| 00:00 | <paul_irish> | kennyluck: know who to contact to ask for that? cc Peter` |
| 00:00 | <kennyluck> | paul_irish, no idea. |
| 00:00 | <Peter`> | For WebKit? |
| 00:00 | <paul_irish> | ya |
| 00:00 | <Peter`> | I'd try webkit-dev (the mailing list) |
| 00:00 | <Peter`> | But bugs at WebKit are a big mess |
| 00:00 | <Peter`> | the component/OS/platform/etc fields are rarely accurate |
| 00:01 | <kennyluck> | Peter`, that's true. |
| 00:05 | <boblet> | HEY PPLZ, what was your most “mind = blown” website for 2011 for use of HTML5? |
| 00:20 | <zewt> | "ip-based security"? really? |
| 01:22 | <rniwa> | sicking: have you had a chance to look at algorithms to reapply DOM changes? |
| 01:32 | <zewt> | who came up with this terrible idea of having separate mailing lists for "CGs"? i don't want to subscribe to lots of tiny one-spec mailing lists |
| 01:41 | <zewt> | Hixie: i'm confused; of course @lang has an effect on rendering ("User agents may use the element's language to determine proper processing or rendering"), and most browsers implement that (everyone but WebKit, it seems) |
| 01:41 | <zewt> | (asking in here rather than mail since I'm assuming I'm misunderstanding you) |
| 01:44 | <zewt> | by the way, your thesis-emails are causing gmail's editor to hang Firefox :| |
| 01:46 | <rniwa> | zewt: I agree. |
| 01:46 | <rniwa> | zewt: I don't want to end up subscribing to 20+ mailing list |
| 01:46 | <rniwa> | zewt: and having to create a filter for each one of them |
| 01:48 | <zewt> | the end result is going to be fewer interested parties contributing to specs, which is a very bad thing |
| 01:49 | <TabAtkins> | Just add them all to one big filter for "webapps" or something. |
| 01:49 | <TabAtkins> | A tiny effort for each, and it lets you pretend they're all the same. |
| 02:03 | <rniwa> | TabAtkins: just having to add 'em all is quite annoying |
| 02:03 | <rniwa> | zewt: right... |
| 02:04 | <rniwa> | my gmail inbox is already spammed by www-* mailing lists :( |
| 02:04 | <rniwa> | don't want to add more mailing lists |
| 02:04 | <zewt> | er, that's what filters are for |
| 02:05 | <zewt> | set a label, enable "skip inbox" |
| 02:05 | <rniwa> | zewt: you never know which ones are important |
| 02:05 | <zewt> | doesn't mean they should go into inbox |
| 02:05 | <zewt> | also if you explicitly want a thread in inbox, move to -> inbox |
| 02:05 | <rniwa> | zewt: maybe |
| 02:06 | <zewt> | i can't imagine having mailing lists in my inbox :) |
| 02:06 | <rniwa> | zewt: I might just unsubscribe myself from www-style... |
| 02:06 | <rniwa> | that mailing list gets too many emails per day |
| 02:10 | <zewt> | i've been on lkml and other high-volume lists (though I'm not right now); as long as it's stuffed away into a label, i can selectively ignore it when I feel like it, without having to unsubscrube |
| 02:14 | <rniwa> | zewt: I guess. |
| 02:14 | <rniwa> | zewt: though my problem is that it's consuming a lot of my inbox space. |
| 02:16 | <zewt> | well, you can delete mails by date, though in my experience even high-volume mailing lists don't do much of a dent to gmail's ~7gb |
| 02:42 | <zewt> | heh, @microsoft quoting ancient out-of-date text from DOM2 on www-dom |
| 02:43 | <zewt> | heres to versioned specs still causing confusion in 2011 |
| 02:44 | <michaelw> | hi! looking at HTML5 sec 4.3.1/15, 4 item, which scripts end up in the "list of scripts that will execute in order ASAP"? Created by JS with an defer attribute? |
| 02:45 | <michaelw> | s/4 item/4th item/ |
| 02:54 | <zewt> | search for that string |
| 03:17 | <zewt> | wow, didn't know alert() throws an exception sometimes in FF (if the user navigates away while a prompt is shown) |
| 04:30 | <zewt> | highly troubling for an @microsoft to claim that something isn't an interop problem because it's "poor design of the page itself" |
| 07:06 | <hsivonen> | whoa. USA is now at 1.0% on http://www.ie6countdown.com/#list |
| 08:35 | <annevk> | Hixie: it is fixed now, for some reason I had to run svn cleanup on html5.org |
| 08:35 | <annevk> | Hixie: do you know anything about that? |
| 08:47 | <annevk> | Can someone write up a Bugzilla quoting etiquette? One that frowns upon excessive quoting in Bugzilla? |
| 08:47 | <annevk> | Or maybe on that suggests to have email conversations on a mailing list and not in Bugzilla... |
| 08:48 | <annevk> | one* |
| 08:48 | <roc> | it's hard to move conversations from one medium to another |
| 08:49 | <kennyluck> | +1 |
| 08:50 | <kennyluck> | though I really hope people don't use medium like twitter for technical discussions. |
| 08:51 | <annevk> | heh, you must be new here :p |
| 08:52 | <smaug____> | kennyluck: unfortunately people do you twitter and G+ for technical discussions |
| 08:52 | <smaug____> | s/you/use/ |
| 08:53 | <woef> | We need a forum revival. |
| 09:21 | <jgraham> | smaug____: Surprisignly some of the technical discussion on G+ has been quite good. Of course it has other problems with closedness and archiving |
| 09:21 | <jgraham> | Tryign to have technical discussions as a series of tweets isn't something I can comprehend however |
| 09:22 | smaug____ | isn't reading any G+, so just misses or doesn't care about discussions happening there |
| 09:23 | <jgraham> | Well it seems like mostly people stopped using it for that purpose |
| 09:23 | <jgraham> | Or I stopped noticing |
| 09:23 | <jgraham> | But just after it launched it was used reasonably often and the quakity was quite high |
| 09:23 | <jgraham> | That's why I bothered to join |
| 10:22 | <zcorpan> | hmm. what's https://www.w3.org/Bugs/Public/show_bug.cgi?id=15007#c7 ? sounds a bit like "await a stable state" |
| 10:22 | <kennyluck> | social network without interoperable standard simply sucks. |
| 10:25 | <zcorpan> | hey guys! anyone volunteer to hack on http://html5.org/tools/web-apps-tracker to support the topics thing? (http://code.google.com/p/html5/source/browse/#svn%2Ftrunk%2Fweb-apps-tracker ) |
| 10:27 | <annevk> | suggested UI? |
| 10:27 | <jgraham> | kennyluck: For some definition of "simply sucks", yes. I just can't think of any reasonable process that would get a social network built on an interoperable standard |
| 10:27 | <annevk> | zcorpan: just a column? |
| 10:27 | <jgraham> | The value to the providers is all in the lock-in |
| 10:27 | <jgraham> | annevk: Filter by topic |
| 10:28 | <zcorpan> | annevk: yeah ability to filter would be nice |
| 10:30 | <erlehmann> | kennyluck, social networks suck. they are built on lies, on of which is “just click here and your information will stay private” |
| 10:30 | <erlehmann> | one |
| 10:31 | <annevk> | filters again? |
| 10:31 | <annevk> | nobody used that last time |
| 10:31 | <kennyluck> | I am so glad that at least we have mailing lists, IRC, and well… newsgroup. These belong to some definition of social network, although these don't work so well with the Web I guess. |
| 10:31 | <jgraham> | I did |
