| 00:08 | <AryehGregor> | It's uncanny how I've agreed with everything I saw Shelley Powers write in the last two days or so. |
| 00:08 | gsnedders | draws blade |
| 00:09 | <AryehGregor> | Do you think I should e-mail her and warn her that her Bugzilla account may have been hacked by a WHATWG partisan? |
| 00:09 | <boogyman> | lol |
| 00:14 | <annevk> | what she says is echoing what Hixie has been saying for a while |
| 00:15 | <annevk> | suspect #1 |
| 00:15 | <annevk> | (assuming it is about that date handling bug) |
| 00:15 | <AryehGregor> | That plus the bug on allowing cached JavaScript scripts to be shared across sites. |
| 00:16 | <AryehGregor> | (pointing out that the time to load widely-used JavaScript libraries is typically small compared to all the images and stuff) |
| 00:37 | <Hixie> | is the reflection, mask, etc, css stuff implemented without prefixes yet? |
| 00:37 | <Hixie> | or is it still experimental? |
| 00:39 | <paul_irish> | in webkit? they're all still prefixed. |
| 00:39 | <Hixie> | anywhere |
| 00:40 | <Hixie> | how about background-size, etc? |
| 00:41 | <oojacoboo> | most of the newer css3 styles are still prefixed AFAIK |
| 00:41 | <Hixie> | background-size is like 8 years old now |
| 00:41 | <Hixie> | hardly new :-) |
| 00:41 | <oojacoboo> | it is for the browser implementation |
| 00:41 | <oojacoboo> | you can't even rely on it's use |
| 00:41 | <Hixie> | if that's new, what's the html stuff? :_) |
| 00:42 | <oojacoboo> | html stuff? |
| 00:42 | <paul_irish> | background-size is unprefixed. mostly all of borders/backgrounds went unprefixed everywhere in the past 6mo |
| 00:43 | <Hixie> | oojacoboo: like onhashchange and appcache and so on |
| 00:43 | <Hixie> | paul_irish: k |
| 00:43 | <oojacoboo> | Hixie: vapor? |
| 00:43 | <Hixie> | oojacoboo: onhashchange is even in IE8, hardly vapour |
| 00:43 | <Hixie> | paul_irish: implemented widely? or just webkit? (sorry for the dumb questions, i'm so out of touch with what's deployed it's silly) |
| 00:44 | <oojacoboo> | im not familiar with it, so I can't say really |
| 00:46 | <paul_irish> | widely |
| 00:46 | <paul_irish> | Hixie: a few of these are covered here: http://caniuse.com/#cats=CSS |
| 00:46 | <paul_irish> | in other cases the bottom of the MDC page for each thing (like background-size) has compatibility info, typically |
| 00:46 | <oojacoboo> | paul_irish: nice link |
| 00:47 | <Hixie> | paul_irish: cool, thanks |
| 00:48 | <paul_irish> | or ask me about any particular ones.. in many cases i know offhand. |
| 00:48 | <paul_irish> | :) |
| 00:53 | <Hixie> | paul_irish: k :-) |
| 01:08 | <roc> | Hixie: no-one has ever submitted a proposed spec for Webkit's CSS reflections and masks |
| 01:08 | <Hixie> | ah |
| 01:09 | <Hixie> | is there any css-based way to apply a gradient mask to an image? |
| 01:09 | <oojacoboo> | by mask you mean overlay? |
| 01:09 | <Hixie> | i'm playing with doing something where an image fades out as the content starts |
| 01:09 | <Hixie> | i guess i can just edit the image |
| 01:09 | <Hixie> | like everyone else :-) |
| 01:09 | <roc> | Hixie: you can do it in SVG |
| 01:10 | <oojacoboo> | Hixie: you can use opacity: |
| 01:10 | <oojacoboo> | don't forget the prefixes for that though |
| 01:10 | <Hixie> | roc: you mean create an SVG that wraps the JPEG and applies the mask and then link that in using 'background'? I guess I could do that |
| 01:10 | <roc> | Hixie: so you could apply an SVG mask to an SVG image and that should work on all SVG-supporting browsers |
| 01:10 | <Hixie> | seems like an extra level of indirection |
| 01:11 | <roc> | oh, you want to apply the gradient mask to a CSS background image? |
