| 00:00 | <Lachy> | and figcaption is better than fcaption |
| 00:00 | <othermaciej> | I can see an argument for both "label" and "summary" |
| 00:01 | <TabAtkins> | I like the way <summary> quite directly suggests what it should be, for maximum accessibility. |
| 00:01 | <Dashiva> | But it's not necessarily a summary, is it? |
| 00:01 | <TabAtkins> | Though it is clickable like a <label>, <summary> just matches better in my head. |
| 00:01 | <TabAtkins> | It's pretty close. |
| 00:02 | <TabAtkins> | Or rather, *should* be pretty close if <details> is used correctly. |
| 00:02 | <TabAtkins> | Note: I consider heading-like content to be close to a summary. |
| 00:02 | <Lachy> | I think summary sort of works as an alternative to figcaption, though its name would clash semantically with the table summary attribute, which is certainly not a caption |
| 00:03 | <Lachy> | but I don't see how summary works for details |
| 00:03 | <TabAtkins> | Lachy: Check http://www.whatwg.org/specs/web-apps/current-work/multipage/interactive-elements.html#the-details-element |
| 00:04 | <TabAtkins> | Note how the prose uses the term "summary" to describe the toggler bit. |
| 00:04 | <Lachy> | yeah, I saw that. I'm not sure that's a good description for it |
| 00:05 | <Hixie> | http://www.w3.org/mid/1263225612.26294.2.camel@chacal is interesting |
| 00:05 | <othermaciej> | in the spec example pictures (from the Mac OS X Finder), it's probably more accurate to call the short bits a "label" rather than a "summary" |
| 00:05 | <othermaciej> | Hixie: what is interesting about it? |
| 00:05 | <othermaciej> | Hixie: I think what Philippe is saying is: |
| 00:06 | <Hixie> | that the new rdfa group _doesn't_ include the text/html rdfa stuff |
| 00:06 | <othermaciej> | ah |
| 00:06 | <Lachy> | othermaciej, what are those controls called in OS X? |
| 00:06 | <othermaciej> | I was expecting that "RDFa in HTML" would stay an HTML WG deliverable |
| 00:06 | <Hixie> | Lachy: disclosure widgets is what i've heard them called |
| 00:06 | <TabAtkins> | Interesting, doesn't that make some of the statements about the pointed-out lacks in RDFa being addressed in 1.1 moot? |
| 00:07 | <othermaciej> | the term for the triangle part is "disclosure triangle" |
| 00:07 | <Lachy> | oh, yeah, I remember reading "disclosure triangles" in the Apple HIG |
| 00:07 | <othermaciej> | looking up what the control itself is called officially |
| 00:07 | <TabAtkins> | Or does it just mean the group responsible for it is different, but they'll still describe integration with HTML as well as XML? |
| 00:08 | <Lachy> | Hixie, did you make those Pillar magazine graphics? |
| 00:08 | <Hixie> | yes |
| 00:08 | <Hixie> | that's my cat |
| 00:09 | <Lachy> | yeah, I know |
| 00:09 | <Lachy> | I recognised it |
| 00:09 | <AryehGregor> | TabAtkins, RDFa in HTML never did anything but said how to integrate it, right? I assume it will just say how to integrate 1.1 just as much. |
| 00:09 | <othermaciej> | Lachy: "disclosure triangle" is the only term, the control is provided as just the spinny triangle, and not as the whole integrated thing that incorporates the contents |
| 00:09 | <TabAtkins> | AryehGregor: Well, at the very least it'll have to be updated for the various incompatible changes that 1.1 is making. |
| 00:10 | <AryehGregor> | I don't see how that's relevant to the statements you referred to, though. |
| 00:10 | <TabAtkins> | It is probably me just misreading the statement, so don't worry about it. |
| 00:11 | <TabAtkins> | Oh man, I've been off work for 15 minutes! Woo! |
| 00:14 | TabAtkins | goes to play more Saints Row. |
| 01:42 | <othermaciej> | Hixie: I keep accidentally opening http://damowmow.com/playground/htmlwg/chart.html in released Safari instead of trunk |
