| 00:30 | <tantek> | othermaciej - the original web intents w3cmeme: http://w3cmemes.tumblr.com/post/22399681762 (though I suppose it could apply to any number of features proposed by first-time web platform feature requesters ;) ) |
| 00:44 | <grom358> | ah... so looking at the UTF-8 2.4 section it says to replace overlong forms with U+FFFD |
| 00:45 | <grom358> | and I'm wondering.. is overlong form pretty much whenever you have 80 bytes at the start of the sequence |
| 01:25 | <othermaciej> | hober: I made a bad mistake and decided to read that whole www-tag thread |
| 01:37 | smaug____ | just gives up w3 bugmail for now. They are totally unreadable |
| 01:37 | <smaug____> | shepazu: ^ |
| 01:38 | <shepazu> | smaug____: I've filed a bug… sorry about that... |
| 01:39 | <smaug____> | shepazu: do you have the bug# ? |
| 01:39 | <shepazu> | smaug____: sysreq doesn't use bugzilla... |
| 02:40 | <Hixie> | weird |
| 02:41 | <Hixie> | there seems to be a higher proportion of crazy bugs filed using the form on the /TR/ specs than on the whatwg ones |
| 02:56 | <shepazu> | http://www.bugzilla.org/releases/4.2/release-notes.html#v42_feat_email |
| 03:58 | <karlcow> | Hixie higher proportion for spam? → Google karma ? |
| 07:22 | <jgraham> | Sigh. Can someone tell sicking to lay off the flamebait |
| 07:26 | <annevk> | “I wish to not be involved in threads regarding the W3C process any more.” hear hear |
| 07:28 | <annevk> | hober: maybe clean up @w3cmemes a bit? |
| 07:28 | <jgraham> | annevk: Link? |
| 07:28 | <annevk> | or maybe it doesn't matter much |
| 07:29 | <annevk> | my inbox, but also http://lists.w3.org/Archives/Public/www-archive/2012Oct/0012.html it seems |
| 07:29 | <hsivonen> | jgraham: link? |
| 07:31 | <annevk> | http://i.qkme.me/3p5tkf.jpg |
| 07:31 | <jgraham> | hsivonen: On dev.mozilla.platform. I have no idea how to link to it |
| 07:31 | <annevk> | google groups of course |
| 07:31 | <jgraham> | hsivonen: I guess I might have been being unfair, but it kind of read like "Opera people write crappy testsuites" |
| 07:32 | <jgraham> | annevk: How do I link to a specific post? |
| 07:32 | <annevk> | you mean like https://groups.google.com/d/msg/mozilla.dev.platform/AUJaVnuGFKI/q_giMS1einEJ ? |
| 07:33 | <annevk> | press some arrow on the side |
| 07:33 | <jgraham> | Ah |
| 07:33 | <hsivonen> | jgraham: I think you’ve heard the same feedback about the complexity of the harness from various Mozilla devs over the last couple of years |
| 07:34 | <hsivonen> | jgraham: if you want us to write to your harness instead of mochitest, proposing a harness that’s harder to use does not make it an easy sell |
| 07:34 | <jgraham> | hsivonen: Well mostly from you and sicking tbh |
| 07:35 | <jgraham> | But that's not the point |
| 07:36 | <jgraham> | The point is it's easier to work on some compromise if you don't write something that sounds like "their requirements aren't important and they write crappy tests ahyway" |
| 07:36 | <othermaciej> | annevk: not impressed with him saying that on a thread he started by invoking W3C process bs |
| 07:37 | <annevk> | othermaciej: you could see it as a positive, he learned something ;-), but sure |
| 07:38 | <hsivonen> | othermaciej: also poor form not to say a clear yes or no to a yes or no question |
| 07:38 | <hsivonen> | which isn’t a false dichotomy or trick question |
| 07:38 | <annevk> | personally I'm happy to talk to Jeff & co about W3C Process and more importantly copyright, but I suspect my changes of success are rather low |
| 07:39 | <annevk> | chances, even |
| 07:39 | <zcorpan> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=19311 oooh i want an html5 too! |
| 07:39 | <hsivonen> | annevk: the copyright stance of the W3C stinks |
| 07:40 | <ashemedai> | That's some nifty use of HTML5 there: http://www.drawtheline.org/ |
| 07:40 | <othermaciej> | hsivonen: indeed, particularly when the answer could free him from being involved in further w3c process threads |
| 07:41 | <hsivonen> | annevk: after all, the W3C isn’t paying you |
| 07:41 | <othermaciej> | annevk: you might be able to get an audience with him and he is generally sensible in my experience |
| 07:42 | <annevk> | othermaciej: http://lists.w3.org/Archives/Public/www-archive/2012Oct/0008.html |
| 07:42 | <othermaciej> | I would personally greatly prefer for the w3c to use a permissive license for all its spec; indeed I can't think of any good reason not to |
| 07:42 | <othermaciej> | but continuing to press the point is frustrating |
| 07:43 | <othermaciej> | on the other hand, while it is arguably hypocritical for the w3c to deny forking of its own specs while making use of it, it is also poor form to publish something under a permissive license and then complain about people making use of the permissiveness |
| 07:44 | <othermaciej> | annevk: so my prediction about being able to meet with him was right, though it turns out to be a postdiction |
| 07:45 | <jgraham> | Not that poor form, really. Like the quote about free speech "I disapprove of what you say, but I will defend to the death your right to say it" |
| 07:45 | <annevk> | othermaciej: I think "we don't allow you to fork, but we're happy to fork you" is kinda silly too though |
| 07:45 | <annevk> | othermaciej: and I think mocking such a stance is fine, I don't think anyone is really complaining about it |
| 07:46 | <othermaciej> | I don't think Ms2ger is defending the w3c's right to fork to the death :-/ |
| 07:46 | <annevk> | oh Ms2ger |
| 07:46 | <jgraham> | Well he is publishing under a license that allows it |
| 07:46 | <annevk> | thought you were talking about http://w3cmemes.tumblr.com/post/30399969240 |
| 07:46 | <jgraham> | So he is. He might also be grumpy about the fact that they are taking the option :) |
| 07:46 | <othermaciej> | yeah but https://www.w3.org/Bugs/Public/show_bug.cgi?id=11204#c54 |
| 07:47 | <othermaciej> | I can't argue with that particular meme |
| 07:47 | <annevk> | Ms2ger is just trolling |
| 07:47 | <annevk> | cannot really read what he wrote in that bug in any other way :-) |
| 07:48 | <othermaciej> | I know but it is trolling gone a bit too far (particularly since he already made his point on public-webapps) |
| 07:49 | <hsivonen> | othermaciej: in some cases, having license to fork is like having a credible army. |
| 07:50 | <hsivonen> | othermaciej: that is, it’s not something you are supposed to use, but the threat is supposed to make others behave reasonably. |
| 07:50 | <othermaciej> | I guess I may have a different perspective because webkit is probably one of the most forked projects ever and many forkers have never contributed back |
| 07:51 | <AryehGregor> | The W3C's position seems to be something like: We and other Established Standards Bodies should be the ones developing standards. Therefore, if someone else writes some type of spec outside of an Established Standards Body, we're doing the world a favor by taking it over. If we write a spec and some non-Established Standards Body tries to take it away from us, however, that's bad. |
| 07:51 | <othermaciej> | if I got upset about it I wouldn't have time to do real work (or chatter semi-pointlessly on irc) |
| 07:51 | <AryehGregor> | I don't agree with the position, but it's not hypocritical. |
| 07:51 | <AryehGregor> | Self-serving, yes. |
| 07:52 | <AryehGregor> | hsivonen, what people have objected to the complexity of testharness.js who have actually written a nontrivial number of tests in it? |
| 07:52 | <AryehGregor> | I've used it a lot, and mochitest, and they seem about equally easy to use. |
| 07:52 | <AryehGregor> | (although mochitest has more features, understandably) |
| 07:52 | <othermaciej> | I don't know if it's about Established per se; the W3C would not favor other Established bodies forking their specs either (unless through a delta spec, which they are totally cool with apparently) |
| 07:53 | <jgraham> | Yeah, in practice people fork W3C specs all over the place |
| 07:53 | <AryehGregor> | othermaciej, no, but they also don't generally fork specs that are already actively maintained by organizations like the IETF or ISO, do they? |
| 07:53 | <jgraham> | For practical definitions of "fork" |
| 07:53 | <othermaciej> | do the IETF or ISO have licenses that would permit doing so? |
| 07:53 | <jgraham> | Like "you can't implement W3C spec A and spec B in the same codebase and conform to both" |
| 07:54 | <AryehGregor> | Doubtful. |
| 07:54 | <othermaciej> | (as far as I am aware the answer is no) |
| 07:54 | <AryehGregor> | Delta specs are perhaps less objectionable because they inherently cannot supplant the original spec, and so they can't really take anything away from the original standards body. |
| 07:54 | <othermaciej> | so we don't have a natural experiment available to us |
| 07:54 | <jgraham> | (well, maybe there is usually some way you could e.g. by switching on some magic marker) |
| 07:55 | <jgraham> | AryehGregor: That's a very standards-body centric view |
| 07:55 | <AryehGregor> | Anyway, I don't think anyone at the W3C has said that forking specs is bad per se, only that other people shouldn't fork *their* specs. |
| 07:55 | <AryehGregor> | jgraham, what do you expect from a standards body? |
| 07:55 | <jgraham> | Well sure |
| 07:55 | <jgraham> | I'm just pointing out that it doesn't really make sense outside the echo chamber |
| 07:56 | <AryehGregor> | Member companies pay good money to be part of the W3C, they're not going to vote to allow forking specs they could have paid influence over. |
| 07:56 | <AryehGregor> | I don't think it doesn't make *sense*, it's just not true. |
| 07:56 | <AryehGregor> | At least for web stuff. |
