| 00:14 | <MikeSmith> | hey, it's boaz |
| 00:16 | <boaz> | MikeSmith! |
| 00:16 | <boaz> | yo, I've been idling in here for a few weeks |
| 00:16 | <boaz> | :D |
| 00:18 | <boaz> | MikeSmith: how's tokyo? |
| 00:19 | <boaz> | hey, by the way: http://code.bocoup.com/audio-data-api/examples/ambient-extraction-mixer/ |
| 00:21 | <MikeSmith> | boaz: tokyo is feeling impoverished by the lack of annevk |
| 00:22 | <boaz> | bummer |
| 00:22 | <MikeSmith> | annevk was visiting for a while but ended his visit a few days back |
| 00:24 | <MikeSmith> | Ambient Audio Extraction looks relevant to some recent discussion about starting a W3C incubator group for audio-related API work |
| 00:24 | <MikeSmith> | http://www.w3.org/2010/04/audio/audio-incubator-charter.html |
| 00:26 | <MikeSmith> | boaz: oh, I see that page is actually one that Al MacDonald made |
| 00:26 | <boaz> | yah |
| 00:27 | <boaz> | it's his, I'm sitting right next to him |
| 00:28 | <llrcombs> | hey |
| 00:28 | <llrcombs> | I'm getting an SQL error when I try to INSERT |
| 00:28 | <llrcombs> | transaction.executeSql("INSERT INTO quotes (id, quote, name, life, response) VALUES (?,?,?,?,\"Type here to respond!\")", |
| 00:28 | <llrcombs> | SQL Error. Error was "constraint failed" (Code 1) |
| 00:28 | <boaz> | mikesmith: we work together at bocoup, and we're all pretty pumped about the audio incubator! |
| 00:28 | <llrcombs> | (after the , comes the other arguments, i.e. the ? values) |
| 00:29 | <boaz> | mikesmith: Al just finished, so I thought I'd share. |
| 00:31 | <MikeSmith> | boaz: cool, thanks.. and tell Al thanks too |
| 00:31 | <boaz> | yeah! definitely! |
| 00:31 | <MikeSmith> | boaz: I've not been following the audio XG plans so closely.. Doug Schepers is driving it |
| 00:32 | <llrcombs> | is it possible to get the text SQL query from a transaction after an executeSQL call? |
| 00:32 | <MikeSmith> | but I'm catching up on it now |
| 00:32 | <MikeSmith> | boaz: I was kind of away for a bit last month |
| 00:32 | <boaz> | right on, nice, yeah, doug is the man! |
| 00:32 | <boaz> | he was up here for the mozilla hack weekend a few weeks ago |
| 00:33 | <boaz> | also, welcome back! |
| 00:33 | <MikeSmith> | llrcombs: sounds like q sqllite error maybe? so checking the sqllite docs might be good, dunno |
| 00:33 | <MikeSmith> | boaz: yeah, Doug is a whirlwind of activity |
| 00:33 | <llrcombs> | it seems that HTML5 provides its own SQLite dialect |
| 00:34 | <llrcombs> | which is... pretty much 100% undocumented |
| 00:35 | <MikeSmith> | llrcombs: at this point it's not likely to ever get documented, because it's not actually in HTML5 any more |
| 00:35 | <llrcombs> | why not? |
| 00:35 | <MikeSmith> | client-side SQL DB is not going to be part of the Web platform cross-browser |
| 00:36 | <llrcombs> | Safari 4 implemented it... |
| 00:36 | <MikeSmith> | yeah, Opera has too, I think |
| 00:36 | <MikeSmith> | but Mozilla and Microsoft IE have no plans to implement it |
| 00:42 | <llrcombs> | the localStorage API |
| 00:43 | <llrcombs> | does it support a table-ish method of doing this? |
| 00:46 | <karlcow> | http://john.jubjubs.net/2010/05/11/whats-next-for-me-but-not-yet/ |
| 00:51 | <MikeSmith> | llrcombs: no, the localstorage api does not have an built-in support like that |
| 00:51 | <MikeSmith> | you can layer something on top of it in your code, though |
| 00:51 | <MikeSmith> | or there may already be some support for that in whatever JS libraries |
| 00:52 | <MikeSmith> | llrcombs: there is also active work going on around a different, non-SQL client-side DB mechanism |
| 00:53 | <Dashiva> | Is there a document that helps developers decide which storage mechanism is the right one for them? |
| 01:22 | <AryehGregor> | There's no way to say something like "fire a script every time any <img> loads", is there? Other than by actually adding an onload handler to every <img> in the HTML source? |
| 01:23 | <AryehGregor> | Or by putting a <script> right after every <img>, or whatever? |
| 01:23 | <AryehGregor> | I mean, so I don't have to alter the HTML markup. |
| 01:31 | <Philip`> | AryehGregor: Add a capturing event listener on the body |
| 01:32 | <AryehGregor> | Oh, cool. Is that interoperably supported? |
| 01:32 | <Philip`> | I don't see why it wouldn't be |
| 01:32 | <Philip`> | (ignoring IE) |
| 01:32 | <Philip`> | and at least it works in Opera |
| 01:33 | <AryehGregor> | Too bad I have to deal with IE. :) |
| 01:33 | <Philip`> | It should work in IE9 :-) |
| 01:33 | <AryehGregor> | Okay, I'll make a note on my calendar to start using it in 2020. |
| 01:35 | <Philip`> | (Maybe IE's event functions have similar features, but I have no idea about that) |
| 01:44 | <boblet> | morning all. anyone ever marked up musical scores? |
| 01:52 | <boblet> | hrm. I’ll check the logs in case anyone has the answer |
| 01:52 | <boblet> | bbl |
| 03:03 | <othermaciej> | this is super sexy: http://0xfe.blogspot.com/2010/05/music-notation-with-html5-canvas.html |
| 03:04 | <othermaciej> | now I want to see the interactive version that lets you edit a score live |
| 03:07 | <AryehGregor> | I'm not a music expert, but the demo seems incorrect. Aren't the little musical symbols supposed to mostly be on the lines, not scattered in random places overlapping each other? |
| 03:08 | <othermaciej> | looks right to me |
| 03:08 | <AryehGregor> | By the way, for those who remember my doctype woes: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/66254 |
| 03:08 | <doublec> | looks right to me too |
| 03:08 | <AryehGregor> | What browsers? I'm on Chrome dev channel. |
| 03:08 | <roc> | I don't know it's right, but it looks cool |
| 03:08 | <othermaciej> | Safari 4.0.5 and WebKit trunk |
| 03:08 | <roc> | Firefox trunk |
| 03:09 | <othermaciej> | it's definitely correct |
| 03:09 | <othermaciej> | I can read music |
| 03:09 | <AryehGregor> | Maybe they are supposed to be randomly floating around like that, then. |
