| 00:02 | <eseidel> | TabAtkins_, Hixie: it's unclear to me if seamless (http://www.whatwg.org/specs/web-apps/current-work/#attr-iframe-seamless) is supposed to cause the parent document's stylesheets to show up in the child document's "document.styleSheets" accessor: http://dev.w3.org/csswg/cssom/#document-style-sheets |
| 00:03 | <eseidel> | Not having them appear there, could be a bit of implementation work... :) |
| 00:03 | <eseidel> | so I'm tempted to have them appear |
| 00:03 | <eseidel> | but I wanted to make sure there was no objection before I go down that path |
| 00:03 | <eseidel> | no (big) objection :) |
| 00:09 | <eseidel> | TabAtkins_, Hixie: actually, on second thought. I'm not sure which approach makes more sense in webkit. Will investigate and get bakc to you |
| 00:12 | <TabAtkins_> | eseidel: I *think* they're not supposed to show up. The rules just get inserted into the cascade. |
| 00:52 | <Hixie> | eseidel: they're not supposed to currently, but i don't feel strongly one way or the other, so please don't hesitate to mail the list suggesting that the spec be clarified one way or the other and we'll see what people think |
| 00:52 | <eseidel> | Hixie: k |
| 00:53 | <eseidel> | Hixie: my current patch does not make them appear. |
| 00:55 | <Hixie> | not making appear seems safest long-term |
| 00:56 | <othermaciej> | yay I'm actually in whatcg now |
| 00:57 | <Hixie> | i'm amused at all the people wanting to join the cg |
| 00:57 | <nessy> | Hixie: I see you've started going through WebVTT stuff! Cool! |
| 00:57 | <Hixie> | since it makes no difference currently :-) |
| 00:57 | <Hixie> | nessy: yeah, finally got to that on my "urgent todo" list |
| 00:57 | <Hixie> | nessy: i half think i should find someone to take over that spec since it took me so long, i feel bad |
| 00:58 | <nessy> | Hixie: we got through a lot of discussion in that time and basically I have collected all the feedback in the bugs |
| 00:59 | <Hixie> | cool |
| 00:59 | <Hixie> | i noticed some of them |
| 00:59 | <nessy> | Hixie: you might walk through it bug by bug (just search for WebVTT in the W3C bug tracker) then you shouldn't miss anything |
| 00:59 | <Hixie> | if they're all assigned to me and open then i won't miss them |
| 01:00 | <nessy> | Hixie: they are, but if you go through bugs by their age, you'll miss the newer ones ;-) |
| 01:00 | <nessy> | Hixie: there's "only" 23 of them |
| 01:00 | <Hixie> | my plan is to work on them for a while then go back to FIFO handling, so i expect i won't get through everything right away |
| 01:00 | <Hixie> | k |
| 01:01 | <nessy> | and some of them are even only for <track>, not WebVTT specific |
| 01:01 | <nessy> | no worries - do what you can |
| 01:01 | <othermaciej> | Hixie: it makes a difference in that anyone joining has to agree to the CLA |
| 01:01 | <othermaciej> | though I'm not sure if "contribution" only applies to text or also ideas |
| 01:02 | <nessy> | I assume you're not so interested in, e.g. the specs for the metadata header or so - if you want to flick some over to me to write patches for the spec, I could give it a try |
| 01:03 | <Hixie> | othermaciej: as i understand it, a "contribution" is text you send to the public-something-contrib list |
| 01:03 | <Hixie> | othermaciej: which i expect will only rarely be used for us |
| 01:03 | <Hixie> | othermaciej: maybe e.g. if hober writes a chunk of text that i import verbatim |
| 01:03 | <nessy> | Hixie: I do like keeping the style of the spec consistent, so would prefer if your editing continued, may with some assistance though |
| 01:03 | <Hixie> | othermaciej: but that's happened almost never so far |
| 01:04 | <othermaciej> | so even sending proposed text to the non-contrib list doesn't count? |
| 01:04 | <Hixie> | nessy: i'm not planning on quitting, just a bit swamped |
| 01:05 | <Hixie> | othermaciej: ianal, but that's my understanding. in practice i would tell people to send any to the contrib list. not that that's ever come up before. |
| 01:06 | <nessy> | Hixie: good! (/me wipes off sweat) ;-) |
| 01:06 | <nessy> | Hixie, othermaciej: the FAQ says "Each group may establish its own mechanisms for recording contributions.." |
| 01:07 | <nessy> | Hixie, othermaciej: the problem is about "recording contributions" that seems to require some formal list of what ppl contributed and how it was integrated into the spec |
| 01:08 | <Hixie> | othermaciej: btw since we don't really have "chairs" in the whatwg i thought maybe i should assign all the whatwg "members" (the ones on the private list) as "chairs" in the cg. as far as i can tell it just means you're an admin on the wordpress side which doesn't really mean anything. |