| 01:11 | <roc> | and not the rest of the contents of the element? |
| 01:11 | <Hixie> | oojacoboo: opacity is a solid opacity all the way through, i want it to apply non-uniformly (as a gradient) |
| 01:11 | <Hixie> | roc: yeah this is just an image in css |
| 01:11 | <oojacoboo> | Hixie: use another container with background gradient |
| 01:12 | <roc> | Webkit's mask wouldn't do what you want then, that masks the whole element |
| 01:12 | <Hixie> | roc: ah |
| 01:12 | <Hixie> | i wish we had something like xbl already, except inline in css |
| 01:12 | <roc> | Gecko lets you apply SVG masks to any kind of element, which I think is the best way forward standards-wise, but is apparently still not what you want |
| 01:13 | <oojacoboo> | I would just use two containers, with 2 backgrounds |
| 01:13 | <Hixie> | maybe instead of changing xbl2 to be an html language i should have changed it to be a css language |
| 01:13 | <roc> | I don't see how that would work |
| 01:13 | <roc> | we don't want to introduce CSS syntax for DOM subtrees |
| 01:14 | <Hixie> | well we'd use the existing xml or html syntax |
| 01:14 | <Hixie> | but in the text/css file |
| 01:14 | <Hixie> | we really need svg in text/css too for similar reasons |
| 01:15 | <roc> | hmmmmmmm |
| 01:15 | <roc> | that could work |
| 01:15 | <roc> | in fact |
| 01:15 | <roc> | that would solve a lot of problems!!! |
| 01:16 | <roc> | kinda freakish though |
| 01:16 | <Hixie> | yeah |
| 01:16 | <roc> | a style sheet sort of becomes a document |
| 01:16 | <roc> | does the style sheet apply to itself? |
| 01:17 | <roc> | can it contain script? |
| 01:18 | <roc> | can the script set location.href? |
| 01:19 | <roc> | if we disable script and say that the stylesheet applies to its own document, it sounds good |
| 01:20 | <Hixie> | those are good questions. dunno what the best answers are. |
| 01:29 | <erlehmann> | im a right if i assume that no one bothered to implement ::outside as of yet? |
| 01:30 | <erlehmann> | (in CSS, that is) |
| 01:31 | <Hixie> | is box-sizing widely implemented? |
| 01:32 | <Hixie> | i guess not |
| 01:32 | <boogyman> | Hixie: it's easy enough to create a testcase :) |
| 01:33 | <Hixie> | i do, but i don't have any non-beta browsers to test on |
| 01:33 | <Hixie> | all my browsers are like nightly builds and stuff |
| 01:34 | <boogyman> | well that puts a damper on testcases :P |
| 01:35 | <boogyman> | give me a sec and ill test on stable webkit,gecko,trident rendering-engines |
| 01:35 | <Hixie> | nah, don't worry about it |
| 01:35 | <Hixie> | i found another solution :-) |
| 01:36 | <Hixie> | man i wish ::outside was available |
| 01:46 | <nessy> | only opera supports box-sizing faik |
| 01:49 | <oojacoboo> | nessy: -webkit-box-sizing |
| 01:49 | <erlehmann> | others have prefixes http://caniuse.com/css3-boxsizing |
| 01:49 | <oojacoboo> | right |
| 01:50 | <erlehmann> | IE can into box-sizing? |
| 01:50 | <erlehmann> | i am of confused. |
| 01:50 | <nessy> | I love caniuse :) |
| 01:50 | <oojacoboo> | erlehmann: didn't think IE supported it, that's cool |
| 01:51 | <oojacoboo> | although, I don't find box-sizing to be very useful |
| 01:51 | <erlehmann> | well, it's logical in a sense. |
| 01:51 | <erlehmann> | after all, they have both models already. |
| 01:51 | <erlehmann> | i do. |
| 01:51 | <oojacoboo> | maybe I haven't had a good use case |
| 01:54 | <erlehmann> | yesterday, i friends ex-boyfriend came by while we were sitting in a cafe. he was telling me about how he does websites, with “stylesheets, you know, i use the newest version, version 4 or 5”. |
