| 00:00 | <jamesr_> | tab replied at length to a few of L2L's threads |
| 00:00 | <jamesr_> | but the replies were just more spew of the same form |
| 00:00 | <jamesr_> | i plan to just ignore their emails |
| 00:30 | <jamesr_> | roc: if you can figure out what he's asking (or is he asking something) please let me know, i am curious |
| 01:57 | <MikeSmith> | who's L2L? |
| 02:35 | <roc> | someone sending a lot of incoherent messages to standards groups |
| 02:37 | <caitp> | i never claimed to be a good communicator |
| 02:38 | <roc> | nice try |
| 08:05 | <zcorpan> | tantek: http://krijnhoetmer.nl/irc-logs/css/20140908#l-479 |
| 08:08 | <annevk> | hsivonen: btw, mobile Safari does exactly what we talked about the other show, show only the EV name and not the domain |
| 08:12 | <zcorpan> | Hixie: removing n.whatwg.org seems like it would solve that problem |
| 08:14 | <annevk> | zcorpan: what problem? I think n.whatwg.org is already solved |
| 08:14 | <zcorpan> | annevk: that people will use https in their xmlns |
| 08:19 | <zcorpan> | https://www.google.com/search?q=n+whatwg+work - google adds " - WhatWG" to the titles |
| 08:20 | <annevk> | Somewhat ironic that https://hstspreload.appspot.com/ doesn't use HSTS |
| 08:21 | <annevk> | zcorpan: I wonder where it gets that from |
| 08:22 | <karlcow> | I wonder if zcorpan was searching for https://www.google.co.jp/search?as_epq=n.whatwg.org |
| 08:23 | <zcorpan> | karlcow: no |
| 08:23 | <karlcow> | oh understood |
| 08:25 | <karlcow> | I wonder where they get that from |
| 08:39 | <annevk> | Bah, wget doesn't do certificate checks correctly |
| 08:39 | <MikeSmith> | I always turn off the wget certificate-checking option |
| 08:39 | <MikeSmith> | I think it's on by default and it mostly just gets in the way |
| 08:40 | <annevk> | Might be that DreamHost has an ancient version of wget? https://savannah.gnu.org/bugs/index.php?20421 |
| 08:40 | MikeSmith | reads |
| 08:43 | <MikeSmith> | yeah the current version of wget in debian stable is 1.13.4 |
| 08:44 | <MikeSmith> | but then debian also sometimes patches upstream stuff in ways that actually break things |
| 08:48 | <karlcow> | That would be fun to test all the command line clients: httpie, curl, etc. |
| 08:52 | <darobin> | my experience with cert checking in command line stuff is that it is very often wrong |
| 08:52 | <darobin> | (including PhantomJS) |
| 08:53 | <darobin> | that said, I wouldn't be shocked if it were because they were all doing it somewhat strictly by default and browsers are laxer for compat |
| 08:53 | <jgraham> | Writing in JS not magically leading to implementation quality shocker? |
| 08:59 | <MikeSmith> | since even the browsers don't behave interoperably with cert checking I guess it's not surprising that command-line tools behave the way they do here |
| 09:13 | <darobin> | jgraham: not sure what you meant? Phantom isn't written in JS :) |
| 10:08 | <annevk> | DreamHost has 1.12 per wget --version |
| 11:58 | <smaug____> | hayato: how should older shadow trees work in case they are not in the composed document, but the host is |
| 11:59 | <smaug____> | hayato: so, the newer shadow tree doesn't have <shadow> |
| 12:14 | <smaug____> | hayato: filed couple of bugs related to that |
| 12:21 | <MikeSmith> | annevk: It'd seem like 1.12 should have that bug fix from 2007. unless it's one of those cases where the Debian packager unfixed/regressed it |
| 12:22 | <annevk> | MikeSmith: hmm yeah |
| 12:22 | <annevk> | Mac OS X doesn't ship with wget, and I don't really care I guess |
| 13:20 | <mathiasbynens> | zcorpan: woah, nice work on those <picture> tests! |
| 13:20 | <zcorpan> | mathiasbynens: thx |
| 13:21 | <zcorpan> | mathiasbynens: wanna review? :-) |
| 13:21 | <mathiasbynens> | just went over them |
| 13:21 | <mathiasbynens> | all i could find was a small nitpick |
| 13:21 | <zcorpan> | ah ok. thanks! |
| 13:22 | <mathiasbynens> | (posted as note_ |
| 13:22 | <mathiasbynens> | ) |
| 13:23 | <jgraham> | mathiasbynens++ |