| 01:08 | <nessy> | Hixie, othermaciej: in particular this part of the FAQ "When participants entrust the Editor with the job of recording contributed material, the group must be diligent about recording the true source and date of the Contribution when the Editor is merely transcribing the Contribution." |
| 01:08 | <othermaciej> | Hixie: I think it's up to you - I imagine there won't be much actual active management of the CG required in any case |
| 01:11 | <nessy> | In theory: "The participants of the Group choose their Chair(s)." http://www.w3.org/community/about/agreements/ |
| 01:15 | <nessy> | In practice: whatever you suggest will be ok - I like the idea of the original WHATWG members |
| 06:30 | <zcorpan> | "iso-8859-8 visual" and "logical" are labels for the same encoding? |
| 06:30 | <zcorpan> | does hebrew use the same characters for visual and logical? |
| 06:32 | <zcorpan> | annevk: what about labels for the other encodings? |
| 07:00 | <annevk> | zcorpan: I think that was from when direction details were embedded in the encoding label |
| 07:00 | <annevk> | zcorpan: I've been somewhat more liberal with multi-byte encodings, but I think either we should do more content study or just check browsers manually |
| 07:00 | <annevk> | zcorpan: http://wiki.whatwg.org/wiki/Web_Encodings has some data |
| 07:01 | <annevk> | zcorpan: one area that is currently missing labels is utf-16, because of the new default (le instead of be) browsers have some mismatching going on |
| 07:26 | <annevk> | still not much wiser about isTrusted |
| 07:26 | <annevk> | apart from that it might be needed for the shadow DOM, of which implementations details are less than clear |
| 07:56 | <zcorpan> | annevk: http://simon.html5.org/dump/charsets/charsets-count.txt http://simon.html5.org/dump/charsets/stevef-charsets-count.txt |
| 07:58 | <annevk> | zcorpan: how was that created? |
| 07:58 | <zcorpan> | see http://simon.html5.org/dump/charsets/ |
| 07:58 | <zcorpan> | the stevef dataset is http://www.paciellogroup.com/blog/2012/04/html5-accessibility-chops-data-for-the-masses/ |
| 07:59 | <annevk> | cool |
| 07:59 | <freddyb> | hi |
| 08:00 | <freddyb> | why is it, that the upcoming domcrypt is so little bold? |
| 08:01 | <Ms2ger> | Because it's italic already |
| 08:01 | <zcorpan> | nice to see utf-8 at the top of both lists |
| 08:01 | <annevk> | surprising that gbk is so dominant |
| 08:02 | <annevk> | and actually a few pages labeled gb18030 |
| 08:05 | <zcorpan> | keep in mind that pages may be using multiple labels, so even if a label shows up here, it doesn't necessarily mean that there are the same number of pages actually using that as their first label |
| 08:07 | <zcorpan> | also, i forgot ";" in the regexp |
| 08:07 | <freddyb> | Ms2ger: no, seriously. why does it provide random numbers but no other neat crypto features? is it planned to provide this before release or will people have to wait for another version? |
| 08:08 | <Ms2ger> | Oh, what's specced on https://dvcs.w3.org/hg/domcrypt/raw-file/tip/Overview.html ? |
| 08:09 | <freddyb> | yes, sorry. wasn't that clear? ;P |
| 08:09 | <annevk> | freddyb: features evolve incrementally |
| 08:09 | <Ms2ger> | That's just because it's the only part that had a spec |
| 08:10 | <annevk> | freddyb: there's not really such a thing as "release" |
| 08:10 | <Ms2ger> | So I just put that in the spec |
| 08:10 | <freddyb> | I mean, I know it says work in progress. but I was wondering what to expect :) |
| 08:10 | <Ms2ger> | While waiting for the editors to start working on it |
| 08:10 | <freddyb> | I see |
| 08:11 | <Ms2ger> | I think the rest had something of a description at https://wiki.mozilla.org/Privacy/Features/DOMCryptAPISpec/Latest |
| 08:11 | <Ms2ger> | has* |
| 08:12 | <zcorpan> | annevk: fixed ";" |
| 08:12 | <freddyb> | Aah! thanks, Ms2ger |
| 08:12 | <Ms2ger> | Np |
| 08:13 | <zcorpan> | charset=iso-8859-1 is interesting, wonder if we should make the encoding sniffing algorithm pick it up |
| 08:14 | <zcorpan> | i.e. <meta http-equiv=content-type content="text/html; charset=charset=iso-8859-1"> |
| 08:15 | <Ms2ger> | windows-1252, you mean? |
| 08:15 | <Ms2ger> | (Did I get that number right?) |
| 08:15 | <zcorpan> | i mean "charset=iso-8859-1" turned up as a used encoding label |
| 08:16 | <Ms2ger> | I did, now I feel really bad |
| 08:16 | <Ms2ger> | Oh |
| 08:16 | <Ms2ger> | Still, you mean charset=charset=windows-1252? :) |
| 08:17 | <zcorpan> | that didn't turn up :-P |
| 08:22 | <foolip> | fantasai, we (Opera) are interested in implementing balanced line wrapping for WebVTT as the default, is there anything we can do to help speed up the spec process? |
| 08:50 | <annevk> | MikeSmith: http://www.w3.org/community/whatwg/ "Contact Group" still points to public-whatwg |
| 08:51 | <annevk> | windows-1234 cute |