| 01:55 | <erlehmann> | and i looked it up: apparently he really makes web sites. in photoshop. which is the reason why the text on one site is part of the background picture. |
| 01:55 | <erlehmann> | D: |
| 01:58 | <oojacoboo> | :/ |
| 01:59 | <oojacoboo> | erlehmann: I'm actually glad that these people exist in large numbers ;) |
| 02:00 | <oojacoboo> | anyone in here working with the @font-face for webkit or gecko? |
| 02:01 | <erlehmann> | oojacoboo, i am not amused by every single site that i can not use without images turned on on my phone. |
| 02:01 | <oojacoboo> | I wanted to mention how I lost a day's worth of work over some miserable experiences with the engines on windows |
| 02:01 | <oojacoboo> | erlehmann, you better just get over that |
| 02:02 | <erlehmann> | i do, i do. |
| 02:04 | <erlehmann> | oojacoboo, it is just that before that experience i thought it would be hilarious to meet a cliché bad web worker. afterwards, it is more like “there is no hope for humanity” |
| 02:05 | <oojacoboo> | erlehmann: these people exist in every facet of life |
| 02:06 | <oojacoboo> | before long you will get the "elitist" stereotype though |
| 02:06 | <oojacoboo> | so, you are damned if you do and damned if you don't |
| 02:08 | <erlehmann> | no, i'll be fine, :) |
| 02:10 | <erlehmann> | it was just … weird, in a way. in my usual social context, people with at-best-mediocre skillsets don't boast about it in an annoying way. |
| 02:10 | <erlehmann> | there should be a “there, i fixed it” for web sites |
| 02:12 | <oojacoboo> | google should make use of the author meta data in their algo :) |
| 02:12 | <erlehmann> | harr harr |
| 02:13 | <erlehmann> | google should use the outline algorithm and only consider web sites whose section headings suggest a good plot! |
| 02:14 | <oojacoboo> | maybe, except in my case there isn't a plot |
| 02:25 | <Hixie> | how about rem units, anyone know what their status is in implementations? |
| 02:25 | <Hixie> | i don't see it on whencaniuse |
| 02:26 | <erlehmann> | i tried to use them in gecko and failed miserably. maybe your nightlies have more luck with it. |
| 02:31 | <roc> | rem works in Gecko |
| 02:31 | <oojacoboo> | roc: what version? |
| 02:31 | <roc> | trunk |
| 02:31 | <roc> | data:text/html,<div style="font-size:300px; height:1rem; width:1rem; background:blue;"> |
| 02:32 | <roc> | I think it worked in 3.6 as well |
| 02:32 | <oojacoboo> | hardly reliable though |
| 02:32 | <jamesr_> | what's a rem? |
| 02:32 | <oojacoboo> | root em |
| 02:32 | <roc> | ems for the root element font size |
| 02:32 | <roc> | although I still think "rem" should be a unit of radiaton |
| 02:33 | <oojacoboo> | or some damn good sleep |
| 02:33 | <jamesr_> | for styling vault dwellers? |
| 02:34 | <roc> | .hulk { gamma-radiation:1000rem; } |
| 02:35 | <oojacoboo> | ha |
| 02:36 | <roc> | mmmm, this is a much better unit: http://en.wikipedia.org/wiki/Banana_equivalent_dose |
| 02:37 | <oojacoboo> | ha, wikipedia is now using letters from authors for donations |
| 02:38 | <roc> | yeah, FF3.6 supported rem |
| 02:38 | <erlehmann> | hehe, maybe the current spec should have “a message from HTML5 editor ian hickson” |
| 02:39 | <Hixie> | heh |
| 02:39 | <Hixie> | ok, i've updated my blog and other sites to a new even-worse-than-before style sheet. |
| 02:39 | <Hixie> | i'll fix the various problems the update has caused in the next few weeks |
| 02:40 | <erlehmann> | i am doing the same thing right now. a:visited:after { content: […] } has to go, unfortunately |
| 02:41 | <erlehmann> | oh, and the fix is the last thing mentioned on http://ln.hixie.ch/ |
| 02:41 | <oojacoboo> | erlehmann: I thought multiple psedu selectors threw IE7 for a loop |
| 02:41 | <oojacoboo> | pseudo* |
