatically NoInterfaceObject these days? |
| 16:41 | <AryehGregor> | Nope. |
| 16:41 | <AryehGregor> | Seems not. |
| 16:41 | <AryehGregor> | So that's a difference, yeah, but is it a desirable one? |
| 16:41 | <annevk> | not really |
| 16:42 | <annevk> | but implements can be useful if you define something like Workers and you want to expose a bunch of existing stuff |
| 16:43 | <annevk> | maybe they should be defined as "implementable interface Foo { }" |
| 16:43 | <annevk> | and then implements Foo gives the same semantics as partial |
| 16:44 | <annevk> | and Foo is just some name |
| 17:24 | <Hixie> | AryehGregor: the [OverrideBuiltins] should be on the DOM Document interface |
| 17:24 | <AryehGregor> | Hixie, so DOM4 should add [OverrideBuiltins] to the original interface, and you should remove it from the partial interface? |
| 17:24 | <AryehGregor> | Incidentally, preliminary HTML IDL tests: http://w3c-test.org/html/tests/submission/AryehGregor/interfaces.html |
| 17:27 | <Hixie> | AryehGregor: yeah |
| 17:28 | <AryehGregor> | Hixie, except the getter is defined in HTML, and OverrideBuiltins is only allowed to appear on an interface with a getter . . . |
| 17:28 | AryehGregor | isn't sure what the right way is to handle that |
| 17:29 | <dglazkov> | good morning, Whatwg! |
| 17:30 | <annevk> | AryehGregor: the right way to handle that would be to file a bug on heycam|away :) |
| 17:30 | AryehGregor | does so |
| 17:30 | <Hixie> | AryehGregor: partial interfaces are just an editorial detail, Document has a getter, whether it's defined on the block with 'partial' or not |
| 17:31 | <AryehGregor> | Editorially, that violates the fiction that DOM4 doesn't depend on HTML. |
| 17:32 | <Hixie> | why? |
| 17:32 | <Hixie> | oh because you need a getter |
| 17:32 | <Hixie> | hm |
| 17:33 | <AryehGregor> | Right. |
| 17:33 | <Hixie> | let it be heycam's roblem :-) |
| 17:34 | <annevk> | I looked at CORS again today to fix various issues |
| 17:34 | <annevk> | it really needs to be rewritten at some point :( |
| 17:35 | <annevk> | the positive thing here is that I've learned something the past couple of years |
| 17:35 | <annevk> | so there's that |
| 17:35 | <Hixie> | annevk: we should merge things like fetch, cors, maybe mimesniff, and various others at some point, but i highly recommend waiting a few years |
| 17:36 | <annevk> | yeah merging fetch was coming to mind but indeed |
| 17:36 | <annevk> | patience |
| 17:36 | <AryehGregor> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=14913 |
| 17:36 | <annevk> | AryehGregor++ |
| 17:39 | <AryehGregor> | Hixie, you have partial interface URL, but I don't see any reference to the full interface. |
| 17:39 | <Hixie> | yeah |
| 17:39 | <Hixie> | it doesn't exist yet |
| 17:40 | <AryehGregor> | That seems like a bit of a spec bug. |
| 17:40 | AryehGregor | files it so he can put the bug number in his tests as a comment |
| 17:41 | <annevk> | AryehGregor: http://dvcs.w3.org/hg/url/raw-file/tip/Overview.html |
| 17:41 | <annevk> | Hixie: ^^ |
| 17:41 | <annevk> | I asked MikeSmith to define URL parsing in that spec too |
| 17:42 | <AryehGregor> | LGTM, but we need a reference. Also, that IDL itself doesn't parse, but that's a separate issue. |
| 17:42 | <annevk> | it's abarth's Google Docs document converted to HTML by me |
| 17:42 | <annevk> | I might have made some mistakes |
| 17:42 | <AryehGregor> | interface [Replaceable] URL { |
| 17:42 | <AryehGregor> | That looks like it's not your mistake. :) |
| 17:43 | <AryehGregor> | (easily fixable, though) |
| 17:43 | <Hixie> | ah, cool, it exists now |
| 17:44 | <Hixie> | if you've filed a bug on me to update the html spec to point to that spec, thanks |
| 17:45 | <annevk> | I think that spec should eventually define Syntax/Parsing/Model/API |
| 17:45 | <AryehGregor> | I did, yeah. |
| 17:45 | <annevk> | for URLs |
| 17:45 | <AryehGregor> | I didn't give the spec URL, though. |
| 17:48 | <AryehGregor> | So if I'm testing that document.body is an HTMLBodyElement and document.head is an HTMLHeadElement, does it make sense to repeat all the tests for Node/Element/HTMLElement for both objects? |
| 17:48 | <AryehGregor> | (for instance) |
| 17:48 | <AryehGregor> | That means they'd all be run once for every interface that extends HTMLElement. |