| 13:23 | <jgraham> | Also zcorpan++ but that's implied |
| 13:29 | <zcorpan> | gotta do something useful at a csswg f2f |
| 13:30 | <jgraham> | mathiasbynens: If you reviewed the tests you should mark them as reviewed |
| 13:33 | <jgraham> | mathiasbynens: (and now I changed the review so that you can ;) |
| 13:36 | <mathiasbynens> | jgraham: ty |
| 13:40 | <jgraham> | mathiasbynens: It looks like you pressed "I will review this" (which isn't necessary) but didn't mark anything as reviewed yet. Intentional? |
| 13:41 | <zcorpan> | mathiasbynens: click "everything" in https://critic.hoppipolla.co.uk/showcommit?first=9a149860&last=87823b51&review=2529 |
| 13:43 | <mathiasbynens> | jgraham: nah, just a critic n00b |
| 13:43 | <jgraham> | mathiasbynens: OK, np, welcome to the learning curve ;) |
| 13:43 | <mathiasbynens> | zcorpan: thanks |
| 13:45 | <jgraham> | For future reference "I will review this" basically just means "don't let anyone else mark this as reviewed". To actually mark something as reviewed you a) have to be a reviewed for that path and b) click the checkbox next to the file that you have reviewed |
| 13:45 | <jgraham> | (and then submit like for comments) |
| 13:45 | <foolip> | jgraham: zcorpan says that testharness.js has a "Not run" result. how does one use that and when is it appropriate? |
| 13:46 | <jgraham> | foolip: You don't use it. You get it if a test has been declared using async_test() but no .step() has been called when the overall test times out |
| 13:46 | <jgraham> | If a .step() has been called you get Timeout |
| 13:46 | <jgraham> | (I hope :) |
| 13:47 | <foolip> | ok, doesn't sound like the right tool for the job then :) |
| 13:47 | <foolip> | thanks jgraham! |
| 13:47 | <jgraham> | foolip: What are you trying to achieve? |
| 13:48 | <foolip> | jgraham: I'm trying to test for a media element GC bug and it's unavoidably racy, and it would be nice to distinguish between "passed" and "couldn't really test" |
| 13:49 | <foolip> | jgraham: https://codereview.chromium.org/552303006/diff2/150001:170001/LayoutTests/media/gc-while-seeking.html |
| 13:51 | <jgraham> | foolip: Oh interesting. I guess we could make done() take a human-readable message or introduce a new status, or something, but I'm not sure what has the right cost/effort tradeoff |
| 13:52 | <foolip> | jgraham: I wouldn't suggest that either |
| 13:53 | <foolip> | Obviously the risk with a test like this is that it ends up always taking the shortcut t.done(), but I don't see a way to change that. A unit test at a lower level seems like the best answer so far. |
| 13:56 | <jgraham> | foolip: Yeah, I tend to agree |
| 14:03 | <MikeSmith> | is there anybody good and active in the CSS WG who speaks Mandarin? |
| 14:18 | <roc> | what's the problem? |
| 14:18 | <roc> | I don't speak Mandarin, but my wife does, which is sometimes useful... |
| 14:21 | <MikeSmith> | roc: possible speaking opportunity in Beijing |
| 14:25 | <roc> | haha ok no |
| 14:29 | <MikeSmith> | roc: too late you already volunteered :) |
| 14:30 | <boogyman> | haha |
| 14:56 | <jgraham> | hsivonen: Did you get the chance to update those fragment tests? |
| 14:56 | <jgraham> | (for the review comment[s]) |
| 14:57 | <MikeSmith> | does anybody here use Symphony CMS? |
| 14:57 | <MikeSmith> | asking for a friend |
| 14:59 | <MikeSmith> | nm |
| 14:59 | <MikeSmith> | was confused with symfony |
| 15:03 | <darobin> | lol |
| 15:05 | <MikeSmith> | darobin: I'm hurt my your pointing and laughing at my ignorance |
| 15:06 | <MikeSmith> | I've tarnished my reputation as a PHP expert |
| 15:07 | <darobin> | MikeSmith: if it's any consolation I have now tarnished my reputation by demonstrating some degree of PHP expertise :) |
| 15:08 | <MikeSmith> | haha you took the bait |
| 15:08 | <MikeSmith> | my next trick question is about Wordpress |
| 15:08 | <MikeSmith> | be ready |
| 15:08 | <darobin> | I'm glad most of the related explanation took place on a secret team channel |
| 15:09 | <darobin> | who said transparency was a good thing? |