| 02:41 | <Hixie> | (one thing i expect broke is the "axiomatic proof" post) |
| 02:42 | <erlehmann> | oojacoboo, i cannot possibly test that. |
| 02:42 | <oojacoboo> | ha |
| 02:43 | <oojacoboo> | unfortunately that's something I have to deal with |
| 02:43 | <erlehmann> | Hixie, what is with the picture of the train on the bottom? it makes some text hard to read. |
| 02:43 | <Hixie> | it shouldn't touch any of the text |
| 02:43 | <Hixie> | if it's touching the text there's either a bug in your browser or in my style sheet |
| 02:44 | <oojacoboo> | he means the depth of field |
| 02:45 | <erlehmann> | no, in gecko 20100916 the train is HUGE |
| 02:46 | <erlehmann> | in chromium 6.0.472.62 all is well |
| 02:48 | <jamesr_> | chromium _6_? |
| 02:48 | <jamesr_> | wtf |
| 02:48 | <jamesr_> | you are fully two stable versions behind |
| 02:48 | <erlehmann> | also, the image is sent as Content-Type: text/css |
| 02:49 | <erlehmann> | with Content-Language: en-GB-x-Hixie. the funney, it burns. |
| 02:49 | <erlehmann> | jamesr_, you mean, like 6 weeks or something ;) |
| 02:49 | <jamesr_> | erlehmann: wheels of progress man |
| 02:50 | <jamesr_> | but seriously you should update. 2 major versions = security disclosures have been made public |
| 02:50 | <erlehmann> | oh well. |
| 02:50 | <erlehmann> | consequences will never be the same. |
| 02:50 | <oojacoboo> | erlehmann: I'd stay on it for debugging purposes |
| 02:50 | <oojacoboo> | I am on 7, but I need a copy of 6 myself |
| 02:50 | <jamesr_> | but there are no users on 6 |
| 02:50 | <jamesr_> | like literally _none_ |
| 02:50 | <jamesr_> | so why do you need to debug on 6? |
| 02:51 | <oojacoboo> | jamesr_: I beg to differ |
| 02:51 | oojacoboo | opens analytics |
| 02:51 | <jamesr_> | what % of your chrome visitors in the past week are on 6? |
| 02:51 | <erlehmann> | i installed it some time ago so i have a reference webkit browser that does not suck |
| 02:52 | <oojacoboo> | jamesr_: 3.5% |
| 02:52 | <oojacoboo> | of chrome users |
| 02:52 | <erlehmann> | ;_; |
| 02:53 | <jamesr_> | is that enough to care about? |
| 02:53 | <oojacoboo> | that's 3.5% of 24% |
| 02:53 | <oojacoboo> | so, pretty small |
| 02:53 | <erlehmann> | newest chromium in debian sid/squeeze is 6.0.472.63 |
| 02:53 | <oojacoboo> | does look like 7+ is most users |
| 02:53 | <erlehmann> | maybe the maintainer sleeps |
| 02:54 | <jamesr_> | those are either people who haven't restarted chrome in the past 6 weeks (and haven't had it crash), or that are not on official Google Chrome builds, or who have some crazy bug |
| 02:54 | <oojacoboo> | except for the couple Chromium 4 visits... :/ |
| 02:54 | <jamesr_> | i'd be amazed if debian sid users were 3.5% of chrome users |
| 02:54 | <jamesr_> | on any website :) |
| 02:54 | <oojacoboo> | jamesr_: most likely mac/win |
| 02:55 | <oojacoboo> | mac is lagging a version behind, right? |
| 02:55 | <erlehmann> | wat. they are doing a new version every 14 weeks? |
| 02:55 | <erlehmann> | U MAD. |
| 02:55 | <oojacoboo> | I wish UAs would start building in auto updaters |
| 02:55 | <oojacoboo> | relying on users to update is a fucking joke |
| 02:56 | <jamesr_> | oojacoboo: all OSes update at the same rate. as of today we're pushing 8 to people |
| 02:56 | <erlehmann> | i wish UAs would do exactly the thing google chrome does. use the packaging system. |
| 02:56 | <oojacoboo> | jamesr_: yea, just saw that 8 update for myself after searching for it |
| 03:01 | <oojacoboo> | what's the motive behind the insane release schedule for chromium? |
| 03:03 | <erlehmann> | oojacoboo, that was the goedel sentence that made him quit. |
| 03:04 | <oojacoboo> | goedel? |