| 07:56 | <jgraham> | Depends which Member companies |
| 07:56 | <AryehGregor> | Because mostly it's just a few organizations and they move really quickly and can change things easily. |
| 07:57 | <AryehGregor> | For, I dunno, electrical wiring standards, probably standards bodies are a lot more essential. |
| 07:57 | <othermaciej> | available evidence would indicate that most Members with an opinion do not favor allowing forking |
| 07:57 | <AryehGregor> | I think it's a quite clear majority, no? |
| 07:57 | <othermaciej> | (though the Members with opinion are not necessarily the ones contributing to the specs in question) |
| 07:57 | <jgraham> | Right, that was going to be my point |
| 07:58 | <AryehGregor> | I recall the result of the HTMLWG tally being roughly: members that heavily contribute to web standards mostly favor allowing forks, other members overwhelmingly oppose. |
| 07:58 | <jgraham> | I have no idea what the results of that vote were, because they are Super Secret |
| 07:58 | <AryehGregor> | There was a public-ish vote within the HTMLWG a few years ago. |
| 07:58 | <jgraham> | But yes, exactly, the people who care are the people that are least affected |
| 07:58 | <othermaciej> | HTML WG participants survey had very different results from the AC survey |
| 07:59 | <jgraham> | Which doesn't back up the theory that it's about having paid influence |
| 07:59 | <AryehGregor> | Yes, probably because it was so heavily weighted toward contributing organizations. |
| 07:59 | <AryehGregor> | jgraham, the organizations that write the spec control what it says anyway, so they don't need influence through the W3C. The ones who don't only have influence through the W3C, and in fact the W3C is their only real way of having influence on the web platform. |
| 08:00 | <jgraham> | I suspect it's more about having a theoretical model of the guarantees that a standards body should provide, and a limited understanding of how stuff actually works in the case of the Web Platform |
| 08:01 | <othermaciej> | for those who can see Member stuff: https://www.w3.org/2002/09/wbs/33280/htmllicense2010/results |
| 08:02 | <AryehGregor> | Then why should they care either way? What's their stake in it? Some of these companies have no direct involvement in any part of things like HTML, so why should they bother even voicing an opinion? |
| 08:02 | <jgraham> | You get people saying things like "we need the standard to be stable so that we can deploy things to our multi-thousand seat intranet and then leave it alone for a decade" |
| 08:02 | <jgraham> | When of course there is no such guarantee |
| 08:02 | <AryehGregor> | othermaciej, didn't Sam do a breakdown by organization at the time? |
| 08:02 | <AryehGregor> | Oh, I'm thinking of a different vote. |
| 08:04 | <othermaciej> | this is the html wg survey: http://lists.w3.org/Archives/Public/public-html/2011May/0121.html |
| 08:05 | <othermaciej> | I believe w3c management refused to forward the results to the AC because the survey included licenses that allow forking |
| 08:09 | <AryehGregor> | Not much point in pestering the AC with things it already decided decisively. |
| 08:34 | <annevk> | if not for @BrendanEich I would have thought NaCl to be dead |
| 08:36 | <darobin> | AryehGregor: I would contend that the push to get the AC to agree to a fork-friendly license was poorly made last time around |
| 08:37 | <AryehGregor> | Maybe, but do you think they could realistically be convinced regardless? |
| 08:37 | <othermaciej> | darobin: I definitely agree with that - the w3c team did not frame it in a way that was likely to get a good reception |
| 08:37 | <darobin> | if people are interested in making a fork-friendly license happen, I think it's possible but it requires a will to make it happen through the AC, with a decent plan, and the energy to match |
| 08:37 | <darobin> | AryehGregor: I think there's a chance, but it should be driven from the floor, and open with strong political momentum |
| 08:38 | <darobin> | in other words, it would require an AC rep to organise a statement and come to the table with the goods already written up in a convincing fashion |
| 08:38 | <darobin> | someone would have to |
| 08:38 | <darobin> | 1) draft a license |
| 08:38 | <darobin> | 2) draft a statement of support |
| 08:38 | <darobin> | 3) get as many supporters offline as possible to already sign off on it |
| 08:38 | <darobin> | and 4) *then* submit that package to the AC |
| 08:39 | <darobin> | AC votes don't typically involve the whole membership, so you don't need 200 members to carry the day on this |
| 08:40 | <darobin> | during the last attempt, at least two things went wrong IMHO |
| 08:40 | <darobin> | one is that it was pitched as an exception granted to the HTML WG |
| 08:40 | <darobin> | that's just bad politics, and it doesn't make sense to make changes for one WG — it should be a WG-level decision which license to use |
| 08:41 | <darobin> | and two much of the criticism levelled at the forkable options was that it wasn't clear whether they enabled forking or not |
| 08:41 | <darobin> | some of the votes against those cite that as the reason — which is a shame |
| 08:41 | <othermaciej> | I think pitching it as an exception was partly to make people less scared that it was changing how all WGs would work |
| 08:42 | <othermaciej> | it's possible this was counterproductive though |
| 08:42 | <darobin> | othermaciej: yeah, but a) the HTML WG triggers knee-jerk antagonistic reactions and b) it's unfair to other WGs |
| 08:42 | <othermaciej> | I personally would like all WGs to have the option |
| 08:42 | <annevk> | for those with access (hey darobin) you might to check how many members actually voted; it's fascinating how democrazy works |
| 08:43 | <othermaciej> | the way it was managed with the AC was not in any way in accordance with my preferences |
| 08:43 | <darobin> | annevk: well, votes are won by those who show up — nothing new there |
| 08:44 | <darobin> | so I guess my overall point here is: there's a shot at making this happen, but it can only happen from an organised action within the AC |
| 08:44 | <annevk> | here's a way to do it: "We're gonna do CC0 unless a two-third majority objects." |
| 08:45 | <annevk> | or you know, one-third would already cut it, or even less... |
| 08:45 | <darobin> | I'd be happy to sit down with whoever is interested to chat out the details at TPAC |
| 08:46 | <darobin> | annevk: nah, proposing a brutal vote like that will just backfire. Besides, there's no support in Process for such a modality |
| 08:47 | <othermaciej> | I don't even know what the official Process is for changing the W3C Document License or adding an alternative |
| 08:48 | <othermaciej> | as far as I can tell, the W3C Process and the W3C Membership Agreement both entitle the W3C to use whatever license it wants |
| 08:49 | <othermaciej> | (though oddly the membership agreement mentions the software license by name) |
| 08:51 | <othermaciej> | In fact, if you read the W3C Membership Agreement literally, it seems to require that all copyrightable materials produced by the w3c must be published under <http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231>; |
| 08:51 | <othermaciej> | (IANAL) |
| 08:51 | <othermaciej> | "Specific exceptions may be made upon approval of the Director, with the advice of the Advisory Committee." |
| 08:52 | <darobin> | othermaciej: if the AC votes on changing that, then it becomes effective after the vote |
| 08:52 | <darobin> | my process point was that you can't introduce a 2/3rds majority rule out of the blue |
| 08:52 | <darobin> | you'd have to have a vote on that first, and then we're in process hell :) |
| 08:52 | <othermaciej> | yeah, but it seems like in principle the w3c is allowed to publish specs under the (GPL-compatible, forking-permitting) W3C Software License without consulting the AC at all |
| 08:53 | <darobin> | mmmmmm |
| 08:53 | <othermaciej> | in fact, publishing a spec under anything else (such as the W3C Document License) requires a Director exception |
| 08:53 | <othermaciej> | http://www.w3.org/2009/12/Member-Agreement |
| 08:53 | <AryehGregor> | I wonder how closely organizations are legally held to this sort of thing. |
| 08:54 | <darobin> | that would make for an interesting loophole; though if it's true it'll be a shitstorm to apply it :) |
| 08:54 | <othermaciej> | well, one could presume the Director is making an exception for each and every spec published |
| 08:54 | <othermaciej> | though I don't see the AC giving advice on each specific exception |
| 08:55 | <othermaciej> | seriously, read the "b. Ownership of Copyrights and Patents" section |
| 08:55 | darobin | reading |
| 08:55 | <othermaciej> | I should ask someone at w3c who is in the know whether the w3c believes they are in compliance with that section |
| 08:55 | <othermaciej> | (it is also possible that this membership agreement is not the operative one as it is labeled "draft") |
| 08:55 | <darobin> | wait, that document says ***DRAFT*** |
| 08:56 | <annevk> | find me a Member Agreement that does not say DRAFT |
| 08:56 | <annevk> | could not find one last time |
| 08:56 | <othermaciej> | yes, but it it;s the latest version linked from here: http://www.w3.org/Consortium/Agreement/ |
| 08:57 | <darobin> | ah I know |
| 08:57 | <darobin> | this is a draft contract — but it is the actual agreement |
| 08:57 | <darobin> | you remove the draft after parties have agreed to the content |
| 08:57 | <othermaciej> | the latest version that does not say DRAFT is here: <http://www.w3.org/2003/01/Member-Agreement>; but it says it has been superseded by the 2005 version |
| 08:58 | <othermaciej> | (and it also has the bit about the w3c software license) |