| 03:09 | <AryehGregor> | I can't. :) |
| 03:09 | <othermaciej> | I mean, correct is relative |
| 03:09 | <othermaciej> | some of the notes are supposed to be above or below the staff |
| 03:10 | <doublec> | I look forward to a tool like 'guitar pro' being available as a web app |
| 03:10 | <roc> | it would be slightly nicer if it output SVG, but whatever |
| 03:11 | <othermaciej> | so you could scale your score? |
| 03:11 | <roc> | yeah, and for printing |
| 03:13 | <othermaciej> | some of the guitar tablature doesn't match the notes |
| 03:24 | <Hixie> | that is indeed correctly written music |
| 03:24 | <Hixie> | syntactically correct, that is |
| 03:25 | <Hixie> | well, almost |
| 03:25 | <othermaciej> | would the TAB not matching the score count as a syntax error or a semantic error? |
| 03:26 | <Hixie> | the last measure of the first line is short and missing an end marker |
| 03:26 | <othermaciej> | some of it does match, so I don't think the mismatch is deliberate |
| 03:28 | <Hixie> | oh, it's just line wrapped half-way through a bar |
| 03:32 | <Hixie> | still, not at all bad for a demo |
| 03:55 | <othermaciej> | yeah, it's quite nice |
| 03:55 | <othermaciej> | the note shapes look good |
| 05:10 | <boblet> | well that’s inconvenient. logs are down so if anyone gave me info about marking up musical scores I’d appreciate a repeat |
| 05:21 | <erlehmann> | boblet, lilypond ? |
| 05:23 | Hixie | ponders how to make timed tracks load |
| 05:24 | <boblet> | erlehmann: hmm, thanks. so this is a custom notation for use by Lilypond software, rather than a way to display music scores on web pages right? (if you know) |
| 05:25 | <erlehmann> | boblet, the lilypond authors have written a paper about musical notation. i recommend you read it, it explains why e.g. XML is unsuited for that purpose. |
| 05:29 | <Hixie> | the lilypond guys are basically the world leaders in music typography as far as i can tell |
| 05:29 | <boblet> | erlehmann: do you have a link? I see it can output SVG though, that’s good |
| 05:31 | <erlehmann> | boblet, http://lilypond.org/about/automated-engraving/ |
| 05:32 | <erlehmann> | read everything, be enlightened |
| 05:32 | <erlehmann> | or, wait. start here: http://lilypond.org/about/automated-engraving/problem-statement |
| 05:32 | <boblet> | hehe, read about/intro, was reading about/faq… |
| 05:33 | <boblet> | thanks |
| 05:49 | <dannytran> | hi everyone ... question ... i'm writing a web socket server ... and for some reason google chrome's WebSocket connection is not sending 8 random bytes |
| 05:49 | <dannytran> | what am I doing wrong? |
| 05:49 | <dannytran> | it's not sending Sec-WebSocket-Key1 or Sec-WebSocket-Key2 either |
| 05:51 | <erlehmann> | dannytran, i do not know about websocket. why is this random byte thingy needed ? |
| 05:51 | <dannytran> | http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol-76#page-32 |
| 05:52 | <dannytran> | html5? |
| 05:53 | <Hixie> | dannytran: chrome doesn't yet implement the new handshake |
| 05:53 | <Hixie> | dannytran: tis still very early days for websocket |
| 05:53 | <dannytran> | ahh |
| 05:54 | <dannytran> | so where would i find documentation on how chrome does handle web sockets |
| 05:55 | <Hixie> | i recommend waiting a few weeks until they update to the new handshake |
| 05:56 | <Hixie> | but for the server the old handshake is basically the same as the new one, just without the key part |
| 05:56 | <Hixie> | and you can ignore the client's handshake |
| 05:56 | <dannytran> | Hixie: very cool ... thank you |
| 05:56 | <Hixie> | np |
| 05:56 | Hixie | is waiting for chrome to update too |
| 06:03 | <nog_lorp> | The major httpd vendors should work together to ban old browsers from the internet |
| 06:03 | <ment> | like lynx and stuff |
| 06:03 | <nog_lorp> | Serve anyone with a browser older than 5 years old a sorry error |
| 06:04 | <nog_lorp> | yeah, and tell them a UA they can spoof to get by |
| 06:04 | <nog_lorp> | if they really want it |
| 06:04 | <nog_lorp> | since that is the bottom line of detection |
| 06:04 | <nog_lorp> | the point is to oust the entreched IE4 population |
| 06:06 | <nog_lorp> | And there are modern builds of lynx I think |
| 06:06 | <nog_lorp> | last stale build was 10 months ago |
| 06:06 | <nog_lorp> | stable* |
| 06:12 | <dannytran> | Hixie: sorry to bother you again ... but which version of the protocol should i expect google chrome to be using? http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol-75#page-25 |
| 06:15 | <nog_lorp> | I say screw security and give javascript access to raw sockets |
| 06:16 | <nog_lorp> | ooh maybe that'll be doable threw the device tag? |
| 06:19 | <MikeSmith> | http://www.pcworld.com/article/196023/facebook_privacy.html |
| 06:19 | <MikeSmith> | dude has an inspiring bio |
| 06:19 | <MikeSmith> | search in that page for "Author Dan Tynan" |
| 06:22 | <boblet> | MikeSmith: you could use that bio yourself! |
| 06:32 | <wirepair> | MikeSmith nice article, i love the blue graphs ;> |
| 08:03 | <annevk> | MikeSmith, heh, you're too kind :) |
| 08:33 | <MikeSmith> | annevk: hopefully you can get back for another visit sooner rather than later |
| 08:40 | <MikeSmith> | I'm trying http://code.bocoup.com/audio-data-api/examples/ambient-extraction-mixer/ in latest Minefield but I don't get any sound at all when I try it |
| 08:41 | MikeSmith | wonders if boaz is still around |
| 08:42 | <MikeSmith> | ah, OK, I see I need a special build |
| 08:58 | <annevk> | is the IE blog just trying very hard to sound silly with their "same markup"? |
| 09:04 | <MikeSmithX> | annevk: I guess they're just trying to coin a term and see if others start to pick it up |
| 09:05 | <MikeSmith> | of course it's not an accurate term |
| 09:06 | <othermaciej> | is that their buzzword attempt to stamp out IE-specific code paths? |
| 09:06 | <MikeSmith> | yeah |
| 09:06 | <MikeSmith> | seems so |
| 09:06 | <othermaciej> | silly buzzword, good motive |