| 03:04 | <erlehmann> | http://c2.com/cgi/wiki?HowGoedelSentencesWork |
| 03:05 | <erlehmann> | >The idea behind Goedel sentences is that they say something like "This statement cannot be proved by formal system such and so". |
| 03:05 | <jacobolus> | goedel = gödel, he of the famous incompleteness theorem. presumably erlehmann is talking about self reference |
| 03:05 | <erlehmann> | okay, so chrome gets like 4 version number bumbs in a year |
| 03:05 | <jacobolus> | self reference + contradiciton |
| 03:05 | <jacobolus> | *contradiction |
| 03:06 | <oojacoboo> | am I missing something here? |
| 03:07 | <erlehmann> | so in 2020, we will all be happily surf with chrome 48 |
| 03:08 | <oojacoboo> | yea, well, I don't know that that's going to be the case, but I figured there might be an internal motive behind it |
| 03:08 | <oojacoboo> | something like, people will adopt something that's version 10 faster than version 5 |
| 03:08 | <oojacoboo> | or, if we update it a lot, people will update faster, pushing web standards and adoption forward |
| 03:08 | <erlehmann> | or they switch to an idiosyncratic numbering scheme, where they re-use version numbers. |
| 03:09 | <erlehmann> | like no one uses chrome 1 anymore. so why have chrome 10? recycle the version number! |
| 03:10 | <oojacoboo> | or screw versions that are public knowledge |
| 03:10 | <oojacoboo> | and just auto-update it in the background |
| 04:24 | <aboodman> | oojacoboo: that is exactly what chrome does |
| 04:25 | <oojacoboo> | aboodman: not my chrome |
| 04:26 | <aboodman> | what i mean is, chrome updates in the background, and Google doesn't refer to version numbers publicly |
| 04:26 | <aboodman> | we just say 'google chrome now does xyz' |
| 04:26 | <aboodman> | in our user-facing pr |
| 04:26 | <oojacoboo> | ah ok, gotcha |
| 04:27 | <oojacoboo> | glad to hear thats the case |
| 04:38 | <oojacoboo> | is it being recommended that you don't style the html5 semantic elements? |
| 04:38 | <oojacoboo> | or purely not using them for styling purposes |
| 04:39 | <oojacoboo> | in other words, styling them when they are being used properly, is still the intention, correct? Not just using them as semantic wrappers and injecting more elements into the DOM to handle styling purposes |
| 04:39 | <oojacoboo> | for instance, background, border, margin, padding, etc |
| 06:32 | <Hixie> | anyone know what happened with http://lists.w3.org/Archives/Public/public-html/2009Jan/0308.html ? |
| 06:43 | <mhausenblas> | Hixie you surely read http://www.w3.org/QA/2010/06/thanks_for_a_great_15_years_at.html no? |
| 06:44 | <Hixie> | i had not |
| 06:44 | <mhausenblas> | ;) |
| 06:44 | <Hixie> | what happened to the stuff he'd done in the year and a half before that? |
| 06:44 | <Hixie> | the first e-mail only talks about "6 months to a year" |
| 06:45 | <mhausenblas> | tbh, dunno |
| 06:45 | <mhausenblas> | best you ask him directly ;) |
| 06:45 | <Hixie> | well if he's left the standards world altogether he presumably doesn't care about that anymore :-) |
| 06:45 | <Hixie> | can't blame him |
| 06:46 | mhausenblas | nods |
| 06:46 | webr3 | notes he is around and giving valuable feedback on many standardization efforts often |
| 06:48 | <mhausenblas> | true, for example http://www.w3.org/2001/sw/rdb2rdf/track/actions/78 |
| 06:49 | <hsivonen> | http://google-chrome-browser.com/version-number-dead-google-barely-whispers-launch-chrome-8 So will Chrome now actaully get rid of the version number like WHATWG post-5 HTML? |
| 06:52 | <webr3> | nah because people need milestones and something to lookforward to, to countdown to, to remember and refer to as the previous thing/state |