| 17:48 | <AryehGregor> | Since I'm testing one of each element type. |
| 17:48 | <AryehGregor> | That seems like a lot of repetition. |
| 17:50 | <AryehGregor> | It seems like it makes more sense to only do the tests for the bottommost interface if I've tested the higher ones, at least for most of the element types. |
| 17:51 | <jgraham> | AryehGregor: Yeah, having huge numbers of tests that every HTML*Element is also an Element is just a waste of entropy |
| 17:52 | <AryehGregor> | And that it's an HTMLElement, and a Node. |
| 17:52 | <jgraham> | Indeed |
| 17:52 | <AryehGregor> | However, we do want to test that both Document and Text are Nodes, for instance. |
| 17:53 | <AryehGregor> | So I'll have to make it configurable somehow. |
| 17:54 | <AryehGregor> | Oh, so things like WindowBase64 are so that workers can implement them too, right? Interesting. |
| 17:54 | AryehGregor | will have to support that |
| 17:57 | <Hixie> | AryehGregor: re your mail about my crappy checkin comments, I've added a line to the checkin comment that lists the affected topics |
| 17:57 | <Hixie> | AryehGregor: it won't be perfect, certainly at first where a lot of the time it'll just say "Affected topics: HTML" |
| 17:57 | <Hixie> | AryehGregor: but it's a start |
| 17:57 | <AryehGregor> | Hixie, is it on the first line? If not, it won't be visible to anyone reading commit-watchers unless they look at the actual e-mail instead of just the subject. |
| 17:58 | <Hixie> | last line. If I make it the first line it'll be the only thing in the e-mail subject. |
| 17:58 | <AryehGregor> | For non-editorial changes, you can usually figure out the general subject from the first line of the commit summary. |
| 17:59 | <Hixie> | yeah |
| 18:00 | <AryehGregor> | As I said, it's not a big deal for me, I just glance at all of them. |
| 18:10 | <AryehGregor> | You know what's fun? When my tests call alert() with no arguments because that's supposed to throw, and WebKit doesn't throw but instead alerts the empty string, interrupting the tests. |
| 18:11 | <AryehGregor> | . . . then it hangs. I wonder what the last method it called was. |
| 18:13 | <AryehGregor> | window.open is fun too. |
| 18:14 | <AryehGregor> | ("pop-up blocked") |
| 18:23 | <zewt> | sort of need browser help to test some stuff properly... |
| 18:24 | <zewt> | eg. a (restricted) api to fire keyboard/mouse input as if it was real user input, replace modal things (alert, file pickers) with a callback, that sort of thing |
| 18:24 | <zewt> | i assume browsers do all that for their internal testing anyway |
| 18:30 | <AryehGregor> | HTML defines interface Function? o_O |
| 18:30 | <AryehGregor> | WTF? |
| 18:37 | <Hixie> | that predates WebIDL, i think. As far as I know, though, WebIDL still doesn't provide a type equivalent to that Function interface. |
| 18:40 | <dglazkov> | heycam: what's the new way to describe [Supplemental] in WebIDL? |
| 18:41 | <dglazkov> | Window implements SupplementalInterface? |
| 18:41 | <Hixie> | dglazkov: depends what you want exactly |
| 18:41 | <Hixie> | dglazkov: what are you trying to do? |
| 18:41 | <dglazkov> | I want to add "is" attribute to HTMLElement |
| 18:42 | <Hixie> | then you want a partial interface |
| 18:42 | <dglazkov> | oh cool |
| 18:42 | <dglazkov> | thanks |
| 18:51 | <AryehGregor> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=14916 |
| 19:19 | <dglazkov> | does w3 have a way to create localized bug trackers for specs? |
| 19:19 | <dglazkov> | it would be nice to have all component model todos organized as bugs. |
| 19:21 | <AryehGregor> | dglazkov, you can request a component in their Bugzilla. |
| 19:22 | <dglazkov> | a component model component |
| 19:22 | <dglazkov> | will adding bugs not spam mailing lists? |
| 19:23 | <dglazkov> | or is that only for certain components? |
| 19:23 | <AryehGregor> | I don't know the details. |
| 19:24 | <AryehGregor> | I think it's configurable per-component which lists get notified. |
| 19:27 | <AryehGregor> | . . . location.hasOwnProperty is undefined in Gecko? Really? |
| 19:31 | <smaug____> | AryehGregor: .location is so bizarre object that I don't expect much sanity in any implementation :) |
| 20:02 | <Hixie> | i rather like this new notification system |
| 21:14 | <sicking> | annevk: ping |