| 15:09 | <darobin> | luckily, I actually know pretty much fuck-all about Wordpress :) |
| 15:10 | MikeSmith | scrambles to change the name of the secret team-only php channel |
| 15:11 | <Hixie> | christ, dreamhost support is moronic these days |
| 15:14 | <MikeSmith> | Hixie: I don't understand why you guys don't switch to virtual hosting |
| 15:15 | <boogyman> | Hixie: try their twitter handle. phone is non-existent, and email always seems to take days. |
| 15:16 | <Hixie> | MikeSmith: how do you mean? who would you recommend? we're on virtual hosting at dreamhost |
| 15:16 | <Hixie> | boogyman: the speed is not the issue |
| 15:16 | <Hixie> | boogyman: they just seem to only respond in platitudes |
| 15:16 | <Hixie> | brb |
| 15:19 | <jgraham> | Hixie: I thought you were on a shared host not a VPS? Or did that change? |
| 15:24 | <MikeSmith> | Hixie: yeah, I meant what jgraham said -- a separate virtual machine that you have root access to and control, not shared hosting |
| 15:26 | <annevk> | MikeSmith: but then, maintenance |
| 15:27 | <MikeSmith> | annevk: I think it'd be less maintenance in the long run |
| 15:28 | <jgraham> | Yeah, my experience is that VPS is a huge win over shared hosting |
| 15:28 | <jgraham> | I mean unless you want to run email or something |
| 15:29 | <jgraham> | But for websites |
| 15:30 | <mounir> | annevk: I'm looking at your comments at https://github.com/w3c/screen-orientation/issues/72 |
| 15:30 | <mounir> | annevk: I wonder what you mean by "They do not queue tasks to synchronize but rather just dispatch events" |
| 15:31 | <mounir> | annevk: do you want me to queue a task before resolving the promise and before fire the events? |
| 15:31 | <annevk> | mounir: firing events needs to happen from a task |
| 15:31 | <annevk> | MikeSmith: hmm yeah perhaps |
| 15:32 | <mounir> | annevk: ideally, 4.4 would run on the animation task source you want to add |
| 15:32 | <annevk> | MikeSmith: I've never run my own server software, I could give it a shot I suppose... Recommendations? |
| 15:32 | <mounir> | annevk: but it's not defined yet |
| 15:34 | <annevk> | mounir: then you need to add some kind of issue marker |
| 15:34 | <annevk> | mounir: pretending everything is okay seems bad |
| 15:35 | <annevk> | MikeSmith: also, recommendations for a hosting provider? |
| 15:35 | <annevk> | MikeSmith: I was considering to give https://www.transip.eu/vps/ a try as I have domains registered there |
| 15:36 | <mounir> | annevk: issue #40 is there for that |
| 15:36 | <annevk> | mounir: in the spec |
| 15:37 | <MikeSmith> | jgraham: my experience is that even running a mail server is not a lot of work -- for the MTA just use exim and everything pretty just works. For POP/IMAP, pretty much all the options suck, just in different ways, but it's more just annoyances than it is about being a lot of maintenance work |
| 15:38 | <annevk> | MikeSmith: what about spam? |
| 15:38 | <MikeSmith> | annevk: I think http://www.bytemark.co.uk is good |
| 15:39 | <MikeSmith> | annevk: yeah spam is another "all options suck" case. You just choose one and live with it |
| 15:39 | <MikeSmith> | but I guess jgraham is right you're better off not trying to run a mail server if you can avoid it |
| 15:42 | <mounir> | annevk: I could have a look and see if respec has a way to have issue markers that disappear for publication documents |
| 15:42 | <mounir> | annevk: it's less of a trouble for living document obviously |
| 15:42 | <jgraham> | annevk: As I said I have used Linode, and I have heard people say that Digital Ocean is OK. Gandi also have something. |
| 15:45 | <Hixie> | MikeSmith: oh we've been on a virtual host for years |
| 15:45 | <Hixie> | MikeSmith: virtual hosts are still shared, that's why they're virtual :-) |
| 15:45 | <MikeSmith> | hair splitter |
| 15:45 | <Hixie> | it's not that fine a hair to split |
| 15:45 | <annevk> | Hixie: but it's not like you're running our own Apache server installation |