| 10:31 | <erlehmann> | jgraham is absolutely spot-on regarding the lock-in. just compare google+ with the rest of googles offerings. hell, you could get your email via ATOM last time i checked. |
| 10:31 | <jgraham> | Until you took it away |
| 10:31 | <annevk> | interesting |
| 10:32 | <annevk> | what did you use it for? |
| 10:33 | <erlehmann> | kennyluck, mailing lists, blogs and feedreaders can do most of the things social networks are used for. but then people start to complain that no one ever reads their blog. in a related note, my girlfriend recently deleted facebook friends until she was definitely under dunbar's number. |
| 10:34 | <jgraham> | … filtering? |
| 10:36 | <jgraham> | Although I guess I mainly used the editorial/non editorial bit, which is still there |
| 10:36 | <jgraham> | Pretty sure I did use the other thing sometimes though |
| 10:36 | <jgraham> | I would use topics for sure |
| 10:37 | <erlehmann> | jgraham, i think the notion of a “social network” should be overcome, if people want to overcome the lock-in. sadly, few developers realize that – just look at diaspora, whose paramount achievement is being a less usable facebook. |
| 10:38 | <annevk> | jgraham: yeah I remember you saying you mostly used the editorial thing, which is why kept that |
| 10:38 | zcorpan | uses the editorial filter too |
| 10:39 | <annevk> | so what are people doing with Growl? paying for it? |
| 10:39 | <erlehmann> | wat |
| 10:40 | <annevk> | it keeps notifying me I need to buy it |
| 10:40 | <erlehmann> | annevk, i was of the impression that growl was the (lib)notify thingy OS X uses. so it is not part of the base system? |
| 10:40 | <annevk> | no it's not |
| 10:40 | <annevk> | it's a download |
| 10:41 | <zcorpan> | it's fucked that OS X still doesn't have a native notification system |
| 10:42 | <erlehmann> | annevk, i suggest you use something built on libnotify/notifyd then? |
| 10:42 | <erlehmann> | in before OS X does not have something like DBUS |
| 10:42 | <erlehmann> | i seriously do not know |
| 10:42 | <jgraham> | Seems that growl is pretty standard on OSX |
| 10:42 | <erlehmann> | but i am always amazed when people – geeks – say they want something that *just works* and then it doesn't. |
| 10:43 | <erlehmann> | jgraham, wouldn't it be easy to hook into the growl message bus and make a simple notification replacement? |
| 10:43 | <erlehmann> | or does something prevent that? |
| 10:43 | <jgraham> | I am not sure how the arch works |
| 10:45 | <erlehmann> | annevk, what benefits do you get upon purchase? i mean … popups. can there be a premium popup? |
| 10:46 | <jgraham> | Maybe you get a valet who brings you the messages on a silver platter |
| 10:46 | <tomasf> | http://growl.info/documentation/developer/gntp.php |
| 10:46 | <annevk> | erlehmann: no idea |
| 10:47 | <annevk> | erlehmann: maintained software I guess |
| 10:47 | <erlehmann> | jgraham, i chuckled a mit. |
| 10:48 | <erlehmann> | annevk, it is open source. patch the nag screen away. |
| 10:50 | <annevk> | you can't tell end users that |
| 10:50 | <annevk> | it's open source, it doesn't do what you want, but you can make it so and feel special |
| 10:51 | <annevk> | that might work for some people, but not me |
| 10:51 | <erlehmann> | annevk, that is true. however, i use an operating system having a) a software repository b) a notification system. but i know it would not be considered helping if i told you to switch to that. |
| 10:53 | <annevk> | Ubuntu is cool, whenever someone writes hardware that's better than a MacBook Pro I'll happily switch back |
| 10:53 | <erlehmann> | in the absence of repositories of curated software, patching remains the only option. it is what people do, even non-programmers occasionally. |
| 10:53 | <annevk> | as long as the OS has a somewhat decent terminal by default I'm good |
| 10:54 | <erlehmann> | well, i use debian. ubuntu breaks every 6 months ;) |
| 10:54 | <erlehmann> | though recently i accidentally installed gnome 3. but the debian gods had mercy and the default session had been magically switched to XFCE. |
| 10:55 | <erlehmann> | can't you run stuff on mac books? on hacker conferences i see lots of people running something with proper X11 (i suspecd gnu/linux or *BSD) on their shiny metal devices. |
| 10:56 | <zcorpan> | Hixie: i hit the 40kb limit on whatwg⊙wo |
| 10:56 | <zcorpan> | "Reason: Message body is too big: 66449 bytes with a limit of 40 KB" |
| 10:57 | zcorpan | sends part of it to www-archive |
| 10:57 | <erlehmann> | >40kb |
| 10:57 | <erlehmann> | wat |
| 10:57 | <zcorpan> | i pasted in some SRT timing lines |
| 10:58 | <zcorpan> | hmm, it went through to public-texttracks |
| 10:59 | <zcorpan> | hmm. http://www.w3.org/mid/op.v5uwpvvtidj3kv⊙sl says 404. has Archived-At stopped working? |
| 10:59 | <annevk> | no |
| 10:59 | <zcorpan> | MikeSmith? |
| 10:59 | <annevk> | it usually takes a couple of minutes |
| 10:59 | <annevk> | sometimes more |
| 10:59 | <zcorpan> | ah |
| 11:00 | <annevk> | "This may mean that the message-id you have entered was not valid, or that the message has not yet been added to the database." |
| 11:01 | <annevk> | heh, could point on "language" :) |
| 11:01 | <annevk> | good* |
| 11:02 | <annevk> | hmm |
| 11:04 | <zcorpan> | maybe i should disable the "send" shortcut in opera... |
| 11:05 | <annevk> | WHATWG should use W3C mailing list software |
| 11:05 | <annevk> | if we could somehow |
| 11:05 | <annevk> | it's much nicer |
| 11:06 | <kennyluck> | +1 |
| 11:06 | <annevk> | kennyluck: can't you fix that for us? :) |
| 11:07 | <kennyluck> | annevk, what exactly is the problem why WHATWG can't use the W3C mailing list software? |
| 11:07 | <erlehmann> | would W3C mailing list software mean i could finally sent GPG signed messages? |
| 11:12 | <annevk> | kennyluck: I think we use the default setup from Dreamhost |
| 11:12 | <annevk> | kennyluck: and I guess doing anything more complicated than that is too much work |
| 11:14 | <kennyluck> | :( |
| 11:20 | <annevk> | kennyluck: maybe someone at the W3C can be convinced to create an alternative archive? e.g. something like whatwg⊙wo ? |
| 11:20 | <annevk> | kennyluck: import all the old stuff and subscribe whatwg⊙wo to whatwg⊙wo and make that a readonly list |
| 11:23 | <kennyluck> | annevk, that sounds quite interesting. |
| 11:23 | kennyluck | will write to site-comments⊙wo |
| 11:24 | <annevk> | thanks man |
| 11:25 | <zcorpan> | does the webrtc group use bugzilla? |
| 11:26 | <eli__> | hello, Can I ask about develop.opera this channel? |
| 11:29 | <annevk> | eli__: dev.opera.com ? maybe |
| 11:30 | <eli__> | thank you. I hope to ask about video and stream. |
| 11:31 | <annevk> | it's best to just ask a question in this channel |
| 11:31 | <erlehmann> | ask ask! |
| 11:31 | <erlehmann> | ASK |
| 11:31 | <erlehmann> | DO IT |
| 11:31 | <erlehmann> | NOW DO IT |
| 11:31 | <eli__> | :) thank you |
| 11:32 | <eli__> | in first, please this URL (http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2011-July/032417.html) |
