| 00:44 | <erlehmann> | Workshiva, that looks good :) http://redbot.org/?uri=http%3A%2F%2Fgsoc2010.dieweltistgarnichtso.net%2Fwp-content-x%2Fplugins-x%2Fwordpress-cc-plugin%2Fembed-helper.php%3Fid%3D49 |
| 02:05 | <jamesr> | Hixie: entities support is inconsistent |
| 02:05 | <jamesr> | Hixie: frex WebKit supports % but IE and FFx do not |
| 02:06 | <Hixie> | once everyone implements the spec, problem solved |
| 02:06 | <Hixie> | (you are right that today it's a mess) |
| 02:07 | <jamesr> | yeah, but it's causing compat issues today |
| 02:07 | <jamesr> | for example if you save a document in chrome and then open it in IE it's broken |
| 02:07 | <jamesr> | so right now we exclude some entities when you save a page in chrome |
| 02:07 | <jamesr> | which is not really good either |
| 02:07 | <Hixie> | why wouls saving have any effect |
| 02:07 | <jamesr> | and if you load up data:text/html;charset=utf-8,<html>%0A<body>%0Aamp:%20%26amp;%0A<br>%0Agt:%20%26gt;%0A<br>%0Apercnt:%20%26percnt;%0A<br>%0Ansup:%20%26nsup;%0A<br>%0Asupl:%20%26supl;%0A</body>%0A</html>%0A it looks different |
| 02:08 | <Hixie> | the HTML serisalisation algorithm never serialises entities |
| 02:08 | <jamesr> | i'm not sure why the serialization code does that |
| 02:08 | <Hixie> | fix it :-) |
| 02:08 | <jamesr> | sure. but that page still looks different |
| 02:10 | <Hixie> | yes |
| 02:10 | <Hixie> | that's why we defined it once and for all |
| 02:12 | <jamesr> | so you took the superset of supported entities? |
| 02:14 | <Hixie> | no, i took whatever the mathml working group came up with -- my build script just slurps in their table and converts it to HTML each time i regen the spec |
| 02:14 | <Hixie> | but it is, i believe, a superset |
| 02:14 | <Hixie> | if there are any missing we can get them added |
| 09:16 | <annevk5> | AryehGregor, are you testing invalid values as well for your reflecting tests? |
| 09:18 | <annevk5> | AryehGregor, e.g. WebKit seems to just check if enctype contains "text" or "multipart" rather than check the full media type... |
| 09:18 | annevk5 | was playing around with form.encoding |
| 11:09 | <AryehGregor> | annevk5, I'm testing some invalid values, but it's hardly comprehensive. For enums that are limited only to known values, I test the two values "" and "Something that's hopefully not defined here". |
| 11:15 | <annevk5> | AryehGregor, if it's easy to add I suggest testing "xtext/plain" and "xtext" and "xmultipart" or some such |
| 11:16 | <annevk5> | AryehGregor, if you are building a test suite that is |
| 11:16 | <AryehGregor> | annevk5, I don't see any extra failures when I do that, on Chrome dev. |
| 11:17 | <annevk5> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cform%20enctype%3D%22xtext%22%3E%3C/form%3E%0A%0A%3Cscript%3E%0A%20w(document.forms[0].enctype)%0A%3C/script%3E |
| 11:17 | <AryehGregor> | Hmm. |
| 11:17 | <AryehGregor> | Yeah, I see it when I test manually. |
| 11:17 | <annevk5> | gives text/plain rather than application/x-... for me |
| 11:17 | AryehGregor | double-checks his text code |
| 11:19 | <AryehGregor> | s/text/test/ |
| 11:19 | <AryehGregor> | Oh, whoops. |
| 11:19 | <AryehGregor> | How silly of me. |
| 11:19 | <AryehGregor> | I was using .concat() instead of .push(). |
| 11:24 | <AryehGregor> | There we go. |
| 11:24 | <AryehGregor> | http://aryeh.name/tests/reflection.html |
| 11:26 | AryehGregor | thinks his tests might be wrong, though |
| 11:26 | AryehGregor | will look later |
| 11:26 | <annevk5> | sweet test suite |
| 11:27 | <micheil> | annevk5: did my _idea_ email actually reach the hybi mailing list? |
| 11:27 | <Philip`> | micheil: http://www.ietf.org/mail-archive/web/hybi/current/msg03269.html ? |
| 11:28 | <annevk5> | micheil, yes |
| 11:28 | <AryehGregor> | Yeah, something more is wrong with my enum tests. I'll look later. |