| 08:51 | <annevk> | euc-jup hahaha |
| 08:55 | <zcorpan> | annevk: the script misses labels where the grep output a match that spans several lines... not sure how to fix that |
| 08:56 | <annevk> | we should get someone from Google to do some crawling |
| 08:59 | <annevk> | okay, I think I'll add all the labels where len(browsers) > 1 |
| 08:59 | <annevk> | based on spectable.html |
| 09:00 | <annevk> | and counting Safar/Chrome as 1 |
| 09:07 | <Ms2ger> | Sounds good |
| 09:09 | <annevk> | I wonder how "OOnFIC: iso-8859-8 visual" happened |
| 09:09 | <annevk> | "iso-8859-8 visual" is not a label in Opera at least, though visual is |
| 09:12 | <annevk> | seems Steve Faulkner has been carefully redacting his past tweets to make him seem less silly |
| 09:16 | <annevk> | okay so "iso-8859-8 visual" was an input label |
| 09:17 | <annevk> | and the PHP does header("Content-Type:text/plain;charset=".$_GET["label"]); |
| 09:18 | <annevk> | which seems to indicate label sniffing in HTTP might be somewhat interesting in all browsers but Safari? |
| 09:18 | <jgraham> | annevk: I am sad to say I can't quite bring myself to suggest character encoding names as the naming scheme for the new Opera meeting rooms |
| 09:18 | <annevk> | you should use label names |
| 09:19 | <annevk> | they're better |
| 09:19 | <annevk> | x-x-big5 |
| 09:19 | <annevk> | because one x just wasn't enough |
| 09:19 | <jgraham> | "legacy character encodings, like meetings, are a fact of life. No one likes them but we can't get rid of them" - see I even have a good justification |
| 09:19 | <annevk> | or have visual and logical but since they're the same you never know which room to go to |
| 09:22 | <AryehGregor> | It amuses me to no end when I find stuff like this on a W3C webpage: |
| 09:22 | <AryehGregor> | <body class="page page-id-1225 page-child parent-pageid-79 page-template page-template-threecolumn-page-php logged-in admin-bar clearfix"> |
| 09:22 | <AryehGregor> | <!--[if lte IE 8 ]> |
| 09:22 | <AryehGregor> | <noscript><strong>JavaScript is required for this website to be displayed correctly. Please enable JavaScript before continuing...</strong></noscript> |
| 09:22 | <AryehGregor> | <![endif]--> |
| 09:26 | <AryehGregor> | It's probably from Wordpress or something, to be fair. |
| 09:43 | <annevk> | I should probably run the single-byte test again with an updated label set |
| 09:48 | <zcorpan> | text/html; charset= charset= charset=iso-8859-1 |
| 09:49 | <zcorpan> | http://simon.html5.org/dump/charsets/charset-charsets.txt |
| 09:49 | <annevk> | hey Steve Faulkner, thanks for reading along, first you post https://twitter.com/stevefaulkner/status/194715165059592193 then you post your so-called "optimistic tweets" where you apparently suddenly changed your mind without any reason whatsoever (about what I've no idea... what would be the better situation here?) |
| 09:50 | <AryehGregor> | So apparently my editing draft is a Community Group Report now! Yay! http://www.w3.org/community/reports/#editing |
| 09:51 | <annevk> | zcorpan: I guess heuristics will take care of those |
| 09:51 | <zcorpan> | annevk: what heuristics? |
| 09:51 | <annevk> | AryehGregor: I guess now we should see what companies sign off |
| 09:52 | <AryehGregor> | annevk, no, this is just a draft, not a final specification. |
| 09:52 | <AryehGregor> | It wasn't actually linked from anywhere before, because the W3C hadn't finished all the machinery to support "official" draft publication in CGs. |
| 09:52 | <AryehGregor> | I just asked Ian about publishing a snapshot as a Final Specification. |
| 09:52 | <annevk> | AryehGregor: oh so the draft is continually edited? |
| 09:53 | <AryehGregor> | Yeah, look at the URL. |
| 09:53 | <annevk> | AryehGregor: I thought you had CLA snapshots and FSA snapshots |
| 09:53 | <annevk> | and an editor's draft |
| 09:53 | <annevk> | hmm |
| 09:53 | <AryehGregor> | Nope, CLA has no snapshots, thankfully. |
| 09:53 | <AryehGregor> | The publication form says that Draft links are allowed to change, only Final Specification links have to be static. |
| 09:54 | <AryehGregor> | So basically we have only ED and REC. |
| 09:54 | <AryehGregor> | And REC is only for patent-policy purposes if we want. |
| 09:54 | <AryehGregor> | With no technical requirements. |
| 09:54 | <AryehGregor> | Which is *exactly* what we want. |
| 09:54 | <AryehGregor> | Three cheers for the W3C! |
| 09:54 | <annevk> | well sort of |
| 09:54 | <zcorpan> | AryehGregor: having the w3c logo but the whatwg stylesheet looks....weird :-) |
| 09:54 | <annevk> | you don't get WG mandatory patent disclosures |
| 09:54 | <AryehGregor> | zcorpan, the publication requirements said I needed the W3C logo. They didn't say it had to look pretty. :) |
