| 02:01 | <Yuhong> | on script elements, Netscape once suggested this to hide from older browsers: <script><!-- ... <!-- --></script> |
| 02:01 | <Yuhong> | Of course, one problem is how -- inside a comment work in SGML. |
| 06:01 | <annevk> | the plot thickens |
| 06:11 | <karlcow> | then the slick top |
| 06:59 | <hsivonen> | AryehGregor: I think you WONTFIX rationale for bug 13057 sucks. |
| 06:59 | <hsivonen> | AryehGregor: saying they don't serve a common need as a reason to *keep* them??? |
| 07:00 | <hsivonen> | *your |
| 07:01 | <annevk> | that bug makes me long for TPAC |
| 07:01 | <hsivonen> | annevk: why? |
| 07:02 | <annevk> | in a sarcastic way |
| 07:03 | <annevk> | lots of bitching about HTML by people who have not participated in years |
| 07:07 | <othermaciej> | hsivonen: seems like a plausible reason not to replace them |
| 07:07 | <othermaciej> | otoh you could argue that the use case is still valid and the rarity of use just shows the bad design |
| 07:09 | <hsivonen> | othermaciej: it could be used as an argument for removing them and not providing a replacement |
| 07:09 | <annevk> | I don't understand why he can't have "block" and "inline" level delete and insertion in his editor |
| 07:09 | <hsivonen> | othermaciej: doesn't make sense as a reason to keep them |
| 07:09 | <annevk> | he seems to think UI and markup has to match 1:1 |
| 07:09 | <annevk> | which strikes me as odd for a WYSIWYG editor |
| 07:10 | <hsivonen> | I'm inclined to think HTML shouldn't try to track deletions |
| 07:14 | <annevk> | so what if you want to publish how a document was redacted? |
| 07:14 | <othermaciej> | hsivonen: what do you think is the correct resolution? drop entirely with no replacement? |
| 07:15 | <hsivonen> | othermaciej: I'd do the following: |
| 07:15 | <hsivonen> | 1) Define <ins> as inline-only and a synonym for <u> |
| 07:15 | <hsivonen> | 2) Define <del> as inline-only and a synonym of <s> |
| 07:16 | <hsivonen> | 3) Introduce an attribute that takes a list of edit entries that contain a time stamp and an optional user identifier |
| 07:16 | <hsivonen> | and indicates that the text was added by that user at that time |
| 07:17 | <hsivonen> | hmm. maybe it shouldn't take a list |
| 07:18 | <hsivonen> | anyway, dropping <ins> and <del> even as inline underline and strike markup would annoy a lot of people who use them manually on blogs as inline-only |
| 07:18 | <hsivonen> | and <del> across blocks sucks badly and is rarely used |
| 07:18 | <othermaciej> | removing their semantics would likely also annoy people |
| 07:19 | <hsivonen> | othermaciej: yeah, maybe |
| 07:19 | <hsivonen> | othermaciej: maybe the right thing is to make them inline-only and add attributes that you can put on any element |
| 07:19 | <othermaciej> | for the sort of people who bother to use them, anyway |
| 07:19 | <hsivonen> | othermaciej: anyway, <del> as block sucks big time |
| 07:19 | <othermaciej> | what is the difficulty with using them at block level? |
| 07:19 | <hsivonen> | and isn't really used by anyone, AFAICT |
| 07:20 | <hsivonen> | othermaciej: even without glazou's wysiwyg case, in the block case you'd want to style them based on the blockness of their content |
| 07:21 | <othermaciej> | I see, so they either need a special class or you need non-existing inside-out selectors |
| 07:21 | <hsivonen> | othermaciej: and if they are rarely used, it doesn't make sense to design and engineer a solution that'd make the block duality nature not suck |
| 07:22 | <othermaciej> | anyway if you disagree with Aryeh's rationale, you could reopen the bug and let Hixie sort it out |
| 07:22 | <hsivonen> | ok |
| 07:23 | <annevk> | how would you address the styling problem with attributes? |
| 07:24 | <annevk> | note also that the Selector problem is being solved |
| 07:24 | <othermaciej> | your selector could look at the element and the attribute |
| 07:24 | <hsivonen> | annevk: the problem of being able to select on the computed style of children? |
| 07:24 | <annevk> | I mean what would be a good way to style it |
| 07:24 | <annevk> | default-style wise |
| 07:25 | <hsivonen> | annevk: for stuff that's display:block; border-right or outline-right |
| 07:26 | <othermaciej> | you'd need different default rules for different elements depending on their default display type |
