| 01:37 | gsnedders | shakes head at trying to check one or two things about Lolita by googling |
| 01:37 | <gsnedders> | Half of what I'm getting is stuff like "Free YOUNG Lolita PORN MOVIES", which really isn't helpful. |
| 02:30 | <gsnedders> | My physics makes no sense. |
| 02:30 | <gsnedders> | Apparently t = 1/2g for any object to fall any distance |
| 02:34 | <karlcow> | gsnedders: g(earth gravity?) => m/s2, t(time) => s ;) |
| 02:35 | <gsnedders> | g = gravitational field strength |
| 02:37 | <jcranmer> | 1/2g * t^2 = d -> t = sqrt(2d/g) if I am not mistaken |
| 02:37 | <karlcow> | so a good rule of thumb, checking units on each side. It will give you an idea if you messed up |
| 02:37 | gsnedders | realizes he missed the sqrt in his calculation |
| 02:37 | <karlcow> | heh |
| 02:37 | <jcranmer> | karlcow: that's called dimensional analysis |
| 02:38 | <jcranmer> | best thing that I ever learned in physics |
| 02:38 | gsnedders | wonders why he was doing this anyway |
| 02:38 | <jcranmer> | now, if only that could help with my habit of randomly swapping signs |
| 02:38 | <karlcow> | jcranmer: aaaah sorry. I know nothing in physics. I'm really dumb. Just know only sex and poetry |
| 02:39 | gsnedders | is trying to write his English dissertation, so it doesn't seem overly useful |
| 02:44 | <gsnedders> | MikeSmith: Challenge: When is the second time Lolita seduces Humbert? |
| 02:45 | <karlcow> | hmmm lolita is still in my boxes |
| 02:45 | <karlcow> | I mean the book |
| 02:45 | <MikeSmith> | gsnedders: I think it was the time after the first time, but before the third time |
| 02:46 | <gsnedders> | MikeSmith: There is no third time. |
| 02:46 | gsnedders | can't remember what the second time is, just there is a second time |
| 02:47 | <karlcow> | gsnedders: how was it? |
| 02:47 | <gsnedders> | I have three Nabokov novels around me, and a Boyd criticism of Ada too |
| 02:47 | <gsnedders> | karlcow: how was what? |
| 02:48 | <Hixie> | man i'd forgotten how annoying mario galaxy is |
| 02:49 | <karlcow> | the second time :p |
| 02:49 | <gsnedders> | karlcow: "was it" in what way? I can't remember it! |
| 02:50 | <karlcow> | heh. people always remember the first time. Very rare the second time |
| 02:50 | <gsnedders> | Oh, duh |
| 02:50 | gsnedders | is half asleep |
| 02:53 | <gsnedders> | ``You chump,'' she said, sweetly smiling at me. ``You revolting creature. I was a daisy-fresh girl, and look what you've done to me. I ought to call the police and tell them you raped me. Oh you dirty, dirty old man.'' |
| 02:59 | <karlcow> | gsnedders: found? |
| 02:59 | <gsnedders> | no |
| 03:00 | gsnedders | wonders if it has any connection to 342 |
| 03:02 | <jcranmer> | Hixie: what made you remember |
| 03:02 | <Hixie> | playing it |
| 03:04 | <jcranmer> | I don't have that fear |
| 03:04 | <jcranmer> | I don't have the game :-) |
| 03:05 | <MikeSmith> | https://bugzilla.mozilla.org/show_bug.cgi?id=485663 |
| 03:05 | <MikeSmith> | [[ |
| 03:05 | <MikeSmith> | You know how sometimes you don't have a network connection, but you have a |
| 03:05 | <MikeSmith> | source tree, and you want to see a slideshow of ten little dog pictures of |
| 03:05 | <MikeSmith> | different sizes changing rapidly with no attempt to keep the title from jumping |
| 03:05 | <MikeSmith> | up and down? |
| 03:05 | <MikeSmith> | Yeah, me neither. |
| 03:05 | <MikeSmith> | ]] |
| 03:06 | <karlcow> | ahaha |
| 03:09 | <takkaria> | http://thefuckingweather.com/ |
| 03:10 | jcranmer | looks at the slideshow |
| 03:10 | <jcranmer> | OMG, that IS bad |
| 03:13 | <azazul`> | no no |
| 03:13 | <azazul`> | there is global warming |
| 03:13 | <azazul`> | shut up and ignore the snow |
| 03:14 | <azazul`> | .. last time there was snow "on" in thos time of year , was in 1961 ... thats what i has been told |
| 03:14 | <azazul`> | =/ |
| 03:16 | <gsnedders> | Hmm, the more I think about it the less I believe HH's claim that L seduced him in The Enchanted Hunters |