| 01:42 | <othermaciej> | Hixie: I gotta get that bug into a software update for my own sanity |
| 01:43 | <Hixie> | hah |
| 02:05 | <othermaciej> | Hixie: is it useful to have separate "HTML5 spec bugs" and "HTML5 spec proposals" components in w3c bugzilla? |
| 02:05 | <othermaciej> | I'm thinking of suggesting they be folded into one |
| 02:06 | <Hixie> | no, but please don't change that because that would break all my queries |
| 02:06 | <Hixie> | all my queries, bookmarks, scripts, etc, use those two components |
| 02:06 | <othermaciej> | I see |
| 02:06 | <Hixie> | (similarly, please don't add new components) |
| 02:06 | <othermaciej> | is it possible to make, say, the "proposals" one closed to new bugs without removing it fully? |
| 02:06 | <Hixie> | dunno |
| 02:06 | <Hixie> | is it causing a problem? |
| 02:07 | <othermaciej> | not really, just minor annoyance when forming a query by hand |
| 02:07 | <Hixie> | ah |
| 02:07 | <othermaciej> | I do think separate drafts should have separate components, cause otherwise it's pretty hard to separately track when each gets to 0 bugs in a non-terminal state |
| 02:08 | <Hixie> | well if there are other editors, they should get components, sure, but for drafts i edit i treat them all as one draft anyway |
| 02:08 | <Hixie> | so it's moot |
| 02:09 | <othermaciej> | so I'm seeing 231 bugs resolved since the start of the month |
| 02:09 | <othermaciej> | 67 still open |
| 02:09 | <othermaciej> | incoming rate must be out of control |
| 02:09 | <othermaciej> | since there were fewer than 200 open on the 1st |
| 02:09 | <Hixie> | incoming rate grew significantly higher than i expected, yes |
| 02:09 | <Hixie> | though i don't track that |
| 02:10 | <Hixie> | so i don't have numbers |
| 02:11 | <othermaciej> | 115 created since the start of the month |
| 02:12 | <TabAtkins> | Jeez, that's 10 a day. |
| 02:13 | <othermaciej> | FWIW Hixie's accept/reject rate is still around 50% of the non-garbage bugs |
| 02:14 | <othermaciej> | (where by "garbage" I mean duplicate or invalid) |
| 02:14 | <othermaciej> | I looked at some past months and the rate seems to be slowly going down over time |
| 02:14 | <othermaciej> | the highest I found was 70% |
| 02:14 | <Hixie> | as in i'm rejecting more? |
| 02:14 | <othermaciej> | that seems like pretty much whatI would expect for a maturing spec |
| 02:14 | <othermaciej> | yes, you seem to be rejecting more percentage-wise than in, say, July |
| 02:15 | <Hixie> | i'd expect the rate to go the other way, actually, because the problems should be getting more trivial (e.g. reviewers should be moving from "this doesn't work" to "this is misspelt") |
| 02:16 | <othermaciej> | well some reviewers are still submitting requests for new features |
| 02:16 | <Hixie> | there might be a better correlation of rate vs reporter than rate vs time |
| 02:16 | <othermaciej> | but you're right, I could imagine it going the other way if the proportion of editorial issues was increasing as one might expect |
| 02:16 | <othermaciej> | possible - I did not try to measure that cause it would be too much work :-) |
| 03:11 | <cardona507> | since we are moving into a fluid versionless HTML - will we also move into a versionless web applications? |
| 07:25 | <hsivonen> | Is the Director going to appoint three so far unannounced individuals to the TAG in addition to the re-elected and the continuing participants |
| 07:25 | hsivonen | tries to understand http://www.w3.org/News/2010#entry-8694 |
| 07:29 | <othermaciej> | hsivonen: that's what it sounds like |
| 07:30 | <othermaciej> | hsivonen: you should run for TAG next time |
| 07:30 | <othermaciej> | I'd tell my AC rep to vote for you :-) |
| 07:30 | <othermaciej> | and it would make their minutes fun to read I bet |
| 08:27 | <zcorpan__> | Hixie: http://www.whatwg.org/specs/web-apps/html5/ gives me 500 |