| 11:35 | <Philip`> | http://www.ietf.org/mail-archive/web/hybi/current/msg03233.html - it's good to know that security can be proven from fundamental theorems of computer science and the laws of thermodynamics |
| 11:40 | <annevk5> | in Opera <button>.type is limited to known values per your test |
| 11:40 | <annevk5> | in Firefox too |
| 11:40 | annevk5 | checks HTML5 |
| 11:41 | <annevk5> | seems your test is wrong |
| 11:43 | <AryehGregor> | Anything under the "enum" column is limited to only known values. |
| 11:43 | <AryehGregor> | Other enumerated attributes reflect as strings, so they're classified as strings. |
| 11:43 | <AryehGregor> | Yeah, there's something buggy there. |
| 11:43 | <AryehGregor> | I'll look at it later. |
| 11:43 | <AryehGregor> | I don't see offhand from the code how it could possibly be expecting anything not on the list of known values. |
| 11:45 | <annevk5> | it seems to be expecting the input value for some |
| 11:45 | <AryehGregor> | Yeah, I dunno why. |
| 11:45 | <AryehGregor> | Oh, wait. |
| 11:45 | <AryehGregor> | Dur. |
| 11:45 | <AryehGregor> | No. |
| 11:45 | <AryehGregor> | Hmm. |
| 11:47 | <annevk5> | unsigned long for <canvas> also seems wrong |
| 11:47 | <AryehGregor> | In what way? |
| 11:47 | <annevk5> | or maybe not |
| 11:47 | <annevk5> | nm |
| 11:48 | <AryehGregor> | Browsers are inconsistent on that one. |
| 11:48 | <AryehGregor> | But unsigned long seems to make the most sense. |
| 11:48 | <AryehGregor> | var tests = values; <-- Does this mean changes to tests will affect values too, in JavaScript? |
| 11:48 | <AryehGregor> | If so, that's my problem. |
| 11:49 | <AryehGregor> | (then how do I get around it?) |
| 11:49 | AryehGregor | doesn't know much JS |
| 11:49 | <AryehGregor> | Oh, that would explain the infinite loop too. |
| 11:49 | <AryehGregor> | var tests = values.slice(0)? |
| 11:50 | <annevk5> | what do you want? |
| 11:50 | <Philip`> | AryehGregor: "=" just copies the reference, it doesn't deep-copy the data structure |
| 11:51 | <AryehGregor> | I want an independent copy. |
| 11:51 | <AryehGregor> | .slice(0) seems to work. |
| 11:51 | <Philip`> | Like values[:] in Python? |
| 11:51 | <annevk5> | http://my.opera.com/GreyWyvern/blog/show.dml/1725165 is the first hit on Google that explains this :) |
| 11:51 | AryehGregor | found that, yeah |
| 11:52 | <AryehGregor> | Okay, but now this uncovers the fact that my enum tests are wrong in all sorts of other ways. |
| 11:52 | <AryehGregor> | I'll handle that later. |
| 11:53 | <annevk5> | heh |
| 11:54 | <AryehGregor> | I guess I need to handle invalid value defaults and missing value defaults. |
| 11:57 | <AryehGregor> | (as well as floats, token lists, and a whole bunch of other TODOs) |
| 11:58 | annevk5 | needs to fill his fridge |
| 16:01 | <kennyluck> | Mike Smith: "when you push the back button, it should go back to ... God Damn it!" |
| 16:01 | <kennyluck> | Mike Smith: "When you push this button in most browsers, it should do... oh shit!" |
| 18:18 | <gsnedders> | http://www.google.com/webhp?hl=xx-bork |
| 18:18 | <gsnedders> | (probably old) |
| 18:20 | <Philip`> | gsnedders: That's at least eight years old |
| 18:20 | <gsnedders> | Ah, old enough I missed it :) |
| 18:21 | <Philip`> | At least nine years, even |
| 18:21 | <Philip`> | judging by http://www.xent.com/pipermail/fork/2001-October/005126.html |
| 18:21 | <Rik`> | annevk5: your problems on w() and live dom viewer, arent' they related to the XSS auditor ? |
| 18:35 | <variable> | gsnedders, |
| 18:35 | <variable> | there is also xx-hacker |
| 18:39 | <gsnedders> | variable: That I have seen |
| 18:40 | <variable> | gsnedders, there is also piglattin - which is surprsingly not machine generated |
| 19:57 | <Hixie> | what adjective can we use to describe document.write in the cases where it implies a document.open()? |