| 08:58 | <annevk> | darobin: what is btw with W3C people and their tendency to suggest "lets discuss this at this later date" instead of actually fixing it? |
| 08:59 | <darobin> | annevk: have I done that? |
| 08:59 | <annevk> | you just said you're happy to discuss this at TPAC :-) |
| 09:02 | <darobin> | annevk: heh, fair enough :) |
| 09:02 | <darobin> | I was suggesting TPAC because it's pretty soon and that I'll have some time to devote to that then |
| 09:02 | <darobin> | but if you want to get the ball rolling for this, I'm happy to help you now |
| 09:03 | <darobin> | just now in a synchronous fashion because I have a bunch of other things to attend to |
| 09:03 | <darobin> | so email is fine! |
| 09:03 | <annevk> | not really sure how, and Jeff could only devote time to it around that time too |
| 09:04 | <annevk> | I was mostly curious about this postponing strategy, some managers at Opera did the same thing, even for particularly pressing issues |
| 09:04 | <annevk> | (not saying licensing is pressing) |
| 09:04 | <darobin> | I don't think anyone is deliberately postponing anything — in Jeff's case I don't want to speak for him but I suspect that he just wanted to chat with you f2f |
| 09:05 | <darobin> | in this case, if you have the time to devote to it, I reckon the first thing you need to do is select a licence |
| 09:05 | <darobin> | one that's reasonable and all |
| 09:05 | <darobin> | as part of the license, explain how it interacts with IP — it's important that members aren't scared of changing the patent policy just because a document is forkable |
| 09:05 | <darobin> | (that's possibly the hard part, you'll need a lawyer friend) |
| 09:06 | <darobin> | then draft a grandiose statement about how forking will save the world and is the future of the web |
| 09:06 | <darobin> | (and I mean that) |
| 09:06 | <darobin> | the sort of thing that companies would be proud to put their names at the bottom of |
| 09:06 | <darobin> | (if you draft I can help with the language) |
| 09:07 | <annevk> | ah, Date.now() and performance.now() are not at all compatible... :/ |
| 09:07 | <darobin> | then we can start putting together a list of supporters |
| 09:07 | <darobin> | I don't have the bandwidth to lead on this (otherwise I'd be doing it) but I can help |
| 09:07 | <darobin> | I know the AC decently well :) |
| 09:08 | <annevk> | sounds like we need lawyers really :/ |
| 09:10 | <darobin> | annevk: there are quite a few people here who work for companies that have lawyers — maybe one of them could help there |
| 09:11 | darobin | nails anne to the channel so he stops dropping out |
| 09:13 | <AryehGregor> | I wrote a mini essay for the HTMLWG vote, if anyone is interested. |
| 09:13 | <darobin> | AryehGregor: sure |
| 09:14 | <AryehGregor> | http://lists.w3.org/Archives/Public/www-archive/2011May/0008.html |
| 09:14 | <AryehGregor> | I doubt it's ideal for convincing anyone. |
| 09:14 | <AryehGregor> | Probably too long, and likely some of the points should be dropped or revised depending on the audience. |
| 09:15 | <AryehGregor> | Also, some parts are quite HTML5-specific. |
| 09:15 | <darobin> | I'll read that later today, thanks |
| 09:19 | <AryehGregor> | And the end could definitely use tightening up -- the last two paragraphs are tacked on. |
| 09:19 | <AryehGregor> | Possibly could use some more sources, too. |
| 09:25 | <Ms2ger> | AryehGregor, does http://pastebin.mozilla.org/1862848 look better to you? |
| 09:25 | <annevk> | darobin: sorry about that, was home -> cycle -> train (with tethering) -> walking -> with RobbertAtWork |
| 09:30 | <annevk> | darobin: I guess I should talk to tantek; he and Mozilla care |
| 09:30 | <AryehGregor> | Ms2ger, 1) FYI, the patch conflicts with a change I pushed a few hours ago. this.inheritance is now this.base, which is either string or null. Should be easy to fix. 2) Don't we have at least one other place already in the file where we try to figure out if an interface prototype object is correct? I'd unify the checks into a helper function. But that doesn't make the patch wrong, it could be done later. At least add a TODO, though. 3) |
| 09:30 | <AryehGregor> | Why did you move up the assert_class_string check? Any specific reason? |
| 09:30 | <AryehGregor> | Other than that, LGTM. |
| 09:30 | <annevk> | hey look, it's Ms2ger |
| 09:30 | <Ms2ger> | Hi annevk :) |
| 09:30 | <annevk> | Ms2ger: going to add Event.systemTime later today |
| 09:30 | <Ms2ger> | Busy week :) |
| 09:31 | <AryehGregor> | If you want to push that, I'm fine with it. Good catch. I think that code can use more work, but I'm happy to be the one to do that. |
| 09:31 | <AryehGregor> | (we can really do all the same checks for NoInterfaceObject interface prototype objects as for any other interface prototype objects) |
| 09:32 | <AryehGregor> | (I didn't prioritize it because DOM doesn't contain any NoInterfaceObject -- right? -- and that's what I was focusing on) |
| 09:32 | <Ms2ger> | 1) Yay, 2) don't immediately find one, 3) because I wanted the NoInterfaceObject mess to not get in the way :) |
| 09:32 | <Ms2ger> | Class starts, see you |
| 09:34 | <AryehGregor> | See you. |
| 09:42 | <annevk> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=19402#c0 I'm not even seeing the problem :-( |
| 09:43 | <annevk> | oh, I might now |
| 10:04 | <annevk> | wtf happened to W3C bugmail? |
| 10:04 | <annevk> | ow sorry |
| 10:04 | <annevk> | wtwtf happened |
| 10:05 | <darobin> | annevk: bugzilla has been upgraded |
| 10:05 | <darobin> | if you're seeing issues -> sysreq⊙wo is a good place to report |
| 10:10 | <annevk> | ta |
| 10:15 | <annevk_> | I think Nightly just froze my entire system |
| 10:22 | <hsivonen> | the W3C bugmail doesn’t look that bad to me |
| 10:22 | <hsivonen> | I think I even like the HTML table in place of the ASCII art table |
| 10:24 | <hsivonen> | darobin: who is ledahulevogyre? |
| 10:24 | <darobin> | hsivonen: some guy on Twitter — why? |
| 10:24 | <darobin> | with a cool nick I may add |
| 10:26 | <annevk> | anyone know where heycam is? |
| 10:26 | <hsivonen> | darobin: well, the nick looked like it might be your Twitter alter ego, but then I saw him comment on github, too |
| 10:26 | <darobin> | hsivonen: heh, no, I only go by one identity :) |
| 10:27 | <hsivonen> | darobin: since you follow him, I though you might know who he is |
| 10:27 | <darobin> | the cult of the dahut is far more widespread than you realise hsivonen, we are everywhere |
| 10:27 | <jgraham> | darobin: So you deny beingg Mr Last Week> :p |
| 10:27 | <darobin> | I follow him because he has a cool nick, and we've exchanged a few times |
| 10:27 | <jgraham> | s/>/?/ |
| 10:28 | <karlcow> | Le Dahu Levogyre… lévogyre seems a reference to Goldorak. Probably someone in between 35 and 45 |
| 10:28 | <darobin> | jgraham: indeed, I can say "it waddunt me" :) |
| 10:28 | <darobin> | karlcow: I hadn't seen the Goldorak reference there |
| 10:29 | <hsivonen> | I suspected he trolled me on Twitter instead of really thinking I was confused, so I didn’t reply. Maybe I was impolite. |
| 10:29 | <karlcow> | darobin: Come out of this body robinCornofulgure |
| 10:30 | <darobin> | karlcow: dahuts can be either lévogyre or dextrogyre, it's part of their nature |
| 10:30 | darobin | is unlikely to get Goldorak references, didn't watch that much tv as a kid :) |
| 10:30 | <darobin> | hsivonen: I suspect I've probably met him IRL, but I tend to have a very poor memory of people I've seen only a few times at busy conferences.... |
| 10:30 | <hsivonen> | karlcow: I thought lévogyre was a joke on making polarizing comments |
| 10:31 | <jgraham> | hsivonen: You writing a HTML parser in Rust? |
| 10:32 | <hsivonen> | jgraham: I’m probably translating the one I’ve already written to Rust |
| 10:32 | hsivonen | is a one-trick pony |
| 10:32 | <darobin> | I suspect lévogyre is actually an indication of political orientation |
| 10:32 | <annevk> | hsivonen: I think he was confused |
| 10:33 | <darobin> | at least, the tradition in dahut jokes is that it is |
| 10:33 | <karlcow> | aaaaah no, darobin, au temps pour moi, it is Clavicogyre |
| 10:33 | <jgraham> | hsivonen: That… might or might not be a good idea. I guess I hadn't considered the possibility. |
| 10:33 | <karlcow> | http://membres.multimania.fr/armelanuel/armes.html |
| 10:33 | <annevk> | hsivonen: I doubt many people realise that effectively UTF-16 is the only encoding that breaks ASCII |
| 10:33 | <karlcow> | I wonder if it's the first time that goldorak is mentioned on #whatwg |
| 10:33 | <karlcow> | There is an achievement there |
| 10:33 | <darobin> | karlcow: you are a fount of wisdom about things I didn't even know existed :) |
| 10:34 | <hsivonen> | jgraham: today, I’ve mostly been writing a Java to Java transform that removes fall-through in switch |
| 10:35 | <jgraham> | I see |
| 10:35 | <jgraham> | Rust seems nice enough that it would be fun to write from-scratch to get it idiomatically correct |
| 10:36 | <jgraham> | (I have only played with it a very little bit though) |
| 10:41 | <hsivonen> | darobin: so how do lévogyre and dextrogyre map to politics? Left and right respectively? |
| 10:41 | <darobin> | yeah, turning towards left or right |
| 10:45 | <karlcow> | ☺ and not the same level of turn depending on the country |
| 10:46 | <hsivonen> | levels of turn famously vary by language :-) |
| 10:51 | <annevk> | I always get this impression that sicking (or maybe someone he works with) just implements something and then bothers to tell everyone else involved in the standards process after the fact |
| 10:51 | <annevk> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=17242#c9 |