| 09:55 | <AryehGregor> | They expect to have their own stylesheets ready sometime or other, then I can use those. |
| 09:55 | <annevk> | but if companies sign off then it's nice |
| 09:56 | <AryehGregor> | annevk, yes, true, but what do we care? We're tech people, not lawyers. As long as the lawyers tell us we can use the spec, we can be happy. |
| 09:56 | <AryehGregor> | I guess it remains to be seen who signs off on the specs. |
| 09:56 | <AryehGregor> | Since CG membership doesn't require it. |
| 09:56 | AryehGregor | refreshes his memory |
| 09:56 | <annevk> | well my employer cares |
| 09:57 | <annevk> | but I guess we'll see |
| 09:57 | <AryehGregor> | There are also a lot fewer members in the Editing CG than in the HTMLWG, say. |
| 09:58 | <AryehGregor> | So even if all the members sign off, it's not a big deal. |
| 09:58 | <AryehGregor> | As big a deal. |
| 09:58 | <AryehGregor> | Oh well. |
| 10:00 | <AryehGregor> | Ian Jacobs seemed to think the patent license for CGs was good. Apparently there were lots of lawyers involved, including from various key web companies. |
| 10:01 | <AryehGregor> | We'll see. |
| 10:49 | <smaug____> | hsivonen: did you send the email about readystate |
| 10:53 | <smaug> | hsivonen: any comments to http://lists.w3.org/Archives/Public/public-web-perf/2012Apr/0062.html ? |
| 11:14 | <annevk> | reply to i18n-core email: http://lists.w3.org/Archives/Public/public-i18n-core/2012AprJun/0029.html |
| 11:31 | <jgraham> | smaug____: You have to love 'Creating a test that calls out this discrepancy seems odd to me. I think given that, I think we should remove the test from the suite.' |
| 11:32 | <jgraham> | Nothing like removing a test for demonstrating bugs |
| 11:33 | <zcorpan> | jgraham: where's that from? |
| 11:33 | <zcorpan> | oh |
| 11:34 | <jgraham> | It's a rather crappy test anyeay since it allows so much variation in behaviour |
| 11:35 | <smaug____> | performance wg is known to have plenty of buggy tests before reviewing |
| 14:30 | <hsivonen> | smaug____: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-April/035521.html |
| 14:30 | <smaug____> | hsivonen: thanks. Will read later today |
| 14:31 | <hsivonen> | smaug____: I guess that would be my response to http://lists.w3.org/Archives/Public/public-web-perf/2012Apr/0062.html Can you post a pointer to the list so that I don't need to subscribe? |
| 14:31 | <smaug____> | ok, I'll do that |
| 14:31 | <hsivonen> | thanks |
| 14:33 | <hsivonen> | who is writing Fake TAG? |
| 14:33 | <hsivonen> | Marcos seemed to be a proposer for TACG. Any correlation? |
| 15:03 | <hober> | fake tag is so awesome |
| 15:13 | <hsivonen> | "Neil Archstrong" has Marcos' photo in the comments as avatar... |
| 15:19 | <kennyluck> | FTAG |
| 15:20 | <gsnedders> | hsivonen: http://www.w3.org/mid/op.wc9s77zq4p7avi⊙ll if you hadn't seen it |
| 15:20 | <gsnedders> | (MO) |
| 15:22 | <hsivonen> | gsnedders: I hadn't seen that. thanks |
| 15:25 | <hsivonen> | I wonder why that message was Member-only |
| 15:25 | <Ms2ger> | Shitstorms? |
| 15:26 | <gsnedders> | Yeah, exactly. |
| 15:27 | <hsivonen> | I expect one anyway |
| 15:27 | <hsivonen> | including comments that lack "chair hat off" |
| 15:30 | <miketaylr> | funny how that ended up in a blog article |
| 15:32 | <jgraham> | It is times like this that I really appreciate the subtle nuances of communication afforded by twitter |
| 15:35 | <hsivonen> | jgraham: any particularly nuanced tweets? |
| 15:35 | <miketaylr> | heh |
| 15:36 | <miketaylr> | this one is my favorite so far, http://twitter.com/joshbroton/status/195166561806462976 |
| 15:36 | <hsivonen> | :-( |
| 15:37 | <gsnedders> | The real fun on mobile is how long Android/iOS WebKit release-cycles are. |
| 15:37 | <hsivonen> | good thing OO.o developers weren't afraid to support microsoft.com XML namespaces |
| 15:38 | <gsnedders> | Android esp. |
| 15:38 | <zewt> | gsnedders: rather, the fact that there are no release cycles on android ... |
| 15:38 | <gsnedders> | zewt: Too often, yes :( |
| 15:38 | <zewt> | old browsers on many phones simply never being upgraded until the phone dies |
| 15:39 | <zewt> | the only possible saving grace being that phones don't last as long as pcs, heh |
| 15:39 | <gsnedders> | Also plenty of Android releases haven't done much to WebKit. |
| 15:39 | <gsnedders> | So even being up to date with Android doesn't help massively. |
| 15:39 | <gsnedders> | The time-to-market of WebKit ToT to shipping on an Android device is insane. |
| 15:40 | <zewt> | one of my biggest gripes with google is that, despite being a company based almost entirely on the web, they created a platform that out-IE's IE |
| 15:41 | <hsivonen> | Firefox and Opera to rescue |
| 15:41 | <hsivonen> | (as with IE) |