| 19:58 | <Workshiva> | Destructive |
| 19:58 | <Hixie> | thanks |
| 19:58 | <Workshiva> | I seem to recall 'blow away the current document' being common, so maybe something inspired by that |
| 19:59 | <Hixie> | destructive is good |
| 20:00 | <cheeser> | makes me think of: xml is like violence; if it's not working, you're not using enough of it. |
| 21:18 | <annevk5> | Rik`, Safari has one of those? |
| 21:18 | <Rik`> | annevk5: yep, since Safari 5 I think |
| 21:18 | <annevk5> | oh god |
| 21:19 | <jgraham> | Chrome has long had problems with W() |
| 21:19 | <jgraham> | s/W/w/ |
| 21:19 | <jgraham> | I find if you write a constant string first it generally works |
| 21:19 | <Rik`> | so maybe that's this you're seeing, maybe that's note |
| 21:19 | <jgraham> | like w("a") |
| 21:19 | <Rik`> | -e |
| 21:19 | <AryehGregor> | The live DOM viewer totally failed to work for me on IE8. |
| 21:20 | <AryehGregor> | Because it reported reflected XSS. |
| 21:20 | <AryehGregor> | (which is accurate, obviously) |
| 21:20 | <jgraham> | Yeah Live DOM Viewer + IE === disaster |
| 21:20 | <gsnedders> | And IE9? |
| 21:20 | jgraham | finds an annoying Opera bug which causes the cursor to disappear |
| 21:21 | <gsnedders> | jgraham: file a bug, kthxbai :P |
| 21:21 | <jgraham> | gsnedders: I have a feeling it is known |
| 21:21 | <jgraham> | But I can check, obviously |
| 21:22 | <gsnedders> | Many bugs are known. That doesn't mean you shouldn't file another. |
| 21:23 | <annevk5> | did Hixie cave to the citing of number of bugs criticism? ;p |
| 21:23 | annevk5 | sees bugs being resolved |
| 21:23 | annevk5 | is now somewhat unhappy he e-mailed ;p |
| 21:24 | <annevk5> | though admittedly I also have some bugs filed that could do with an answer |
| 21:24 | <jgraham> | gsnedders: You appear to be encouraging the filing of duplicates, which is nonsensical |
| 21:24 | <jamesr> | refiling a known bug is great |
| 21:24 | <jamesr> | it means that it's not yet fixed |
| 21:24 | <AryehGregor> | gsnedders, IIRC, Live DOM Viewer works in IE9, maybe because it's pre-beta. |
| 21:24 | <jamesr> | which is always a question when i'm looking through our bug tracker at old bugs that are difficult for me to reproduce |
| 21:25 | <AryehGregor> | jamesr, surely you should just update the existing bug to say it's still present. |
| 21:25 | <gsnedders> | jgraham: It tends to be less effort for someone who knows a duplicate exists to find the dulicate than for you |
| 21:25 | <jamesr> | that's even better |
| 21:25 | <Rik`> | AryehGregor: live dom viewer used to work on IE8, when typing the code |
| 21:25 | <jamesr> | but most people don't want to try to find the correct bug |
| 21:25 | AryehGregor | agrees with gsnedders on that point |
| 21:25 | <gsnedders> | jamesr: Or simply it's hard to know what to search for in a lot of cases |
| 21:25 | <jamesr> | yeah |
| 21:26 | gsnedders | spends around 30s looking for a dupe before filing |
| 21:26 | <gsnedders> | The number of extra dupes I find myself by spending 60x that amount of time is neglible |
| 21:26 | AryehGregor | too, maybe a couple of minutes |
| 21:26 | <jgraham> | gsnedders: I think that depends on the bug tracker though. Some have better search than others |
| 21:27 | <gsnedders> | What annoys me is when I see a bug I _know_ is a duplicate but can't find the original |
| 21:27 | <AryehGregor> | Totally. |
| 21:27 | <annevk5> | Rik`, so I guess I would have to disable that then? |
| 21:27 | <annevk5> | that sounds annoying |
| 21:27 | <annevk5> | I guess I can just use Chrome |
| 21:29 | <jamesr> | also it really bothers me when i'm trying to fix a bug and someone has attempted to helpfully filter down some bug reporting data but left out some important bit of data |