| 07:26 | <othermaciej> | but there'd be no way for that to update automatically if someone restyles an element to a different display type |
| 07:27 | <annevk> | Is the mean reason for transferring ArrayBuffer rather than cloning preserving memory? |
| 07:27 | <othermaciej> | I presume the reason is performance |
| 07:28 | <othermaciej> | you don't want to incur the cost of a copy each time you transfer cross-thread |
| 07:28 | <othermaciej> | and you don't want to use copy-on-write since that adds cost to every write, even when not using threads |
| 07:52 | <zewt> | othermaciej: copy-on-write makes reusing buffers efficiently hard, iirc |
| 07:53 | <zewt> | eg. video double-buffering |
| 07:57 | <annevk> | thanks othermaciej |
| 07:57 | <othermaciej> | zewt: yeah, and there are probably also good cases for creating a buffer on the main thread, passing it to a Worker, modifying it in-place in the Worker, and passing it back |
| 07:58 | zcorpan | put Hixie's bookmarklets as in-page links with userjs |
| 07:58 | <zcorpan> | hmm, i should have a link for FIXED as well, for the differences doc |
| 07:58 | <annevk> | oh sweet, you'll continue editing that? |
| 08:00 | <zcorpan> | yeah |
| 08:01 | <zcorpan> | except in july |
| 08:02 | <annevk> | cool |
| 08:03 | <annevk> | guess I owe bratell a beer |
| 08:03 | <annevk> | well, and you |
| 08:03 | <annevk> | :) |
| 08:06 | <zcorpan> | i was thinking of switching to anolis for it |
| 08:06 | <zcorpan> | to get xspec xrefs |
| 08:06 | <zcorpan> | how do i do that? |
| 08:08 | <annevk> | better to ask Ms2ger |
| 08:08 | <annevk> | I have not got that to work yet |
| 08:08 | <annevk> | I want to switch all my specs too |
| 08:09 | <annevk> | xsxr would be so great |
| 08:09 | <annevk> | Ms2ger has some instructions on https://bitbucket.org/ms2ger/dom-core (see bottom) but I did not get Anolis properly installed |
| 08:10 | <annevk> | I suspect jgraham can help you out if you are in the office |
| 08:10 | <zcorpan> | btw i love new Event('foo') |
| 08:10 | <zcorpan> | unsursprisingly :) |
| 08:10 | <zcorpan> | will be in an hour |
| 08:14 | <zcorpan> | can you transfer an object back and forth with postMessage? |
| 08:14 | <annevk> | for WHATWG Weekly I have spec/whatwg mailing list summary + dom core + new approach exceptions |
| 08:14 | <annevk> | zcorpan, since http://html5.org/tools/web-apps-tracker?from=6272&to=6273 you can do it for MessagePort |
| 08:15 | <annevk> | should I cover anything else in the weekly? |
| 08:15 | <annevk> | nothing much happened last week but maybe I misremember? |
| 08:16 | <zcorpan> | yah i'm reading that. it says once you transfer an object you can't transer it again. wondered if it means that you can't transfer it *back* (or further) on the receiving side |
| 08:16 | <zewt> | zcorpan: the transfer api could be used to transfer other complex objects, eventually |
| 08:16 | <annevk> | zcorpan, not being able to transfer it back would be weird |
| 08:16 | <zcorpan> | indeed |
| 08:17 | <annevk> | oh, maybe there is just data equality but not object equality? |
| 08:18 | <annevk> | so you would have a new object when you transfer it back that holds the same data in memory? |
| 08:18 | <zewt> | yeah, transferring back doesn't "repopulate" the old object or anything |
| 08:19 | <zcorpan> | makes sense |
| 08:19 | <annevk> | anyway, nothing? |
| 08:19 | <annevk> | should I mention the Editorial Assistents? |
| 08:19 | <zewt> | havn't looked over the new spec language yet but the general idea is that transfer is indistinguishable from structured clone, from the perspective of the receiver |
| 08:19 | <annevk> | I guess I should mention the timeline othermaciej posted |
| 08:20 | <zewt> | (eg. a postMessage might be received by more than one receiver, in which case it's forced to clone anyway) |
| 08:20 | <zewt> | (iirc--still need to look over the spec language, too) |
| 08:21 | <zcorpan> | annevk: assistants seems mentionworthy |
| 08:22 | <zcorpan> | zewt: it'd throw if you try to transfer it twice |
| 08:23 | <annevk> | EUR 0.16 per MiB, does not work abroad. http://www.kpn.com/prive/mobiel/aanbiedingen/nieuw-in-nederland/chromebook.htm Chromebook is going to be a huge success. |