| 11:33 | <eli__> | I try "video user" |
| 11:33 | <eli__> | i tried "video user" |
| 11:34 | <eli__> | but my opera browser catch "video environment" always. |
| 11:35 | <eli__> | this is it. (http://dev.0-0a.me/test.php) |
| 11:36 | <annevk> | mpt: where is your blog? |
| 11:37 | <eli__> | do you know reason about opera cannot catch "face-view-camera"? |
| 11:37 | <mpt> | annevk, on an HD waiting for me to put it online again |
| 11:38 | <annevk> | mpt: I wanted to point out your "perfection is" post to someone |
| 11:38 | <annevk> | mpt: it's not even on Google anymore |
| 11:38 | <annevk> | mpt: I can offer you mpt.html5.org or some such if you need a domain and some space |
| 11:39 | <mpt> | annevk, <http://web.archive.org/web/20100711072841/http://mpt.net.nz/archive/2006/11/27/perfection>;? |
| 11:40 | <annevk> | sweet |
| 11:41 | <mpt> | annevk, thanks for the offer, but I have a domain and space, I just need to hook them together |
| 11:42 | <annevk> | eli__: richt can maybe help you |
| 11:43 | <eli__> | please help me. |
| 11:43 | <richt> | eli__: just reading back on the logs... :) |
| 11:43 | <eli__> | where logs? |
| 11:44 | <eli__> | is this bug? |
| 11:55 | <eli__> | please help me.. |
| 11:58 | <richt> | eli__: camera hints are coming to the Opera implementation soon with UI stuff. Currently, mobile implementations select the back-facing camera by default. |
| 11:58 | <richt> | eli__: We will implement hints though. There's a reason this stuff is released in Opera Labs :) |
| 11:59 | <eli__> | yet I can not use it? |
| 12:02 | <eli__> | richt: Can I not use it yet? |
| 12:06 | <MikeSmith> | zcorpan: yeah, seems not to be working laterly |
| 12:06 | <MikeSmith> | dunno why |
| 12:09 | <zcorpan> | it's still 404 |
| 12:09 | <MikeSmith> | yeah |
| 12:23 | <jgraham> | Sigh, device orienation points to DOM 2 events |
| 12:23 | <jgraham> | *orientation |
| 12:24 | <jgraham> | And uses init*Event |
| 12:24 | <annevk> | I pointed that out to them during TPAC :( |
| 12:24 | <annevk> | they said they would fix it |
| 12:25 | <jgraham> | Oh, so I don't need to send feedback? good :) |
| 12:25 | <eli__> | richt? |
| 12:26 | <zcorpan> | does orientation events still say to fire an event when a listener is registered instead of having the information available on navigator or so? |
| 12:27 | <annevk> | jgraham: yes you should |
| 12:27 | <jgraham> | It says to fire an event if you think the page's data isn't fresh |
| 12:27 | <jgraham> | I think the event based design is a mistake :( |
| 12:27 | <annevk> | jgraham: by queuing a task? |
| 12:27 | <jgraham> | annevk: It doesn't mention tasks anywhere afaict |
| 12:28 | <annevk> | o_O |
| 12:29 | <annevk> | <3 vacuum groups |
| 12:33 | <jgraham> | Yeah, I hate the fact that every time I have to look at this spec it sucks up a bunch of my time on stuff that people familar with other specs would just get right |
| 12:39 | <richt> | eli__: I popped out for lunch. You can't use hints yet. I'll follow up to see if we can get it in a build soon. |
| 12:39 | <richt> | eli__: so no, you can't get the front-facing camera on mobile. Desktop provides the web cam which is front-facing though. |
| 12:39 | <annevk> | jgraham: turns out that's true for most specs developed outside WebApps / WHATWG |
| 12:39 | <annevk> | jgraham: see e.g. the battery API stuff |
| 12:40 | <jgraham> | Do I have to? I think it will just make me sad :( |
| 12:40 | <jgraham> | I thought sicking was involved with that stuff |
| 12:40 | <jgraham> | Does it also use events? |
| 12:43 | <richt> | s/ you can't get the front-facing camera on mobile./ you can't get the front-facing camera on mobile...yet. |
| 12:44 | <jgraham> | annevk: What's the best example to point them to for the queue as task to fire an event stuff? |
| 12:44 | <zcorpan> | hmm. i'm not too happy with webvtt being axed from the html spec. |
| 12:45 | <zcorpan> | now i can't conveniently file bugs on it, and the xrefs are broken |
| 12:45 | <zcorpan> | Hixie: ^ |
| 12:47 | <annevk> | jgraham: WebSocket API maybe? |
| 12:48 | <zcorpan> | i was going to suggest websocket, too |
| 12:49 | <zcorpan> | <video> also queues events, but it updates some stuff sync but queues events, which causes raciness |
| 12:50 | <jgraham> | WebSocket might be good. I found http://www.whatwg.org/specs/web-apps/current-work/#common-event-behaviors which is pretty OK |
| 12:50 | jgraham | checks websocket |
| 12:56 | <bga> | heh. to avoid wait when github's 300k css will load and 5 sec rendering w/ cpu & browser freeze will end just add github's css to urlfilter.ini |
| 12:56 | <bga> | annoying modern web |
| 13:09 | <hsivonen> | \o/ responseType == "json" |
| 13:09 | <hsivonen> | annevk: does it differ from "moz-json" except for the prefix? |
| 13:09 | <annevk> | maybe |
| 13:09 | <annevk> | I went with fighting encoding proliferation |
| 13:09 | <annevk> | enforcing UTF-8 |
| 13:10 | <annevk> | not sure if I can get agreement on that, but it really seems better if we can get everyone on board with UTF-8 |
| 13:10 | <bga> | XHR? |
| 13:11 | <hsivonen> | bga: yes |
| 13:12 | <zcorpan> | annevk: just prepare to be laughed at by bjoern |
| 13:12 | <Ms2ger> | News at ten |
| 13:13 | <jgraham> | I thought json was designed to work with utf8 only |
| 13:13 | <jgraham> | anyway |
| 13:13 | <hsivonen> | jgraham: the spec even requires UTF-32, IIRC! |
| 13:14 | <jgraham> | "JSON text shall be encoded in unicode" |
| 13:14 | <hsivonen> | jgraham: yay. IETF spec writingg |
| 13:14 | <hsivonen> | s/g// |
| 13:15 | <hsivonen> | jgraham: yeah, not clear if the spec requires UTF-32 from impls |
| 13:15 | <hsivonen> | annevk: I'm OK with UTF-8-only, FWIW. |
| 13:15 | <jgraham> | You would have thought the IETF people would be super picky about the difference between a characterset and an encoding |
| 13:16 | <hsivonen> | annevk: looks like Gecko supports the charset parameter on the HTTP level for "moz-json" |
| 13:17 | <zcorpan> | hsivonen: is there a bug about forcing utf-8 for workers in gecko? |
| 13:19 | <hsivonen> | zcorpan: I don't know |
| 13:21 | <zcorpan> | speaking of which, i should get around to converting our worker tests to testharness and get them released... |
| 13:23 | <zcorpan> | my favorite test is a windows-1252 text/html page that opens a worker to itself and checks that the worker is interpreted as utf-8 |
| 13:23 | <gsnedders> | hsivonen: Practically everyone bases their JSON impl for ES5.1 |
| 13:23 | <gsnedders> | *from |
| 13:28 | <jgraham> | For the benefit of relevant authorities: 07:26 < zcorpan> my favorite test is a windows-1252 text/html page that opens a worker to itself <-- this is the proof that zcorpan is evil, if you ever want him locked up for any reason |
| 13:29 | <gsnedders> | jgraham: Were there not more evil things in his testsuite than that? |
| 13:29 | <zcorpan> | probably, that was just my favorite |
| 13:30 | <jgraham> | Note how he takes pleasure from the pain |