| 10:51 | <annevk> | it's getting kinda annoying |
| 10:51 | <annevk> | especially if you read elsewhere about Mozilla "doing the right thing" all the time |
| 10:52 | <hsivonen> | \o/ my Java to Java transform still passes the test suite |
| 10:55 | <hsivonen> | annevk: always or just after B2G started? |
| 10:55 | <doublec> | annevk: mostly since b2g ramped up I suspect |
| 10:55 | <annevk> | I had it with CORS too |
| 10:55 | <annevk> | and I've been doing my best to follow up in bug reports and such |
| 10:56 | <annevk> | but I don't get the same courtesy it seems |
| 10:57 | <annevk> | you can't both complain about NaCl and just extend XMLHttpRequest without telling other vendors about it |
| 10:57 | <annevk> | well, apparently you can |
| 10:58 | <hsivonen> | hooray. ibm864 removed from Gecko |
| 10:58 | <annevk> | nice |
| 10:58 | <annevk> | can't stay mad at Gecko too long it seems :p |
| 11:00 | <annevk> | AryehGregor: so yeah, e.g. shift_jis is Shift_JIS |
| 11:00 | <annevk> | AryehGregor: seems like a more annoying API, but you're right that we should probably just go with it and advice people to use toLowercase() or some such |
| 11:01 | <AryehGregor> | annevk, yes, I don't think it's worth trying to get browsers to change from the plurality on little things like this. |
| 11:01 | <annevk> | Opera will need to change though |
| 11:01 | <annevk> | because Opera has a nice API |
| 11:01 | <annevk> | and the web wants it ugly |
| 11:01 | <annevk> | booo |
| 11:01 | <annevk> | boooooooooo |
| 11:02 | <AryehGregor> | I don't actually think all-lowercase is inherently nicer than sensible mixed-case. |
| 11:02 | <AryehGregor> | The only thing that's really ugly is all-uppercase. |
| 11:04 | <darobin> | says the guy called AryehGregor |
| 11:04 | <darobin> | maybe if you were annevk you'd see the world differently |
| 11:04 | <AryehGregor> | :) |
| 11:06 | <annevk> | hsivonen: you still think we should define x-user-defined right? http://krijnhoetmer.nl/irc-logs/whatwg/20121003#l-461 |
| 11:06 | <annevk> | hsivonen: and x-user-defined is also its only label correct? |
| 11:09 | <hsivonen> | annevk: I think we should define the way it’s defined in Gecko (and WebKit if SO is to be believed) |
| 11:10 | <hsivonen> | let’s see about the labels… |
| 11:10 | <hsivonen> | annevk: it’s the only label |
| 11:11 | <hsivonen> | that is, lower half maps to Basic Latin and the upper half to PUA |
| 11:11 | <annevk> | okay, I'll try to fix that today as it seems the other things I wanted to fix require some confirmation from others |
| 11:11 | <annevk> | but first the casing sadness |
| 11:11 | <hsivonen> | casing sadness? |
| 11:11 | <annevk> | thanks hsivonen |
| 11:11 | <annevk> | well utf-8 -> UTF-8, shift_jis -> Shift_JIS |
| 11:12 | <hsivonen> | annevk: in what context? |
| 11:12 | <annevk> | windows-1252 stays the same though, but iso-8859-2 becomes ISO-8859-2 |
| 11:12 | <annevk> | hsivonen: the encoding name is exposed through document.characterSet |
| 11:12 | <hsivonen> | annevk: does the Web depend on the case exposed by Gecko? |
| 11:12 | <hsivonen> | annevk: or is it IE’s fault and Gecko just does the same? |
| 11:12 | <annevk> | I'm not sure (Opera has lowercase and no bugs), but AryehGregor argued that we should not change casing as Gecko/WebKit/IE agree |
| 11:13 | <AryehGregor> | Does IE agree? |
| 11:13 | <annevk> | not sure IE agrees |
| 11:13 | <annevk> | I don't have IE |
| 11:13 | <AryehGregor> | I only tested UTF-8, and IE seemed to call that "unicode". |
| 11:13 | <annevk> | oh |
| 11:13 | <annevk> | http://dump.testsuite.org/encoding/label-test.html can be used to test others |
| 11:13 | <hsivonen> | AryehGregor: really? I thought IE called UTF-16LE unicode |
| 11:13 | <annevk> | just type in a label, hit enter, and it'll give you the name back |
| 11:13 | <AryehGregor> | Hmm. |
| 11:14 | <annevk> | I wish I had IE |
| 11:14 | <AryehGregor> | According to that page, IE says "utf-8". |
| 11:14 | <annevk> | oooh |
| 11:14 | <AryehGregor> | So I don't have a problem with that, if it's matching IE/Opera instead of Gecko/WebKit. |
| 11:14 | <AryehGregor> | Yeah, seems to be all lowercase. |
| 11:14 | annevk | makes a happy dance |
| 11:15 | <AryehGregor> | annevk, I use Windows 8 Developer Preview, which I downloaded for free and run in VirtualBox. |
| 11:15 | <AryehGregor> | Works well for me. |
| 11:15 | <AryehGregor> | You use some type of Linux, right? |
| 11:15 | <annevk> | euh no, not enough room |
| 11:15 | <AryehGregor> | Oh, really? Oh well. |
| 11:15 | <annevk> | not enough room to dance |
| 11:15 | <annevk> | I'll try that preview |
| 11:16 | <hsivonen> | IE10 standards mode might not be Web-compatible |
| 11:18 | <AryehGregor> | That's true. |
| 11:18 | <Ms2ger> | othermaciej, my comments on bug 11204 are unrelated to forking, they are about the HTMLWG Process |
| 11:18 | <AryehGregor> | IE9 mode also lowercases. The page doesn't seem to work at all in IE7 or IE8 mode. |
| 11:21 | <hsivonen> | I think I’m better at remembering the canonical casing for charset labels in Gecko than I am at remembering namespace URLs |
| 11:21 | <annevk> | maybe IE8 does not support characterSet? |
| 11:21 | <annevk> | if namespaces were short strings instead of something silly like URLs they might have succeeded |
| 11:21 | <Ms2ger> | othermaciej, I don't deny that the W3C has the right the fork the spec, but I think it doesn't align with the W3C's policy; that's why I objected to the publication in WebApps |
| 11:23 | <AryehGregor> | Which spec is this, parsing and serialization? |
| 11:25 | <Ms2ger> | Yeah |
| 11:26 | <Ms2ger> | annevk, heycam appears to be away until 5 November |
| 11:26 | <annevk> | hsivonen: should I define x-user-defined using an index like the other single-byte encodings or should I define it as an algorithmic encoding somehow? |
| 11:26 | <annevk> | thanks Ms2ger |
| 11:26 | <Ms2ger> | Np |
| 11:29 | Ms2ger | goes off for food |
| 11:31 | <annevk> | hsivonen: going with algorithmic I think |
| 11:32 | <AryehGregor> | annevk, so are you going to stick with lowercase for everything? hsivonen, are we okay with changing to that? |
| 11:33 | <annevk> | AryehGregor: if I can, yes; if Gecko objects I'd like to know from Travis or Adrian what they think |
| 11:33 | <AryehGregor> | Sounds good. |
| 11:38 | <hsivonen> | annevk: algorithmic bytes to UTF-16 code points is super-simple to define at least |
| 11:39 | <hsivonen> | AryehGregor: I have no idea of the Web compat impact. |
| 11:39 | <AryehGregor> | Hmm. |
| 11:39 | <hsivonen> | AryehGregor: in terms of elegance and sanity, I’d prefer lower case |
| 11:40 | <AryehGregor> | Are we willing to try in principle? Really I just need to know whether I can fairly test that spec requirement -- I don't want to test it if it's not clear what we want the spec to say. |
| 11:40 | <hsivonen> | AryehGregor: but if we break something, it will be really hard to justify to product management |
| 11:40 | <AryehGregor> | Sure. |
| 11:40 | <AryehGregor> | We could ask Opera if they've hit any compat issues. |
| 11:41 | <hsivonen> | AryehGregor: I’d we willing to *try*, but you should ping smontagu and perhaps bz. |
| 11:41 | <hsivonen> | AryehGregor: I almost filed the bug already. :-) |
| 11:41 | <annevk> | AryehGregor: Opera has no registered compat issues |
| 11:42 | <annevk> | AryehGregor: I would have been cc'd on any such bugs and also researched this when writing the standard |
| 11:43 | <AryehGregor> | So, anyone want to guess what doc.URL does if doc is something other than the current document? |
| 11:43 | <AryehGregor> | (in browsers) |
| 11:44 | <AryehGregor> | IE: The page's current URL (dunno what happens with iframes). Chrome: Empty string. Firefox: The page's current URL if doc is an HTML document, undefined if it's XML. Opera: about:blank if the doc is an HTML document, undefined if it's XML. |
| 11:44 | <AryehGregor> | Spec: about:blank. |
| 11:44 | AryehGregor | scratches head |
| 11:45 | <annevk> | documents and URLs are a big mess |
| 11:46 | <annevk> | I think the cleanup the spec did here makes sense |
| 11:46 | <AryehGregor> | I'd hope at least *you* think what you wrote makes sense. :) |
| 11:46 | <AryehGregor> | What's the point of .URL anyway? |
| 11:46 | <AryehGregor> | Is it different from .documentURI in implementations in any useful way? |
| 11:47 | <AryehGregor> | Is there any way to change a document's URL if it didn't start its life attached to a window? |
| 11:48 | <AryehGregor> | I guess about:blank makes as much sense as anything. |
| 11:48 | <AryehGregor> | (I hope Gecko and Opera make .URL defined for XML documents if the current page is an XML document.) |
| 11:48 | <annevk> | .URL was first |
| 11:48 | AryehGregor | leaves the test matching the spec |
| 11:48 | <annevk> | and for some reason they added documentURI |
| 11:48 | <annevk> | and the spec makes both the same |
| 11:48 | <AryehGregor> | Sounds good to me. |
| 11:49 | <annevk> | oh, and makes them readonly, which they were in most implementations already |
| 11:50 | <annevk> | AryehGregor: thanks for writing these tests! |
| 11:50 | <AryehGregor> | Most? |
| 11:51 | <annevk> | I think WebKit implemented setting but then removed it because Gecko didn't allow it |
| 11:51 | <annevk> | dunno about others |
| 11:51 | <annevk> | well, don't remember |
| 11:51 | <annevk> | but they aligned on that point to some extent already |
| 11:52 | <AryehGregor> | interfaces.html should already test that, in principle. |
| 11:52 | <AryehGregor> | Whether it's read-only, I mean. |
| 11:53 | <annevk> | hsivonen: do you happen to know how the encoder works for x-user-defined? |