| 08:25 | <zcorpan> | i think |
| 08:25 | <othermaciej> | but.. but… the cloud! |
| 08:26 | <annevk> | It is time someone takes over the provider market the way Apple took over the phone market |
| 08:29 | <annevk> | I watched that 1997 closing note by Steve Jobs gruber linked to and in it he makes quite a nice point. That there is so much opportunity not necessarily to innovate, but to improve existing services and products a whole lot. Applies perfectly here, except that the startup costs are like way high compared to software :/ |
| 08:30 | <annevk> | (He also goes on and on how Apple should license their software and allow people to compete with Apple on hardware.) |
| 08:31 | <othermaciej> | you can get a good deal on data plans for iDevices but not for Apple laptops yet |
| 08:32 | <othermaciej> | I thought Apple was already licensing their software and letting people compete on hardware in '97, and stopped as soon as SJ was in charge |
| 08:32 | <annevk> | The hardware had to be bought from Apple or some such? Not sure, he wanted to change the way that was done... |
| 08:33 | <annevk> | I am quite surprised there is no SIM-slot in my MacBook Air |
| 08:34 | <annevk> | But data plans for e.g. the iPhone suck when I go abroad |
| 08:34 | <annevk> | Within the Netherlands it is okay, but I travel a lot... |
| 08:35 | <othermaciej> | no, there used to be non-Apple vendors selling Apple hardware |
| 08:35 | <othermaciej> | er, not Apple hardware |
| 08:36 | <othermaciej> | but Mac-compatible hardware |
| 08:46 | <annevk> | I wish my German was better |
| 08:46 | <annevk> | http://muenz.wordpress.com/2011/02/25/der-lange-abschied-vom-„toten“-standard/ |
| 08:46 | <annevk> | Also, I posted http://blog.whatwg.org/weekly-event-constructors just now |
| 08:47 | <annevk> | Not that any of you ever give feedback, but I keep trying :) |
| 08:54 | <hsivonen> | annevk: based on Google translate, the German article doesn't seem positive |
| 08:54 | <annevk> | RDF, Semantic Web, Web 3.0, Linked Data |
| 08:54 | <annevk> | RDF/XML, N3, Turtle, RDFa |
| 08:54 | <annevk> | rebranding and resyntaxing forever and ever, to no avail |
| 08:55 | <annevk> | hsivonen, my Google translate was fairly positive |
| 08:55 | <annevk> | :) |
| 09:00 | <annevk> | Microdata follows RDFa in that list I guess |
| 09:01 | <zcorpan> | what's after microdata? |
| 09:02 | <zcorpan> | RDFa 2.0? |
| 09:02 | <annevk> | I'm sure we'll find out |
| 09:04 | <gsnedders> | othermaciej: It wasn't quite straight away, it was they stopped licensing Mac OS as of 8.0 for the clones. |
| 09:05 | <zcorpan> | maybe RDFb |
| 09:05 | <othermaciej> | it was a little before I joined Apple so I don't necessarily remember the details |
| 09:05 | <zcorpan> | "RDF bastards" |
| 09:06 | <hsivonen> | zcorpan: that's not nice |
| 09:10 | <annevk> | Maybe Web IDL should define "initialize" so specifications can "initialize" attributes whereas authors can set them (and maybe sometimes specifications need to set them too) |
| 09:12 | <hsivonen> | othermaciej: if one is to believe what a Finnish operator says, Apple requires the data plans for iPad to suck equally all over the world |
| 09:13 | <hsivonen> | othermaciej: their data plans were already uncapped and they say Apple wanted them to add a capped plan in order to sell iPad |
| 09:13 | <hsivonen> | and they said no |
| 09:15 | <othermaciej> | I have no idea about that stuff |
| 09:15 | <annevk> | iPhone sells here without cap, no idea about iPad |
| 09:15 | <othermaciej> | the Verizon data plan for my iPad 2 in the US is just fine |
| 09:16 | <hsivonen> | I find it slightly stressful that I have no idea how close I am to the data cap of my German prepaid |
| 09:16 | <hsivonen> | (capped at 200 MB per month) |
| 09:17 | <hsivonen> | it would really suck to reach the cap when I have an acute use case for Google Maps going on |
| 09:18 | <othermaciej> | that would suck |
| 09:23 | <othermaciej> | bug resolve rate on Hixie-edited drafts over the past 4 weeks: 18, 19, 50, 81 |
| 09:27 | <othermaciej> | incoming over the past week is 44, so I believe outgoing now exceeds incoming by a healthy margin |
| 10:12 | <nessy> | wow, that's quite some bug replying rate there! |