| 09:07 | <annevk> | at some point in blogging history we used to have debates on "tag" vs "element" |
| 09:07 | <annevk> | it seems everyone has moved on from that or otherwise there'd be huge outcry over "same markup" |
| 09:10 | <ment> | what does "same markup" mean? |
| 09:10 | <othermaciej> | people are ok with using "HTML5" to refer to anything cool you can do on the Web without plugins |
| 09:12 | <karlcow> | HTML5 is the new black for now, and will become the new hell in a few years from now (cf XHTML). Marketing/Branding Cycles. |
| 09:12 | <annevk> | ment, ask Microsoft |
| 09:12 | <othermaciej> | I dunno - "Web 2.0" and "AJAX" both still get plenty of pla, and they are years out of date |
| 09:13 | <othermaciej> | ment: I think it's supposed to mean "you don't need an IE-specific code path any more cause IE supports standards and stuff" |
| 09:15 | <karlcow> | othermaciej: I start to see/hear "Web 2.0" as a forbidden term in commercial propositions for websites. Or maybe my own working environment distortion field |
| 09:16 | <karlcow> | http://www.google.com/trends?q=html5%2C+xhtml%2C+web+2.0 |
| 09:16 | <othermaciej> | http://www.google.com/trends?q=html5%2C+xhtml%2C+web+2.0%2C+ajax&ctab=0&geo=all&date=all&sort=0 |
| 09:16 | <othermaciej> | I think most of the ajax hits are non-web |
| 09:18 | <karlcow> | ajax: cleaning product, football team, greek mythology, etc |
| 09:18 | <mut> | whoop |
| 09:18 | <karlcow> | very common indeed http://en.wikipedia.org/wiki/Ajax |
| 09:18 | <mut> | I actually finished this sodding website. |
| 09:18 | <ment> | othermaciej: i assume that means something different than "we support standards" |
| 09:19 | <ment> | also as i'm reading the ie blog, what does GPU-powered HTML5 mean, other than running h.264 decoder in hw? |
| 09:19 | <ment> | s/in hw/on gpu/ |
| 09:20 | <othermaciej> | I believe they mean that their browser engine will do drawing using GPU-accelerated APIs |
| 09:20 | <othermaciej> | though of course most drawing is not HTML per se |
| 09:28 | <ment> | hmm, for a moment i was excited about the idea of microsoft's parallelizing javascript compiler |
| 09:29 | <Peter`> | ment: what changed your mind? |
| 09:33 | <ment> | Peter`: video cards ten to crash often when running CUDA code and i don't really want my os to be taken down by some compiled js script |
| 09:34 | <Peter`> | From what I understand they parallize compiling (while running the interpreter) on multiple CPU cores, not using the GPU/CUDA |
| 09:37 | <ment> | Peter`: i think that's different kind of parallelization than i have in mind - they probably just run multiple javascript states on multiple cores, not one javascript state on multiple cores concurrently |
| 09:38 | <Peter`> | I understand, but where exactly does CUDA come around? Do they actually compile (bits) of JS code to CUDA? |
| 09:40 | <ment> | i've imagined that ie would compile js with parallelizing compiler into cuda code and run on GPU when i heard about "GPU-accelerated HTML5", but that's a silly idea |
| 09:43 | <hsivonen> | It's really more like GPU-accelerated SVG and CSS painting |
| 09:43 | <hsivonen> | but HTML5 is cooler than SVG or CSS |
| 09:43 | <hsivonen> | as a marketing buzzword that is |
| 09:45 | <hsivonen> | "First, SVG is part of the HTML 5 family of technologies." -- http://googlecode.blogspot.com/2009/10/svg-at-google-and-in-internet-explorer.html |
| 09:45 | <Peter`> | With various companies claiming "HTML5" and "CSS3" support combined with a really small percentage of people who actually understand it's fairly hard to stop that |
| 09:45 | <Peter`> | It does put more weight behind development of the more prominent features however, so I recon it brings good things as well |
| 09:46 | <annevk> | back then still with a space |
| 09:47 | <hsivonen> | OTOH: https://twitter.com/bradneuberg/status/13161144493 |
| 09:48 | <hsivonen> | SVG is lacking a cool number like 3 or 5 |
| 09:49 | <hsivonen> | It's well-known that 1.1 and 2 are not cool numbers :-) |
| 09:49 | <ment> | if any, it definitely has to be an odd number, preferably prime |
| 09:49 | <karlcow> | Peter `: first people will be happy. HTML5 wiiiiz effect, then they will realize there are many holes, interoperability issues creating arguments with their clients, etc. and because there is only one word even if it's about a specific part, they will start to hate it. |
| 09:51 | <Peter`> | karlcow: Quite likely indeed, in the end we might even end up with a "good browser" being defined as "the browser who most advertised their HTML5 compliance" |
| 09:51 | <zcorpan_> | what's happened to the logs lately? |
| 09:51 | <zcorpan_> | krijnh? |
| 09:53 | <ment> | btw does html5 specify where are line-breaks allowed when layouting/rendering? (for example, where are line breaks allowed in "some text<b>in bold</b>the end"?) |
| 09:53 | <hsivonen> | ment: that belongs in the CSS land |
| 09:53 | <annevk> | oh hey, <track> is no longer in the W3C copy, saves some editing work in html5-diff I guess... |
| 09:54 | <zcorpan_> | i thought the chairs only wanted websrt removed |
| 09:54 | <annevk> | zcorpan_, according to krijn we're still being logged but the server changed IP and DNS was not updated or something |
| 09:54 | <zcorpan_> | ok |
| 09:55 | <ment> | hsivonen: css is not particulary verbose about line breaks either (Note. CSS 2.1 does not fully define where line breaking opportunities occur.) |
| 09:55 | <krijn> | Ah, my IP changed, grmbl |
| 09:55 | <zcorpan_> | ment: maybe you need to reverse engineer browsers and write a spec |
| 09:56 | <hsivonen> | ment: it belongs in CSS but CSS hasn't quite drank the "well-defined" kool aid to the fullest yet |
| 09:56 | <annevk> | zcorpan_, thought so too, but it seems Hixie just took the whole thing out |
| 09:56 | <zcorpan_> | ment: i think unicode has some rules about line breaking opportunities also |
| 09:56 | annevk | wonders if a html5-diff based on the WHATWG spec would be more useful |