| 11:54 | <AryehGregor> | hsivonen, are you going to file a bug on case-folding here, or shall I? |
| 12:06 | AryehGregor | finds three slightly different implementations of the same basic algorithm in nsDocument::GetDocumentURI, nsHTMLDocument::GetURL, and nsSVGDocument::GetURL |
| 12:06 | AryehGregor | unifies! |
| 12:20 | <hsivonen> | AryehGregor: please go ahead and file |
| 12:20 | <hsivonen> | annevk: let’s see… |
| 12:22 | <hsivonen> | Whoa. We have nsUnicodeToZapfDingbat.cpp |
| 12:25 | <jgraham> | Isn't that from the legay practice of doing <font face="Dingbats">A</font> and expecting an alpha character, or something |
| 12:25 | <hsivonen> | annevk: sorry, the interesting part seems to be buried somewhere outside the encoder itself |
| 12:25 | <hsivonen> | annevk: I don’t know what happens to unencodable characters |
| 12:25 | <jgraham> | Which worked in IE and Netscape 4, but not in Mozilla who insisted finding an actual A glyph (and ignoring the font, since it didn't contain one) |
| 12:26 | <hsivonen> | jgraham: more likely for rendering Unicode on OS/2, Classic Mac OS or legacy X11 |
| 12:26 | <jgraham> | Oh, it's from the inverse problem |
| 12:26 | <jgraham> | I see |
| 12:29 | <annevk> | AryehGregor: could you cc me on your DOM and Encoding-related activities? |
| 12:31 | <annevk> | hsivonen: okay, I'll dig a bit |
| 12:35 | <hsivonen> | wow crazy legacy. Gecko knows about Zapf Dingbats. So does FreeType. And PDF.js. |
| 12:35 | <hsivonen> | yay magic fonts |
| 12:37 | <AryehGregor> | annevk, sure. |
| 12:39 | <hsivonen> | Added in 1999 with the helpful comment “add unicode encoders” |
| 12:41 | <hsivonen> | jgraham: it’s quite possible this thing has never solved a problem but exists merely for completeness |
| 12:42 | <hsivonen> | to cover everything someone has bothered to contribute to unicode.org data |
| 12:43 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=799913 |
| 12:44 | <AryehGregor> | annevk, why does this output the empty string in Opera? <!DOCTYPE html><meta charset="Utf-8"><script>document.documentElement.textContent = document.characterSet</script> |
| 12:45 | AryehGregor | is trying to get a test-case to illustrate the Gecko bug he's filing |
| 12:45 | <AryehGregor> | Also, IE seems to say "unicode" here. |
| 12:45 | <AryehGregor> | Do they not pick up the charset declaration properly somehow? |
| 12:46 | <hsivonen> | AryehGregor: are you testing with document.open() or with real network? |
| 12:46 | <AryehGregor> | Live DOM Viewer. |
| 12:46 | <hsivonen> | AryehGregor: there’s your problem |
| 12:46 | <hsivonen> | AryehGregor: it uses document.open() |
| 12:47 | <AryehGregor> | Oh. |
| 12:47 | <hsivonen> | so always "unicode" in IE |
| 12:47 | <AryehGregor> | I'd try data: URLs, but those don't work in IE for this. |
| 12:50 | <annevk> | AryehGregor: seems Opera is affected by the document.open() thing too |
| 12:50 | <annevk> | AryehGregor: data URL in an <iframe> maybe? |
| 12:50 | <AryehGregor> | Does IE support that? |
| 12:51 | <AryehGregor> | Nope, it doesn't. |
| 12:53 | AryehGregor | just submits the bug without a testcase |
| 12:54 | <hsivonen> | In Gecko, <meta charset> in a document.open()ed doc affects subsequent CSS and JS loads and iframes |
| 12:54 | <hsivonen> | but the doc itself is really UTF-16 |
| 12:54 | <annevk> | oh so weird |
| 12:54 | <annevk> | is that even specced? |
| 12:54 | <hsivonen> | dunno |
| 12:55 | <annevk> | AryehGregor: you could just attach a testcase right?, i.e. the document you just described |
| 12:55 | <hsivonen> | if you ask the doc for its charset, you get the charset inherited into CSS, JS and frames |
| 12:55 | <AryehGregor> | I could, but doesn't seem necessary. |
| 12:55 | <hsivonen> | in Gecko that is |
| 12:56 | <hsivonen> | document.opened docs don’t have a converter, of course |
| 12:56 | <annevk> | fair enough |
| 13:01 | <annevk> | hsivonen: oh, x-user-defined is not supported in <meta>? |
| 13:01 | <AryehGregor> | Whoa, IE doesn't support .documentURI? I wouldn't have guessed that. |
| 13:01 | <annevk> | hsivonen: I need HTTP to test this? |
| 13:01 | <hsivonen> | annevk: right |
| 13:02 | <annevk> | that's a bummer |
| 13:02 | <hsivonen> | annevk: the patch for allowing it in meta hasn’t landed yet |
| 13:02 | <annevk> | oh you're gonna do that? does not seem to work in Chrome either fwiw |
| 13:02 | <annevk> | but consistency wfm |
| 13:03 | <annevk> | it works in Opera btw, except Opera uses the wrong encoder/decoder |
| 13:10 | AryehGregor | discovers a DOM spec bug, yay |
| 13:11 | <annevk> | just one? :-) |
| 13:11 | <annevk> | hsivonen: afaict x-user-defined just emits decoder errors for anything non-ASCII |
| 13:11 | <annevk> | hsivonen: I'm gonna write that down for now |
| 13:12 | <AryehGregor> | annevk, so far! |
| 13:16 | <hsivonen> | annevk: *decoder* errors? |
| 13:16 | <hsivonen> | an is REPLACEMENT CHARACTERS? |
| 13:16 | <annevk> | hsivonen: sorry, encoder errors |
| 13:16 | <annevk> | hsivonen: I keep reversing those terms :-( |
| 13:17 | <hsivonen> | annevk: testing how? |
| 13:17 | <annevk> | hsivonen: URLs and <form> submission |
| 13:17 | <hsivonen> | annevk: OK |
| 13:18 | <hsivonen> | annevk: I’d have expected it to round trip for the first 128 characters of the PUA |
| 13:18 | <annevk> | me too, but the browsers emit the fallback for those instead (? / utf-8 encoded form in URLs depending on the browser and &#x... for <form>) |
| 13:19 | <annevk> | hsivonen: you're hsivonen on github? |
| 13:19 | <hsivonen> | annevk: the encoder definition is supposed to have a defined range from F780 to F7FF |
| 13:19 | <hsivonen> | annevk: yes |
| 13:21 | <hsivonen> | curious. that’s not even near the start of the PUA |
| 13:21 | <hsivonen> | but not quite at the end either |
| 13:22 | <annevk> | hsivonen: well the encoder in practice does not appear to do anything special with that range |
| 13:22 | <hsivonen> | huh. Unicode has wasted two entire planes as astral PUAs |
| 13:22 | <hsivonen> | ok. always trust testing over code inspection |
| 13:22 | <zcorpan> | AryehGregor: i thought this was something we wanted to change, hence the spec. https://www.w3.org/Bugs/Public/show_bug.cgi?id=19431 or is there content relying on this? |
| 13:22 | <hsivonen> | or, rather, comment inspection in this case |
| 13:23 | <AryehGregor> | zcorpan, I assumed it was a spec bug. Why would we want to change it when all browsers agree? |
| 13:23 | <AryehGregor> | If the spec disagrees with all browsers, IMO it should have a note saying so. |
| 13:23 | <AryehGregor> | Until at least one or two browsers change to match. |
| 13:24 | <AryehGregor> | Is there some strong reason that all browsers should change? |
| 13:24 | <zcorpan> | AryehGregor: because scripts should work even when the document is not an html document, iirc |
| 13:24 | <AryehGregor> | Why is that related to namespaceURI? |
| 13:25 | AryehGregor | notices that Opera doesn't implement .parentElement on very many interfaces |
| 13:31 | <annevk> | hsivonen: http://encoding.spec.whatwg.org/#x-user-defined |
| 13:39 | <AryehGregor> | Oh, I think it just doesn't implement .parentElement at all. |
| 13:41 | <AryehGregor> | http://w3c-test.org/webapps/DOMCore/tests/approved/Node-properties.html |
| 13:41 | <AryehGregor> | Okay, what should I test next? |
| 13:41 | <AryehGregor> | Any suggestions? |
| 13:42 | <AryehGregor> | Encodings? |
| 13:43 | <annevk> | AryehGregor: for your createElement test, you might want to check what happens in browsers when the root element does have a namespace |
| 13:43 | <annevk> | AryehGregor: that's why createElement just defaults to HTML, because that's what it most commonly is |
| 13:43 | <AryehGregor> | Mm. |
| 13:43 | <annevk> | AryehGregor: and because defaulting to the root element namespace seemed really yucky |
| 13:43 | <AryehGregor> | Oh, because the Document itself can't have a namespace, right? |
| 13:44 | <AryehGregor> | That does seem lame. |
| 13:44 | <annevk> | AryehGregor: or defaulting to whatever the MIME type is that was in use (as Gecko does) |
| 13:44 | <AryehGregor> | But nobody cares about namespaces, so I see no reason not to match browsers even if they're totally arbitrary. |
| 13:44 | <annevk> | e.g. if you load application/xhtml+xml createElement will create in the HTML namespace in Gecko I think |
| 13:44 | <annevk> | because you get an HTMLDocument and not an XMLDocument |
| 13:44 | <AryehGregor> | Browsers shouldn't have to expend effort to change this stuff so that it's prettier, if we can avoid it. |
| 13:44 | <annevk> | kinda dependent too on the big Document object convergence project |
| 13:44 | <AryehGregor> | Which is what? |
| 13:45 | <annevk> | the main thing is that all browsers have a different strategy for dealing with this stuff |
| 13:45 | <annevk> | and that there's several steps of an agreed upon solution that people haven't really gotten around to implementing because nobody cares about XML |
| 13:46 | <annevk> | one of those things is that there's only 1 Document object |
| 13:46 | <annevk> | and not different objects for SVG / HTML / XML depending on some MIME type |
| 13:47 | <AryehGregor> | "To find the pointers and their corresponding code points in an index, let lines be the result of splitting the resource's contents on U+000A. Then remove each item in lines that is the empty string or starts with U+0023. Then . . ." |
| 13:48 | <AryehGregor> | That seems like an overly explicit way of describing a line-based data format with tab used as a delimiter and # as a comment character. |
| 13:49 | <AryehGregor> | It reminds me of Hixie spelling out all the code points in "<!doctype html>". |
| 13:50 | <jgraham> | Tells you that # must be the first character in the line |