| 15:41 | <zewt> | that would be irresponsible from any company--but Google doing it boggles the mind |
| 15:47 | <gsnedders> | https://twitter.com/#!/patrick_h_lauke/status/195174823230586881 is cute |
| 15:47 | <hsivonen> | My favorite so far is https://twitter.com/#!/alexmuller/status/195170931205873665 |
| 15:49 | <jgraham> | hsivonen: That's a joke, right? |
| 15:50 | <Philip`> | zewt: It also seems odd that they've released a whole new browser app (Chrome on Android) that competes with the default browser on their own platform |
| 15:50 | <hsivonen> | jgraham: twitter nuance is hard, but the next tweet suggests not |
| 15:52 | <miketaylr> | i'm still angry that the HTML5 parser transforms my markup. i wanted <i><b>THIS</i></b>. |
| 15:53 | <zewt> | Philip`: i'm less surprised at google being uncoordinated |
| 15:54 | <zewt> | i suppose that may be a vague attempt at fixing the damage (but as long as it's just an app in the market that people have to install manually, it probably can't do any more than third-party browsers) |
| 15:54 | <jgraham> | Isn't the plan to make Chrome the default from Android.next? |
| 15:54 | <Philip`> | (Also it's not very helpful that Chrome currently only works on a version of Android that 3% of users have) |
| 15:55 | <jgraham> | And then to have it autoupdate? |
| 15:55 | <zewt> | well, the real problem is that the browser can only be upgraded with a full OS upgrade, not with a market update like an app |
| 15:56 | <zewt> | jgraham: don't know anything about their plans, but ... hopefully :) |
| 15:57 | <jgraham> | Well I only know what I thought I had read |
| 15:57 | <jgraham> | So it's possible I imagined the whole thing |
| 15:57 | <zewt> | now I only know what I read on IRC about what someone else thought he had read |
| 15:58 | <Philip`> | https://developers.google.com/chrome/mobile/docs/faq seems quite vague, e.g. "Are you still working on the Android browser, or are you dropping support in favor of Chrome?" says "... We will continue to evaluate where it makes sense to harmonize our efforts ..." which is pretty meaningless |
| 15:59 | <zewt> | one of those sort of insulting answers that mean "we're not answering this, and we won't just say that we won't answer it" |
| 16:00 | <hsivonen> | I expect them to ship Chrome as a bundled app that updates independently of OS and leaving the other WebKit for apps to embed. Fun times for Dolphin HD. |
| 16:01 | Philip` | tried Chrome for about two minutes and then abandoned it because it doesn't automatically re-wrap text like Opera Mobile does so that it fits on the screen (which is hugely useful) |
| 16:02 | <zewt> | i've tried to make web apps work on android's browser, and it really, honestly did feel the same as making pages behave in ie6 (okay, maybe 7; it's not *that* broken) |
| 16:04 | <zewt> | got it working eventually, and it promptly broke in the next version, so I gave up |
| 16:11 | <hsivonen> | my attempts to advocate Opera-like wrapping for Firefox have failed |
| 16:11 | hsivonen | agrees that it's extremely useful |
| 16:30 | <Hixie> | hsivonen: wow (re that tweet) |
| 16:30 | <Hixie> | hsivonen: easier to say where we _didn't_ use that reason... it'd be a shorter list |
| 16:45 | <jgraham> | How have we ended up with people top-posting on public-webapps? |
| 16:45 | <jgraham> | Are google engineers all using outlook now or something? |
| 16:46 | <jgraham> | (this is not a theoretical complaint, I have actually lost track of what's been said in that thread) |
| 16:48 | <Hixie> | gmail encourages it |
| 16:48 | <Hixie> | drives me crazy |
| 16:49 | <zewt> | it's pretty depressing how bad people have become at mailing lists; not editing quotes at all has become common, which gives me migraines when i try to reply to anything |
| 16:49 | <zcorpan> | why isn't there a way to change gmail to use bottom-posting by default? |
| 16:50 | <zewt> | gmail's "quote masking" gives everyone really nasty habits; sometimes I'll try to reply to a mail, only to find that the actual text is nested between two gigantic blocks of quotes and I have to dig it out |
| 16:53 | <zewt> | zcorpan: what I really want (aside from the unreadable "new look" going away) is an option to simply ignore font colors and sizes |
| 17:36 | <jgraham> | Ms2ger: So how much would it break for you if testharness.js tried to notify opener windows about test status as well as parent windows? |
| 17:40 | <Ms2ger> | Nothing, I hope |
| 17:40 | <Hixie> | oh sweet kittens |
| 17:40 | <Hixie> | i get cc'ed every time someone joins the cg |
| 17:40 | Hixie | sets up a filter |
| 17:41 | <othermaciej> | you can't turn that off? |
| 17:42 | <Hixie> | not that i can see |
| 17:42 | <Hixie> | but doesn't matter |
| 17:42 | <Hixie> | filters are easy |