| 10:13 | <nessy> | a few that I saw were by co-editors, so maybe that is starting to have an effect |
| 10:40 | <karlcow> | annevk: typo in "Earlier Maciej also posted on editorial assistents that will help Ian out with dealing with the Last Call feedback." |
| 10:40 | <karlcow> | assitants in http://blog.whatwg.org/weekly-event-constructors |
| 10:41 | <karlcow> | also "WHATWG develops a parellel edition" |
| 10:42 | <karlcow> | * parallel |
| 11:49 | <karlcow> | https://bugzilla.mozilla.org/show_bug.cgi?id=553888 |
| 11:49 | <karlcow> | "Additional XHR request after Redirect response doesn't forward non standard headers" FIXED |
| 11:51 | <hsivonen> | roc: btw, regarding the IRC log: I wouldn't blame the IBM CIO for not having someone follow Mozilla blogs about EOL when the main Mozilla blog is silent about EOL. |
| 11:52 | <roc> | I was thinking dev.planning |
| 11:52 | <hsivonen> | roc: OK. I still don't get it why a major change like this didn't go on the official blog. :-( |
| 11:53 | <roc> | oh I agree |
| 11:54 | <roc> | blame is not a zero-sum game |
| 11:54 | <Ms2ger> | Still a fun game ;) |
| 11:55 | <gsnedders> | Is there anything in the DOM with a custom toSring impl? |
| 11:55 | <annevk> | hmm, krijn?! |
| 11:55 | <annevk> | krijn, offline again? |
| 11:55 | <Ms2ger> | gsnedders, Selection, a elements, location? |
| 11:56 | <krijn> | Yeah, no idea what's wrong |
| 11:56 | <krijn> | Sigh |
| 11:57 | <gsnedders> | Ms2ger: Okay, so nothing as evil as Date. (Which touches external state when getting.) |
| 11:58 | <Ms2ger> | Nothing is as evil as Date, period :) |
| 12:00 | <annevk> | yay, time to rewrite the event chapter once again |
| 12:02 | <bga_> | annevk btw. not {new Event(alot of args >_<)}. {var e = new Event(); e.target = foo; e.x = bar} imho |
| 12:02 | <bga_> | argless creation, post fill |
| 12:03 | <krijn> | annevk: connection here sucks atm :/ |
| 12:03 | <annevk> | bga_, you can just do |
| 12:04 | <annevk> | obj = {}; obj.target = foo; obj.x = bar; var e = new Event("yay", obj) |
| 12:04 | <annevk> | same thing |
| 12:05 | <bga_> | not so pure but ok |
| 12:06 | <Ms2ger> | Purity on the web? Ha. Ha. Ha. |
| 12:06 | <bga_> | annevk my approach is fully based on setters/getters |
| 12:07 | <bga_> | some "options" hash table arg is bad imho |
| 12:07 | <annevk> | zcorpan suggested the same thing |
| 12:07 | <annevk> | it requires making readonly attributes read-write |
| 12:07 | <bga_> | look like jq plugin :( |
| 12:07 | <annevk> | and only sometimes readonly |
| 12:08 | <annevk> | and dictionary-style APIs are introduced all over so it made sense to do that |
| 12:08 | <bga_> | annevk heh. not r/w and r/o |
| 12:09 | <bga_> | just mutable and immutable state |
| 12:09 | <bga_> | foo.onclick = function(e /* immutable */) |
| 12:09 | <bga_> | var e = Event() /* mutable */ |
| 12:10 | <annevk> | sure, I get that |
| 12:11 | <bga_> | i know that ES does not differ immutable and mutable state on spec level |
| 12:11 | <bga_> | but you can emualte |
| 12:12 | <bga_> | splitting interface to 2 intefaces, mutable inherits immutable |
| 12:16 | <Ms2ger> | And what's the gain for developers? |
| 12:17 | <bga_> | stricter safer oop |
| 12:19 | <karlcow> | annevk: there are typos in the blog post |
| 12:19 | <Ms2ger> | Excuse me? |
| 12:19 | <karlcow> | the weekly summary |
| 12:19 | <annevk> | karlcow, what are they? |
| 12:19 | <karlcow> | I put them on irc before you arrived but I can't access krijn right now |
| 12:19 | <karlcow> | let me try to find them |
| 12:20 | <krijn> | Yeah, I'm just not that open, sorry |
| 12:20 | <karlcow> | assistents |
| 12:20 | <karlcow> | parellel |
| 12:20 | <krijn> | Why don't you all use IRCCloud? |
| 12:20 | <karlcow> | krijn: I had a friend who was telling me it is a question of breathing |
| 12:21 | <krijn> | I have two invites left if you want |
| 12:21 | <zcorpan> | krijn: your server is closing the connection for me :( |
| 12:22 | <krijn> | I cannot even reach it from within the same building.. |
| 12:22 | <krijn> | Network cables are the crappiest invention ever |
| 12:23 | <zcorpan> | go plug in the cable again and throw out the cat |
| 12:24 | <hsivonen> | Someone from Opera might be interested to know that http://my.opera.com/haavard/ is 404 |
| 12:24 | <karlcow> | eat the cat! |