| 13:31 | <hsivonen> | this is interesting: https://bugzilla.mozilla.org/show_bug.cgi?id=707108 |
| 13:31 | <hsivonen> | I wonder how Apple's Mail and Opera's mail client deal |
| 13:32 | <gsnedders> | What. Just what. |
| 13:35 | <wilhelm> | That attachment is priceless. |
| 13:36 | <Ms2ger> | <DIV dir=3Dltr align=3Dleft><FONT face=3DVerdana></FONT> </DIV> |
| 13:36 | <Ms2ger> | \o/ |
| 13:36 | <jgraham> | I like my emails in 3D! |
| 13:39 | <erlehmann> | hahaha oh wow |
| 13:40 | <zcorpan> | i doubt opera deals with that |
| 13:52 | <annevk> | classic: http://torrentfreak.com/copyright-corruption-scandal-surrounds-anti-piracy-campaign-111201/ |
| 13:54 | <Ms2ger> | Also, typical |
| 14:02 | <hsivonen> | yay UA sniffing https://bugzilla.mozilla.org/show_bug.cgi?id=699289 |
| 14:03 | <gsnedders> | hsivonen: Well, we already know that Opera only makes a mobile browser. |
| 14:10 | <hsivonen> | is Opera 11.60 going to ship Ragnarök? |
| 14:10 | hsivonen | is too lazy to get the release candidate before the release makes it to the .deb repo |
| 14:11 | <hsivonen> | first it seemed to be an Opera 12 feature. then it looked like it was going into 11.60 |
| 14:12 | <jgraham> | Yes |
| 14:12 | <jgraham> | At least it is in the snapshots I believe |
| 14:12 | <jgraham> | I suppose things could change before release |
| 14:12 | <hsivonen> | jgraham: it was in 11.60 beta |
| 14:13 | <jgraham> | Right, that's the same information I have more or less |
| 14:13 | <hsivonen> | I'm slightly annoyed that 11.60 beta hasn't autoupdated to RC via the repo |
| 14:14 | <hsivonen> | but I realize that this is a glass house I'm in, considering that Mozilla doesn't even have a .deb repo |
| 14:14 | <jgraham> | Well there is a PPA fro Mozilla. |
| 14:15 | <jgraham> | The Opera builds typically aren't pushed to autoupdate for a little while after release in case there are horrible problems |
| 14:15 | <jgraham> | The deb counts as autoupdate in this case I believe |
| 14:16 | jgraham | doesn't really know much about desktop |
| 14:18 | <hsivonen> | the PPA for Firefox contains Canonical's builds--not Mozilla Corporation's builds, though |
| 14:18 | <gsnedders> | hsivonen: First there wasn't an 11.60 :) |
| 14:18 | <hsivonen> | gsnedders: ooh. agile release management |
| 14:19 | hsivonen | wonders if it's true that the major version of Opera is the year of the final release |
| 14:19 | <jgraham> | Haha, I wish\ |
| 14:19 | <gsnedders> | 10 was 2009 |
| 14:19 | <hsivonen> | OK |
| 14:20 | <gsnedders> | No, the major version is arbitrary. |
| 14:20 | <jgraham> | Afaict all elemnts of the Opera version number are arbitary apart from the core id |
| 14:20 | <hsivonen> | I wonder if Opera users complain about version number decisions |
| 14:20 | <gsnedders> | 11.60 and 12 are basically identical with the exception of 12 having hardware acceleration |
| 14:20 | <zcorpan> | it's not arbitrary! it's always greater than the previous version |
| 14:21 | <miketaylr> | and CORS |
| 14:21 | <jgraham> | It can be monotonically increasing and arbitary! |
| 14:21 | <hsivonen> | jgraham: what's the logic in the core id? |
| 14:21 | <zcorpan> | except, of course, between mobile and desktop |
| 14:21 | <gsnedders> | miketaylr: Right, 11.60 has now branched off |
| 14:21 | Ms2ger | points at the topic |
| 14:21 | <gsnedders> | miketaylr: But for a while they were almost identical |
| 14:21 | <hsivonen> | It seems that there's a build number and a Presto number |
| 14:21 | <miketaylr> | gsnedders: yeah |
| 14:22 | <jgraham> | hsivonen: The build number is a desktop thing that gets reset every so often |
| 14:22 | <jgraham> | So that is both arbitary and non monotonic |
| 14:22 | <jgraham> | the X.YY.ZZZ is the important number |
| 14:22 | <jgraham> | Specifically the ZZZ is the core version |
| 14:22 | <hsivonen> | jgraham: it's slightly unfair if your users aren't complaining at least about non-monotonic numbers :-) |
| 14:23 | <gsnedders> | hsivonen: It's basically x.y.z, x is a constant (2), y relates to API stability (at a platform-interface level), and z is just an ever increasing integer which is the revision, more or less |
| 14:23 | zcorpan | -> gym |
| 14:23 | <hsivonen> | gsnedders: so after 2.9.x there will be 2.10.y? |
| 14:23 | <gsnedders> | hsivonen: After 2.9.100 there will be 2.10.101 |
| 14:23 | <Ms2ger> | Or 2.X.y? |
| 14:24 | <jgraham> | In theory 2 isn't constant |
| 14:24 | <gsnedders> | "In theory". |
| 14:24 | <gsnedders> | Very much in theory, though. |
| 14:24 | <jgraham> | Well gecko changed to version 2 eventually |
| 14:24 | <hsivonen> | gsnedders: umm. it's already at 2.9.220 |
| 14:24 | <gsnedders> | hsivonen: I just took 100/101 as a random example |
| 14:24 | <hsivonen> | gsnedders: ah |
| 14:25 | <gsnedders> | hsivonen: The point is it increments regardless of the rest of it |
| 14:26 | <jgraham> | Everything is irrelevent apart from the last number which is a monotonically increasing integer |
| 14:26 | <jgraham> | This is the basic rule of thumb |
| 14:26 | <hsivonen> | yay for putting irrelevant numbers in the UA string |
| 14:26 | <jgraham> | Yeah :( |
| 14:27 | <gsnedders> | Indeed, there is actually 2.10.229 and 2.11.229, which differ in terms of whether new code can land which introduces breakage in the platform interfaces. |
| 14:28 | <gsnedders> | So they can differ, both in a black-box observable way and in a non-black-box observable way |
| 14:29 | <gsnedders> | (and for example, even something major such as Carakan had no affect, and went into both 2.4.17 and 2.5.17) |
| 14:29 | gsnedders | blinks realizing he can still remember that version |
| 14:30 | <gsnedders> | Oh, no, that was before we changed the release number format, so was 2.5.28 |
| 14:30 | <gsnedders> | (where all three numbers are essentially arbitary) |
| 14:31 | <gsnedders> | (the only difference from what we have now is that the final number just incremented every fortnight, bsaically) |
| 14:49 | karlcow | wants UA string becomes from arcane poetry to "Hey I'm a browser", "Hey I'm a bot" :) |
| 14:49 | <karlcow> | or something along that |
| 14:49 | <karlcow> | "Yo man, browser here" |
| 14:50 | <karlcow> | hmm maybe it should include a fortune too. |
| 14:50 | <karlcow> | and the weather forecast for the day. |
| 14:51 | <karlcow> | That would trigger not that much more errors than we have now |
| 14:51 | <jgraham> | HTTP should be done in haiku. I am a browser / Visiting your nice web site / So give me the page |
| 14:51 | <jgraham> | (pedants will note this is not a haiku) |
| 14:54 | <erlehmann> | karlcow, i use program/version for my own projects. |
| 14:54 | <erlehmann> | like redokast/11-22-3333 |
| 15:07 | <Workshiva> | jgraham: It would feel a bit weird if my browser started talking about the seasons and weather |
| 15:09 | Philip` | wonders to what extent the simplistic syllable-counting understanding of haiku in English is caused by schools using them as a way to teach people how to count syllables |