| 03:37 | <gsnedders> | hah! "You talk like a book, Dad." |
| 03:47 | gsnedders | opens up Last.fm to listen to System Fault, takkaria ;) |
| 03:53 | gsnedders | is listening to Death Row by System Fault from Perfect Gentlemen EP |
| 04:14 | <Hixie> | anybody know anything about SVG's |
| 04:14 | <Hixie> | 'load' events? |
| 04:17 | <gsnedders> | No, but I'm sure I can tell you a fair amount about Nabokov's work instead. |
| 05:46 | Hixie | cannot find whiteforest inn commentary node 4 |
| 05:50 | <Hixie> | or 8 for that matter |
| 06:01 | <MikeSmith> | does chatzilla do completion on names? |
| 06:10 | <gsnedders> | Hixie: huh? |
| 06:37 | <jwalden> | MikeSmith: tab-complete, yes |
| 06:37 | <MikeSmith> | jwalden: thanks |
| 06:38 | <MikeSmith> | was trying to tell somebody else to use it, but she said it didn't seem to be work. will follow up with her |
| 07:02 | gsnedders | has sore throat |
| 07:29 | <Hixie> | i wonder what mutation events fire with innerHTML |
| 08:03 | gsnedders | finds more bits of his computing project need working on |
| 09:04 | jgraham | wwonders if anyone sa his question about foreign content parsing from yesterday |
| 09:05 | <jgraham> | wow, that was really illegible |
| 09:05 | <jgraham> | And anyway, I'm sure somebody saw it, but I'm more wondering if anyone can answer it |
| 09:11 | <zcorpan> | jgraham: what was the question? |
| 09:12 | <jgraham> | 06:28 < jgraham> So, I feel like I am being really dumb but, given a tree like <math><annotation-xml><svg><p> what prevents the secondary insertion mode being set to InForeignContent when the <svg> element is processed as if it were In Body? |
| 09:12 | <jgraham> | zcorpan: ^ |
| 09:14 | <zcorpan> | jgraham: ok |
| 09:28 | <hsivonen> | jgraham: I guess the answer is that it would really help if the spec made "in foreign" into a flag like the V.nu parser does |
| 09:34 | <jgraham> | hsivonen: What implications would that have? Just that once you were inforeign setting the flag again would be a noop? |
| 09:40 | <hsivonen> | jgraham: I think the implication is that having SVG in <annotation-xml> or MathML in <foreignObject> makes sense (if it doesn't already) |
| 09:42 | <hsivonen> | IIRC, I did run use cases from Jacques Distler with the flag-based impl. |
| 09:44 | <hsivonen> | so does twitter still have known XSS holes? |
| 10:28 | <jgraham> | hsivonen: I _think_ those cases are infinite loops in the spec currently. Although I would like someone to verify that because I could be wrong |
| 10:28 | <roc> | annevk3: we don't seem to have any CSSOM API that would allow determining the screen coordinates of elements |
| 10:29 | <annevk3> | screen coordinates? |
| 10:30 | <jgraham> | (html5lib doesn't pass the relevent tests if I remove the infinite loops in a naive way sdo I don't trust that I have read the spec correctly) |
| 10:30 | <roc> | like screenX and screenY on ScreenView and MouseEventView |
| 10:31 | <annevk3> | why is that needed? |
| 10:33 | <roc> | I don't have any good use cases |
| 10:33 | <annevk3> | well then I won't add it :) seems like screenX/Y don't even work interoperably between Opera and Firefox |
| 10:33 | <roc> | but it's weird that we have screen coordinates for events and "client windows" |
| 10:33 | <roc> | but not elements |
| 10:34 | <roc> | it might be useful for that multi-window bouncing ball app |
| 10:34 | <annevk3> | :) |
| 10:34 | <roc> | they seem to get by by assuming that all windows have the same toolbar/urlbar/etc height |
| 10:34 | <roc> | and then just compare window.screenX/Y |
| 10:35 | <roc> | it's useful to coordinate with some other code that's running on the desktop |
| 10:35 | <annevk3> | I suppose screenX/Y would be more useful if they were from viewport origin to screen origin |
| 10:35 | <roc> | but that isn't really a good use case for Web content |
| 10:35 | <roc> | yeah |
| 10:36 | <roc> | I'm not sure exactly what would break if we changed them |
| 10:36 | <roc> | it's a bit scary |