| 13:51 | <annevk> | pull requests welcome |
| 13:51 | <AryehGregor> | Ah, of course, github. |
| 13:51 | <AryehGregor> | Noted. |
| 13:51 | <AryehGregor> | Wow, the encoding spec is extremely long and boring. Kudos on the good work! |
| 13:52 | <annevk> | haha, I'll read that as "fascinating" |
| 13:54 | <zcorpan> | AryehGregor: i think the html spec and dom spec's way of dealing with Document and createElement is the only sane and compatible-enough end point. each browser has not-ideal behavior like looking at the root element namespace or looking at the mime type, or both, and we don't want that |
| 13:54 | <AryehGregor> | I see. |
| 13:55 | <zcorpan> | implementing it has been low prio, though |
| 13:55 | <AryehGregor> | So you're saying that implementers will all have to change anyway, and what's specced is simpler than what all of them do now and believed to be compatible enough? |
| 13:55 | <AryehGregor> | In that case, it makes sense. |
| 13:55 | AryehGregor | looks at what Gecko does |
| 13:55 | <annevk> | right |
| 13:56 | <zcorpan> | yeah |
| 13:58 | AryehGregor | does not actually see what code sets the default namespace for anything but HTML and XUL documents in Gecko |
| 14:00 | <zcorpan> | annevk: i don't understand step 6 of x-user-defined decoder |
| 14:00 | <annevk> | copypasta |
| 14:00 | <annevk> | if I remove that, does it work for you? |
| 14:00 | <zcorpan> | yep |
| 14:01 | <zcorpan> | also, isn't "0xF780 + byte − 0x80" equivalent to 0xF700 + byte ? |
| 14:01 | <annevk> | yes, but the style of the draft is the former since it provides more information |
| 14:01 | <SimonSapin> | zcorpan: I guess this formulation that the results are in the range starting at U+F780 |
| 14:02 | <zcorpan> | ok |
| 14:05 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1832 - seems chrome doesn't support x-user-defined in xml |
| 14:06 | <ashemedai> | Sad thought it might be, it made me laugh: http://www.zdnet.com/the-do-not-track-standard-has-crossed-into-crazy-territory-7000005502/ -- how you guys maintain your sanity is beyond me |
| 14:07 | <zcorpan> | also seems like opera uses U+0080 for 0x80 rather than U+F780 |
| 14:07 | <zcorpan> | ashemedai: what makes you think we maintain our sanity? |
| 14:08 | <annevk> | zcorpan: yeah, we have bugs |
| 14:08 | <annevk> | wait, not we |
| 14:08 | <annevk> | :-) |
| 14:08 | ashemedai | backs slowly away |
| 14:08 | <AryehGregor> | ashemedai, DNT was a ridiculous idea from day one. |
| 14:09 | <ashemedai> | Sure, but dealing with the marketing fellas ain't easy either ;) |
| 14:12 | <zcorpan> | AryehGregor: no we just need more. like DNTAIRMITT:1 |
| 14:12 | <AryehGregor> | ? |
| 14:13 | <zcorpan> | ...and i really mean it this time |
| 14:14 | <tomasf> | DNSM, Do Not Serve Malware. a nice way to opt out from all malware. it's the final solution. |
| 14:15 | <zcorpan> | tomasf: now we're talking |
| 14:15 | <ashemedai> | tomasf: +1 |
| 14:18 | <jgraham> | I am totally up for joining the DNT group if I get to spend time in a house by the beach rather than in Sweden in the winter |
| 14:19 | <karlcow> | AryehGregor: DNT is not a technological solution. It doesn't work by itself basically and is completely ineffective without the legal framework going with it. ☺ |
| 14:19 | <karlcow> | After it is just a question of faith in techno or in society (Contrat Social). |
| 14:19 | <AryehGregor> | What legal framework is this, exactly? |
| 14:19 | <AryehGregor> | If it doesn't work without a legal framework, why was it implemented without a legal framework? |
| 14:20 | <karlcow> | because it is independent. Technology != Society rules. |
| 14:20 | <AryehGregor> | Why would any hypothetical legal framework not just restrict tracking directly instead of making it dependent on a switch that almost all users would set to true if they knew about it? |
| 14:20 | <tomasf> | maybe DNT was just a April fools joke, like the evil bit, but no-one realized? |
| 14:20 | <zewt> | annevk: i asked that guy for a code sample, and he basically said "no", heh |
| 14:20 | <jgraham> | Mozilla said they were against a legal framework |
| 14:20 | <AryehGregor> | Maybe it's simpler to postulate that it's a political move by companies that don't benefit much from tracking or ads to make competitors who depend on ad revenue look bad? |
| 14:21 | <jgraham> | And iirc it was Mitchell, so actually Mozilla, not just "someone at Mozilla" |
| 14:21 | <karlcow> | AryehGregor: nope |
| 14:21 | karlcow | is again in the weird position of explaining something which he is not advocating :) |
| 14:21 | <karlcow> | It's not a political move |
| 14:22 | <SamB_MacG5> | I don't think Mozilla has any say in whether or not a legal framework is actually created ;-P |
| 14:22 | <jgraham> | Ah, it was Gerv http://blog.gerv.net/2012/09/do-not-track-and-the-role-of-government/ |
| 14:23 | <karlcow> | He is an attempt of providing a way for users to show that you are against being tracked. So a big sign in a demo for example. Now for the fact of being tracked or not, it relies on local legal frameworks. |
| 14:23 | <jgraham> | He just linked to Mitchell's blog |
| 14:23 | <karlcow> | And it's where it becomes difficult. |
| 14:23 | <zewt> | AryehGregor: i don't think DNT makes the slightest bit of sense, but I doubt there are ulterior motives |
| 14:24 | <jgraham> | FWIW I imagine "a legal framework" might encompass "companies are not required to honour DNT, but if they say they will and don't they can be punished" |
| 14:24 | <SamB_MacG5> | though of course they could serve as experts in any government work on the subject... |
| 14:24 | <AryehGregor> | zewt, not on everyone's part, but on some supporters' part . . . |
| 14:24 | <zewt> | AryehGregor: perhaps, not sure that means much |
| 14:24 | <karlcow> | as a personal point of view, DNT being on or off, I will still used Ghostery and other Ads blockers. But that would not also protect of everything in life. ;) |
| 14:24 | <SamB_MacG5> | jgraham: for some values of "say", the framework exists already |
| 14:24 | <karlcow> | what jgraham said |
| 14:25 | <SamB_MacG5> | that certainly seems the simplest |
| 14:26 | <zewt> | jgraham: there are already mechanisms to do that (eg. logo testing via trademarks), though they're weak (don't think trademark infringement in the US gets you much more than an order to stop) |
| 14:26 | <karlcow> | samB, there is no "THE framework" |
| 14:26 | <SamB_MacG5> | well, okay, a multitude of twisty frameworks, all different |
| 14:26 | <karlcow> | there will be frameworkS. It is dependent on countries legal systems |
| 14:27 | <SamB_MacG5> | but I believe most legal systems allow for certain kinds of promises to be legally binding? |
| 14:27 | <karlcow> | SamB_MacG5: nope. :) |
| 14:27 | <AryehGregor> | A legal system that didn't allow that wouldn't work very well. |
| 14:28 | <AryehGregor> | The question is which types. :) |
| 14:28 | <karlcow> | or not in the sense of being tracked in the context of Internet |
| 14:32 | <AryehGregor> | annevk, why is .namespaceURI only defined on Element in DOM? Browsers seem to define it on all nodes but make it null except on elements. |
| 14:32 | <zewt> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=16304 wow, seriously? |
| 14:32 | <AryehGregor> | Is there any point in making === null checks stop working unnecessarily? |
| 14:34 | <zcorpan> | zewt: yeah. i actually started writing a comment on that bug explaining why he was being silly, but then i realized what i was about to do and closed the tab |
| 14:35 | <zewt> | zcorpan: yeah I think I won't be able to retrain myself heh |
| 15:02 | <annevk_> | AryehGregor: we wanted to slim down objects |
| 15:03 | <annevk_> | zewt: yeah I'm not sure why he got so upset |
| 15:13 | <annevk_> | AryehGregor: a bigger change is Attr no longer inheriting from Node |
| 15:14 | <annevk_> | AryehGregor: we're not quite sure whether any of those changes still stick long term, but Gecko is in the process of experimenting with them |
| 15:17 | <annevk_> | hmm, http://xhr.spec.whatwg.org/#data:-urls-and-http does not define what happens if the data: URL fails to parse |
| 15:27 | <annevk> | hmm the navigation timing stuff is kinda ill-defined |
| 15:32 | <annevk> | Hixie_: haven't made much progress with URLs this week |
| 15:32 | <annevk> | Hixie_: hopefully later, have been catching up on some Encoding/DOM stuff |
| 15:40 | <annevk> | Ms2ger? |
| 15:40 | <annevk> | why does new Event("x").isTrusted yield true in Gecko? |
| 15:43 | <annevk> | https://bugzilla.mozilla.org/show_bug.cgi?id=799982 |
| 15:43 | <annevk> | pretty close to 800k |
| 15:46 | <annevk> | Anyone here taking MediaWiki bug reports? http://wiki.whatwg.org/index.php?title=Special:UserLogin&type=signup has a bug in that it requires the password fields even if I hit "By e-mail" where it will randomly generate such a password |
| 16:03 | <AryehGregor> | annevk, for one thing, the version is 1.17alpha (1.17.0 being released in June 2011), while the latest stable is 1.19.0 (from April 2012). |
| 16:03 | <AryehGregor> | In 1.17alpha, $wgHtml5 was disabled by default, so probably things like this only got sufficient testing later. |
| 16:04 | <AryehGregor> | I think it's now enabled on Wikipedia, so probably stuff like that has been ironed out (assuming that feature is in use on Wikipedia). |
| 16:04 | <annevk> | ah okay, I wonder if someone can update wiki.whatwg.org then |
| 16:05 | <AryehGregor> | I was the last one to do it, but I don't think I'm going to spend my time on that. |
| 16:06 | <AryehGregor> | Well, certainly I won't spend *my* time on it. But I don't think it's worth my employer's time either. |