| 15:20 | <karlcow> | luv https://gist.github.com/1410787 |
| 15:21 | <karlcow> | Philip`: Haikus are complicated even in Japanese :) |
| 15:22 | <karlcow> | one kanji != one syllabus |
| 15:32 | <annevk> | http://jena.sourceforge.net/iri/javadoc/com/hp/hpl/jena/iri/IRIFactory.html hahahaha |
| 15:47 | <annevk> | I guess I should step up on my encoding work again... |
| 15:47 | <annevk> | I fear the multi-byte encodings... |
| 15:47 | <Velmont> | It is fear inducing, I agree. |
| 16:03 | <hsivonen> | I can tell that annevk is reading my emails |
| 16:16 | <Hixie> | just got an e-mail from someone who wanted to register on the wiki but couldn't |
| 16:16 | <Hixie> | do we have plans for reopening the wiki? |
| 16:19 | <annevk> | TIL: https://www.w3.org/2011/12/01-tag-nominations |
| 16:19 | <annevk> | (W3C member-only btw) |
| 16:19 | <annevk> | Hixie: we don't have good anti-spam measures :( |
| 16:20 | <Hixie> | not having a wiki at all seems a bit drastic as a solution :-) |
| 16:21 | <annevk> | it's a non-starter really, but I got tired of deleting a lot of pages and users |
| 16:21 | <annevk> | Lachy and AryehGregor have access and can maybe patch it somehow |
| 16:21 | <annevk> | but AryehGregor is occupied for the next two-three weeks I think |
| 16:21 | <annevk> | not sure about Lachy |
| 16:22 | <annevk> | I think I may have access too, but I no next to nothing about wiki software and there's other things I can be doing |
| 16:22 | <annevk> | know* |
| 16:22 | <Hixie> | yeah, join the club :-/ |
| 16:23 | <annevk> | I can look into it on Monday |
| 16:23 | <annevk> | tonight and weekend is somewhat occupied |
| 16:28 | <AryehGregor> | You can always create accounts manually for the time being. |
| 16:29 | <AryehGregor> | While logged in, go to Special:Userlogin. |
| 16:29 | <AryehGregor> | You might need to be an admin. |
| 16:29 | <AryehGregor> | Alternatively, enable some kind of confirmation for account creation, if there isn't already. |
| 16:33 | <Hixie> | hm, that's not too hard |
| 16:33 | <Hixie> | i guess we can do that like we do with the blog |
| 16:33 | <Hixie> | i'll put some notes on the wiki to say to contact us for an account |
| 16:34 | <Hixie> | I'll put in my e-mail for now, if it gets too much traffic I'll start rotating y'all in ;-) |
| 16:48 | <AryehGregor> | agraragarf. Typed Arrays as implemented in browsers are seriously not architecture-independent? |
| 16:49 | <AryehGregor> | W? T? F? |
| 16:49 | <AryehGregor> | I can understand the Khronos Group not getting web compat requirements, but is that how browsers actually implemented it? |
| 16:54 | <jgraham> | I guess that people wanted to follow the spec and get the perf benefits |
| 16:54 | <jgraham> | Not least because aiui Khronos does some sort of certification |
| 16:55 | <AryehGregor> | Right, perf benefits. It's really nice to have your code run as fast as possible when it isn't going to work anyway. |
| 16:57 | <jgraham> | Well how many browser vendors care about BE systems? |
| 16:58 | <AryehGregor> | How many browser vendors support BE systems? |
| 16:58 | <jgraham> | All I know is >= 1 |
| 16:58 | <AryehGregor> | That one being what, Opera? |
| 16:58 | <jgraham> | Yup |
| 16:59 | <AryehGregor> | I assume Gecko and WebKit at least compile on big-endian platforms, for their use as rendering libraries if nothing else. |
| 16:59 | <zewt> | well, if (read: when) code starts assuming the LE architecture-dependant choice, BE browser(s) will end up implementing it as LE for compat anyway; if/when that happens, we can point at it and go "see, this is why you were being silly, now fix your spec" |
| 17:00 | <AryehGregor> | Assuming there's such a thing as a BE browser, to within standard margins of error. |
| 17:00 | <zewt> | obscure mobile stuff, i suspect |
| 17:01 | <jgraham> | TVs mostly afaik |
| 17:01 | <jgraham> | Gecko seems to run on BE systems |
| 17:01 | <zewt> | at least, rare enough that it's highly unlikely that we'll end up with a silo of BE-dependant code |
| 17:01 | <zewt> | (which would be the "bad end" outcome) |
| 17:02 | <AryehGregor> | It's self-evident any BE browsers that want to support TypedArray will have to emulate LE, it's just really annoying that people are even pretending otherwise. It confuses people. |
| 17:02 | <AryehGregor> | Message to world: TypedArray is LE. Ignore the spec. |
| 17:03 | <zewt> | i've gradually been trying to nudge the WebGL guys towards something resembling modern web API design |
| 17:03 | <zewt> | only so much I can do alone, though. heh |
| 17:03 | <AryehGregor> | Oh well, not my problem. |
| 17:03 | <AryehGregor> | I've seen WebGL. It's horrifying. It looks like a straight port of a C API to JS. That's what it is, yes? |
| 17:03 | <zewt> | (trying to convince them to use prefixing in the Web way, not the OpenGL way; suggesting--without much hope on this one--that versioning their specs is not a good idea) |
| 17:04 | <zewt> | not exactly, but it definitely mimics OpenGL too directly |
| 17:04 | <AryehGregor> | That's what happens when we let non-web standards group make up web standards, I guess. |
| 17:04 | <AryehGregor> | Oh well. |
| 17:04 | <zewt> | that's mostly just a style issue, though, the API is perfectly usable despite that |
| 17:05 | <zewt> | the more fundamental problems are related to interop (hardware differences) and context loss |
| 17:06 | <zewt> | it doesn't abstract that stuff away like other APIs (because it basically can't, especially for context loss) |
| 17:09 | <jgraham> | To be fair, given my experience with smaller groups at W3C, it is not even slightly clear that Web Standards groups are much good at coming up with good web apis |
| 17:09 | <jgraham> | Or at least they typically have a ton of issues that someone needs to fix |
| 17:10 | <zewt> | i don't think webgl is horrifying; it definitely doesn't follow all of the same guidelines as other web APIs, which is unfortunate, but they're intrinsic to trying to expose something like OpenGL without inventing something entirely new (which is orders of magnitude harder) |
| 17:20 | <zewt> | am i the only one that find the concept of using language tags in WebVTT to allow correct TTS amusing |
| 17:21 | <zewt> | (the real use case is narrating translations, but it makes me think of people transcribing a voice track, then TTSing the transcription of something they just transcribed) |
| 17:24 | <zewt> | (i'm probably just easily amused) |
| 17:25 | <Hixie> | zewt: audio descriptions are a valid use case for VTT |
| 17:25 | <Hixie> | zewt: (for blind users) |
| 17:25 | <Hixie> | zewt: (or users like me, who are listening to a tv show on their ipod while cycling) |
| 17:26 | <zewt> | yeah, I get it, but my first mental image was listening to TV with the voices replaced with dr sbaitso |
| 17:29 | <zewt> | anyone know if any browsers actually vary their word wrapping based on @lang, out of curiosity? |