| 12:24 | <karlcow> | hsivonen: all my.opera.com |
| 12:24 | <Philip`> | There is no cat |
| 12:24 | <karlcow> | it seems |
| 12:26 | <karlcow> | and the cat's name is Erwin |
| 12:27 | <erlehmann> | so opera has jumped the shark now? |
| 12:32 | <annevk> | thanks karlcow |
| 12:32 | <annevk> | fixed |
| 12:37 | <gsnedders> | hsivonen: People poked. |
| 12:37 | <hsivonen> | gsnedders: thanks |
| 12:37 | hsivonen | is interested in seeing what haavard has to say about Microsoft & WebGL |
| 12:39 | <gsnedders> | hsivonen: Fairly unsurprisingly, already known issue |
| 12:49 | <hasather> | hsivonen: cached: http://webcache.googleusercontent.com/search?q=cache:Y_sXnE9vvJwJ:my.opera.com/haavard/blog/2011/06/22/microsoft |
| 12:50 | <hsivonen> | hasather: thanks |
| 12:52 | <hsivonen> | hasather: ok. nothing new there |
| 12:56 | <doublec> | irccloud has issues with freenode cloaks last I used it |
| 13:19 | <annevk> | Ms2ger, https://bitbucket.org/ms2ger/dom-core/changeset/d9098cf6a41b |
| 13:19 | <annevk> | Ms2ger, I fiddled with some external WebIDL defined terms, might create issues for you I suppose |
| 13:21 | <karlcow> | seo hell and titles in the eyes of google and bing http://searchengineland.com/writing-html-title-tags-humans-google-bing-59384 |
| 13:24 | <annevk> | Ms2ger, also, if you agree this new approach makes sense I'll update Progress Events too |
| 13:24 | <annevk> | again... |
| 13:37 | <Ms2ger> | How about waiting a day or two before updating PE? :) |
| 13:39 | <Ms2ger> | (and pushed) |
| 13:40 | <annevk> | ta |
| 13:40 | <annevk> | and I guess that's better |
| 13:46 | <scor> | can microdata work with HTML < 5 without having to change their doctypes? |
| 13:46 | <Ms2ger> | Define "can" and "work" |
| 13:47 | <zcorpan> | generally stuff works regardless of doctype |
| 13:47 | <scor> | I'm assuming the consequence is that it will not validate any longer right? |
| 13:48 | <scor> | (validate against their doctype validator) |
| 13:48 | <Ms2ger> | True |
| 13:48 | <zcorpan> | if you validate as html5 then it it'll validate (assuming you use a doctype that html5 allows, e.g. html4 strict) |
| 13:49 | <scor> | will it have any impact on the browser? don't browser have different behaviors depending on the doctype? |
| 13:49 | <scor> | zcorpan: interesting, so some old doctype will still validate in html5? |
| 13:49 | <zcorpan> | yeah |
| 13:49 | <scor> | but some others will not validate? |
| 13:49 | <scor> | is there a list? or a rule? |
| 13:49 | <zcorpan> | there's a list in the spec |
| 13:50 | <zcorpan> | see "writing html documents" |
| 13:50 | <Philip`> | http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#obsolete-permitted-doctype-string |
| 13:50 | <scor> | ok, got it |
| 13:51 | <zcorpan> | well, it only allows them in that you "should not" use them (instead of "must not") |
| 13:52 | <scor> | how about the browsers? |
| 13:52 | <scor> | will they behave differently based on the doctype? |
| 13:52 | <zcorpan> | not if you use a permitted one |
| 13:52 | <scor> | I'm not expecting them to do anything with the microdata attributes, just wondering if some would complain or break |
| 13:53 | <zcorpan> | the thing that can happen is that the browser would go in to quirks mode or almost standards mode if you use a doctype that triggers that |
| 13:53 | <zcorpan> | but microdata would work even then |
| 13:54 | <zcorpan> | (at least per spec, it's not implemented in browsers yet) |
| 13:54 | <scor> | ok, thanks zcorpan |
| 13:54 | <scor> | yes |
| 13:54 | <Philip`> | (Wouldn't work in IE quirks even if they do implement it) |
| 13:54 | <zcorpan> | (and ie has a different policy where new features are not enabled in old rendering modes) |
| 13:54 | <zcorpan> | ARIA is an exception |
| 13:55 | <scor> | including IE 7 and 8? |
| 13:55 | <zcorpan> | ie7 and 8 don't get new features |
| 13:55 | <annevk> | aah, you added that period |
| 13:56 | <annevk> | hg diff yielded nothing and I had only seen the previous checkin :) |
| 15:03 | <zcorpan> | onload={handleEvent:...} seems nice. should we make it work? |
| 15:05 | <Ms2ger> | smaug wanted that to work, iirc |
| 15:05 | <zcorpan> | fine with me! |
| 15:06 | <smaug____> | Ms2ger: I care more about addEventListener listeners |