| 10:36 | <annevk3> | in Opera they currently return negative values o_O |
| 10:37 | <roc> | do you happen to have a dual-screen setup? |
| 10:39 | <annevk3> | yeah, but it involves two separate laptops :) |
| 10:45 | <hsivonen> | jgraham: like I've said before, the spec has the modeness and flagness of 'in foreign' and 'frameset-ok' backwards :-) |
| 10:51 | <jgraham> | hsivonen: OK, well I am joining your crusade to get it changed :) |
| 10:51 | <annevk3> | Does rubys mean that he wants the HTML syntax to require quotes and such for SVG fragments? |
| 10:52 | <annevk3> | Or that he's just bored discussing it? :) |
| 10:57 | <jgraham> | annevk3: Or that he doesn't care about quotes and stuff because it is easy for a machine to correct those things |
| 10:57 | <annevk3> | Yeah, but why bring XHTML + SVG up then? |
| 10:57 | <annevk3> | meh |
| 11:00 | <annevk3> | Is Julian just playing dumb and am I wasting my time or is this discussion actually leading somewhere? *sighs* |
| 11:00 | <jgraham> | Maybe he is saying that the focus on making things copy/pastable is misguided because fixing the type of differences that people have focussed on for the copy/paste issue is trivial |
| 11:00 | <jgraham> | s/he/Sam/ |
| 11:05 | <Philip`> | Unless he's arguing that we should discuss how SVG DOMs are processed and rendered, I'm not sure what the HTML WG could discuss other than the trivial syntax issues |
| 11:06 | Philip` | wonders if he should bother making his html-to-xhtml service a bit more professional |
| 11:08 | <jgraham> | Philip`: More professionsal as in, doesn't work as well but charges you through the nose? |
| 11:12 | <Philip`> | jgraham: More professional as in not saying "rubbish" and "Wowzers!" quite as much as it does |
| 11:12 | <annevk3> | krijnh, I had this idea for the IRC logs which will make them a bit slower, in case of a rename, if one the strings is an ASCII case-insensitive substring of the other, ignore it |
| 11:12 | <Dashiva> | Philip`: You should make it more document-ish |
| 11:13 | <krijnh> | annevk3: a rename? |
| 11:13 | <annevk3> | krijnh, of a nick |
| 11:14 | krijnh | doesn't follow |
| 11:14 | <annevk3> | e.g. "* slightlyoff is now known as slightlyoff_afk" can be ignored because slightlyoff is a substring of slightlyoff_afk |
| 11:15 | <krijnh> | How do I not ignore it now? |
| 11:15 | <annevk3> | but e.g. "* foo is now known as bar" cannot because that's more significant |
| 11:15 | <annevk3> | krijnh, http://krijnhoetmer.nl/irc-logs/whatwg/20090328#l-193 |
| 11:15 | <krijnh> | Ah! |
| 11:17 | <annevk3> | and maybe strip trailing digits so it works for me too :p |
| 11:17 | annevk3 | is making this too complicated |
| 11:17 | <krijnh> | Yeah, a bit :) |
| 11:18 | <krijnh> | slightlyoff_afk back to slightlyoff - how would you do that? |
| 11:18 | <krijnh> | Ow, wait |
| 11:18 | <Philip`> | Sounds more complicated than moving one's eyes downwards to skip over those lines :-) |
| 11:18 | <krijnh> | Hmm |
| 11:18 | krijnh | agrees with Philip` :) |
| 11:19 | <krijnh> | It gives the logs a personal touch ;) People are actually alive and go afk once in a while |
| 11:19 | <krijnh> | So, I need more use cases :) |
| 11:20 | <annevk3> | my main use case is actually line count on the frontpage |
| 11:20 | <annevk3> | so maybe it is enough to ignore renames altogether for that |
| 11:21 | <krijnh> | You could also educate people to not change their nicks |
| 11:21 | <krijnh> | Nobody cares about afk'ness |
| 11:21 | <zcorpan> | maybe another thing: if i enter "zcorpan" as nick it should recognize me as "zcorpan_" and "zcorpan_afk" and "zcorpan3", etc, too |
| 11:22 | <zcorpan> | although i just use zcorpan and zcorpan_ so it's not so hard to enter both |
| 11:22 | <krijnh> | Please remember there are like 5 users of these logs ;) |
| 11:23 | <zcorpan> | so? :) |
| 11:23 | <krijnh> | So it's not always worth it :) |
| 11:23 | <annevk3> | krijnh, this is your most popular site to date, stop complaining and pray we keep using it! |