| 21:16 | <TabAtkins> | jarek: The grouping doesn't actually *do* anything. I don't know why it exists, but it's meaningless, since they're all using the | separator. |
| 21:17 | <jarek> | TabAtkins: I see, thanks letting me know |
| 21:19 | <TabAtkins> | jarek: No problem. All of the grammar combinators like that are associative, so grouping only has an effect when you're mixing combinators. |
| 21:29 | <annevk> | sicking: sort of here |
| 21:29 | <annevk> | sicking: gaming |
| 21:30 | <sicking> | annevk: so, ContentType is always added to the list of custom headers |
| 21:31 | <sicking> | annevk: but the server still doesn't need to send "Content-Type" in ac-response-headers if the value is "text/plain" "multipart/form-data" etc, right? |
| 21:31 | <sicking> | annevk: nor do we pre-flight if it's one of those values, right? |
| 21:31 | <annevk> | correct |
| 21:32 | <annevk> | the only requirement that changed is that the server needs to set Content-Type if it wants to handle it |
| 21:32 | <annevk> | in so far you can make server requirements |
| 21:33 | <annevk> | it's more advice I guess |
| 21:34 | <rniwa> | AryehGregor: ping |
| 21:38 | <annevk> | sicking: I don't think we want to change any UA requirements in the spec |
| 21:46 | <Hixie> | MikeSmith: i'm not going to touch the bugs that aren't assigned to me (counting contributor@ bugs as being mine) |
| 21:47 | <divya> | MikeSmith: Lemme know when you are here. got to ask you smthing |
| 21:49 | <sicking> | annevk: what do you mean by "if it wants to handle it"? |
| 21:51 | <annevk> | the server section is currently written assuming a list of header names |
| 21:52 | <annevk> | we could have the weird content-type special case there too |
| 21:52 | <annevk> | but currently it's not there |
| 21:53 | <annevk> | just read the diff or the section, that's easier prolly :) |
| 21:56 | <Ms2ger> | annevk, web-apps-tracker frontpage seems to be out of date |
| 22:00 | <sicking> | annevk: oh, you're saying that the spec now recommends that if the server want's the client to be able to set "Content-Type", then it should add it to the acah-header. No matter what values it wants to allow |
| 22:01 | <annevk> | sicking: basically |
| 22:01 | <annevk> | Ms2ger: did source change? |
| 22:01 | <Ms2ger> | Yes |
| 22:01 | <Ms2ger> | 6829 [Authors] [Conformance Checkers] [Gecko] [Internet Explorer] [Opera] [Webkit] [Tools] extend timezone format to also allow omitting the colon 2011-11-18 23:22 |
| 22:01 | <Ms2ger> | Is the newest I get |
| 22:02 | <sicking> | annevk: hmm.. we should probably add the exception there too |
| 22:02 | <sicking> | annevk: for people using the spec as documentation. Also so that people don't think they are safe if they don't add it |
| 22:03 | <annevk> | feel free to suggest a rewrite |
| 22:04 | <annevk> | few people did give feedback on that section |
| 22:04 | <annevk> | so I guess it's of use to some |
| 22:04 | <annevk> | but I don't really like it |
| 22:05 | <annevk> | Ms2ger: no idea what that can be |
| 22:05 | Ms2ger | hopes it'll go away on itself, then |
| 22:05 | <annevk> | doubt it |
| 22:08 | <annevk> | Ms2ger: http://svn.whatwg.org/webapps/ says the same as web-apps-tracker |
| 22:10 | <annevk> | sicking: also, I think the current text is fine |
| 22:10 | <annevk> | sicking: it explains the whole thing |
| 22:28 | <timeless> | Ms2ger / annevk : if i stuck up minutes for tpac webapps mon/tue + comp(wed), would any of you proof them? |
| 22:29 | <Ms2ger> | Perhaps |
| 22:29 | <Ms2ger> | But not tonight |
| 22:30 | <timeless> | so, should i send them direct to you? |
| 22:30 | timeless | doesn't have particularly handy web servers |
| 22:30 | <Ms2ger> | Sure |
| 22:33 | <timeless> | ok, i presume you won't be doing scribe edits, so i'm just sending the files as i expect to commit them |
| 22:34 | <timeless> | enjoy |
| 22:34 | timeless | crosses task off todo list for a bit |
| 22:34 | <timeless> | that just leaves expenses |
| 22:56 | <annevk> | http://www.reuters.com/article/2011/11/18/us-hp-lane-idUSTRE7AH2DK20111118 lol |
| 22:56 | <annevk> | that photo |
| 22:58 | <TabAtkins> | annevk: What, the HP guy using an Apple? |
| 23:01 | <timeless> | annevk: he should use a WebOS pad? |
| 23:02 | <timeless> | it looks like he's trying not to laugh |
| 23:02 | <timeless> | MikeSmith: ping ping |
| 23:43 | <MikeSmith> | timeless: here now |