| 08:28 | <Hixie> | try now |
| 08:29 | hsivonen | never remembers how Windows "NT" version numbers map to branding names |
| 08:31 | <zcorpan__> | Hixie: oh, it's removed? |
| 08:31 | <Hixie> | read your whatwg mail :-) |
| 08:32 | <zcorpan__> | not today |
| 08:33 | <Hixie> | http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-January/024708.html |
| 08:35 | <hsivonen> | hmm. still 3 columns at http://wiki.whatwg.org/wiki/FAQ#What_are_the_various_versions_of_the_spec.3F |
| 08:35 | <hsivonen> | If this trend continues, I guess I'll have to migrate html5.validator.nu to validator.nu and migrate validator.nu to custom.validator.nu |
| 08:35 | <hsivonen> | or something |
| 08:36 | <hsivonen> | calling the primary UI "html5" was probably a bad idea in the first place |
| 08:36 | <Hixie> | heh |
| 08:36 | <hsivonen> | of course, "html5" wasn't the primary UI when the "html5" UI was minted |
| 08:37 | <hsivonen> | changing the meaning of the validator.nu UI sucks for the APIs and for bookmarks... |
| 08:37 | <hsivonen> | and decoupling the API endpoints from the UI URLs would suck, too |
| 08:37 | <hsivonen> | well, I think I'm going to wait and see how things develop |
| 08:38 | <hsivonen> | although by HTML6, I need to rearrange the UI anyway |
| 08:39 | <Hixie> | what is this "html6" you speak of? :-P |
| 08:40 | <hsivonen> | well, HTML5.next |
| 08:41 | <zcorpan__> | HTML5 is HTML5.next |
| 09:05 | <MikeSmith> | Hixie: have you run into cvs locking problem with the last few checkins? |
| 09:06 | <Hixie> | no |
| 09:07 | <MikeSmith> | ok |
| 09:15 | <MikeSmith> | Hixie: revision number no longer shows up correctly in the line above the copyright |
| 09:16 | <MikeSmith> | I think because you have to change it to $Revision: $ |
| 09:16 | <MikeSmith> | with a colon |
| 09:16 | <MikeSmith> | maybe |
| 09:16 | <Hixie> | oops |
| 09:16 | <Hixie> | will fix |
| 09:48 | <roc> | is it true that the lists.w3.org archives aren't indexed by Google? |
| 09:49 | <Hixie> | http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=site:lists.w3.org |
| 09:49 | <Hixie> | Results 1 - 10 of about 1,080,000 from lists.w3.org |
| 09:50 | <zcorpan__> | how come i have about 1,520,000 |
| 09:51 | <Hixie> | it's an approximation |
| 09:51 | <Hixie> | probably based on how many servers responded, and you probably hit a different data center than i do |
| 09:51 | <jgraham> | Quite a bad one it seems :) |
| 09:51 | <roc> | hmm |
| 09:51 | <roc> | how come my searches never find anything |
| 09:52 | <Hixie> | (i actually have no idea what it's based on) |
| 09:52 | <Lachy> | I get 1,090,000. |
| 09:52 | <Hixie> | roc: what are you looking or? |
| 09:52 | <jgraham> | It should just say results 1-10 of about 1E6 |
| 09:52 | <roc> | e.g. |
| 09:52 | <roc> | www-style making pt a physical unit |
| 09:52 | <Lachy> | zcorpan__, also, google.se returns different resutls from google.com |
| 09:52 | <roc> | that query doesn't even bring up lists.w3.org in the top-10 |
| 09:52 | <Hixie> | http://www.google.com/search?hl=en&safe=off&q=site:lists.w3.org+www-style+making+pt+a+physical+unit&aq=f&oq=&aqi= |
| 09:53 | <Hixie> | i guess lists.w3.org just doesn't seem very useful to google's algorithms |
| 09:53 | <Hixie> | hard to argue with that position :-P |
| 09:53 | <roc> | also |
| 09:53 | <roc> | "www-style archive" brings up lists.w3.org as the top hit |
| 09:54 | <roc> | but just http://lists.w3.org, not, say, "www-style⊙wo Mail Archives" (http://lists.w3.org/Archives/Public/www-style/) |
| 09:54 | <Hixie> | weird |
| 09:55 | <Hixie> | you can if you search for that a lot, click "add a result" at the bottom, and enter that URL for that query |
| 09:55 | <zcorpan__> | http://www.bing.com/search?q=www-style+making+pt+a+physical+unit&go=&form=QBLH&filt=all |