| 06:53 | <webr3> | tis why we have time, years, birthdays, anniversaries etc - without them you're just all over the place, especially after a few years :) |
| 06:53 | <mhausenblas> | yey! looking forward to webr4 :D |
| 06:54 | <webr3> | lol that comes afer revision 3 is done - (r3..) |
| 11:13 | <annevk> | http://heideri.ch/jso/ really gives the wrong kind of advice |
| 11:14 | <annevk> | "Don't allow users to submit markup containing "form" and "formaction" attributes or transform them to bogus attributes." -- you really want to not allow any untrusted attributes otherwise this is going to happen again in the future |
| 11:19 | <zcorpan> | Hixie: your site is about as ugly as before :P |
| 11:24 | <annevk> | aah, styling in Opera does not work because of unrecognized elements |
| 11:24 | <annevk> | doh |
| 11:28 | <annevk> | I think the only HTML5 element I use is <footer> |
| 11:28 | <annevk> | well, only new element |
| 11:35 | <charlvn> | what's that thing in the top-right corner? a traffic light? |
| 11:40 | <hsivonen> | looks like a railway traffic light |
| 11:40 | <hsivonen> | (which probably has a fancier term of art for it) |
| 11:41 | <jgraham> | I thought they were called "signals" on railways |
| 11:42 | <jgraham> | But IANATS |
| 11:42 | <zcorpan> | I am not a train station? |
| 11:42 | <jgraham> | Close |
| 11:43 | <jgraham> | (I guess Train Spotter is really a single word, but that is harder to guess) |
| 11:45 | <zcorpan> | never heard of 'trainspotter' |
| 11:46 | <jgraham> | Have you heard of the film "trainspotting"? It has very little to do with actual trainspotting, but nevertheless |
| 11:47 | <jgraham> | hsivonen: Are you aware of any spec for the case shifting in XPath-in-text/html? |
| 11:54 | jgraham | wonders why //some-prefix-that-resolves-to-null:div throws rather than just returning nothing |
| 11:56 | <jgraham> | Oh well, I guess I can deal with that |
| 12:08 | <hsivonen> | jgraham: AFAIK, there's no spec, Gecko does different things depending on API entry point and my opinion of what to do differs from the module owner's |
| 12:09 | <hsivonen> | my opinion of what we should do that is |
| 12:10 | <jgraham> | Yay! |
| 12:22 | <jgraham> | hsivonen: What different API entrance points are there? document.evaluate, obviously |
| 12:23 | <hsivonen> | I can't recall. Let's see if Awesomebar tells me. |
| 12:25 | <hsivonen> | jgraham: (new XPathEvaluator).evalute |
| 12:25 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=376740 is the bug |
| 12:32 | <jgraham> | hsivonen: I think everything should work like document.evaluate |
| 12:34 | <annevk> | why can XPath not work like Selectors? |
| 12:36 | <jgraham> | In what way? |
| 12:37 | <annevk> | case-sensitive except for elements in the HTML namespace |
| 12:37 | <jgraham> | That is how document.evaluate works afaict |
| 12:37 | <jgraham> | and attributes |
| 12:37 | <jgraham> | that is, attributes on elements in the HTML namespace are not case sensitive |
| 12:39 | <annevk> | not per HTML5 |
| 12:39 | <jgraham> | Yeah, HTML5 is wrong |
| 12:39 | <jgraham> | It is a web compat issue |
| 12:44 | <hsivonen> | annevk: AFAICT, making XPath work like Selectors is not strictly needed for compat but would complicate the XPath matcher compared to doing case-sensitive matching and requiring names that are meant to match HTML local names to be in the lower case in the expressions |
| 12:44 | <hsivonen> | jgraham: AFAICT, that's not how document.evaluate works in WebKit or Gecko |
| 12:44 | <hsivonen> | IIRC |
| 12:45 | <hsivonen> | jgraham: how can it be a Web compat issue when WebKit doesn't case fold? |