| 16:06 | <AryehGregor> | Especially since it might break as much as it fixes. |
| 16:06 | <AryehGregor> | MediaWiki is not known for its great quality control or stability. |
| 16:09 | <annevk> | :/ |
| 16:13 | <dglazkov> | good morning, Whatwg! |
| 16:13 | <tantek> | good morning dglazkov |
| 16:21 | <annevk> | arv: could you take a look at my suggestion here regarding DOMTokenList.enable(): https://www.w3.org/Bugs/Public/show_bug.cgi?id=18463#c1 (i.e. to give .toggle() enable as argument) |
| 16:22 | <arv> | annevk: That seems OK to me too |
| 16:23 | <annevk> | arv: thanks |
| 16:43 | <annevk> | that was surprisingly complicated |
| 17:26 | <annevk> | TabAtkins: re twitter: changing the terminology yet again, now between Level 3 and 4 |
| 17:27 | <TabAtkins> | annevk: I don't think he's still objecting, though he would be the one to say so. |
| 17:27 | <annevk> | TabAtkins: although Level 4 does mention "group of selectors" somewhere, but that seems like a bug |
| 17:27 | <annevk> | TabAtkins: nobody said objecting ;) |
| 17:27 | <TabAtkins> | annevk: Ah, yes, that should be "selector list". |
| 17:29 | <annevk> | and fwiw, I have been guilty of renaming certain things too terminology wise, though often with the DOM specs there was not really much terminology present to begin with |
| 17:29 | <TabAtkins> | I at least believe we're renaming toward consistency and better readability. |
| 17:30 | <TabAtkins> | A lot of our terminology was pretty bad in the old specs. :/ |
| 17:32 | <zcorpan> | annevk: wow, reading the toggle algorithm i have no idea what it's doing :-P |
| 17:33 | <annevk> | zcorpan: I had to walk through the 6 variants to make sure it made sense |
| 17:34 | <zcorpan> | annevk: is toggle('foo', false) not supposed to be different from toggle('foo') ? |
| 17:37 | <TabAtkins> | annevk: I don't understand how toggle('foo', true) is different from add('foo'). Will add() add duplicates? |
| 17:38 | <zcorpan> | TabAtkins: it's not supposed to be different |
| 17:38 | <TabAtkins> | Oh. In that case I don't understand what it's for. ;_; |
| 17:39 | <zcorpan> | http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Aug/0017.html |
| 17:39 | <hober> | it's for toggle('foo', nonLiteralBoolean) |
| 17:39 | <TabAtkins> | Ah, I see. add/remove based on a boolean, to avoid having to explicitly do a trinary to get the effect. |
| 17:39 | <hober> | right |
| 17:41 | <TabAtkins> | Hrm. I don't like naked booleans like that, but switching to a string enum removes much of the point. |
| 17:42 | <zcorpan> | {enable:bool} better? |
| 17:43 | <zcorpan> | or separate method after all? |
| 17:43 | <TabAtkins> | Dunno how best to name it. :/ 'enable' doesn't sound great. |
| 17:43 | <jwalden> | frobnicate |
| 17:44 | <annevk> | would be great if people read their www-dom bugmail and bikeshedded at the point of filing rather than at the point of fixing |
| 17:44 | <TabAtkins> | .set()? |
| 17:44 | <TabAtkins> | annevk: I'm not subscribed to www-dom. :/ |
| 17:45 | <zcorpan> | annevk: dude you fix things before i read the email :-) |
| 17:45 | <annevk> | zcorpan: how much behind are you? |
| 17:45 | <TabAtkins> | I'm not necessarily opposed to what you've done. I can't think of anything better, after all. |
| 17:46 | <zcorpan> | annevk: oh, sorry, i thought it was raised today |
| 17:47 | <annevk> | maybe you guys want to go through https://www.w3.org/Bugs/Public/buglist.cgi?product=WebAppsWG&component=DOM&resolution=--- before I fix more things :-) |
| 17:47 | <annevk> | all *.spec.whatwg.org that I edit have an "open bugs" list at the top I go through every now and then to fix things from |
| 17:49 | <TabAtkins> | Actually, I'm cool with what you've done. It's actually reasonably readable. You toggle it to true, or toggle it to false. Or just toggle it, if the second is omitted. |
| 17:49 | <TabAtkins> | Perhaps a different argument name would make it more understandable. Like 'forcedState'? |
| 17:50 | <annevk> | forcedPresence? |
| 17:50 | <annevk> | hmm no |
| 17:50 | <TabAtkins> | I just had no clue what "enable" was supposed to mean. |
| 17:50 | <annevk> | yeah that came from the original method name proposal |
| 17:51 | <zcorpan> | annevk: i just picked a random bug from the list and concluded that i agree with https://www.w3.org/Bugs/Public/show_bug.cgi?id=13971#c8 :-P |
| 17:52 | <annevk> | heh |
| 17:53 | <zcorpan> | annevk: how about just "force"? |
| 17:53 | <TabAtkins> | Yeah, "force" works for me. |
| 17:53 | <annevk> | mkay |
| 17:53 | <TabAtkins> | Or "forceTo". |
| 17:53 | <zcorpan> | ttyl |
| 17:56 | <annevk> | from now on I'll just generate UUIDs for argument names |
| 17:56 | <hober> | just name them "bikeshed" |
| 17:56 | <hober> | and wait for people to suggest alternatives |
| 17:57 | <jsbell> | lol |
| 17:57 | <annevk> | not a bad idea |
| 17:57 | <TabAtkins> | That's what we do in CSS. |
| 18:00 | <annevk> | jsbell: the decodeStream() thing works for me, but then you probably want some state so that invoking decode() no longer works after invoking that and maybe even vice versa? |
| 18:00 | <TabAtkins> | annevk: Where should I go looking for the actual doc? The github doesn't help very much. |
| 18:01 | <TabAtkins> | Ah, got it. |
| 18:02 | <TabAtkins> | annevk: I wouldn't care about what you named your arguments if you'd put in explanatory paragraphs for each function, rather than relying solely on people grokking the algorithm to know what's going on. ^_^ |
| 18:02 | <odinho> | I see I'm not alone in thinking [TreatUndefinedAs=Missing] is cool, and not having it is painful. |
| 18:02 | <jsbell> | annevk: yeah, unless decode() doesn't touch the state at all which would also be weird. I'm not rushing to change it. |
| 18:02 | <annevk> | TabAtkins: there are such paragraphs actually |
| 18:02 | <odinho> | So the "this is only for legacy" language in webidl seems a bit language puristic to me. |
| 18:03 | <odinho> | I'd rather have nice, expected behaviour when decorating functions. |
| 18:03 | <TabAtkins> | annevk: Ah yeah, up in the big green block. Sorry, easy to miss when you're looking at the actual algorithm. |
| 18:03 | <TabAtkins> | annevk: Mind if I suggest a rewording of the toggle() explanation? |
| 18:03 | <jsbell> | odinho: step 1: invent time machine... |
| 18:03 | <annevk> | TabAtkins: not at all |
| 18:04 | <Hixie_> | annevk: what's funny about that www-archive thread is that there's no reason they should be asking the question in the first place. It's typical of how they've operated, though -- ask a question over and over until the person you're asking stops arguing back, and then claim you've won. |
| 18:06 | <annevk> | you could also do what the PFWG did with my formal objection, simply forget about it |
| 18:06 | <odinho> | jsbell: I'm actually interested in what step 2 will be, even though I probably won't spend much time on any of the steps :P |
| 18:06 | <TabAtkins> | annevk: "If force is not provided, 'toggles' the token, removing it if it's present and adding it if it's not. If force is true, adds the token (same as add()). If force is false, removes the token (same as remove())." |
| 18:08 | <TabAtkins> | Separating the "force is not present" from the "force is present" cases makes it easier to read, because you don't have to reason through a tri-state boolean expressed via negations. |
| 18:08 | <annevk> | jsbell: kinda feel like it's going to be a hack either way, because in implementations utf-16 and utf-16le mean the same, but you want them to be different |
| 18:11 | <jsbell> | annevk: having "utf-16" as endian-agnostic is from ICU c/o suggestion by zewt. We could come up with a different way to signal that to the decoder constructor than special-casing the label. |
| 18:11 | <annevk> | TabAtkins: you're tabatkins on github? |
| 18:12 | <TabAtkins> | annevk: Yeah. Should I just pr? |
| 18:13 | <annevk> | no, it's for acknowledging you in the checkin, it's a thing I'm playing with |
| 18:13 | <TabAtkins> | kk |
| 18:13 | <jsbell> | But when dealing with web content, utf-16/utf-16be end up being agnostic anyway since the BOM wins. So perhaps we want to expose a flag which turns BOM handling on/off... |
| 18:13 | <jsbell> | (in the API) |
| 18:14 | <TabAtkins> | Pretty sure I'm the only tabatkins on the web. My biological dad only barely uses computers, and afaict never uses his full name like I do. |
| 18:14 | <TabAtkins> | And I think we're the only two "Tab Atkins"s in the world. |
| 18:14 | <annevk> | jsbell: yeah, utf-16/utf-16be is only for bomless fallback at the moment |
| 18:14 | <annevk> | jsbell: we could have ignorebom which defaults to false? |
| 18:15 | <annevk> | jsbell: if you set it to true we skip the step where it looks for the bom and does something with it |
| 18:15 | <jsbell> | annevk: let me dig up the threads and see if that was proposed/rejected... |
| 18:15 | <annevk> | jsbell: because technically valid utf-16le/utf-16be content is not allowed to contain a bom anyway so that would work |
| 18:16 | <annevk> | jsbell: unrelated, should I integrate this or do you want to be involved somehow? either way I'll make sure to acknowledge the fact you designed the thing |
| 18:18 | <jsbell> | annevk: go for it, but being involved would be swell if we can figure out the "somehow" |
| 18:19 | <annevk> | jsbell: well the whole thing is hosted here: https://github.com/whatwg/encoding do you have a github account? |
| 18:20 | <annevk> | jsbell: I can just add you to the WHATWG team and then you can commit to drafts (including the Encoding Standard) as you please |
| 18:20 | <annevk> | jsbell: once you commit it will be automatically synced with encoding.spec.whatwg.org |
| 18:21 | <jsbell> | annevk: I'm inexorabletash on github - thanks |