| 15:06 | <smaug____> | but for consistency, onfoo should work the same way |
| 15:06 | <Ms2ger> | But who uses those?! :) |
| 15:07 | <smaug____> | atm everyone how wants to use { handleEvent: function() {} } uses addEventListener |
| 15:08 | <smaug____> | if we could then change setTimeout/Interval a bit, we could get rid of FunctionOnly |
| 15:08 | <Ms2ger> | Feel like bugging gal about it? (Or whoever is rewriting onfoo) |
| 15:09 | <smaug____> | bz is rewriting onfoo |
| 15:44 | <bga_> | http://techcrunch.com/2011/06/24/opera-founder-jon-s-von-tetzchner-resigns-over-differences-with-board/ |
| 16:38 | <zcorpan> | i'm disappointed nobody has implemented https://twitter.com/#!/zcorpan/status/85118589711040513 yet |
| 16:39 | <Ms2ger> | I'm disappointed by a lot of things, but that's not one of them |
| 16:40 | <zcorpan> | i'm disappointed that you're not disappointed |
| 16:41 | <Ms2ger> | I'm sorry to hear that |
| 16:41 | <Ms2ger> | Actually, not *that* sorry |
| 16:41 | <MikeSmith> | I am relatively appointed about it |
| 16:41 | Ms2ger | appoints MikeSmith |
| 16:41 | Philip` | points at MikeSmith |
| 16:42 | MikeSmith | disaccepts the appointment |
| 16:42 | charlvn | needs a pint after this discussion |
| 16:43 | Ms2ger | points at the pint |
| 16:43 | <zcorpan> | guys, if you'd put the energy on implementing it instead! |
| 16:43 | Philip` | punts around the point |
| 16:44 | charlvn | points at the punts |
| 16:44 | Ms2ger | implements blink |
| 16:45 | <charlvn> | ok now this is official - each of us needs to get a life |
| 16:45 | <Ms2ger> | I know right |
| 16:45 | Philip` | falls off the punt and contracts Weil's disease |
| 16:47 | <MikeSmith> | https://twitter.com/#!/html5grind is posting some good stuff |
| 16:47 | <MikeSmith> | Ian Langworth |
| 16:48 | <bga_> | html5grindcore |
| 17:16 | <AryehGregor> | hsivonen, the point of the bug was to remove <ins> and <del> and replace them with attributes. I mostly focused on we don't want to add new standardized attributes without strong need. I didn't address the question of removing them, but Hixie's approach has been to keep preexisting features even if they have very marginal utility (e.g., <kbd>) so as not to gratuitously make pages invalid. If you want to reopen and let Hixie handle it, by a |
| 17:16 | <AryehGregor> | ll means feel free. |
| 18:39 | <smaug____> | annevk: it is a bit hard to understand when you want to spec what the existing implementations do, and when you want to remove already implemented features |
| 18:40 | <smaug____> | perhaps you have some "implemented in less than 2 years ago, can be removed" |
| 18:41 | <smaug____> | (Note, I don't care too much about initProgressEvent, but would just like to understand the logic behind removing existing features from the specs) |
| 18:43 | <Ms2ger> | "If we can get away with it", I suppose |
| 18:45 | <smaug____> | that has changed a lot since last TPAC, when I said that we can remove features from the web and I was laughed at :) |
| 18:45 | <Ms2ger> | Meh, F2F :) |
| 19:50 | <AryehGregor> | Hmm, why did zcorpan ask me to spec queryCommandSupported() back in March? It seems like a very trivial method. |
| 20:31 | <The_8472> | <AryehGregor> It seems like a very trivial method. <- famous last words? |
| 20:31 | <AryehGregor> | The_8472, I mean based on my observation. |
| 20:31 | <AryehGregor> | It looks like what sane browsers do is return true if they support the command and false otherwise. |
| 20:31 | <AryehGregor> | At least WebKit does that. |
| 20:33 | <AryehGregor> | Gecko always just throws an exception, and Opera seems to return true for things it supports if there's something editable on the page and false otherwise. |
| 20:33 | <AryehGregor> | IE matches WebKit. |
| 20:33 | <AryehGregor> | So I really don't see the issue. |
| 20:34 | <The_8472> | well, is there any spec at all for it? |
| 20:34 | <AryehGregor> | Well, yes, I'm writing it. :) |
| 20:34 | <AryehGregor> | There's HTML5, but that's extremely sparse and often doesn't match reality on this subject. |
| 20:34 | <The_8472> | then he probably just wanted it to be put somewhere, even if the spec consists of a one-liner |
| 20:34 | <AryehGregor> | I dunno, I'll ask him when he comes back on if I remember. |