| 17:45 | <zcorpan> | jgraham: i've set up a virtualenv now. how do i install lxml and html5lib in it? |
| 17:46 | <zcorpan> | and anolis |
| 17:46 | <jgraham> | I guess anolis doesn't have a useful setup.py file |
| 17:47 | <jgraham> | Nope |
| 17:47 | <jgraham> | Something like |
| 17:47 | <jgraham> | pip install lxml |
| 17:48 | <jgraham> | pip install html5lib (unless you want the dvcs version, which is a little faster) |
| 17:48 | <jgraham> | pip install hg+https://bitbucket.org/ms2ger/anolis |
| 17:48 | <jgraham> | I think |
| 17:48 | <jgraham> | That's from memory though |
| 17:48 | <Ms2ger> | Some old anolis version is on pip |
| 17:48 | <Ms2ger> | I need to update it at some point |
| 17:49 | <jgraham> | You mean pypi |
| 17:50 | <Philip`> | Python developers ought to discover that there are consonants other than "p" |
| 17:50 | <jgraham> | zcorpan: (also, I assume you activated the virtualenv) |
| 17:51 | <Philip`> | (PyPI, PyPy, pip, pyip, ...) |
| 17:52 | <zcorpan> | jgraham: i hadn't :-) |
| 17:53 | <zcorpan> | ok this looks promising, now installing lxml |
| 17:58 | <zcorpan> | great success! i needed pip install simplejson also |
| 17:58 | <zcorpan> | thanks jgraham |
| 17:59 | <jgraham> | Oh, needing simplejson is silly |
| 17:59 | <jgraham> | That's in the stdlib |
| 17:59 | <jgraham> | Or, a json decoder is and I can't imagine that anolis cares too much about speed |
| 18:00 | <Ms2ger> | How old is your python? |
| 18:00 | <zcorpan> | $ python --version |
| 18:00 | <zcorpan> | Python 2.5.6 |
| 18:00 | <zewt> | the json module in python has a native module anyway, so it's fast |
| 18:00 | <zewt> | yuck--at least get to 2.6 |
| 18:01 | <zewt> | i always have trouble finding 2.7 modules in ubuntian for some reason, even in backports... |
| 18:11 | <annevk> | oh, I thought we would have announced the -webkit- thingy for a while now |
| 18:28 | <zcorpan> | how should i invoke anolis to make references work? |
| 18:29 | <annevk> | see the Makefile's checked in for various specs that use Anolis in that way |
| 18:30 | <Ms2ger> | --enable=xspecxref |
| 18:43 | <zcorpan> | yay |
| 18:45 | <zcorpan> | "CSS Syntax Module Level 3, Tab Atkins Jr.. W3C." hmm, maybe one of the dots should go |
| 18:48 | <Philip`> | "CSS Syntax Module Level 3, Tab Atkıns Jr.. W3C." - how about that dot? |
| 18:48 | <Ms2ger> | Philip`++ |
| 18:50 | <Ms2ger> | TabAtkins_ just shouldn't have so silly a name :) |
| 18:53 | <zcorpan> | "CSS, Bert Bos, Tantek Çelik, Ian Hickson et al.. W3C." has the same thing, heh |
| 18:58 | <kennyluck> | hober, did you mistakenly send http://lists.w3.org/Archives/Public/www-archive/2012Apr/0061 without adding Steve in the "To:" list? |
| 18:59 | <Hixie> | hmm |
| 19:00 | <Hixie> | cues with end times before their start times |
| 19:00 | <Hixie> | what to do, what to do |
| 19:02 | <annevk> | Hixie: something tells me to just flip them |
| 19:03 | <Hixie> | hadn't considered that |
| 19:03 | <Hixie> | seems unlikely to be right though |
| 19:03 | <zewt> | better off just clamping them to zero time |
| 19:03 | <Hixie> | i mean it's probably a typo in one of the times, not the times backwards |
| 19:03 | <Hixie> | clamping the end time to the start time is an option |
| 19:03 | <zewt> | duration |
| 19:03 | <Hixie> | i was thinking of just dropping the cue altogether |
| 19:05 | <zewt> | i sort of wish vtt was start/duration instead of start/end |
| 19:05 | <zcorpan> | i think i'd drop them |
| 19:05 | <zcorpan> | we drop cues for less :-) |
| 19:05 | <Hixie> | yeah |
| 19:05 | <zewt> | probably doesn't make much difference either way except for metadata |
| 19:06 | <Hixie> | we don't drop them in the TextTrackCue constructor either |
| 19:06 | <zcorpan> | maybe that should throw |
| 19:08 | <zewt> | lot of arguments for that ctor ... might be better to have the optional args in a dictionary |
| 19:09 | <zewt> | startTime/endTime aren't readonly anymore, and it wouldn't really make sense to have assigning to those fail |
| 19:10 | <zcorpan> | the arguments could be dropped altogether for any properties that can be set |
| 19:10 | <zcorpan> | good point that it doesn't make sense to make assignments fail |
| 19:10 | <zewt> | well, they can all be set |
| 19:11 | <zcorpan> | very well then :-) |
| 19:11 | <Hixie> | i wonder if maybe i should just check all the algorithms to make sure they can handle negative durations, and then just let them be |
| 19:11 | <zcorpan> | i think i've filed a bug about the arguments for this ctor |
| 19:11 | <zewt> | sort of odd that id is first and required; in most cases it'll be empty, right? |
| 19:12 | <Hixie> | are there impls yet? |
| 19:12 | <Hixie> | we can definitely add an overloaded version with a dict |