| 11:23 | <krijnh> | Haha |
| 11:28 | <Philip`> | "I, for one, was not aware that service [html-to-xhtml] existed until 4 days ago as part of this discussion" - I suppose that's to be expected, since it didn't exist until 4 days ago |
| 11:29 | <Philip`> | (at least not in a publicly-accessible form) |
| 11:30 | Philip` | should probably make it use the latest parser version |
| 11:31 | krijnh | adds a new subtitle to the logs :) |
| 11:31 | Philip` | should probably also find a better way to run Java stuff, rather than having it be a standalone application with its own HTTP server running in screen |
| 11:33 | <annevk3> | krijnh, hehe, how do we get to them again? |
| 11:34 | <Philip`> | annevk3: http://krijnhoetmer.nl/irc-logs/ and then press F5 a lot of times |
| 11:35 | <krijnh> | http://krijnhoetmer.nl/irc-logs/all-subtitles |
| 11:37 | <annevk3> | that list has some superb obscure references to recent HTML history :) |
| 11:42 | <Philip`> | That's an effective demonstration of how publicness is a separate issue from openness - it's not an open process that anyone can participate in because the conversations rely on years of undocumented shared context :-) |
| 11:45 | <annevk3> | Philip`, actually, I think you can get all of that by reading the IRC logs and mailing list archives |
| 11:45 | <annevk3> | Philip`, not that it makes it more likely that people can participate per that definition though |
| 11:45 | <annevk3> | Philip`, but it's not a black/white thing anyway |
| 11:47 | <Philip`> | annevk3: That might be technically true, but reading a third of a million lines of IRC logs and twenty five thousand mailing list posts is hardly a practical idea for people wanting to get involved |
| 11:48 | <annevk3> | right, I already said as much in the following line |
| 11:48 | <krijnh> | Where are the IRC logs from before 2007? |
| 11:48 | <Philip`> | krijnh: They're at http://krijnhoetmer.nl/irc-logs/whatwg/200612 |
| 11:49 | <Philip`> | annevk3: Indeed, but I started counting before your following line :-p |
| 11:49 | jgraham | reads a weblog article from somebody pontificating about technical solutions that will allow him to twitter and have his tweets pushed to facebook, without annoying people on facebook by having too many status updates, sighs |
| 11:49 | <krijnh> | That's about when the W3C and the HTMLWG got involved, I mean from before that |
| 11:50 | <hsivonen> | jgraham: Bob Sutor? |
| 11:50 | <jgraham> | hsivonen: Yes, apparently. I didn't actually look it was just something on planet intertwingly |
| 11:51 | hsivonen | saw it on Planet Intertwingly, too. |
| 11:51 | <annevk3> | krijnh, Charl van Niekerk had logs, not sure until how far back, and I thought the plan was that he'd give them to you |
| 11:52 | <annevk3> | krijnh, they might have gone lost by now though |
| 11:52 | <annevk3> | krijnh, it took a while before we got actual logs though |
| 11:52 | <krijnh> | It shouldn't really all depend on me though :) |
| 11:53 | <krijnh> | Was http://whatbot.charlvn.za.net/ btw |
| 11:53 | <Philip`> | Planet Intertwingly caused me to start using user stylesheets for the first time ever, a few days ago |
| 11:53 | <Philip`> | (to remove posts from a certain author) |
| 11:53 | <Philip`> | I wish browsers had a button that said "edit user CSS for this site", instead of making me mess around with filesystems and stuff |
| 11:54 | <jgraham> | Philip`: I think that functionality is avaliable |
| 11:54 | <Philip`> | If it it, I wish they made it more obvious :-) |
| 11:54 | <Philip`> | *is |
| 11:56 | <annevk3> | krijnh, if I had a server somewhere that allowed me to run bots I'd help you out :) |
| 11:56 | <krijnh> | Bots are overrated |
| 11:57 | <krijnh> | You just need your irc client to log the files ;] |
| 11:57 | <annevk3> | ok, a server that allowed me to run IRC |
| 11:58 | <jgraham> | annevk3: Will dreamhost not let you do that? |
| 11:59 | <annevk3> | not last time I checked |