| 17:30 | <zewt> | wow, someone posting messages with a 90s BEGIN PHP SIGNATURE, heh |
| 17:30 | <zewt> | nostalgia |
| 17:31 | <michaelw> | zewt: closure implements TeX's paragraph formatting, IIRC, including hyphenation, which would be language dependent |
| 17:32 | <zewt> | michaelw: i mean actually using @lang, though, not just heuristic language detection |
| 17:32 | <michaelw> | zewt: ah |
| 17:33 | <zewt> | i know (most) browsers use it for font selection, i'm just not sure what else, if anything |
| 17:35 | <zewt> | ("most" seems to be everyone except webkit) |
| 19:24 | <mikenson> | Hello. Is this a good place to ask a question about parsing HTML5? I'm not sure on something in the spec. |
| 19:24 | <AryehGregor> | mikenson, yes, exactly the right place. |
| 19:24 | <mikenson> | Cool, thank you. |
| 19:25 | <mikenson> | The WHATWG spec, section 12.2.4.69 (Tokenizing character references) "Anything else" is confusing me. |
| 19:27 | <mikenson> | Are the named characters without semicolons in the table (12.5) the only characters accepted without semicolons? |
| 19:28 | <mikenson> | For example the spec uses "I'm ¬it; " which is parsed as "I'm ¬it;" (and a parse error). |
| 19:29 | <mikenson> | But would "I'm &neever;" be parsed as "I'm ≠ever;" (with a parse error)? |
| 19:29 | <mikenson> | I'm asking because the table doesn't have an entry for &ne (without the semicolon) but it does for ¬ (also without the semicolon). |
| 19:34 | <jgraham> | mikenson: Right. You just use the table verbatium |
| 19:37 | <mikenson> | jgraham: Thanks. That clears things up. |
| 19:38 | <mikenson> | Is there an explanation anywhere for which named characters don't require semicolons? |
| 19:50 | <jgraham> | Not as far as I know. It is basically just the list of things that browsers accepted without semicolons |
| 20:03 | <kennyluck> | Where is this list from then? |
| 20:04 | <zewt> | (i'd imagine by testing browsers) |
| 20:08 | <mikenson> | Another question: The 3rd from last paragraph says, "If the characters after the & consist of one or more characters … followed by a semicolon then this is a parse error." But what if they don't? For example, &alpha (no semicolon) -- not in the table, not consumed, and no parse error? |
| 20:22 | <Hixie> | mikenson: correct |
| 20:22 | <Hixie> | mikenson: if there's no semicolon, then there's no risk of there being ambiguity in the future |
| 20:23 | <Hixie> | mikenson: since we'll never add any without semicolons |
| 20:24 | <mikenson> | Hixie: Ok. That makes sense. |
| 20:29 | <mikenson> | Just to be sure I've got it, ""One α One &beta" (missing semicolon at the end) gets parsed as "One α One &beta" with the literal text & b e t a, and a parse error. Is that correct? |
| 20:30 | <Hixie> | no parse error, unless i'm mistaken |
| 20:31 | <mikenson> | Right. Because no match can be made, and it doesn't end in a semicolon. |
| 20:31 | <bga> | somebody should say web developers to not make json sites :/ |
| 20:32 | <martndemus> | any particular reason why? |
| 20:32 | <Hixie> | hsivonen: btw, looks like validator.nu does show an error in that case |
| 20:33 | <bga> | macpherson no indexing, not accesebility from old browsers |
| 20:34 | <zewt> | so everyone should stick to static pages forever |
| 20:34 | <bga> | yes |
| 20:34 | <martndemus> | ah, that's not jsons fault, thats the fault of the devs not making indexable fallbacks |
| 20:34 | <zewt> | no. |
| 20:34 | <bga> | grasefull degradation |
| 20:34 | <karlcow> | zewt: defines static pages ;) |
| 20:35 | <zewt> | pages that navigate to new URLs rather than updating in-place, in this context :) |
| 20:35 | <karlcow> | I think bga is talking about server rendered pages which is orthogonal to static pages |
| 20:36 | <Philip`> | Given that Google employs zillions of people, why haven't they solved the problem of indexing JS-generated pages yet? |
| 20:36 | <Philip`> | (assuming they haven't) |
| 20:36 | <bga> | zewt i mean sites which consist of whole js and json |
| 20:36 | <bga> | 0% html |
| 20:36 | <martndemus> | it should be static pages and json with pushState, not or |
| 20:37 | <zewt> | basically you'll need to get over the fact that different pages have different needs and priorities |
| 20:37 | <Philip`> | Surely you could just run a headless Chrome and simulate clicking on every link and seeing where you end up - how hard can it be? |
| 20:37 | <zewt> | making server-rendered indexable pages in addition to a dynamic JSON interface can be a *lot* of work; whether it's worth it is entirely up to the developer |
| 20:38 | <zewt> | all you can do is make sure developers are aware of the costs in both directions and make an informed choice |
| 20:38 | <bga> | i guess its break idea of acceseble web |
| 20:38 | <karlcow> | zewt: it is more complex than that. It's not only the choice of the developer |
| 20:38 | <karlcow> | because we live in society and some sites do not have the same social requirements |
| 20:38 | <zewt> | it's always entirely up to the developer. |
| 20:38 | <martndemus> | its not a lot of work when your back end is set up right (RoR works pretty awesome with the concept) |
| 20:39 | <zewt> | ("developer" being "the owner of the page"; obviously the choice may be management rather than a programmer) |
| 20:39 | <karlcow> | ahah 180º ;) |
| 20:39 | <zewt> | no? heh |
| 20:40 | <bga> | finally we will have webgl web. everything will be draw by webgl and no dom, html, css |
| 20:41 | <bga> | i dont see any differences between this future and Flash |
| 20:41 | <zewt> | you don't see any difference between open standards and closed, proprietary browser plugins? |
| 20:41 | <zewt> | i'm not sure i can help you then :) |
| 20:41 | <bga> | dont troll me plz |
| 20:41 | <zewt> | uh, i'm not |
| 20:41 | <zewt> | you're saying silly things, so I'm pointing them out |
| 20:42 | <karlcow> | nah the plug behind the head directly into the cortex. cortexJson |
| 20:42 | <bga> | zewt i mean who need dom, css when i can make 'awesome' site fully based of canvas? |
| 20:43 | <bga> | and ok no user text selection |
| 20:43 | <bga> | tab stop |
| 20:43 | <bga> | etc |
| 20:43 | <zewt> | again that's up to the developer; all you can do is make sure people know the costs |
| 20:43 | <martndemus> | you can, but not everyone owns pc/phones/we that can render those sites that fast |
| 20:44 | <bga> | zewt my idea thatstandards should limit developer, not give freedom to make silly things |
| 20:44 | <zewt> | you want sites to be restricted to doing things that *you* think are important |
| 20:45 | <bga> | yeah |
| 20:45 | <zewt> | it's imposing your priorities on everyone else |
| 20:45 | <zewt> | people have different priorities and needs and resources and goals |
| 20:45 | <bga> | zewt look at Java politics |
| 20:45 | <bga> | thay restict users |
| 20:46 | <zewt> | java is one of the worst languages I've ever used |
| 20:46 | <karlcow> | ThoughI mean it's amazing to see how much time companies spend to develop useless web sites just to sell ads |