| 20:34 | <jamesr> | i think speccing every very trivial function in the web platform is a non-trivial but useful job |
| 20:35 | <AryehGregor> | Definitely, yeah. |
| 20:35 | <AryehGregor> | I just wondered why he specifically asked me to spec that. |
| 20:43 | <Hixie> | hmmm |
| 20:43 | <Hixie> | so if AudioTrackList returns a bunch of AudioTrack objects |
| 20:43 | <Hixie> | and a multi-audio-track <video> element has a changing number of audio tracks |
| 20:43 | <Hixie> | e.g. a stream gets a new audio track dynamically |
| 20:44 | <Hixie> | and then that track goes away |
| 20:44 | <rubys> | Hixie: if I remember correctly you had a website where your unprocessed feedback queue could be examined. Does this still exist? |
| 20:44 | <Hixie> | or even simpler, if the <video> gets a whole new media element |
| 20:44 | AryehGregor | waves to rubys |
| 20:44 | rubys | waves back |
| 20:45 | <Hixie> | i wonder what should happen if you keep hold an AudioTrack element and then manipulate it while the track no longer applies |
| 20:45 | <Hixie> | hmmm |
| 20:45 | <hober> | rubys: http://www.whatwg.org/issues/ ? |
| 20:45 | <Hixie> | rubys: you mean whatwg.org/issues ? |
| 20:45 | <Hixie> | it's linked to from the top of the spec |
| 20:46 | <Hixie> | hm, actually, it's not, i removed the link a while back |
| 20:46 | <rubys> | hober/hixie: that's the page. Thanks! |
| 20:46 | <Hixie> | rubys: philip has a static version which may be more convenient depending on what you want to do |
| 20:47 | <Hixie> | rubys: there's also an api if you want to do something more complicated than just look at it |
| 20:47 | <othermaciej> | Hixie: is that list reasonably complete? |
| 20:47 | <Hixie> | it's automatically generated from the master list that is my imap folders, daily |
| 20:48 | <Hixie> | it doesn't include feedback that wasn't sent to a public list my script knows about, but otherwise it's complete |
| 20:48 | <othermaciej> | I was going to say it's surprisingly small but there is a *lot* under "processing model" |
| 20:49 | <Hixie> | "processing model" is where all the feedback goes by default |
| 20:49 | <Hixie> | the name is historical |
| 20:49 | <Hixie> | whatwg.org/issues/data.html has a data view |
| 20:50 | <Hixie> | which may be more convenient if you're just worried about numbers |
| 20:50 | <Hixie> | whatwg.org/issues/data.csv has historical data |
| 20:51 | <Hixie> | i also have per-folder summary data if you want that too, but it's not currently anywhere public |
| 20:52 | <Hixie> | maybe each AudioTrack should have a link (implicit or explicit) back to the AudioTrackList, and once the track goes away, the link is cut, at which point it keeps returning its old data but is otherwise dead |
| 20:52 | <Hixie> | i guess that's the closest to what HTMLOptionElement does |
| 20:55 | <othermaciej> | the count on the whatwg issues page doesn't quite line up with the email count |
| 20:55 | <othermaciej> | I expanded everything and took a plaintext dump, it only has around 1000 lines |
| 20:55 | <Hixie> | if you look at the csv file it breaks down the count into the things you can see on the issues/ page and the things it's not showing because they're hidden for some reason or other |
| 20:55 | <othermaciej> | whereas the graph seems to say 1600 |
| 20:55 | <othermaciej> | ok |
| 20:56 | <Hixie> | i haven't updated the list of public lists so there might be some lists that are being included that are treated as private |
| 20:56 | <othermaciej> | 1066,346,1628 |
| 20:56 | <Hixie> | let me check that now |
| 20:56 | <othermaciej> | the public and private don't seem to add up to the total |
| 20:56 | <Hixie> | the total isn't the sum of the others |
| 20:57 | <othermaciej> | are there email categories other than public and private? |
| 20:57 | <Hixie> | yeah there's like 6 |
| 20:57 | <Hixie> | hold on, let me update the list first and then look at what the categories are |
| 20:58 | <Hixie> | hah, my script in fact is outputting "Miscount!!! Total 1554 != 1057 visible + privateCount hidden" which doesn't bode well |
| 20:58 | Hixie | investigates further |
| 20:58 | <othermaciej> | part of the reason for asking this is to estimate how much effort it would take for someone to put this data into bugzilla, should anyone care to do so (not that we're planning to do that without consulting) |