| 11:59 | <krijnh> | Perhaps it would be smart to upload the logfiles somewhere once in a while? |
| 11:59 | <krijnh> | (I don't do backups) |
| 12:00 | <krijnh> | http://irc-logs.whatwg.org/ or something |
| 12:01 | <krijnh> | It's only 60MB of text files atm |
| 12:01 | <annevk3> | get an account on http://code.google.com/p/html5/ and use that? |
| 12:02 | <annevk3> | or will that get us into license issues? |
| 12:02 | annevk3 | sighs |
| 12:02 | <krijnh> | Don't know, why? |
| 12:02 | <annevk3> | is the content under the Apache license? |
| 12:03 | <krijnh> | The logs? |
| 12:03 | <Philip`> | Why not upload the log files to Dreamhost? |
| 12:03 | <krijnh> | Yeah, why not? |
| 12:03 | <annevk3> | I can make logs.html5.org if we want do something now |
| 12:03 | <annevk3> | and give krijnh an account so he can do stuff |
| 12:04 | <krijnh> | To use as a backup, or as a primary place for the logs (both fine by me) |
| 12:04 | <krijnh> | ? |
| 12:04 | <annevk3> | both are fine by me as well :) |
| 12:04 | <Philip`> | It's nice for the logs to be live |
| 12:04 | <Philip`> | so they presumably have to be on the same machine as the IRC client |
| 12:04 | <krijnh> | Yeah, I don't think I can export the logs to Dreamhost immediately |
| 12:04 | <annevk3> | yeah |
| 12:05 | <krijnh> | So that could be a daily cronjob or something |
| 12:05 | <annevk3> | that sounds fine |
| 12:05 | <annevk3> | i'll make an account |
| 12:05 | <krijnh> | Kewl |
| 12:05 | <annevk3> | anyone prefer irc-logs over logs? |
| 12:05 | krijnh | does |
| 12:05 | <annevk3> | k |
| 12:06 | <krijnh> | Or krijnhoetmers-irc-logs ;) |
| 12:06 | <krijnh> | Or I prefer krijnhoetmers-you-know-that-irc-logging-guys-kickass-irc-logs.html5.org |
| 12:07 | <annevk3> | right, we can set those up later |
| 12:07 | <krijnh> | :) |
| 12:07 | <annevk3> | do you need more than FTP? |
| 12:07 | <krijnh> | Depends |
| 12:07 | <annevk3> | what do you want? |
| 12:07 | <krijnh> | I use mod_rewrite, php and mysql now |
| 12:07 | <krijnh> | But mod_rewrite and php is okay enough |
| 12:08 | <krijnh> | For a backup |
| 12:08 | <annevk3> | you store the raw text in MySQL |
| 12:08 | <annevk3> | ? |
| 12:08 | <krijnh> | No, only the important linenumbers |
| 12:08 | <krijnh> | And the irc nicks |
| 12:08 | <annevk3> | i actually meant whether you wanted SFTP/FTP/ssh |
| 12:08 | <krijnh> | Ow |
| 12:08 | <Philip`> | I thought the policy was to always have <firstname>.html5.org, and I don't think krijnh's first name is irclogs :-) |
| 12:09 | <krijnh> | krijn.html5.org would be cool too :) |
| 12:09 | <krijnh> | annevk3: FTP is okay |
| 12:10 | <krijnh> | http://irc-logs.html5.org/ \o/ |
| 12:10 | <annevk3> | http://krijn.html5.org/ should start working too shortly |
| 12:11 | <virtuelv> | to answer annevk's assertion further up: no, you can't run bots on dreamhost (and they don't really want you to have a persistent irc client there either) |
| 12:11 | <annevk3> | I think that's nicer :) |
| 12:11 | <virtuelv> | if you want something like that, I'd suggest slicehost |
| 12:11 | <virtuelv> | (or whatever they are called now) |
| 12:11 | <virtuelv> | http://www.slicehost.com/ |
| 12:17 | <annevk3> | Not going to invest USD 240 yearly for logging, sorry :) |
| 12:19 | <krijnh> | I am already |
| 12:19 | <krijnh> | At least! :) |
| 13:21 | <zcorpan> | hmm, maybe <textarea placeholder> should have different requirements than <input placeholder> wrt linebreaks |
| 13:29 | <MikeSmith> | zcorpan: I was thinking about the same thing |
| 13:30 | <hsivonen> | MikeSmith: I redeployed V.nu |
| 13:30 | <MikeSmith> | hsivonen: thanks. I'll test the <optgroup> fix and close that bug if it's passing |
| 13:31 | <hsivonen> | MikeSmith: thanks |
| 13:31 | <MikeSmith> | I'm pretty sure the reason for the bug is some regression I introduced with an earlier change |
| 14:16 | <MikeSmith> | hsivonen: so I've been messing around with Assertions.java code for checking for the case of label@for input child with mismatched ID. I can see how to check for an input@id instance that doesn't match any label@for in the entire document, but I have not figured out how to restrict to checking just in the stack |