| 09:56 | <doublec> | search for 'w3c dysfunctional' |
| 09:56 | <doublec> | that brings up lists.w3c.org right away |
| 09:56 | <zcorpan__> | clearly bing is better than google |
| 09:56 | <doublec> | lists.w3.org that is - public-html |
| 09:56 | <hsivonen> | I tried to use Bing as my default engine for a couple of hours and then went back to Google due to search quality issues |
| 09:57 | <hsivonen> | though it could be that only Bing's Finland region sucks |
| 09:57 | <hsivonen> | Bing seems to insist on making the search results locale-sensitive--badly |
| 09:58 | <zcorpan__> | there should be an <input type=range> for locale-senitiveness |
| 09:59 | <zcorpan__> | when i search for an article about some topic, i don't care about locale at all |
| 09:59 | <zcorpan__> | but when i search for a restaurant, i do |
| 10:00 | <hsivonen> | GPS-assisted restaurant search on Google Maps Mobile is awesome |
| 10:00 | <hsivonen> | (when outside Helsinki, that is. for Helsinki, there's a specialized service that beats Google on restaurants) |
| 10:00 | <zcorpan__> | maybe search engines could figure out from the query whether to be locale-sensitive or not |
| 10:06 | <hsivonen> | I believe they try with varying levels of success |
| 10:07 | <roc> | hmm |
| 10:07 | <roc> | Bing IS better than Google |
| 10:07 | <roc> | I should pay more attention to Asa |
| 10:08 | <roc> | I think Google tries to be locale-sensitive with great success |
| 10:09 | <roc> | my search for "empress garden" returns the correct Auckland restaurant as the top hit (actually the top several hits) ... I haven't searched for it before |
| 10:10 | <hsivonen> | roc: I tried to pay attention to Asa's recommendation, but the results were bad for me. |
| 10:10 | <roc> | I was being sarcastic |
| 10:11 | <roc> | I don't have an "add a result" option on Google |
| 10:12 | roc | sends feedback |
| 10:13 | <hsivonen> | Y! outsourcing their search to Bing doesn't improve the state of competition :-( |
| 10:19 | <zcorpan__> | what's Asa? |
| 10:20 | <hsivonen> | s/what/who/ |
| 10:20 | <hsivonen> | zcorpan__: http://weblogs.mozillazine.org/asa/archives/2009/12/if_you_have_nothing.html |
| 10:21 | <hsivonen> | aargh. are data: URLs that don't have <script> elements synchronously navigated to in WebKit? |
| 10:28 | <Hixie> | roc: are you logged into google? |
| 10:28 | <Hixie> | roc: if you're logged in, it should be the first link on the left under the Goooooooogle paginator |
| 10:29 | <roc> | ahah |
| 10:29 | <roc> | thanks |
| 10:29 | <zcorpan__> | hmm, seems MessageChannel is supported in webkit but not in firefox |
| 10:29 | <hsivonen> | Hixie: are you actively following the about:blank discussion in dev.platform? |
| 10:30 | <Hixie> | hsivonen: not actively |
| 10:30 | <hsivonen> | so far it seems that about:blank navigation isn't fully synchronous in Gecko and WebKit |
| 10:31 | <hsivonen> | but might be in Opera and IE8 |
| 10:31 | <Hixie> | when is it not? |
| 10:31 | <hsivonen> | Hixie: http://hsivonen.iki.fi/test/bz-about-blank.html |
| 10:32 | <hsivonen> | Hixie: if the http URL is changed to a data URL, then WebKit acts like Opera |
| 10:35 | <Hixie> | bring it up in #webkit |
| 10:35 | <Hixie> | or on the whatwg list |
| 11:41 | <Hixie> | nn |
| 11:50 | <zcorpan__> | text/sandboxed-html doesn't have the same backcompat as <iframe sandbox> |
| 11:51 | <zcorpan__> | it should be a parameter for text/html instead |
| 11:55 | <Philip`> | zcorpan__: Wouldn't that defeat the point? |
| 11:56 | <Philip`> | If a user could visit the supposedly-sandboxed content by a direct URL, rather than via the iframe, and their (legacy) browser rendered the page, the user would be vulnerable to malicious content |
| 11:57 | <Philip`> | Oh, but I suppose legacy browsers will ignore the sandbox attribute on iframe anyway, so maybe we don't care about them |