| 21:00 | <Hixie> | sounds like a terrible idea |
| 21:00 | <Hixie> | mail feedback and bug feedback are quite different |
| 21:00 | <Hixie> | mail feedback is where things get discussed, threads fork, etc |
| 21:00 | <Hixie> | bug feedback is only good for simple issues without discussion |
| 21:00 | <rubys> | Hixie: it is easier to prove that it is a terrible idea with hard data; that's what we are gathering here. |
| 21:01 | <othermaciej> | like rubys says, we are only gathering data at this point |
| 21:02 | <Hixie> | looks like some of the difference in counts is that a bunch of e-mails are in the folders twice |
| 21:06 | <Hixie> | ok i updated my list of public lists, regenning now |
| 21:23 | <zcorpan> | aw man krijn's cat is still messing with the cords |
| 21:26 | <othermaciej> | wow, email count went way down and the private email count is very low |
| 21:28 | <Hixie> | othermaciej: ok i fixed the script to not count bugmails, which was skewing the count badly |
| 21:28 | <Hixie> | othermaciej: and fixed the various bugs that were making things count double |
| 21:28 | <othermaciej> | I can imagine... |
| 21:29 | <othermaciej> | are any of the 35 private emails things that need to be addressed urgently? and do they really need to be private? |
| 21:29 | <othermaciej> | hope this is not too much distraction |
| 21:29 | <othermaciej> | I'm asking b/c 35 seems like a pretty small number |
| 21:29 | <Hixie> | they're just things people mailed to me directly |
| 21:29 | <Hixie> | usually rants that don't result in the spec changing :-) |
| 21:30 | <othermaciej> | k |
| 21:33 | <Hixie> | (no idea if they _need_ to be private, but i don't want my script making private mails public so i make it err on the side of caution) |
| 21:34 | <zcorpan> | maybe i should start sending my feedback to Hixie directly |
| 21:34 | <zcorpan> | and just rant in public |
| 21:40 | <AryehGregor> | zcorpan, why did you specifically ask me back in March to spec queryCommandSupported()? It seems like a pretty trivial thing. |
| 21:41 | <zcorpan> | AryehGregor: dunno, was probably looking at some opera bug about it |
| 21:41 | <yuhong> | Just found this: http://support.microsoft.com/kb/974322 |
| 21:43 | <yuhong> | Short version: they added support for doing appendChild to ancestors beyond the parent elements in a security update! |
| 21:44 | <zcorpan> | appending children to its ancestors can be dangerous! |
| 21:44 | <zcorpan> | s/its/their/ |
| 21:46 | <yuhong> | zcorpan: what do you mean? |
| 21:47 | <zcorpan> | i was just rambling |
| 21:47 | <yuhong> | Example: http://blogs.msdn.com/b/ie/archive/2008/04/23/what-happened-to-operation-aborted.aspx\ |
| 21:48 | <yuhong> | In fact, found this out when I tried it myself. |
| 21:51 | <zcorpan> | thank you aria for doing this. http://www.sitepoint.com/forums/html-xhtml-52/nav-vs-div-role%3D-navigation-766658.html |
| 21:52 | <yuhong> | I wonder if there is any way to detect from JS. |
| 22:09 | <Hixie> | does http://www.whatwg.org/specs/web-apps/current-work/complete.html#audiotracklist-and-videotracklist-objects seem like the right direction? |
| 22:10 | <Hixie> | (ignore the inheritance chain, i might change that. not sure whether it's worth making them inherit from the same thing given how trivial the interfaces are otherwise) |
| 22:11 | <Hixie> | ok back in a bit to finish this |
| 22:44 | <zewt> | bleh, can't find a workaround for an annoying chrome bug http://code.google.com/p/chromium/issues/detail?id=86730 |
| 22:47 | <bga_> | annevk http://www.twitlonger.com/show/bd2994 |
| 22:47 | <bga_> | i guess s/Event/DOMEvent/ is solution |
| 22:50 | <zcorpan> | so what about new Event('foo', {get cancelable:function(){alert('lol')}}) |
| 22:51 | <bga_> | throw TypeError('value required')? |
| 22:52 | <zcorpan> | because it's a getter? |
| 22:53 | <bga_> | yeah |
| 22:53 | <zcorpan> | throwing would probably reduce impl pain, so yeah |
| 22:56 | <zcorpan> | new Event('foo', {cancelable:{valueOf:function(){alert('lol'), toString:function(){alert('wut')}}}) |
| 22:56 | <bga_> | hehe |
| 22:56 | <zcorpan> | bedtime |
| 23:01 | <heycam> | bga_, Web IDL defines when and in which order those properties are got, and the getters will run |