| 12:45 | <jgraham> | hsivonen: Which part? It certainly seems to be case-insensitive for elements in the HTML namespace and their attributes |
| 12:46 | <jgraham> | hsivonen: Dojo has a special webkit codepath |
| 12:46 | <hsivonen> | jgraham: I see. special code paths are sadness |
| 12:46 | <hsivonen> | jgraham: oh well. If Gecko can't become like WebKit without breaking Dojo, I guess we should do the same thing Selectors do |
| 12:46 | <hsivonen> | :-( |
| 12:47 | <jgraham> | This is brining consistency to the platform :) |
| 12:47 | <jgraham> | *bringing |
| 12:47 | <hsivonen> | jgraham: is the special code path behavior sniffed or UA sniffed? |
| 12:47 | <jgraham> | I mean it's a weird, screwed up, consistency, but it is consistent in being so |
| 12:47 | hsivonen | guesses the latter |
| 12:48 | <jgraham> | Seems to be UA string based |
| 12:48 | <hsivonen> | :-( |
| 12:49 | <hsivonen> | given the new data, I change my opinion |
| 12:49 | <jgraham> | (at least, I think this is what is going on) |
| 13:55 | <jgraham> | Hmm, gecko seems to always match attributes in XPath case insensitively using document.evaluate |
| 14:08 | <jgraham> | Actually it just seems to be broken: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/728 |
| 14:08 | <jgraham> | hsivonen: ^ |
| 14:46 | <smaug____> | jgraham: I vaguely remember that that is done on purpose |
| 14:46 | <smaug____> | jgraham: sicking would know for sure |
| 14:47 | <smaug____> | let me try to find the relevant code |
| 14:48 | <smaug____> | jgraham: hmm |
| 14:48 | <smaug____> | what should I see in the log? |
| 14:49 | <jgraham> | I would expect 1 |
| 14:49 | <jgraham> | Since it should match one element with a refX attribute |
| 14:50 | <smaug____> | but the result is 0 |
| 14:50 | <smaug____> | so it matches case sensitively |
| 14:51 | <smaug____> | jgraham: or am I missing something. You said "case insensitively" |
| 14:51 | <jgraham> | The attribute gets case shifted to "refX" by the HTML parser |
| 14:52 | <smaug____> | ah |
| 14:52 | <smaug____> | right |
| 14:57 | <hsivonen> | jgraham: I was aware that XPath behavior with document.evaluate in Gecko was just broken, but I had no idea it was broken in that particular way |
| 14:58 | <hsivonen> | smaug____: the code may have been written on purpose, but SVG-in-text/html has changed the scenarios the code needs to respond to |
| 14:58 | <hsivonen> | smaug____: that is, IIRC, we don't keep around dual atoms in XPath expressions yet |
| 14:58 | <hsivonen> | we should |
| 15:00 | <hsivonen> | jgraham: IIRC, what happens is that the XPath expression compiler lowercases all name expressions when invoked on an HTML document |
| 15:01 | <jgraham> | hsivonen: Any idea what it does to non-ascii |
| 15:01 | <jgraham> | ? |
| 15:01 | <jgraham> | I can't make that work at all |
| 15:02 | <hsivonen> | jgraham: I'm not reading the code here, just dumping stuff from memory |
| 15:02 | <hsivonen> | jgraham: So I don't know how non-ascii behaves |
| 15:02 | <jgraham> | hsivonen: Sure, I was just wondering if you had any idea |
| 15:02 | <hsivonen> | jgraham: anyway, I'd advice against cloning/speccing what Gecko does now |
| 15:03 | <jgraham> | My plan was to aim for something that made sense, since clearly the current gecko behaviour is buggy |
| 15:03 | <jgraham> | That is ascii-lowercase HTML elements names and their attribute names in HTML documents |
| 15:04 | <hsivonen> | jgraham: given what you told me about Dojo, I think we should do https://bugzilla.mozilla.org/show_bug.cgi?id=499655 for XPath, too. |
| 15:05 | <jgraham> | hsivonen: I think that is equivalent to what I was thinking |