| 21:29 | <jamesr> | it's pretty hard to predict what data will actually be useful for fixing a bug. i often don't even know what data i want until i've spent a few hours digging through the code |
| 21:30 | <jamesr> | so i'd rather have a raw dump of way more data than i need and pick out the bits i really need. it's not hard to ignore extraneous data, much harder to reconstruct missing data. /rant |
| 21:30 | <gsnedders> | jamesr: Deal with more immature code, then it's easier. :) |
| 21:31 | <AryehGregor> | The latest IEBlog post is kind of amusing. "We think Google's plugin to disable Analytics for your machine is such a great idea that we'll tell you about it if you didn't know, and also tell you an even better way to do it without installing yet more of the plugins that make our browser slow and crashy and that we obviously hate." |
| 21:34 | <Rik`> | AryehGregor: I find the fact that Microsoft explains how to disable a Google service much more interesting |
| 21:35 | <AryehGregor> | Yeah, that was part of the interesting bit. |
| 21:41 | <AryehGregor> | I don't get why people are so upset about Google's net neutrality thing. It's a proposal for a compromise that (if you favor net neutrality) is an unambiguous win over the current situation, and a proposal that goes too much further would evidently be unlikely to pass, so why is anyone viewing it as some kind of betrayal or reversal? It's just a compromise, no? |
| 21:41 | AryehGregor | thinks legislated net neutrality is probably a bad idea anyway, though, so . . . |
| 21:42 | <AryehGregor> | Oh, cool. Wikipedia says Vint Cerf is in favor of legislated net neutrality, and Bob Kahn is against it. Fight! |
| 21:42 | <Hixie> | is html5lib up to date? |
| 21:42 | <Hixie> | notwithstanding the changes i made today, i mean |
| 21:45 | <annevk5> | AryehGregor, if you go from net neutrality everywhere to net neutrality just for non-wireless it seems quite clear why some people are upset |
| 21:45 | <annevk5> | AryehGregor, especially if it seems to be just because of a business deal and not some realization that it is actually better |
| 21:47 | <AryehGregor> | annevk5, they didn't say they don't support net neutrality on wireless in principle. They explicitly stated that they want Congress to pass their proposal, so there's no reason to think they actually support it as an ideal solution or anything close, just an improvement. |
| 21:47 | <AryehGregor> | It's in the right direction, if you support net neutrality. |
| 21:48 | <annevk5> | they did not state |
| 21:48 | <annevk5> | that |
| 21:48 | <AryehGregor> | "Verizon and Google are pleased to discuss the principled compromise our companies have developed . . . a proposal that we hope will make a constructive contribution to the dialogue . . . Crafting a compromise proposal has not been an easy process, and we have certainly had our differences along the way. . . ." |
| 21:49 | <AryehGregor> | Clearly billed as a compromise proposal to get the telecoms on their side, not something they view as ideal. |
| 21:51 | <AryehGregor> | Which they clarified even more explicitly in the follow-up blog post. |
| 21:51 | <AryehGregor> | "But given political realities, this particular issue has been intractable in Washington for several years now. . . . We’re not saying this solution is perfect, but we believe that a proposal that locks in key enforceable protections for consumers is preferable to no protection at all." |
| 21:52 | <annevk5> | I read it |
| 21:52 | <annevk5> | it seems like a lot of handwaving to me |
| 21:52 | <annevk5> | nothing concrete |
| 21:52 | <annevk5> | it doesn't state what part they consider to be a compromise and in what way, etc. |
| 21:53 | <annevk5> | anyway, maybe NYT (or who wrote that again?) is right and America should just get more broadband competition |