| 09:56 | hsivonen | mumbles about hyphenation and idolizing performance |
| 09:57 | <annevk> | ment, there's Unicode TR #14 iirc |
| 09:57 | <annevk> | my memory is pretty good: http://unicode.org/reports/tr14/ |
| 09:58 | <annevk> | but then I believe it's not very good and most vendors deviate from it intentionally |
| 09:58 | <zcorpan_> | so we need Line Breaking 5 |
| 09:59 | <annevk> | zcorpan_, so far consensus is that line breaking is something to compete on |
| 09:59 | zcorpan_ | is fine with thtat |
| 09:59 | <ment> | ah this, i'm not interested in hyphenation/word-breaks but rather html/css rules regarding breaking of inline boxes with decoration and stuff like that |
| 09:59 | <krijn> | DNS is being changed, logs will be up in no time! o/ |
| 10:00 | <annevk> | ment, that should be defined in CSS |
| 10:00 | <karlcow> | I do not |
| 10:00 | <karlcow> | understand the hate |
| 10:00 | <karlcow> | about line breaks. |
| 10:01 | <annevk> | that is |
| 10:01 | <annevk> | because you |
| 10:01 | <annevk> | are French |
| 10:01 | <karlcow> | ;) |
| 10:02 | <annevk> | ment, and afaik is |
| 10:03 | <ment> | maybe i should read more between the lines then |
| 10:04 | <jgraham> | I hate Hixie |
| 10:04 | <hsivonen> | jgraham: that's a bit strong. what did Hixie do now? |
| 10:04 | <jgraham> | I have just come across some of his old test cases |
| 10:05 | <jgraham> | main identifying feature: astrophy |
| 10:05 | <jgraham> | They are... not pretty |
| 10:05 | <jgraham> | :) |
| 10:06 | <hsivonen> | jgraham: at least Hixie usually keeps his old test cases around, which is much nicer than finding the whole host name gone |
| 10:06 | <jgraham> | hsivonen: Well yes. Hopefully these ones we imported internally (even Hixie's server goes down sometimes) |
| 10:13 | <hsivonen> | I haven't read the whole timed text thread yet |
| 10:13 | <hsivonen> | is Sean Hayes articulating an IE-oriented position or a Microsoft media framework-oriented position? |
| 10:14 | <MikeSmith> | hsivonen: I think Sean is not part of the IE product-dev team |
| 10:14 | <annevk> | nobody really knows, but he doesn't work for the IE Team that's for sure |
| 10:14 | <Philip`> | Seems to be more of a person-heavily-involved-in-defining-TTML oriented position |
| 10:14 | <roc> | yeah |
| 10:15 | <MikeSmith> | Frank Olivier is part of the IE team |
| 10:15 | <hsivonen> | thanks |
| 10:15 | <othermaciej> | he is listed as one of the authors of TTML and says he is the co-chair of the Timed Text WG |
| 10:15 | <hsivonen> | I think the an IE-oriented view is much more relevant than a person-heavily-involved-in-defining-TTML view |
| 10:16 | <othermaciej> | I don't know to what extent he is representing that position and how much the IE position |
| 10:16 | <hsivonen> | s/the// |
| 10:17 | <Philip`> | A person-heavily-involved-in-defining-TTML view seems useful if you want to know details about TTML or about its history and design decisions |
| 10:18 | <Philip`> | though probably not so much if you want objective comparisons with other formats |
| 10:18 | <jgraham> | It seems less useful if they often say things that are contradicted by the text of the spec |
| 10:19 | <othermaciej> | his claims about the relation to CSS and XSL-FO in context seem like puffery |
| 10:20 | <othermaciej> | in the legal sense |
| 10:20 | <othermaciej> | "Puffery as a http://en.wikipedia.org/wiki/Legal_term refers to http://en.wikipedia.org/wiki/Advertising and claims that express http://en.wikipedia.org/wiki/Subjectivity rather than http://en.wikipedia.org/wiki/Objectivity_(journalism) views, such that no reasonable person would take them literally." |
| 10:21 | <othermaciej> | damn you Colloquy for expanding links! sometimes I just wanna paste plain text |
| 10:22 | <hsivonen> | instead of tacit knowledge, perhaps W3C spec editors should get an actual memo warning them about depending on XML Schema or XSL-FO in Web specs |
| 10:23 | <MikeSmith> | it would be useful to find out what subset of TTML is supported in Flash |
| 10:23 | <hsivonen> | or xml:id or XML Events or XML 1.1 |
| 10:24 | <hsivonen> | MikeSmith: and what the actual layout model is there |
| 10:24 | <MikeSmith> | yeah, that too |
| 10:24 | <MikeSmith> | roc asked about both, iirc |
| 10:24 | <MikeSmith> | and got only a response about TTML support in Silverlight |
| 10:24 | <roc> | that's fair, in a sense |
| 10:25 | <roc> | obviously Larry Masinter should be the one to tell us about TTML in Flash |
| 10:25 | <MikeSmith> | well, it'd be great to have a response from somebody in the group who had specific product knowledge about Flash support |
| 10:25 | <roc> | where's he gone, anyway? |
| 10:26 | <Philip`> | It'd be more great to have test cases to verify their claims |
| 10:26 | <annevk> | anyone else thought of CSS XBL vs SVG XBL, XSL:FO vs CSS, Selectors vs XPath, during this WebSRT vs TTML chatter on the mailing list? |
| 10:26 | <hsivonen> | annevk: yes |
| 10:27 | <MikeSmith> | it would also be great to get a response to the point that roc made about the layout mechanism not being implementable in existing browser layout engines |
| 10:28 | <roc> | Philip`: wanna write some? |
| 10:28 | <roc> | I don't! |
| 10:28 | <othermaciej> | annevk: I kind of would like the HTML WG to be free of format wars to the degree possible :-/ |
| 10:29 | <othermaciej> | but yes, this is reminiscent of many past ones |
| 10:29 | <othermaciej> | one could also cite Microdata vs RDFa |
| 10:30 | <hsivonen> | othermaciej: it seems to me we walk from one format war to another all the time |
| 10:30 | <annevk> | the problem is that there's no real other W3C group with the same amount of browser expertise apart from maybe the WebApps WG |
| 10:30 | <hsivonen> | othermaciej: except MathML inclusion worked nicely |
| 10:30 | <othermaciej> | how about SVG inclusion? |
| 10:30 | <jgraham> | hsivonen: Well I objected to MathML, but not strongly or anything :) |