| 11:57 | <zcorpan__> | or we should have <sandboxediframe> |
| 12:01 | <hsivonen> | Philip`: did you have a page with links to Web content that doesn't declare the charset in any way? |
| 12:05 | <Philip`> | hsivonen: Hmm, I can't find any like that |
| 12:05 | <hsivonen> | Philip`: ok |
| 12:05 | <Philip`> | There's http://philip.html5.org/data/charsets.html#sniffed-not-meta but it sounds like that wasn't looking at HTTP content-type |
| 12:07 | <Philip`> | hsivonen: Oh, there's http://philip.html5.org/data/pages-with-no-explicit-charset-in-tlds-with-funny-alphabets.txt |
| 12:07 | <Philip`> | which might be relevant |
| 12:07 | <hsivonen> | excellent. thanks |
| 12:09 | <gsnedders> | http://code.google.com/p/unladen-swallow/issues/detail?id=105 |
| 12:09 | gsnedders | is seriously impressed |
| 12:10 | <Philip`> | gsnedders: You should wait until it's within an order of magnitude of Java before being that impressed :-p |
| 12:11 | <gsnedders> | The amount of activity is impressive though :) |
| 12:13 | jgraham | prefers to be irreverently impressed |
| 12:15 | <zcorpan__> | /me just learned a new word |
| 12:15 | <zcorpan__> | so will html5lib be on par with lxml? |
| 12:18 | gsnedders | thought he was the irreverent one |
| 12:18 | <gsnedders> | zcorpan__: No |
| 12:20 | <zcorpan__> | gsnedders: then i'm not impressed :P |
| 12:20 | <hsivonen> | does the whatwg wiki require keys? |
| 12:20 | <hsivonen> | or has brucel just not created an account? |
| 12:21 | <Dashiva> | Anyone can register |
| 12:23 | <Lachy> | I don't think he realised it was the whatwg wiki, as opposed to the W3C wiki, which does have restrictions on who can edit it |
| 12:24 | <Lachy> | anyway, I already told him he can edit it |
| 12:53 | <zcorpan__> | Hixie: the web workers spec says "Draft Recommendation" while html5 says "Draft Standard" |
| 13:28 | <Huvet> | oh, nice... isn't this a nice sight? <!DOCTYPE HTML PUBLIC "-//SoftQuad//DTD HoTMetaL PRO 4.0::19970714::extensions to HTML 4.0//EN" "hmpro4.dtd"> |
| 13:28 | <Huvet> | http://nyhetsbrev.nu |
| 13:28 | <Huvet> | horrible! |
| 13:29 | <Huvet> | passes all your code just fine :) |
| 13:29 | Huvet | applauds |
| 13:30 | <Lachy> | Huvet, that's one of the many doctypes that triggers quirks mode. |
| 13:30 | <Huvet> | yep, I'd just never seen one that was that bad :) |
| 13:31 | <Lachy> | do you mean the doctype, or the whole page? |
| 13:31 | <Huvet> | the doctype |
| 13:31 | <Philip`> | http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#quirks-mode-doctypes |
| 13:32 | <Huvet> | I'm parsing swedish newspapers, and have crashed html5lib five times (four of them fixed) |
| 13:32 | <Huvet> | so I have seen horrible stuff :) |
| 13:32 | <Philip`> | These HTML editors are immortalied forever |
| 13:32 | <Philip`> | s//s/ |
| 13:32 | <Huvet> | yeah, the authors will have accomplished something in life |
| 13:33 | <Huvet> | John "HoTMetaL Pro 2.0" Doe |
| 13:34 | <hsivonen> | does Flash have native support for showing spherical images? |
| 13:34 | <hsivonen> | like street view and stuff |
| 13:35 | <hsivonen> | or is it put together from lower level pieces on ActionScript? |
| 13:35 | <zcorpan__> | Huvet: it's worse when people change various parts of the doctype, like <!DOCTYPE php PUBLIC "-//W3C/DTD HTML 4.01//NL" /> |
| 13:35 | hsivonen | wonders if painting <video> frames with WebGL can replicate the demo at http://demos.immersivemedia.com/ |
| 13:36 | <Huvet> | zcorpan__: that's worse, yes :( |
| 13:36 | <Huvet> | I wouldn't be surprised if someone used backslashes in the doctype, cause "they are on windows" |
| 13:37 | <hsivonen> | whoa. I've never seen !DOCTYPE php |
| 13:38 | <Dashiva> | That's a cool effect, hsivonen |