| 19:12 | <Hixie> | if not replace the current one entirely |
| 19:12 | <zewt> | Hixie: if it's useless, that's just making implementors do useless testing |
| 19:12 | <zcorpan> | we have an impl but it's not shipped yet |
| 19:12 | <zcorpan> | and we don't mind making teh api better at this point |
| 19:13 | <zewt> | (still hoping the \n vs. <br> thing can be fixed before anyone ships...) |
| 19:13 | <Hixie> | zewt: well there are some use cases, e.g. being able to set the end time before the start time, or represneting user input in an editor without having to verify its sanity |
| 19:14 | <zcorpan> | the common case is setting start time, end time and text, so the constructor could take zero arguments setting nothing or those three |
| 19:14 | <zcorpan> | and for anything else, you'd set the property afterwards |
| 19:18 | <zewt> | is there a way to get a "settings" string from a TextTrackCue (the inverse of the settings argument to the ctor)? seems useful for editors |
| 19:18 | <Hixie> | annevk, dbaron: fyi i marked you (and maciej) as chairs in the CG, since you're charter "members" in the whatwg |
| 19:18 | <TabAtkins> | Is innerHTML supposed to work for SVG elements? |
| 19:19 | <zcorpan> | zewt: you'd have to build it manually i think |
| 19:19 | <Ms2ger> | Yes |
| 19:19 | <Ms2ger> | Per spec |
| 19:19 | <zewt> | zcorpan: right, avoiding that is what I mean |
| 19:19 | <zewt> | hmm |
| 19:19 | <Hixie> | annevk, dbaron: i believe the only side-effect is that you'll get lots of spam when people join, but a quick filter should take care of that. so, uh, sorry about that. |
| 19:19 | <zcorpan> | except the spec is broken |
| 19:19 | <zcorpan> | (innerHTML) |
| 19:19 | <Ms2ger> | It isn't |
| 19:19 | <Ms2ger> | The hook I call is broken ;) |
| 19:19 | <zcorpan> | yeah i mean the html spec |
| 19:20 | <zewt> | i guess some hoops would need to be jumped no matter what you do, for editors to avoid dropping unknown settings |
| 19:20 | <Ms2ger> | My spec is broken in numerous places, but not in that one :) |
| 19:20 | <zewt> | (eg. so if settings features are added to WebVTT, editors that don't know about them yet don't silently throw them away) |
| 19:21 | <zewt> | does anyone else get nervous when somebody starts a mail with "how about"? heh |
| 19:21 | <zewt> | or rather, starts a thread |
| 19:25 | <Hixie> | looks like only the "rules for constructing the chapter tree from a text track" algorithm needs work to handle negative durations |
| 19:25 | <Hixie> | and it doesn't really need work, i just think it's saner to not have negative duration chapters... |
| 19:26 | <zewt> | seems like allowing it just creates busywork for everyone supporting the file format... |
| 19:26 | <Hixie> | zewt: feel free to point him to the FAQ that says that we should start with a problem description, if he didn't :-) |
| 19:26 | <Hixie> | well it's not allowed |
| 19:26 | <Hixie> | i'm just talking about what to do to handle the error if it occurs |
| 19:26 | <Hixie> | it turns out it would be more work to catch it than not |
| 19:26 | <Hixie> | things just work with negative durations |
| 19:27 | <Hixie> | bbiab, lunch |
| 19:27 | <MikeSmith> | http://lists.w3.org/Archives/Public/www-tag/2012Apr/0197.html |
| 19:28 | <annevk> | more spam, teehee |
| 19:28 | <MikeSmith> | bravo Berjon |
| 19:29 | <zewt> | "not allowed" as in discarded by the parser, or as in just non-conforming? |
| 19:39 | <scott_gonzalez> | MikeSmith: Thanks for working on the validator.nu stuff. The HTML lint task is working great :-) |
| 19:39 | <MikeSmith> | scott_gonzalez: excellent |
| 19:39 | <MikeSmith> | glad to hear it |
| 19:40 | <scott_gonzalez> | The only tricky thing we've run into is that we have a few files which are partial HTML files. |
| 19:40 | <MikeSmith> | oh |
| 19:40 | <scott_gonzalez> | Aside from creating temp files that just prepend "<!doctype html><title></title>" is there anything we can do? |
| 19:40 | <MikeSmith> | that is what you need to do currently |
| 19:41 | <MikeSmith> | there's no other way to avoid the error about the missing title element |
| 19:41 | <MikeSmith> | and the doctype |
| 19:41 | <scott_gonzalez> | ok |
| 19:41 | <MikeSmith> | out of curiosity, can you tell me how long it takes to validate all the files? |
| 19:42 | <MikeSmith> | you said you had about 300 files, right? |
| 19:42 | <scott_gonzalez> | It's pretty fast, let me test. |
| 19:43 | <scott_gonzalez> | 256 files |
| 19:45 | <scott_gonzalez> | I'm running it now, but my laptop isn't very happy at the moment, seems like lots of swapping is going on |
| 19:45 | <scott_gonzalez> | been slow all day |
| 19:46 | <scott_gonzalez> | real 0m21.741s |
| 19:46 | <scott_gonzalez> | user 0m18.571s |
| 19:46 | <scott_gonzalez> | sys 0m2.422s |