| 18:24 | <annevk> | I guess at some point we should actually have separate Owners and Teams but for now just Owners will do :) |
| 18:24 | <annevk> | jsbell: added |
| 18:25 | <jsbell> | swell |
| 18:25 | <annevk> | jsbell: http://wiki.whatwg.org/wiki/Anolis explains the setup we have btw |
| 18:26 | <annevk> | jsbell: if you don't want to bother with all that just change Overview.src.html and I'll generate the spec |
| 18:27 | <annevk> | jsbell: as for the API, I was thinking we could put it between chapters 6 and 7 |
| 18:29 | <annevk> | and that's it for me today I think :-) |
| 18:30 | <Hixie_> | annevk: you should just have the *.spec.whatwg.org domains do the regen for you :-) |
| 18:31 | <annevk> | Hixie_: I have been thinking about setting that up actually |
| 18:31 | <annevk> | there was something that was keeping me back, should look into that again |
| 18:32 | <annevk> | that would actually be ideal because it would remove yet another roadblock for people to contribute |
| 18:35 | <smaug____> | annevk: so how did you create the event? |
| 18:35 | <smaug____> | annevk: could you answer to the bug |
| 18:37 | <annevk> | smaug____: like comment 3 |
| 18:38 | <annevk> | smaug____: sorry, didn't think about the context |
| 18:38 | <smaug____> | yeah, it is odd |
| 18:38 | <smaug____> | but ok, no security bug then :) |
| 18:38 | <smaug____> | I got a bit worried |
| 18:39 | <annevk> | heh |
| 18:39 | <annevk> | oh cool |
| 18:39 | <annevk> | github can has twitter integration for commits |
| 18:40 | <annevk> | I guess I'll create a twitter account for each living standard and link them up |
| 19:12 | <annevk> | haha per http://gavinsharp.com/irc/whatwg.html among our most referenced URLs are the HTML namespace, example.com, and example.org |
| 19:24 | <Hixie_> | foolip: yt? |
| 19:30 | <Hixie_> | foolip: for https://www.w3.org/Bugs/Public/show_bug.cgi?id=17483#c3 - is picking the one with the least overlap sufficiently efficently implementable? |
| 19:34 | <annevk> | ah yeah, the problem with generating specs on spec.whatwg.org is the cross-references stuff |
| 19:34 | <annevk> | need to talk with Ms2ger if we can figure something out for that |
| 19:35 | <Hixie_> | what's the problem? |
| 19:36 | <annevk> | sometimes I need to add new entries for specs (manually) and sometimes the spec I generate adds entries (automatically) to cross-references database; to sync that database I then need to commit these changes |
| 19:36 | <annevk> | every now and then I also need to pull upstream changes in |
| 19:37 | <annevk> | we could probably make most of that work, but the make process fails if something fails in that process (e.g. an entry you reference does not exist) |
| 19:38 | <annevk> | but if Ms2ger and I can work that out somehow, we could make it real easy for people to edit WHATWG-style specs |
| 19:38 | <Hixie_> | ah, right, because it's all still manual |
| 19:38 | <Hixie_> | we should make it automatic or self-contained within the doc, that's what i've been waiting for to do it in html |
| 19:40 | <annevk> | it could still fail at that point, but I guess it could print error messages somewhere |
| 19:40 | <annevk> | also, your automatic process hasn't really been turned into a concrete plan yet |
| 19:41 | <Hixie_> | yeah :-) |
| 19:41 | <Hixie_> | things could fail in the internal cross-refs too |
| 19:43 | <Hixie_> | failure isn't limited to external cross-refs :-) |
| 20:10 | <Hixie_> | foolip: (least overlap with other cues, i mean; as opposed to what you proposed, which is just to consider the least amount that is outside the box) |
| 20:10 | <Hixie_> | foolip: for now i'll just look at the area outside the box, that's pretty easy to implement |
| 21:14 | <aklein> | can any spec authors about think of a spec that uses variable-length WebIDL arrays? |
| 21:14 | <aklein> | I'm curious to see what the WebKit implementation of such a spec looks like... |
| 21:19 | TabAtkins | doesnt' even know what a "WebIDL array" is anymore. |
| 21:19 | <aklein> | TabAtkins: http://dev.w3.org/2006/webapi/WebIDL/#idl-array |
| 21:20 | <aklein> | basically, IDL arrays are pass-by-reference (sorta), sequences are pass-by-value |
| 21:21 | <aklein> | not actually pass-by-reference because they're actually platform objects, so they get copied when assigned from JS |
| 21:23 | <aklein> | but yeah, people seem confused about them in general (cf http://lists.w3.org/Archives/Public/public-script-coord/2012JulSep/0023.html) |
| 21:24 | <TabAtkins> | aklein: Yeah, I don't understand their connection to JS arrays, though. I keep hearing things that sound contradictory, but that may just be Boris correcting miconceptions in other people's specs. |
| 21:57 | <Hixie_> | if anyone can think of anything to add to http://wiki.whatwg.org/wiki/FAQ#Where.27s_the_harm_in_adding.E2.80.94 please be my guest |
| 22:59 | <zewt> | reading bugzilla mails to lists gives me a headache |
| 23:00 | <zewt> | giganto table in every mail is a pain to read and using a fixed-width font for the content makes no sense |
| 23:08 | <zewt> | getting from google to specs is now one step harder :( |
| 23:09 | <zewt> | google for dom4, click first link (TR), click editor's draft, now there's an extra "specification moved" page--can't we get that to be a normal redirect? |
| 23:09 | <Hixie_> | why would you keep clicking the first link if it's not what you want |
| 23:10 | <Hixie_> | teach it that you want something else by clicking on the other thing and it'll become the new first link |
| 23:10 | <Hixie_> | (eventually) |
| 23:10 | <zewt> | because the correct result isn't even in the results |
| 23:10 | <Hixie_> | (though you'll have to first convince it that all the times clicking that bad link are really the past) |
| 23:10 | <Hixie_> | well you cal also just type "dom.spe" and have it autocomplete :-P |
| 23:10 | <Hixie_> | just "dom" is actually enough for my browser at this point |
| 23:12 | <zewt> | i long since gave up on trying to use browser history, every browser I use has made a complete mess of it |
| 23:12 | <zewt> | at least for quick navigation |
| 23:13 | <Hixie_> | time to try again |
| 23:13 | <zewt> | sorry, not changing my everyday usage habits because specs for the HTML spec don't even use redirects properly :) |
| 23:14 | <zewt> | (dom spec, rather) |
| 23:14 | <zewt> | usage habits have a lot of momentum; it takes more than one site to make me try to change them |
| 23:14 | <Hixie_> | i'm just trying to help you out here |
| 23:14 | <Hixie_> | you whined about something, i gave you two ways to make your life way easier, and now your whining about my suggestions |
| 23:15 | <Hixie_> | you're |
| 23:15 | <zewt> | i'm trying to help everyone out--there shouldn't be additional steps in the way of people not using ancient TR snapshots |
| 23:20 | <gavinc> | uh, speaking of which on DOM... based on the Status of Document for http://www.w3.org/TR/dom/ how the heck would I know not to use that? |
| 23:21 | <gavinc> | speaking as someone who made very sure to refer to the "correct" DOM document a few months ago |
| 23:21 | <Hixie_> | gavinc: you would know not to use anything that is in /TR/ |
| 23:24 | <gavinc> | right, since on May 17th annevk said to use it |
| 23:25 | gavinc | gives up |
| 23:25 | <zewt> | well, the /TR/ problem itself isn't specific to DOM or a new problem |
| 23:27 | <Hixie_> | it seems unlikely that anne said to use /TR/ anything |
| 23:27 | <Hixie_> | you sure he didn't mean one of the dev.w3.org servers? |
| 23:29 | <gavinc> | I think the final answer was link to that and wait till contextless parsing is defined and then update |
| 23:29 | <gavinc> | no idea where contextless parsing is |
| 23:32 | <Hixie_> | what's contextless parsing? |
| 23:34 | <gavinc> | What's needed to parse a fragment of HTML that isn't part of a document I'm guessing? |
| 23:34 | <gavinc> | since that's the use case |
| 23:36 | <TabAtkins> | Yeah, parsing that works when you dont' know ahead-of-time what type of document you're in. |
| 23:36 | <TabAtkins> | (Specifically, whether you're in HTML, SVG, or MathML.) |
| 23:37 | <TabAtkins> | The "make jQuery's $('<div>foo</div>') thing work". |
| 23:37 | <TabAtkins> | (jQuery's thing works for a lot of SVG as well.) |
| 23:37 | <TabAtkins> | (with regexes) |
| 23:39 | <Hixie_> | AryehGregor: yt? |
| 23:39 | <Hixie_> | gavinc: not sure what you're talking about or how it relates to the DOM spec |
| 23:41 | <gavinc> | http://www.w3.org/TR/rdf11-concepts/#section-html |
| 23:42 | <Hixie_> | oh, right, you're an rdf guy |
| 23:42 | <gavinc> | or if you prefer not TR :P http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html#section-html |
| 23:42 | <gavinc> | yes yes :P |
| 23:42 | <zewt> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=19414 i'm just going to go and call this guy a troll in here so I don't feel the need to do so in the bug |
| 23:42 | <Hixie_> | gavinc: then anne was probably trolling you when he said to look at /TR/ :-P |
| 23:42 | <gavinc> | geee, thanks |
| 23:43 | <Hixie_> | zewt: i doubt that's a troll |
| 23:44 | <Hixie_> | zewt: probably just confused by the api, it's not like our platform is simple |
| 23:44 | <zewt> | Hixie_: confused is fine, the "what, are you stupid?"-attitude not so much--but either way I'm keeping away from that ticket now :P |
| 23:46 | <Hixie_> | the "what, are you stupid?" attitude isn't trolling, it's just rude :-) |
| 23:46 | <Hixie_> | i should know, i have it regularly :-/ |
| 23:49 | <gavinc> | So currently we reference the HTML fragment parsing algorithm, and then normalize and isEqualNode is there a more correct way of comparing two HTML fragments? |