| 15:47 | <hsivonen> | Wow. the section titled An Authoritative Voice in http://www.webdesignerdepot.com/2010/12/book-review-%E2%80%94-html5-up-and-running/ is odd. |
| 15:48 | jgraham | has no idea who one of the people more famous than Mark is |
| 15:56 | <smaug____> | hsivonen: how should the case sensitiveness work in xpath |
| 15:56 | <smaug____> | is it specified anywhere? |
| 15:58 | <jgraham> | smaug____: (iinh, obviously) it isn't specified anywhere. Or rather it is specified to be case sensitive always, but that is not web compatible unless you are WebKit |
| 15:58 | <smaug____> | k |
| 15:59 | <hsivonen> | smaug____: I think it should work like it works in CSS in Gecko today |
| 16:00 | <hsivonen> | smaug____: I.e. each name expression would have two atoms: original case and ASCII-lower-cased |
| 16:00 | <hsivonen> | smaug____: if the experssion is being compared against a HTML element in an HTML document or an attribute on an HTML element in an HTML document, the lower-case atom would be compared |
| 16:00 | <hsivonen> | otherwise, the original case atom would be compared |
| 16:01 | <hsivonen> | smaug____: and as jgraham said, this hasn't been specced yet. |
| 16:06 | <smaug____> | makes sense to make it work like css |
| 16:15 | <MrWax> | Hi, I am looking for a presentation of HTML5, but a small more summarized one, like 20-30 mins.. does anyone a proper explanation / overview of HTMl5 |
| 16:15 | <MrWax> | ? |
| 16:17 | <hsivonen> | MrWax: I have presentation that's 2 years old and takes thrice as long as you wanted: http://hsivonen.iki.fi/html5-lecture/ |
| 17:05 | <erlehmann> | interesting. i had thought history sniffing was an academic risc as of now http://scienceblogs.com/pharyngula/2010/12/another_reason_to_avoid_visiti.php |
| 17:06 | <erlehmann> | dbaron saved us from the creationists! |
| 17:06 | <TabAtkins> | Yay dbaron! |
| 17:08 | <erlehmann> | TabAtkins, on the downsite, he killed off the nice check marks i was putting after each visited link (._.) |
| 17:09 | <TabAtkins> | Huh? You can't just do that with ::before? |
| 17:12 | <erlehmann> | TabAtkins, i was of the impression that every selector that includes :visited is now limited to color changes regarding text, background and border. |
| 17:12 | <TabAtkins> | Oh, hm, maybe. |
| 17:13 | <oojacoboo> | erlehmann: limited how? |
| 17:14 | <erlehmann> | wait, i'll look it up. |
| 17:15 | <erlehmann> | http://blog.mozilla.com/security/2010/03/31/plugging-the-css-history-leak/ |
| 17:16 | <erlehmann> | oojacoboo, >First of all, we’re limiting what types of styling can be done to visited links to differentiate them from unvisited links. Visited links can only be different in color: foreground, background, outline, border, SVG stroke and fill colors. |
| 17:17 | <erlehmann> | using :before or :after would be “changing position or size of the styled content in the document” |
| 17:17 | <erlehmann> | so if it is not already limited, it should certainly be. |
| 17:30 | <oojacoboo> | erlehmann: I understand the point of this, but it seems a bit draconian |
| 17:31 | <TabAtkins> | Unfortunately, you must either be draconian or not do anything at all. |
| 17:31 | <erlehmann> | oojacoboo, it *seems* but there really is no other easy way. |
| 17:40 | <oojacoboo> | :/ |
| 17:47 | <erlehmann> | oojacoboo, there are more complicated things. see same-origin-restrictions — i even implemented a small web server for a wordpress plugin that shows an embed-button for HTML5 media elements to get the headers right on every possible setup. |
| 17:47 | <erlehmann> | things are complicated and ugly :( |
| 18:21 | <erlehmann> | dbaron, you saved us from the creationists! http://scienceblogs.com/pharyngula/2010/12/another_reason_to_avoid_visiti.php |