| 10:30 | <othermaciej> | annevk: what about CSS WG? |
| 10:30 | <annevk> | and walking away from things in the HTML WG doesn't usually lead to a conclusion elsewhere |
| 10:31 | <hsivonen> | othermaciej: bygones, kumbaya |
| 10:31 | <othermaciej> | I guess SVG vs Canvas was a format war for a while, though that one is particularly crazy |
| 10:32 | <roc> | I think that's died down |
| 10:32 | <othermaciej> | yeah, I think people have accepted that it doesn't need to be a versus situation |
| 10:32 | <roc> | yeah |
| 10:32 | <roc> | the CSS WG is not bad |
| 10:33 | <annevk> | it's not too bad, though the amount of times certain decisions are overturned... |
| 10:33 | <othermaciej> | I would like to say that ECMA TC-39 has similar levels of browser expertise but I don't think it is up to the level of the three aforementioned W3C WGs |
| 10:34 | <othermaciej> | nor do any of the IETF WGs working on highly browser-relevant protocols |
| 10:34 | <annevk> | TC39 is not really addressing platform needs |
| 10:34 | <annevk> | I've been waiting for four years or so now for octet representation |
| 10:34 | <othermaciej> | TC39 has too many of what my colleague Geoff terms "language goobers" |
| 10:35 | <othermaciej> | and not enough people tuned into non-JS-engine parts of the browser |
| 10:35 | <othermaciej> | heck, barely enough people tuned into actual JS engine implementations |
| 10:35 | <jgraham> | The bytearray thing is really really urgent |
| 10:35 | <othermaciej> | as opposed to language theory |
| 10:35 | <othermaciej> | I wish I either had time to work on it or order someone to do so |
| 10:35 | <jgraham> | The ECMA process is also totally broken |
| 10:36 | <othermaciej> | hmmm... maybe I can trick Oliver into it |
| 10:36 | <annevk> | best IETF group so far seems the cookie stuff |
| 10:36 | <othermaciej> | the ECMA process involves the spec going final before there are any implementations |
| 10:36 | <annevk> | mostly thanks to Adam Barth |
| 10:36 | <othermaciej> | which seems absurd by modern standards |
| 10:36 | <hsivonen> | I'm slightly unhappy that TC39 didn't spec the Web compat stuff that's chronicled on the WHATWG wiki |
| 10:36 | <jgraham> | othermaciej: bingo |
| 10:36 | <othermaciej> | I pointed this out and suggested we try to have something more like a "Candidate Rec" / "Draft Standard" stage |
| 10:37 | <othermaciej> | but no one really bought in |
| 10:37 | <annevk> | though even they turned away from addressing certain issues (e.g. publicsuffix) |
| 10:37 | <jgraham> | hsivonen: TC39 have an unshakable belief that old parts of the platform will die off and can be unspecced or even incorrectlyt specced (cough, octals, cough) |
| 10:37 | <othermaciej> | jgraham: I think degree of belief in this varies |
| 10:38 | <othermaciej> | I believe none of the Apple people who participate (mainly olliej and me) are particularly convinced |
| 10:38 | <othermaciej> | Brendan is half-convinced |
| 10:38 | <jgraham> | othermaciej: It seems to be held by a significant number of major stakeholders, at least |
| 10:38 | <othermaciej> | the secure subset / object capability people are quite convinced |
| 10:38 | <jgraham> | Brendan seems pretty convinced |
| 10:39 | <othermaciej> | on some issues he takes that tack, on others not so much |
| 12:04 | <annevk> | hsivonen, it matters somewhat when including WebSRT in the media stream |
| 12:11 | annevk | wonders why Julian didn't reply to his reply |
| 12:18 | hsivonen | wonders how much the Archos thingies with Opera on them cost |
| 12:21 | <hsivonen> | whoa. almost 400 euros |
| 12:23 | <hsivonen> | it seems rather weak to argue for XSL by referring to W3C spec maturity levels |
| 12:23 | <hsivonen> | or to argue for anything at all by referring to the W3C spec maturity levels |
| 12:24 | <annevk> | catching up with email? |
| 12:24 | <jgraham> | Indeed. W3C spec maturity levels are a red herring |
| 12:24 | <annevk> | that thread is funny |
| 12:24 | <hsivonen> | annevk: yeah, I still have a lot unread list email |
| 12:25 | <hsivonen> | I might take the maturity levels more seriously if all WGs required two interoperable implementations *in browsers* and were serious about test suites |
| 12:27 | <Lachy> | which thread? |
| 12:28 | <hsivonen> | Lachy: timed text |
| 12:55 | <hsivonen> | MikeSmith: I'm about to land validator support for HTML5+ARIA+SVG+MathML as the default. are you OK with it becoming the default in the W3C instance of V.nu, too? |
| 12:59 | <MikeSmith> | hsivonen: yep |
| 12:59 | <hsivonen> | ok |
| 12:59 | <MikeSmith> | I think in general the W3C instance should remain in sync with v.nu upstream |
| 13:12 | <zcorpan_> | http://twitter.com/erikdahlstrom/statuses/13841075917 |
| 13:13 | <jgraham> | "Warning: this may kill your browser, as it's quite an unoptimised script." |
| 13:13 | <jgraham> | Seems like an unfair comparison |
| 13:16 | <jgraham> | (although it would be surprising if SVG filters, implemented in C++ and potentially GPU accelerated weren't faster than javascript manipulating pixels on a canvas) |
| 13:16 | <Philip`> | That canvas demo looks like it'll fail if getImageData doesn't return 1 pixel per CSS pixel |
| 13:17 | <Philip`> | Hmm, it even loops over x/y in the wrong order |
| 13:17 | <Philip`> | (Wrong in terms of performance) |
| 13:22 | <hsivonen> | argh. SVN keeps rejecting my login |
| 13:23 | <hsivonen> | I reset my password. twice. still rejecting the new password |
| 13:23 | <zcorpan_> | SVN is having a break eating ice cream |
| 13:25 | <hsivonen> | works on Mac from within Eclipse |
| 13:25 | <hsivonen> | Hmm. MikeSmith left :-( |
| 13:25 | <hsivonen> | IIRC, Mike had experienced trouble with debian svn |
| 13:26 | <hsivonen> | maybe the Ubuntu package inherits brokenness from Debian |
| 13:26 | <annevk> | it's 9:30PM in Tokyo |
| 13:26 | <jgraham> | MikeSmith is barely awake then |
| 13:26 | <annevk> | guess he might be online later due to weird telcon hours |