| 20:46 | <martndemus> | all i wanna say, the web is not java |
| 20:47 | <bga> | but you restrict semanic of tags |
| 20:47 | <bga> | dont allow costom tags |
| 20:47 | <bga> | you restrict users |
| 20:47 | <bga> | restrict more plz |
| 20:47 | <zewt> | actually, you can create custom tags, it's just handwavingly "invalid" (which is completely meaningless, since it works) |
| 20:48 | <bga> | else web will binary |
| 20:48 | <martndemus> | see using html5 tags in ie6, you are creating elements |
| 20:48 | <bga> | and clicking 'view-source' you will see minified json |
| 20:48 | <bga> | or binary code |
| 20:49 | <bga> | it will be death of web |
| 20:49 | <zewt> | http://zewt.org/~glenn/gross.html blink is back, baby |
| 20:49 | <bga> | and html |
| 20:49 | <karlcow> | different models of society |
| 20:49 | <bga> | Ж. |
| 20:49 | <karlcow> | free market, social market, etc. :) old debate |
| 20:49 | <bga> | * :/ |
| 20:51 | <martndemus> | ooooh, thanks for the on = !on, never thought of that possibility =D |
| 20:51 | <bga> | zewt it will end by inventing 'new' web |
| 20:51 | <bga> | text based again |
| 20:51 | <bga> | new fidonet |
| 20:51 | <bga> | but ok |
| 20:51 | <bga> | if you want |
| 20:52 | <zewt> | these levels of hyperbole very rarely convince people of much, fyi |
| 20:52 | <bga> | but currently less and less sited available for me in lynx |
| 20:52 | <bga> | *sites |
| 20:52 | <martndemus> | why would you want to use lynx as your main browser? |
| 20:53 | <martndemus> | unless your browsing on a i368 |
| 20:53 | <martndemus> | or macII |
| 20:54 | <bga> | lynx is good browser, i dont need bloated chrome or ff |
| 20:54 | <zewt> | telling people you use lynx is also a bad idea if you're trying to convince people to change their authoring habits. heh |
| 20:55 | <bga> | zewt i just try ti remember that main idea of web is accesebility |
| 20:55 | <zewt> | no, it isn't. |
| 20:55 | <bga> | css is option |
| 20:55 | <bga> | js is option |
| 20:56 | <bga> | only html is required |
| 20:56 | <zewt> | it's one goal of the web. it's not the driving one, or the only one. |
| 20:56 | <zewt> | and if you think css or javascript are optional, you're way too far out of touch from the web in the real world today. it's nice to say that these things should be optional, but in practice, they just aren't. |
| 20:57 | <martndemus> | css is optional, but a hell lot of people choose to use it, same goes for js |
| 20:57 | <zewt> | you can disable css and javascript, and if you do, you'll break a massive number of sites |
| 20:57 | <bga> | zewt second is make html is new OS? |
| 20:58 | <zewt> | okay? |
| 20:58 | <bga> | zewt authors should use 'graseful degadation' |
| 20:58 | <bga> | its important |
| 20:58 | <zewt> | in the real world, "don't" trumps "should" |
| 20:58 | <martndemus> | no website authors can do whatever the fuck they want |
| 20:59 | <zewt> | s/no/all/ |
| 20:59 | <martndemus> | ^ |
| 20:59 | <zewt> | lots of people make sites that only work with css and javascript; the HTML Police aren't coming to take them away :) |
| 20:59 | <zewt> | we can shake our Fists of Principle at them all day, but it won't change it |
| 20:59 | <martndemus> | its up to the people to browse to the site or not |
| 20:59 | <zewt> | martndemus: and that's exactly it |
| 21:00 | <zewt> | everyone doesn't stop using a site because it requires css |
| 21:00 | <zewt> | 99.99% don't care or notice |
| 21:00 | <zewt> | thus there's no real pressure for sites to work without css |
| 21:00 | <zewt> | (abbreviating "css and javascript" to "css" because repetitive) |
| 21:01 | <bga> | i happy that some sites still have 'mobile' version |
| 21:02 | <bga> | but web way is disappointing me |
| 21:02 | <martndemus> | well, i would say, that the future of mobile is even more xHr/json |
| 21:04 | <martndemus> | well bga, with that opinion, you are the 1%. |
| 21:04 | <bga> | heh |
| 21:04 | <bga> | ok |
| 21:05 | <zewt> | as a lynx user, you need to move the decimal point many places further :) |
| 21:05 | <martndemus> | yeah |
| 21:05 | <martndemus> | its propably far less then 1%, but I always wanted to make a OWS like reference ;) |
| 21:05 | <bga> | macpherson i read same opinion every day |
| 21:06 | <bga> | in forims, chats etc |
| 21:06 | <martndemus> | sources? |
| 21:06 | <bga> | many ppl want html4 back |
| 21:06 | <bga> | michaelrtm mostly linux forums |
| 21:08 | <martndemus> | well, as you can see, the general opinion is different then theirs |
| 21:08 | <bga> | macpherson main reason that web browsers become bigger and bigger, slower and slower, eat more RAM but web content quality not better |
| 21:08 | <martndemus> | correction |
| 21:08 | <martndemus> | browsers are becoming faster |
| 21:08 | <zewt> | (it's rather more complicated than that, heh) |
| 21:09 | <martndemus> | and the RAM consumptions is neglegible |
| 21:09 | <zewt> | (that's just incorrect; FF regularly uses 1.5GB of my 8GB system, which is substantial) |
| 21:10 | <martndemus> | well, there is still 6,5 gb of ram doing nothing |
| 21:10 | <zewt> | um |
| 21:10 | <zewt> | somehow you've assumed that my computer is dedicated entirely to firefox :) |
| 21:11 | <gnarf> | sounds like thats a firefox problem :) |
| 21:11 | <martndemus> | nah |
| 21:11 | <zewt> | chrome isn't better, from what I've heard |
| 21:11 | <martndemus> | my chrome is using 1.1gb atm |
| 21:11 | <zewt> | (i suspect Opera is; no idea about IE) |
| 21:11 | <martndemus> | about 10-15 tabs open |
| 21:11 | <gnarf> | it really isn't but responsible browsing helps ( closing tabs ) - and tbh, plugins like flash hurt a lot more than JS engines |
| 21:12 | <zewt> | gnarf: i object to the idea that i have to close tabs to be "responsible" to work around browser leaks |
| 21:12 | <zewt> | i have ~80 tabs open right now; that's how I use the browser |
| 21:12 | <gnarf> | its not even about leaks, its about caching that much content in memory for display |
| 21:12 | <zewt> | that's how I've used browsers for many many years |
| 21:12 | <gnarf> | 80 tabs is insanity |
| 21:12 | <zewt> | no, not at all |
| 21:13 | <gnarf> | your mobile phone must hate you |
| 21:13 | <gnarf> | :) |
| 21:13 | <martndemus> | the browsers just scale up with what the general pc specs |
| 21:13 | <martndemus> | do |
| 21:13 | <zewt> | what does my mobile phone have to do with my desktop browsing patterns? heh |
| 21:13 | <martndemus> | but eh 1,5gb for 80 tabs |
| 21:13 | <martndemus> | thats impressive |
| 21:13 | <gnarf> | just saying... 80 tabs = 80 documents in memory |
| 21:13 | <zewt> | the same number of tabs was around 300mb around FF3 |
| 21:13 | <gnarf> | and prolly ~200 images |
| 21:14 | <roc> | zewt: but did the content of the tabs change since FF3? |
| 21:14 | <gnarf> | how many flash objects? |
| 21:14 | <zewt> | browsers discard decompressed images in idle tabs (from what I recall) |