| 19:49 | <scott_gonzalez> | I quit Firefox and now real is ~10 seconds |
| 19:58 | <annevk> | MikeSmith: W3C Bugzilla bug: https://www.w3.org/Bugs/Public/show_bug.cgi?id=16697#c3 |
| 19:58 | <annevk> | MikeSmith: might be a general Bugzilla bug I suppose |
| 19:58 | <MikeSmith> | scott_gonzalez: :) |
| 19:58 | <MikeSmith> | scott_gonzalez: yeah, that's right in the range I estimated it would be |
| 19:59 | <MikeSmith> | annevk: taking a look now |
| 19:59 | <MikeSmith> | annevk: hmm, yeah |
| 19:59 | <MikeSmith> | we are just running stock bugzilla |
| 20:00 | <MikeSmith> | annevk: 3.6.2 |
| 20:00 | <MikeSmith> | so if there is a problem with character handling, I think it must be a problem that exists in any 3.6.2 bugzilla instance |
| 20:01 | <MikeSmith> | and hopefully something that has since been fixed in the bugzilla code |
| 20:01 | <MikeSmith> | anyway, I need to drop off for a bit |
| 20:01 | <MikeSmith> | back later |
| 20:07 | <annevk> | ta |
| 20:23 | <annevk> | AryehGregor: btw, if we get rid of Range.detach(); NodeIterator.detach() should maybe go too |
| 20:32 | <jsbell> | annevk: Can you run through your encode / decode steps for euc-kr with code point U+3000 and see if it's doing what you expect? (at the very least, encoder step 7.3 seems like it should be referring to trail not pointer) |
| 20:37 | <annevk> | jsbell: agreed about pointer->trail |
| 20:40 | <annevk> | 0xA1 0xD5? |
| 20:42 | <annevk> | hmm |
| 20:42 | <jsbell> | python yields 0xA1 0xA1 ... but regardless, put 0xA1 0XD5 back through your encoder and I get a different pointer |
| 20:43 | <annevk> | yeah |
| 20:43 | <annevk> | the offsets are wrong :( |
| 20:44 | <annevk> | the second should be the difference between 0x5A and 0x61 + 0x41, the third the second plus the difference between 0x81 and 0x7A |
| 20:46 | <Hixie> | anyone implement TextTrack yet other than opera? |
| 20:46 | <annevk> | jsbell I think the offsets should be 0x41, 0x49, 0x51 |
| 20:47 | <annevk> | hmm that does not give me A1 |
| 20:49 | <annevk> | jsbell: 0x41, 0x47, 0x4D |
| 20:52 | <annevk> | jsbell: 0x47=0x41+0x61-0x5A-1; 0x4D=0x47+0x81-0x7A-1 |
| 20:53 | <jsbell> | annevk: that certainly makes my test happier |
| 20:54 | <annevk> | I hate the index math |
| 20:54 | <annevk> | either I do it completely wrong, or get lost in off-by-one errors |
| 20:54 | <annevk> | spec is fixed now |
| 20:57 | <annevk> | jsbell: btw, at least the shift_jis encoder will change https://www.w3.org/Bugs/Public/show_bug.cgi?id=16839 |
| 21:00 | <annevk> | Hixie: for http://html5.org/r/7071 you should use the string enum from IDL |
| 21:00 | <Hixie> | the what now? |
| 21:00 | <annevk> | Hixie: see http://dvcs.w3.org/hg/xhr/raw-file/tip/Overview.html#xmlhttprequestresponsetype for an example |
| 21:01 | <annevk> | Hixie: that will make the setting behavior consistent |
| 21:02 | <Hixie> | well that's new |
| 21:02 | <Hixie> | cool |
| 21:02 | <annevk> | there might be a bug somewhere about using it more generally in HTML, not sure |
| 21:04 | <Hixie> | first i heard of it :-) |
| 21:05 | <jsbell> | annevk: cool and thanks |
| 21:24 | <annevk> | Hixie: you forgot to update dom-TextTrack-mode; s/DOMString/TextTrackMode/ |
| 21:24 | <fantasai> | foolip: post a proposal for last-line-length: <percentage> to www-style? |
| 21:24 | <Hixie> | woops |
| 21:24 | <Hixie> | thanks for watching anne :-) |
| 21:24 | <annevk> | it's over now, bedtime :) |
| 21:26 | <fantasai> | foolip: http://www.w3.org/Style/CSS/Tracker/issues/241 links to prior discussions |
| 21:27 | <fantasai> | foolip: or at least, the ones I've found so far |
| 21:27 | <fantasai> | foolip: I suspect there's other older ones, that I haven't found |
| 21:53 | <Hixie> | TabAtkins: nessy claims in https://www.w3.org/Bugs/Public/show_bug.cgi?id=16623 that block-flow is now writing-mode? Is that true? I thought we'd gone the other way around. |
| 21:54 | <Hixie> | maybe i just made up block-flow, since the first hit for [css "block-flow" property] is the bug in question |
| 22:00 | <kennyluck> | It's been 'writing-mode' for a while now. |
| 22:09 | <Hixie> | yeah i dunno where block-flow came from |
| 22:09 | <Hixie> | it was writing-mode back in 2001 or so already |
| 23:16 | <zewt> | srt was chosen for vtt because it made a mistake that no other formats did? heh |
| 23:38 | <fantasai> | Hixie: There was a 'block-flow' property in the interim. |
| 23:39 | <fantasai> | Hixie: or rather, proposals for it |
| 23:39 | <fantasai> | Hixie: The May 2003 CR of CSS3 Text added 'block-progression' in between closing LC and publishing CR |
| 23:42 | <fantasai> | Hixie: also you got vertical-lr and vertical-rl backwards, if I'm understanding correctly |
| 23:42 | <fantasai> | Hixie: lr is Left to Right, i.e. growing rightwards |