| 13:38 | <zcorpan__> | can't that be done with 2d canvas? |
| 13:39 | <Philip`> | http://philip.html5.org/data/doctypes.html#%3c%21doctype_php_public_%22-%2f%2fw3c%2f%2fdtd_xphp_1.0_transitional%2f%2fen%22_%22http%3a%2f%2fwww.w3.org%2ftr%2fxphp1%2fdtd%2fxphp1-transitional.dtd%22%3e |
| 13:39 | <Dashiva> | What is the source material like? |
| 13:39 | <Philip`> | (<!doctype php public "-//w3c//dtd xphp 1.0 transitional//en" "http://www.w3.org/tr/xphp1/dtd/xphp1-transitional.dtd">;) |
| 13:39 | <zcorpan__> | hsivonen: probably global s/html/php/ when switching file extensions... |
| 13:39 | <Dashiva> | The sphere mapped to a rectangle? |
| 13:40 | <hsivonen> | zcorpan__: in theory, yes, but is the perf sufficient? |
| 13:40 | <Philip`> | There's a lot of http://philip.html5.org/data/doctypes.html#%3c%21doctype_doctype_public_%22-%2f%2fw3c%2f%2fdtd_html_4.0_transitional%2f%2fen%22%3e |
| 13:41 | <Dashiva> | How would you do a sphere -> rect (or opposite) mapping in 2D canvas? |
| 13:41 | <zcorpan__> | hsivonen: no idea |
| 13:42 | <hsivonen> | Dashiva: by pixel, I guess |
| 13:42 | <hsivonen> | Dashiva: or by slicing the image and drawing approximation with a 2D matrix |
| 13:42 | <Philip`> | I expect it'd be really hard to make it look good in 2D canvas, because you can only do affine transforms, so you can't even do a decent textured polygon mesh |
| 13:43 | <hsivonen> | the above Flash thingy doesn't look so great when there are straight lines in the view |
| 13:43 | <hsivonen> | like the railing of the bridge |
| 13:43 | <Philip`> | That demo already seems to have really dodgy perspective |
| 13:44 | <Philip`> | since the image seems to stretch when it's near the edges |
| 13:47 | <zcorpan__> | 3d effects don't have to be perfect |
| 13:50 | <Philip`> | I think you can draw an arbitrary texture triangle onto an arbitrary screen triangle (by computing an appropriate transform matrix), so I suppose you could pre-compute a spherical triangle mesh that covers the screen, and map each triangle onto texture coordinates depending on view angle, so maybe it wouldn't be too bad |
| 13:50 | <Philip`> | but you'd have to use lots of triangles |
| 13:51 | <zcorpan__> | will <video controls> in firefox gain a full-screen button? |
| 14:11 | hsivonen | wonders when Xiph is going to release XiphQT with Thusnelda |
| 15:02 | <AryehGregor> | zcorpan__, I don't see one in mozilla-central. |
| 15:03 | <AryehGregor> | Just in the context menu. |
| 15:03 | <AryehGregor> | I don't know why not, I imagine people would expect it from YouTube (I would). |
| 15:03 | <AryehGregor> | (except that I specifically know the feature exists so am willing to explore and do things like right-click) |
| 15:07 | <MikeSmith> | AryehGregor: does the MediaWiki API provide any way to extract sections from a page? |
| 15:07 | <AryehGregor> | MikeSmith, um, let me see. |
| 15:08 | <MikeSmith> | OK, to be clear, what I mean is, if I give it a section title, can I get the contents of that section back |
| 15:09 | <AryehGregor> | I'd imagine you could do it by section number. Section title is probably less reliable, those aren't guaranteed to be unique and section 0 has no title. |
| 15:09 | AryehGregor | doesn't usually use the API, so is looking at the docs |
| 15:12 | <AryehGregor> | Hmm. The closest I see is this: http://en.wikipedia.org/w/api.php?action=parse&page=HTML5&prop=sections |
| 15:13 | <Lachy> | text/sandboxed-html doesn't seem to have very nice graceful degradation. |
| 15:13 | <AryehGregor> | I guess you could retrieve that, retrieve the wikitext, and slice out the bytes given. If you want it to be HTML, you can then feed that back to the parser. Of course, that will give weird results if any tags straddle sections. |
| 15:14 | <AryehGregor> | The person to ask in RoanKattouw, he's the API guy, but he seems not to be on IRC right now. |