| 21:14 | <zewt> | roc: i distinctly recall noticing memory usage jumping massively between a couple versions |
| 21:14 | <zewt> | i think 3 -> 3.5 doubled memory usage, and then it doubled again at some "upgrade" after that (4 maybe) |
| 21:15 | <roc> | there have been regressions, but we've fixed most of them |
| 21:15 | <zewt> | it's at 1.1GB right now, but I havn't had the browser open that long (few days) |
| 21:15 | <zewt> | (at least it caps out; it doesn't OOM or anything) |
| 21:15 | <zewt> | (or out-of-address-space; same thing) |
| 21:15 | <roc> | unless you have unusual browsing habits, the content of those tabs has certainly changed since FF3, so at the least some analysis would be required to make memory comparisions "against FF3" |
| 21:16 | <martndemus> | bottomline: this isnt the time anymore where computers had 32mb of RAM and the browser used less then 1mb, get real |
| 21:16 | <zewt> | roc: i'm comparing memory usage at the time of the upgrade |
| 21:16 | <gnarf> | or installing FF3 and testing |
| 21:16 | <zewt> | when I upgraded from 3 to 3.5, memory usage jumped way up |
| 21:16 | <zewt> | (too long ago to remember details) |
| 21:16 | <bga> | macpherson i have 512mb |
| 21:16 | <zewt> | i'm inclined to write bga off as a troll. heh |
| 21:17 | <gnarf> | most people have that on their video card nowadays |
| 21:17 | <martndemus> | yeah i figured you were on some very old pc |
| 21:17 | <roc> | yeah, there have been regressions, but we've fixed many of them |
| 21:17 | <roc> | so it's hard to tell |
| 21:18 | <martndemus> | Just remember this: if you use old tech, dont expect that all the new stuff works on your dinosaur tech |
| 21:18 | <roc> | also, like someone else said earlier, some increased memory usage is directly related to improving performance |
| 21:18 | <martndemus> | it just isnt made for your dinosaur pc |
| 21:18 | <roc> | this is very clear for JS and graphics |
| 21:19 | <zewt> | firefox has always had a tendency to get slower over time, leading to me restarting it eventually; there's always the sense that that's tied to memory usage, but I have no idea if there's actually any relationship (and subjective UI performance is very hard to diagnose, especially when it only happens after a week of use and goes away if I restart it to add diagnostics) |
| 21:19 | <bga> | macpherson as result i will use opera mini servers to browse web |
| 21:20 | <kennyluck> | bga, I think I know what you are saying. The Web I would like to see is the Web where there are lots of nice, xrefed textbooks like the HTML LS, but I don't want HTML4 back and I know the only I can make this happen is to solicit permission from textbook authors or to hack WebKit for better MathML support. |
| 21:20 | <roc> | yes, sluggishness is often tied to GC pauses which is tied ot memory usage |
| 21:20 | <roc> | zewt: BTW in response to your earlier question, Gecko uses language information to control auto-hyphenation, but other than that doesn't use it for line-breaking |
| 21:21 | <zewt> | roc: does it actually use @lang, or is "language information" just language heuristics? |
| 21:21 | <roc> | it actually uses @lang |
| 21:21 | <zewt> | k |
| 21:23 | <bga> | kennyluck yeah. dynamic is option. plain html, svg and other markup formats |
| 21:24 | <bga> | web for me is info, i dont need nice css3 animations to read info |
| 21:26 | <kennyluck> | bga, well, you can think this way. You need css3 animations to make people happy to use html instead of ppt. |
| 21:26 | <kennyluck> | or pdf. |
| 21:26 | <bga> | if author instead show me info what i need, require js support just for get info via XHR - i will read info somewhere else |
| 21:27 | <bga> | kennyluck if ppl want animation - its ok. but, again, its just option |
| 21:27 | <martndemus> | bga you are a statistical 0.0001% which most authors tend to not care about. It might sound harsh, but it is |
| 21:32 | <annevk> | ah, glad we found a solution for the wiki |
| 21:32 | <annevk> | not ideal, but definitely better than no signups |
| 21:46 | <annevk> | AryehGregor: is there some way to find out all users for a MediaWiki installation and then all the users that have never contributed? And maybe a list of those that made a single contribution? |
| 21:46 | <annevk> | AryehGregor: I'm wondering if I can do some cleanup as there are still some spam accounts activated |
| 21:46 | <annevk> | (that have not spammed yet, but probably will at some point in the future) |
| 22:16 | <jgraham> | zewt: It turns out that running a JITing javascript engine takes more memory than a bytecode interpreter. So probably all browsers regressed memory usage as they introduced those |
| 22:17 | <jgraham> | What roc says about X tabs is tremendously important though |
| 22:18 | <jgraham> | Comparing memory usage for unspecified hetrogenous pages is useless |
| 22:21 | <zewt> | jgraham: i'd be very surprised if the memory usage of compiled JIT data is on that order of magnitude |
| 22:21 | <jgraham> | I wouldn't |
| 22:21 | <jgraham> | It's not just compiled code, it's the datastructures that people adopted |
| 22:21 | <zewt> | a gig of memory for (say) 80 pages worth of JIT'd javascript? highly doubtful |
| 22:22 | <zewt> | anyway, i'm not that worried about memory usage; the progressive slowness of firefox is more bothersome |
| 22:22 | <jgraham> | Well sure it's not the only thing on those pages |
| 22:22 | <zewt> | (using more memory is annoying, but memory is cheap; I'll probably jump to 16 GB soon) |
| 22:23 | <zewt> | also, i'd expect browsers to discard regeneratable data from idle pages (like they do--from what I understand--with decompressed images) |
| 22:24 | <zewt> | (of course, some pages just never go idle, but plenty do) |
| 22:52 | <Hixie> | any bash mavens know what i'm doing wrong here? |
| 22:52 | <Hixie> | if [ ( ! -s cldr.inc ) -o ( "`svn info -r HEAD .cldr-data | grep -i "Last Changed Rev"`" != "`svn info .cldr-data | grep -i "Last Changed Rev"`" ) ]; then |
| 22:52 | <Hixie> | i get "bash: syntax error near unexpected token `!'" |
| 22:53 | <Hixie> | seems to work fine (though inefficiently) if i reverse it and remove the ()s: |
| 22:53 | <Hixie> | if [ "`svn info -r HEAD .cldr-data | grep -i "Last Changed Rev"`" != "`svn info .cldr-data | grep -i "Last Changed Rev"`" -o ! -s cldr.inc ]; then |
| 23:01 | <gavinc> | Best commit message of the day in python web app "[solon/feature/admin-tool] Fixing W3C errors in HTML." |
| 23:05 | <jgraham> | Hixie: I believe what you're doing wrong is trying to write scripts in bash |
| 23:06 | <Hixie> | no argument from me there |
| 23:06 | <Hixie> | unfortunately hte bulk of this is just running other apps |
| 23:11 | <bga> | zewt btw i guess opening donateware project 'accesable web' which is proxy servers that run requested site in phantomjs, generate links for js actions ala webtoolkit.eu and serve plain html to client is good solution |
| 23:12 | <bga> | same as opera mini |
| 23:12 | <bga> | may be bit different |