| 21:53 | <AryehGregor> | Yes, obviously. Net neutrality isn't a big problem for me in practice, but customer service is atrocious. I have basically one real choice for broadband here. |
| 21:54 | <AryehGregor> | It took like an hour on the phone to get a new IP address assigned to my account. Seriously. |
| 21:54 | <AryehGregor> | No government regulation's going to improve that. And if you've got more competition, net neutrality is less of an issue too (if it's really an issue to start with). |
| 21:55 | <annevk5> | you might need some regulation to improve competition actually |
| 21:55 | <annevk5> | for instance in the Netherlands network operators are required to allow others to use their market (against some price) so they can be an internet provider without having to have their own network |
| 21:56 | <annevk5> | I think that goes for much of Europe |
| 21:58 | <AryehGregor> | Yeah, that's often true. Like requiring them to allow competitors to use their network at reasonable rates. In my bedroom, I often lose T-Mobile connectivity but have Verizon, so why the heck shouldn't I be able to just use whichever network has the strongest signal right here? |
| 21:58 | <AryehGregor> | Actually, it would make a lot of sense to have all the nearby towers that your phone can detect bid on who will route your call on a call-by-call basis, instead of having to choose in advance. |
| 21:59 | <AryehGregor> | Then you could have a small startup that could provide cell service in only a small area, so you'd use that where it's cheapest and switch to a big network like AT&T when you're elsewhere. |
| 22:00 | <AryehGregor> | Currently you have to set up a ridiculously huge network to get any customers. I'd think it would be easier for wired ISPs, though, since they just have to hook you up to a backbone . . . why aren't people setting up ISPs that initially only cover a few blocks or whatever? |
| 22:00 | AryehGregor | shrugs |
| 22:01 | <annevk5> | startup costs are prolly pretty big |
| 22:02 | <Rik`> | or you could have the network infrastructure built and run by a state company |
| 22:02 | <AryehGregor> | Well, big like ten million or ten billion? It's pretty easy to get a handful million dollars for a startup. That should be enough to set up a small datacenter and hook up some customers. |
| 22:03 | <AryehGregor> | Unless it's too hard to lay lines, which might be a problem. |
| 22:03 | <AryehGregor> | Rik`, unlikely to be a popular idea in the United States. :) |
| 22:03 | <jgraham> | Hixie: Should be |
| 22:03 | <AryehGregor> | (although state monopolies are often less horrible than private monopolies, which are what we sometimes have now) |
| 22:04 | AryehGregor | needs to monitor the ratio of off-topic bytes per minute here posted by him vs. everyone else, and scale output accordingly |
| 22:04 | <AryehGregor> | I'll go do some Wikimedia stuff now and then work on my test suite some more. |
| 22:06 | <jgraham> | Hixie: (python version of course) |
| 22:07 | <annevk5> | AryehGregor, a few customers does not a business make :) might be interesting to see how Google's experiment goes with providing fast internet (what speed was it again?) to a small town |
| 22:07 | <AryehGregor> | Well, you start with a few customers and grow. |
| 22:08 | <Rik`> | AryehGregor: oh, talking about wikimedia, is there an update on the mobile website to display the list of languages for an article ? |
| 22:08 | <AryehGregor> | I think they were laying fiber optics directly to the house, which means basically unlimited last-mile speed, you're limited by routers. |
| 22:08 | <Hixie> | jgraham: k. well, it no longer is, i just made a bunch of spec changes. :-) |
| 22:08 | <jgraham> | Hixie: BTW we should make the parser work with .innerHTML on non-HTML content |
| 22:08 | <jgraham> | Hixie: OK |
| 22:08 | <jgraham> | Hixie: I will likely fix tomorrow |
| 22:08 | <Hixie> | jgraham: the spec parser? |