| 15:15 | <Lachy> | I assume that's probably considered a feature that current UAs won't accidentally interpret the HTML, but Firefox popping up a save dialog to users isn't very nice. |
| 15:15 | <Lachy> | Safari seems to have the best fallback, which is to treat it as text/plain |
| 15:16 | <Philip`> | If your site wants to work in legacy browsers without exposing those users to XSS attacks, the site will have to sanitise the HTML anyway before it can be used in <iframe sandbox> |
| 15:17 | <Philip`> | in which case serving the same content as standalone text/html to legacy browsers wouldn't introduce any new dangers |
| 15:17 | <Philip`> | It only makes a difference for sandbox-supporting browsers, where the sandbox layer of defence would be lost if the content was accessed outside of its iframe |
| 15:17 | <Lachy> | yeah, I realise that. I'm just concerned that this will limit adoption among sites, since they usually need to maintain compat with older browsers for a few years after they've been made obsolete |
| 15:18 | <Philip`> | So I guess something like text/html;sandbox=1 would be better because it will work in legacy browsers without introducing new vulnerabilities, and will allow sandbox-supporting browsers to apply extra defences to it |
| 15:18 | <Lachy> | and I don't know any site that would want users to inadvertently save the file and then possibly open it wondering what it is |
| 15:19 | <Philip`> | (But, on the other hand, people will add ;sandbox=1 to all their pages before anyone's released a sandbox-supporting browser, and therefore prevent it ever being supported) |
| 15:20 | <Lachy> | Perhaps sites will resort to some sort of browser sniffing to determine whether or not text/sandboxed-html is supported |
| 15:28 | <AryehGregor> | Lachy, does nine count as "a few"? |
| 15:34 | <Lachy> | AryehGregor, in the case of IE6, yes, but luckily, upgrade rates do seem to getting better than that these days |
| 15:35 | <AryehGregor> | Thankfully. |
| 15:35 | AryehGregor | is hoping enterprises will switch to Win7 with IE6-in-VM for their intranets |
| 15:35 | <Lachy> | anyone know what the usage of Firefox 2 is like these days, now that it's had 3.0 and 3.5 after it? |
| 15:36 | Lachy | looks it up... |
| 15:36 | <gsnedders> | Lachy: http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm |
| 15:36 | <Philip`> | http://marketshare.hitslink.com/browser-market-share.aspx?qprid=2 says 0.9% |
| 15:38 | <AryehGregor> | So Firefox 2.0 is still 5% or so of Firefox market share, while IE6 is still 20% or so of IE market share. |
| 15:38 | <Philip`> | gsnedders: Looks like it got confused by Opera 10's UA string |
| 15:39 | <AryehGregor> | Well, it's a confusing UA string, isn't it. |
| 15:41 | <Philip`> | All UA strings are confusing |
| 15:42 | <Lachy> | that's not too bad. It's only been ~1.5 years since Firefox 3.0 succeeded it, and it's already at fairly ignorable levels |
| 15:42 | <TabAtkins> | I've been ignoring it for a long time. |
| 15:42 | <TabAtkins> | Though I've also been ignoring IE6 as much as possible for at least 6 months. |
| 15:43 | <Lachy> | yeah, me too. When I was doing web development, I started ignoring older firefox releases fairly soon after the new one was released. |
| 15:43 | <TabAtkins> | I have the benefit of our tech support pushing people onto IE7 and now IE8. We don't get very many IE6 hits now. |
| 15:44 | <Lachy> | which site is that for? |
| 15:45 | <TabAtkins> | igofigure.com |
| 15:45 | <Lachy> | why isn't your tech support pushing people onto Firefox, Opera or Safari? |
| 15:45 | <Lachy> | or Chrome? |
| 15:45 | <TabAtkins> | Because it's much easier to say "Do Windows Update, our software requires it". |