| 13:26 | <annevk> | jgraham, he's awake surprisingly often |
| 13:33 | hsivonen | finds http://krijnhoetmer.nl/irc-logs/whatwg/20090603#l-626 |
| 13:38 | <hsivonen> | command-line svn on Mac worked, too |
| 13:39 | <hsivonen> | though I got a post-commit hook error from the server |
| 13:40 | <hsivonen> | ok. I guess this was a server-side problem |
| 13:40 | <hsivonen> | not I managed to commit from Ubuntu |
| 13:40 | <hsivonen> | and I got the same post-commit hook error |
| 13:45 | <jgraham> | annevk: No, I meant has barely woken up yet :) |
| 13:45 | <jgraham> | He seems to be a night owl |
| 13:46 | <annevk> | fyi, I was in Tokyo the last three weeks staying at his place |
| 13:47 | <hsivonen> | there's something scary and frustrating about being locked out of version control |
| 13:48 | <jgraham> | annevk: I know |
| 13:54 | <hsivonen> | validator.nu went down for a libc update |
| 13:55 | <hsivonen> | up again |
| 13:57 | <hsivonen> | time zones are a huge FAIL |
| 13:57 | <hsivonen> | it's ridiculous how often the tzdata package gets updated |
| 13:58 | <hsivonen> | OK. validator.nu and html5.validator.nu now support HTML5+ARIA+SVG+MathML |
| 13:59 | <zcorpan_> | yay |
| 13:59 | <zcorpan_> | http://validator.nu/?doc=http%3A%2F%2Fhsivonen.iki.fi%2Ftest%2Fmoz%2Fhtml5-hacks-demo.html&showsource=yes |
| 13:59 | <hsivonen> | MikeSmith: I landed HTML5+ARIA+SVG+MathML |
| 14:00 | <MikeSmith> | hsivonen: cool |
| 14:00 | <MikeSmith> | hsivonen: you planning to blog about it? |
| 14:00 | <zcorpan_> | hsivonen: selecting the preset selects the xml parser |
| 14:00 | <hsivonen> | MikeSmith: yes |
| 14:01 | <hsivonen> | zcorpan_: doh. I forgot to update the .js file |
| 14:01 | <hsivonen> | zcorpan_: thanks |
| 14:36 | <hsivonen> | zcorpan: Fixed the preset. |
| 14:37 | <MikeSmith> | hsivonen: everything checked in? |
| 14:38 | <hsivonen> | MikeSmith: everything except script.js |
| 14:38 | <MikeSmith> | OK, I'll wait for that before syncing up the validator.w3.org instances |
| 14:43 | <hsivonen> | MikeSmith: committed |
| 14:43 | <MikeSmith> | thanks |
| 15:07 | <karlcow> | http://html5readiness.com/ |
| 15:12 | <jgraham> | I have never understood the colour sceheme on that site |
| 15:12 | <jgraham> | Firefox should be orange OPera red (it is, good), Chrome and IE should fight over blue and Safari should be grey |
| 15:13 | <hsivonen> | paul_irish: how about adding HTML5 parsing? |
| 15:13 | <jgraham> | Also including Opera 10.10 but IE 9 seems odd |
| 15:13 | <paul_irish> | :) to html5 readiness? |
| 15:13 | <hsivonen> | paul_irish: right |
| 15:14 | <paul_irish> | that's a good call. |
| 15:15 | <tabatkins> | That site yours, paul_irish? |
| 15:15 | <paul_irish> | yup |
| 15:15 | <jgraham> | (actually IE shoud clearly be blue. Dunno what to do with Chrome) |
| 15:16 | jgraham | wonders if people with synesthesia get this feeling of colours being wrong all the time |
| 15:17 | <tabatkins> | IE can take dark blue, we'll take pastel. |
| 15:17 | <Dashiva> | Safari should be green, obviously :P |
| 15:17 | <paul_irish> | its primary intent is to show developers that html5/css3 is ready based on features, not specs. to try and defeat the "i'll wait till it's done" mentality |
| 15:18 | <tabatkins> | I really love the immediate visual impact of clicking between the years. |
| 15:18 | <gsnedders> | Oh, awesome. The new equality minister in Britain has voted against almost all bills for equality with homosexuals in the past decade… |
| 15:18 | <tabatkins> | Perhaps he wants homosexuals to be better than heteros, and equality would be holding them back? |
| 15:19 | <tabatkins> | Hmm. Chrome is insufficiently fast in running these transitions. |
| 15:19 | <Dashiva> | He doesn't want them to suffer the horror that is married life |
| 15:19 | <jgraham> | Oh. I take it he's a Tory then |
| 15:19 | <Dashiva> | It's an act of mercy |
| 15:20 | <gsnedders> | jgraham: Indeed. |
| 15:20 | <karlcow> | jgraham: plus it's confusing because the position makes it harder to read, you are tempted to believe that a circle is one browser until you realize not |
| 15:20 | <jgraham> | Maybe the Lib Dems can beat some sense into them |
| 15:21 | <tabatkins> | karlcow: That just implicitly gives Chrome an edge, so it's okay. ^_^ |
| 15:21 | gsnedders | hopes it doesn't go the way of the Labour/LibDem coalition in Scotland from 1997–2005 |
| 15:21 | <gsnedders> | *2007 |
| 15:22 | gsnedders | is amazed that there is a coalition between Tory/LibDem, but that's another matter |
| 15:23 | <gsnedders> | And a referendum on electral reform? Wow. |
| 15:23 | <gsnedders> | I didn't expect that from the Tories |
| 15:24 | <Dashiva> | Well, it was either that or minority government |
| 15:24 | <jgraham> | Neither did I |
| 15:24 | <jgraham> | But who knows, maybe it will work |
| 15:24 | gsnedders | is also of the understanding that the Lib Dems could not enter a coalition officially until after a poll of all members |
| 15:24 | <gsnedders> | (like, the parlimentary party had to approve it, then the entire party…) |
| 15:25 | <jgraham> | I mean I expect to hate all the Tory policies, but I hope to like many of the Lib Dem ones |
| 15:25 | <gsnedders> | Trident may be replaced. Yay. |
| 15:25 | <gsnedders> | (</sarcasm>) |
| 15:26 | <tabatkins> | Aw, gsnedders, you got me all excited about MS doing something crazy. |
| 15:26 | <gsnedders> | Heh. |
| 15:27 | <gsnedders> | Why do I fail to be surprised under the coalition that the Scottish Secretary (which is now mostly a title and little more) is held by the smaller party? |
| 15:27 | Lachy | looks up what the Tories are. Finds... |
| 15:27 | <Lachy> | "The term was thus originally a term of abuse, "an Irish rebel", before being adopted as a political label in the same way as Whig." -- wikipedia |
| 15:28 | <Lachy> | ... then looks up Whig. |
| 15:28 | <Lachy> | What is it with you british giving your political parties such weird names? |
| 15:30 | <Philip`> | People here 350 years ago spoke funny |