| 22:08 | <jgraham> | Hixie: yeah |
| 22:08 | <AryehGregor> | Rik`, you mean you want the language list on the mobile version? I'm guessing that's just omitted, to save bytes. |
| 22:09 | <jgraham> | Then we can add the API as soon as we can convinve people it is a good idea :) |
| 22:09 | <jgraham> | *convince |
| 22:09 | <Rik`> | AryehGregor: in fact, I just want to see if there is an english or french version |
| 22:10 | <AryehGregor> | Rik`, you can use the full version, I guess. |
| 22:10 | <jgraham> | Because it seems bizzare that <div>.innerHTML = "<math><mo> ..." works but not <math>.innerHTML = "<mo> ..." |
| 22:11 | <annevk5> | yeah, I would expect us to implement <svg>.innerHTML at least at some point |
| 22:11 | <Rik`> | "to save bytes" :) |
| 22:11 | <Rik`> | I don't mind getting the list of languages after an XHR call |
| 22:14 | <AryehGregor> | Rik`, are you willing to write the code? :) |
| 22:20 | <Hixie> | jgraham: what happens today if you do that? Do we not reset the insertion mode appropriately? |
| 22:23 | <Rik`> | AryehGregor: I might yeah |
| 22:24 | <AryehGregor> | Rik`, unfortunately I'm not quite sure where to submit it. I have some recollection that the developer of the mobile site decided it'd be a great idea to write it in Ruby and put it in his own personal git repository rather than in PHP and using Wikimedia's SVN. |
| 22:26 | <Rik`> | AryehGregor: well, if you find a way to contribute, tell me |
| 22:34 | <jgraham> | Hixie: That is the first problem. The second is if you enter something like "</svg>" it all goes to hell |
| 22:35 | <jgraham> | So you need some special cases for the fragment mode |
| 22:38 | <Lachy> | that link relation registration discussion is getting irritating. Hixie, can we just remove the IANA registry crap from the spec, since it's obviously a complete failure, and its advocates clearly have no interest in actually addressing the needs of HTML |
| 22:38 | <seankoole> | What's Wg man |
| 22:39 | <seankoole> | everything wg'in over there |
| 22:40 | <Lachy> | so let's just stop pretending that the IANA registry applies to HTML, and create a better system that provides a simple registration form, some small moderation to prevent spam and the ability to get machine readable link relation data from it in JSON or plain text |
| 22:47 | <Hixie> | jgraham: ah, yeah |
| 22:48 | <Hixie> | jgraham: can we "v2" this issue for now? I'd love to get innerHTML actually interoperable before we start extending it... |
| 22:51 | <Hixie> | Lachy: done |
| 23:01 | <Hixie> | gsnedders: please see mike's comment in bug 10002 |
| 23:14 | <Hixie> | was the decision to use <figcaption> a poll decision? and if so, does anyone remember when it was made? I can't see it on the list of wg decisions, but the list seems out of date |
| 23:15 | <AryehGregor> | I remember it as resolved by consensus. |
| 23:16 | <AryehGregor> | Or, specifically, the people who objected deciding not to bother with a change proposal. |
| 23:19 | <hober> | Hixie: http://lists.w3.org/Archives/Public/public-html/2010Feb/0032.html |
| 23:20 | <hober> | more specifically, http://lists.w3.org/Archives/Public/public-html/2010Feb/0344.html |
| 23:23 | <Hixie> | thanks |
| 23:24 | <Hixie> | there, i resolved a bunch of bugs |
| 23:24 | <Hixie> | hopefully that will get people off my back for a short while and i can go back to dealing with whatwg e-mails |
| 23:26 | <AryehGregor> | Are whatwg e-mails generally more useful or important than bugs? |
| 23:28 | <Hixie> | they're older |
| 23:31 | <othermaciej> | what's the current oldest unanswered email? |
| 23:32 | <othermaciej> | also, thanks for resolving some bugs |
| 23:41 | <Hixie> | Oldest e-mail is from march 2007, but it's in a folder i've put on hold (datagrid). Oldest e-mail that isn't on hold is from March this year. |