| 15:45 | Philip` | wonders how long it takes for an obsolete Firefox release to fall below the combined market share of all versions of Opera |
| 15:46 | <TabAtkins> | And updated IE is a Windows Update. |
| 16:05 | <AryehGregor> | foolip, so Opera is going to support H.264 if the codec is installed on the system? |
| 16:23 | <foolip> | AryehGregor: unless we take steps to prevent it, it will happen to work on Unix, but no other platforms |
| 16:24 | <AryehGregor> | foolip, Unix including OS X or not? |
| 16:24 | <foolip> | no, not Mac |
| 16:24 | <AryehGregor> | Interesting. |
| 16:24 | <foolip> | just platforms where we can use the native GStreamer, which is, basically, Linux and FreeBSD |
| 16:24 | AryehGregor | wishes Apple would end the discussion by supporting Theora already. :( |
| 16:25 | foolip | too |
| 16:25 | foolip | also hope noone goes off claiming that "Opera supports H.264" because GStreamer may happen to |
| 16:26 | Philip` | wonders how many Linuxes include H.264 support |
| 16:27 | <AryehGregor> | Philip`, probably all of them if you click through a "don't click this if you live in the US" button. |
| 16:27 | <foolip> | Philip`: none do by default |
| 16:27 | <AryehGregor> | Well, I can't say "all of them". |
| 16:27 | <AryehGregor> | But I'm pretty sure the major ones have it somewhere in the repos. |
| 16:27 | <Philip`> | I'd assume it's just the ones that are willing to pay licensing fees and the ones that don't care about patent infringement, but I don't know if there's any that aren't either of those categories |
| 16:27 | <AryehGregor> | *Is* it patent infringement in Europe? |
| 16:27 | <AryehGregor> | Or just the US? |
| 16:27 | <foolip> | well, Ubuntu will happily install lots of stuff if you play something in totem |
| 16:35 | <Philip`> | Hmm, it looks like I do have GStreamer installed |
| 16:35 | <Philip`> | but I don't think I ever use anything that uses it |
| 16:35 | <Philip`> | (since I just use mplayer and vlc) |
| 16:49 | <foolip> | ubuntu has it per default anyway |
| 16:57 | <Philip`> | Hooray for Mercurial |
| 16:57 | <Philip`> | (in the W3C) |
| 16:58 | <AryehGregor> | Why not git? |
| 16:58 | <jcranmer> | ? |
| 16:58 | <gsnedders> | Because git is evil. |
| 16:58 | gsnedders | hides |
| 16:59 | <Dashiva> | You git |
| 17:01 | <Philip`> | You mercurial fellow |
| 17:01 | <Philip`> | Hmm, that's not as good an insult |
| 17:02 | <AryehGregor> | Philip`, you're a subversive, mercurial git who . . . hangs out in bazaars. |
| 17:02 | <AryehGregor> | Could use work. |
| 17:04 | <Philip`> | You're such a CVS |
| 17:05 | <Dashiva> | What a load of VCS |
| 17:07 | <AryehGregor> | I know. You're a subversive, mercurial git whose version-control preferences are simply bazaar. |
| 17:07 | <AryehGregor> | (could still use work, but getting there) |
| 17:09 | <Dashiva> | git who VC(t)S bazaars |
| 17:11 | <gsnedders> | What a bazaar conversation. |
| 17:12 | <Philip`> | We should all be committed |
| 18:30 | gsnedders | is hearing rumours about an iniment html5lib release |
| 19:23 | <Huvet> | hi everyone! I was wondering if someone has a couple of minutes to look at a NoReverseMatch error. I've tried to fix this for at least an hour now, trying every combination I can: http://dpaste.com/144265/ |
| 19:23 | <Huvet> | basically, I have a url.py, that includes another url.py, in which I have a name I want to reverse |
| 19:23 | <Huvet> | I can't |
| 19:24 | <Huvet> | it's probably something stupid :( |
| 19:28 | <Huvet> | GAH, I found it |
| 19:29 | <Huvet> | urls.py, I have a $ at the end of the first regexp :/ |
| 20:49 | <crash\> | is it everywhere mentioned how browser must act after removing the selected element? I mean change HTMLSelectElement.value and .selectedIndex? |
| 20:51 | <crash\> | Opera doesn't changes those values after removing the selected option |