| 15:31 | <jgraham> | Or, depending on your point of view, people today do |
| 15:34 | <Philip`> | That's hard to verify, unless you can find a 350 year old person to ask |
| 15:34 | <Philip`> | Maybe they would recognise our modern language as a clear improvement over theirs |
| 15:35 | <hsivonen> | http://intertwingly.net/slides/2010/ws-rest/theses-4 |
| 15:35 | <hsivonen> | I wonder if rubys said something about point #4 that isn't obvious from the slides |
| 15:35 | <Lachy> | Philip`, I think some british people speak funny. Particularly the Welsh and Scottish. |
| 15:36 | <karlcow> | Philip`: not possible british would think that american is an improvement :p |
| 15:38 | <gsnedders> | Lachy: Ah dinnae ken if ya cannae oonderstoond |
| 15:41 | <Lachy> | gsnedders, I can usually understand spoken scottish. But it gets a little harder to figure out what "[I] dinnae ken [if you cannot understand]" translates to |
| 15:41 | <gsnedders> | don't care |
| 15:42 | <Lachy> | wtf? Why would you put an 'n' sound in to "care"? |
| 15:42 | <gsnedders> | Because it's not trying to pronounce the word care. |
| 15:42 | <Lachy> | oh, then what word were you trying to pronounce? |
| 15:43 | <gsnedders> | ken. |
| 15:43 | <Lachy> | in english? |
| 15:43 | <gsnedders> | There is no English word. |
| 15:44 | <gsnedders> | http://en.wiktionary.org/wiki/ken#English |
| 15:44 | <gsnedders> | From the Old English cennan |
| 15:44 | <gsnedders> | It's not really care, though |
| 15:44 | <gsnedders> | But the actual usage of it is somewhat confusing |
| 15:47 | <tabatkins> | Anyone know linear programming? |
| 15:47 | <jgraham> | Ah, I thought it meant "know" |
| 15:47 | <jgraham> | Seems I wasn't entirely wrong |
| 15:50 | <jgraham> | tabatkins: No, but it looks interesting :) |
| 15:50 | <Philip`> | tabatkins: I prefer my programming to be planar |
| 15:51 | <tabatkins> | I'm checking in with the people at work. I'll bet there's a way to one-step the computation of flex units while satisfying min/max constraints, but all I know how to do right now is iterate and solve constraints as they are violated. |
| 15:52 | <tabatkins> | Anyway, off to grab breakfast and then go to class all day. I'll check back tomorrow. |
| 15:53 | <annevk> | what kind of class are you taking? |
| 16:03 | <boaz> | hey all, if you're in boston: http://bocoup.eventwax.com/the-future-of-web-audio |
| 16:51 | <jgraham> | Hmm, randomly closing websockets when resource limits are reached seems bad |
| 16:51 | <gsnedders> | Hey, it's perfectly conforming! |
| 16:52 | <jgraham> | Especially if it would allow a single page to take down all connections made by other pages |
| 16:52 | <jgraham> | s/page.browsing context/ if you prefer |
| 16:54 | <jgraham> | I'm not sure what a good solution would be. You could throw an error but people likwly don't check for that |
| 16:55 | <jgraham> | Or enqueue the connection and hope the application provides UI telling the user about the problem |
| 16:55 | <jgraham> | Which has the slight advantage that pages may start to work "by magic" by closing other pages |
| 16:57 | <Philip`> | Are there any stats on how many users never have more than one page open at a time? |
| 16:57 | <jgraham> | Philip`: Mozilla did some "number of tab" studies |
| 16:57 | <jgraham> | I think the distribution was bimodal at like 3 and 14 |
| 16:57 | <webr3> | can't think of anywhere better to ask, so is anybody aware of browser utils available to JS for crypto functions (for instance sign, seal/encrypt, open/decrypt using client certificates installed in the browser) similar to mozilla/gecko window.crypto ? |
| 16:57 | Philip` | wonders if normal people have a hundred background tabs open and are likely to suffer problems when some unidentified tab blocks the one they're trying to use |
| 16:57 | <jgraham> | but I could be wrong |
| 16:58 | <webr3> | other way around; is there any whatwg work done, mentioned / planned making a spec for js crypto functions? |
| 17:00 | <Philip`> | http://dubroy.com/blog/how-many-tabs-do-people-use-now-with-real-data/ ? |
| 17:00 | <Philip`> | Sounds like it asked for participants who "often use multiple tabs or windows" which would bias the results away from 1 |
| 17:49 | <hsivonen> | asmodai: IIRC, I promised to ping you when v.nu supports MathML in text/html. that would be now |
| 18:28 | <Hixie> | jgraham: yeah if you're dealing with the tests/evil/mixed test cases i apologise :-) |
| 18:28 | <Hixie> | those are ooold |
| 18:29 | <webr3> | widgets run in the browser yes? |
| 18:29 | <gsnedders> | Hixie: It wasn't, it was adhoc, I think |
| 18:29 | <Hixie> | gsnedders: oh dear |
| 18:29 | <gsnedders> | Hixie: He's also not at a computer now :P |
| 18:34 | hsivonen | wonders what http://software.hixie.ch/utilities/js/live-dom-viewer/saved/481 says in IE9 |
| 18:39 | <hsivonen> | isn't pretty much everything from the last few years under adhoc? |
| 18:39 | <Hixie> | not everything |
| 18:39 | <Hixie> | i put things that are more complicated or harder to justify under evil/ |
| 18:40 | <Dashiva> | Like acid3? :) |
| 18:41 | hsivonen | gestures in the general direction of http://hixie.ch/tests/adhoc/dom/level0/write/005.html |
| 18:42 | <Hixie> | pah, that's simple |
| 18:42 | Hixie | runs away |
| 18:42 | <hsivonen> | Hixie: I suspect the test case is bogus |
| 18:42 | <Hixie> | quite possible |
| 18:43 | <Hixie> | i have been quite lax in my maintenance of test cases |
| 18:45 | <hsivonen> | Hixie: specifically, the class attribute present stuff in 005a.js has a negation in script |
| 18:45 | <hsivonen> | so the test claim that a class attibute is present when it is absent |
| 18:46 | <Hixie> | if you e-mail me i'll add it to my large pile of things to fix in test suites |
| 18:46 | <gsnedders> | Can we send patches too and get it fixed more quickly? |
| 18:46 | <hsivonen> | Hixie: it should be already be in your queue |
| 18:46 | <Hixie> | but the odds of me getting to it this decade are near 0 |