| 14:18 | <hsivonen> | MikeSmith: I'd walk up the stack as long as the stack nodes have the label/for bit set on the mask |
| 14:19 | <hsivonen> | MikeSmith: and also having the actual @for values on the stack |
| 14:19 | <MikeSmith> | hsivonen: OK, I see |
| 14:20 | <hsivonen> | the bit mask doesn't help with worse-case time complexity of walking the stack every time, but at least it makes the best case a lot better |
| 14:21 | <hsivonen> | (presumably common case, too) |
| 14:23 | <MikeSmith> | hsivonen: I had just forgotten that the ancestor bits are copied down |
| 14:26 | hsivonen | wonders if two snapshots of the tree builder during a given parse can ever differ in terms of mode, original mode, foreign flag, form pointer or head pointer if the stack and the list of active formatting elements have the same contents in the two snapshots |
| 14:27 | <hsivonen> | that is, if the tree builder goes through balanced pushes and pops, does it always end up in the same mode where it started |
| 14:27 | <hsivonen> | ? |
| 14:28 | <hsivonen> | if one of the snapshots is always taken immediately after </script> has been popped |
| 14:28 | <hsivonen> | at least </script> guarantees that the headPointer has been set (or we are in the fragment case) |
| 14:38 | <hsivonen> | it seems that the form element pointer can be set to null without popping the node it pointed to |
| 14:38 | <hsivonen> | right? |
| 14:46 | <hsivonen> | How do I save the file that I'm viewing in Opera 8.65 for S60 locally? |
| 19:29 | <hsivonen> | http://twitter.com/ilinsky/status/1387916636 |
| 19:29 | <hsivonen> | "he" being Hixie as far as I can tell |
| 19:31 | <annevk3> | it was in reply to http://twitter.com/shepazu/status/1387879523 so yes |
| 19:33 | <annevk3> | I'm not sure if Hixie is a fan of ARIA though :p |
| 19:43 | <annevk3> | the paragraph in section 8.1 no longer points to he next section |
| 19:43 | <annevk3> | guess I need to e-mail that again... |
| 20:44 | <annevk3> | Hixie, you still need an IETF WG of some sorts to actually get to RFC status right? |
| 20:48 | annevk3 | finds http://www.ietf.org/mail-archive/web/hybi/current/msg00000.html |
| 22:39 | <annevk3> | heh, they renamed ES 3.1 to ES 5 |
| 22:42 | <takkaria> | ^_^ |
| 22:48 | <Hixie> | really? |
| 22:48 | <heycam> | surely that announcement is a day too early! |
| 22:49 | <Philip`> | https://mail.mozilla.org/pipermail/es-discuss/2009-March/009111.html |
| 22:55 | <otherarun> | Heh. Yeah, Mark Miller's preempted the press release, but yeah. The name change is because whole numbers were deemed desirable, and ES4 was confusing (ES4 is dead; long live ES4, or at least some of it, as ES6). |
| 22:55 | <Hixie> | http://en.wikipedia.org/wiki/Waka_(protocol) <-- a wikipedia page purely about an idea that roy had |
| 22:55 | <Hixie> | otherarun: hey, arun! you're here! |
| 22:56 | <Hixie> | otherarun: any news on the file/blob api? i'm trying to plan my timetable for html5 work for the coming quarter and need to know whether to punt on the <input type="file"> access API or not |
| 22:57 | <annevk3> | Hixie, you got several of those wiki pages :p |
| 22:58 | <otherarun> | Hixie, yep, I'm here. So the status is: I'm still working on it. I'm going to work on a draft for this quarter. Blob has "problems" that Apple would like to get back to us on; synchronous APIs are out, but some folks are interested in File I/O more than Blobs, etc. and we might want to split that up into separate specs. |
| 22:59 | <annevk3> | i was sort of intrigued by the apple idea of just having .files a simple File interface and supporting File in xhr.send() |
| 22:59 | <annevk3> | maybe not even have local file access in the first api |
| 23:00 | <annevk3> | if we get that shipped all flash uploaders can be shot |
| 23:01 | <otherarun> | annevk3 yeah; I'm coming around to thinking that if anyone is interested in direct-from-script I/O APIs, that part might be something that gets spec'd separately. |
| 23:03 | <annevk3> | besides XHR you want some integration with <canvas> |
| 23:04 | <annevk3> | further in the future i suppose you might want raw data access so you can do text and DOM processing without having to round-trip through the server |
| 23:04 | <otherarun> | annevk3 hmmm... that may be a bit more than originally conceived. good use case scenario? |
| 23:05 | <annevk3> | selecting a image to manipulate, resizing images client-side, etc. |
| 23:06 | <annevk3> | raw data access is for all the other possible use cases; the user has some set of files and the webapps has some script that can do something with the file data and you don't really want to do server round-trip for the files |
| 23:07 | <Hixie> | annevk3: really? |
| 23:07 | <Hixie> | otherarun: k |
| 23:08 | <Hixie> | otherarun: i guess i'll see where we stand at end of Q2 and if the File/Blob API isn't ready yet I'll punt on .files for html5 |
| 23:09 | <annevk3> | Hixie, just kidding; was mainly thinking of HTML5 and WHATWG :) |
| 23:09 | <Hixie> | annevk3: oh ok. At least those are written down somewhere :-P |
| 23:10 | <Hixie> | for file/blob, the way i see it we want to have an object for a readonly file, an object for a mutable file, both of which support being filesystem-backed and support being just memory mapped, and then we'll want some api that exposes streams using a similar mechanism, and finally an API that maps these file/stream objects to document-lifetime-bound URIs |
| 23:10 | <Hixie> | for interaction with things like <video> |
| 23:11 | <Hixie> | and <img> |
| 23:11 | <Hixie> | maybe some protocol like blob://<origin>/<number> or something, accessible as fileObject.URL |
| 23:11 | <Hixie> | anyway |
| 23:13 | <Hixie> | christ, i have a two-line testcase for mutation events and three browsers and they ALL DO SOMETHING DIFFERENT |
| 23:13 | <Hixie> | this was supposed to be the test that checked that i had the write syntax |
| 23:13 | <Hixie> | it wasn't suppoed to find bugs |
| 23:14 | <Hixie> | i hate vague specs |
| 23:14 | <Hixie> | (http://software.hixie.ch/utilities/js/live-dom-viewer/saved/46) |
| 23:16 | <Hixie> | firefox fires the mutation events TWICE when doing innerHTML |
| 23:16 | <Hixie> | -_- |
| 23:17 | <Hixie> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/47 |
| 23:21 | <jwalden> | Hixie: as far as I could tell the one time I fooled around with mutation events, Gecko's method for dealing with them is basically a spec violation that makes it possible to do things without preemptively grabbing strong refs to everything in sight -- nice impl-wise, but doesn't produce behavior any web developer is going to expect |
| 23:21 | jwalden | couldn't bring himself to care, for mutation events |
| 23:22 | <Hixie> | man i hope the dom3 events spec either nukes mutation events altogether or defines how they work in enough detail that they can be implemented |
| 23:22 | <Hixie> | i think what would be ideal is for mutation events to move to web dom core actually |
| 23:22 | <Hixie> | since that's where all the requirements would likely have to be to make the spec sane to write |
| 23:24 | <Lachy> | mutation events would have been nice if they were more widely supported. I know there were a few things that would have been easier for me to do in some projects if they were usable in practice, though I can't remember the details. |
| 23:26 | <Lachy> | although, IIRC, I think using XBL bindings would have helped in some of those cases too |
| 23:41 | <annevk3> | https://mail.mozilla.org/pipermail/es-discuss/2009-March/009112.html o_O |
| 23:41 | <annevk3> | I'd love them to open up those ES meetings |
| 23:47 | <jwalden> | annevk3: not a generalized solution, but could talk to chris pine (?) for scoops, I think |
| 23:49 | <annevk3> | true, I rather not bug him about ES gossip though :) |