| 18:46 | <Hixie> | k |
| 18:46 | <Dashiva> | Where is the class attribute supposed to come from? |
| 18:46 | <Hixie> | gsnedders: no, the time-consuming part is the review |
| 18:47 | <hsivonen> | Hixie: I guess I should arrange things so that my quarterly goals don't depend on that test then... |
| 18:47 | <Dashiva> | I can't find anything defining or setting @class in the main page, 005a, or 005b |
| 18:47 | gsnedders | guesses he should go home soon |
| 18:47 | <hsivonen> | Hixie: is 2010 the first or last year of this decade for you? |
| 18:48 | <Hixie> | first |
| 18:48 | <gsnedders> | wrong answer, kthxbai. |
| 18:48 | <hsivonen> | :-) |
| 18:49 | <Hixie> | gsnedders: it wouldn't make sense for the decade known as "the 90s" to not span 1990-1999 |
| 18:49 | <hsivonen> | Dashiva: indeed! |
| 18:50 | <AryehGregor> | And this decade is "the 10s"? |
| 18:50 | <AryehGregor> | Or is this decade the 201st decade of the calendar? |
| 18:50 | <Philip`> | This is the last year of the 201st decade |
| 18:51 | <Dashiva> | Is the main point of the test that dynamically added scripts should run after load => blow away the document? |
| 18:51 | <Philip`> | but the first year of the 10s decade |
| 18:52 | <AryehGregor> | Just like 2000 was the last year of the second millennium, but the first year of the 2000s. |
| 18:52 | <AryehGregor> | So it's perfectly fair to celebrate the new millennium in both 2000 and 2001. |
| 18:52 | <Philip`> | Indeed |
| 18:52 | <Dashiva> | But why bother when you can celebrate the _current_ millennium any time you want? |
| 18:53 | <Philip`> | though it's wrong to claim that the millennium you're celebrating in 2000 is the start of the 3rd millennium |
| 18:53 | <AryehGregor> | Well, why celebrate anniversaries of anything at all? |
| 18:53 | <Philip`> | AryehGregor: Because it's an excuse to party? |
| 18:53 | <Dashiva> | Because sheeple need excuses |
| 19:10 | <AryehGregor> | Why does <b> say authors "should" use better elements where appropriate, but <i> says authors "are encouraged to consider" using other elements where appropriate? |
| 20:50 | <svl> | hmm. rel="extension" - http://mozillalabs.com/jetpack/2010/05/12/indexing-and-auto-detecting-browser-extensions-on-the-web/ |
| 21:28 | <AryehGregor> | tabatkins, if you aren't familiar with how TeX glue works, I'd highly recommend reading chapter 12 of The TeXbook. It's like flexboxes, but very elegant and powerful, and a core part of how TeX works. For instance, left/right/center alignment is obtained by just putting horizontal glue on one or both sides of the line. There are probably some good ideas there. |
| 21:28 | <AryehGregor> | (Not to say that flexboxes aren't "very elegant and powerful", of course. :P) |
| 21:29 | <AryehGregor> | TeX is very cool. |
| 21:29 | <AryehGregor> | Not practical for variable-size everything like we have on the web, sadly. |
| 21:50 | <TabAtkins_> | AryehGregor: Is the TeXbook available online for free? |
| 21:51 | <AryehGregor> | TabAtkins_, it seems like the TeX source code is available. |
| 21:51 | <TabAtkins_> | Not really interested in reading Knuth's programming. ^_^ |
| 21:52 | <AryehGregor> | Google refuses to accept that "texbook" isn't a typo for "textbook". |
| 21:52 | <AryehGregor> | So that's unhelpful. |
| 21:53 | <AryehGregor> | Wikipedia search isn't as smart, so it just gives me a giant list of typos if I search Wikipedia for "texbook". |
| 21:54 | <pesla> | Anyone here on Colloquy? :) |
| 21:57 | AryehGregor | points pesla to othermaciej |
| 22:01 | <othermaciej> | wait, what? |
| 22:43 | <sicking> | othermaciej: how does safari's implementation of WebSQLDB deal with never-ending transactions? |
| 22:43 | <othermaciej> | sicking: what do you mean by never-ending? |
| 22:43 | <sicking> | othermaciej: i.e. if you just nest executeSql calls in a never-ending fashion? |
| 22:44 | <othermaciej> | you could chain the callbacks forever, yes |
| 22:44 | <othermaciej> | we don't do anything to prevent that |
| 22:44 | <othermaciej> | because the model is asynchronous, it doesn't block the UI or anything |
| 22:44 | <sicking> | right, it just prevents anyone else from getting access to the database |
| 22:44 | <othermaciej> | also it's hard to do accidentally without causing your transaction to error out |
| 22:44 | <sicking> | and presumably uses a whole lot of CPU :) |
| 22:44 | <othermaciej> | anyone using that database on that domain, yes |
| 22:45 | <sicking> | when/how does a transaction error out? |
| 22:45 | <othermaciej> | if you execute a sql statement that would produce an error |
| 22:45 | <othermaciej> | if you exceed the capacity of the database |
| 22:45 | <othermaciej> | you could chain adding and then removing the same item within a transaction |
| 22:45 | <othermaciej> | but you'd really have to go out of your way |
| 22:45 | <sicking> | or just chain "select * from foo", right? |
| 22:46 | <othermaciej> | I don't think I have ever actually seen an infinite transaction |
| 22:46 | <sicking> | yes, i agree it's hard to do this by accident |
| 22:46 | <othermaciej> | well, chaining a read-only transaction is possible but a bit less disruptive since it doesn't block other readers |
| 22:46 | <othermaciej> | anyway - I don't think we have ever seen anyone actually run into a never-ending transaction |
| 22:46 | <othermaciej> | and a web app that does so would only be screwing itself |
| 22:47 | <sicking> | I agree, this isn't a big problem at all |
| 22:47 | <sicking> | imho it's ok as long as it's not easy to do by accident |
| 22:48 | <othermaciej> | so short version, we don't try to do anything specific to address this and it hasn't been a problem |
| 22:49 | <sicking> | great |
| 23:41 | <othermaciej> | oh man another CORS flamewar? |
| 23:47 | <Dashiva> | Just the same one, I think |
| 23:54 | <webr3> | yeah but those headers in CORS need changed to the UMP ones | other than that I'm keeping my mouth shut |