| 15:46 | <Hixie> | i've had to have them move me to a different underlying host a couple of times because of the other people on the hardware |
| 15:46 | <MikeSmith> | forgive me for not being precise I meant what jgraham described |
| 15:46 | <Hixie> | which is the main problem with shared hosting |
| 15:46 | <Hixie> | annevk: sure, it's managed |
| 15:46 | <Hixie> | annevk: you can have dedicated servers that are managed too |
| 15:46 | <Hixie> | annevk: managed vs virtual are entirely orthogonal |
| 15:47 | <annevk> | Hixie: yeah so I guess what people suggest here is to go unmanaged |
| 15:47 | <annevk> | Hixie: I suspect VPS became somewhat synonymous with that |
| 15:47 | <MikeSmith> | Hixie: I'm not sure your taxonomy there actually aligns with how the terms are typically used |
| 15:48 | <jgraham> | Hixie: In theory, perhaps. I think a one dimensional view would capture most of the available options |
| 15:48 | <Hixie> | i'm definitely not going unmanaged |
| 15:49 | <Hixie> | MikeSmith: quite possible. i'm just using the english terms, i don't know if there's terms of art here. |
| 15:49 | <Hixie> | jgraham: well you can always collapse the options, sure |
| 15:49 | <Hixie> | jgraham: managed shared, managed virtual, unmanaged virtual, managed dedicated, unmanaged dedicated |
| 15:49 | <Asterfield> | So I'm sure that this idea has been proposed before, but I'm curious to know what your thoughts on this are. I've been learning the Drag and Drop API over the last few days, and my introductory experience has been baffling and weird. |
| 15:50 | <Hixie> | Asterfield: the API itself is something we inherited from IE4 which inherited it from Microsoft's COM+ |
| 15:50 | <Hixie> | Asterfield: the attributes (dropzone and draggable) are attempts to mask the crazy a little |
| 15:50 | <Hixie> | Asterfield: but yeah, the end result isn't ideal |
| 15:51 | <Asterfield> | That doesn't mean it can't/shouldn't be improved :) Would it make sense to have the event ordering for d-n-d to behave more like the :hover state? http://imgur.com/a/tYpW3 |
| 15:51 | <Hixie> | we can't change it, it's widely shipped and pages depend on it |
| 15:51 | <Hixie> | we can add features, but we can't change things that are shipped |
| 15:51 | <Asterfield> | :| that is unfortunate. Curses, microsoft |
| 15:52 | <Asterfield> | Okay, let's say a new set of events then. I know that would bring the count to 9, but the old ones could be depricated if it works, no? |
| 15:52 | <Asterfield> | dragin and dragout :P |
| 15:53 | <Hixie> | we could make new events entirely, but then we'd have to convince the browser vendors to implement a second redundant drag and drop event processing model, instead of doing whatever it is they are doing now |
| 15:53 | <Asterfield> | Are there any serious problems with that? |
| 15:53 | <Hixie> | so.... what do you think it is they are doing that's less important? :-) |
| 15:55 | <MikeSmith> | annevk: Hixie: anyway http://www.bigv.io is specifically what I was recommending, as far as providers. Among other reasons, they're relatively small and getting support from them amounts to talking one of maybe three people who've actually run/maintain their server hardware/software and have for years and are extremely capable and very responsive |
| 15:55 | <Asterfield> | Hixie: open source, anyone (*cough* me *cough*) could add these changes if they have the time |
| 15:56 | <Asterfield> | It doesn't necessarily have to come at the expense of a current project |
| 15:56 | <Asterfield> | The issue would be getting the vendors to agree to the idea in theory, no? |
| 15:56 | <Hixie> | Asterfield: not all the browsers are open source, and even the open source ones, you'd be taking engineers away from whatever they're doing to orient the new engineer, review their patches, fix security bugs going forward, etc |
| 15:57 | <Asterfield> | Hixie: That's true. Are you saying that this idea should not be attempted, and that the current IE4 model is good as-is? |
| 15:57 | <Hixie> | Asterfield: in practice, unless you're personally volunteering to implement it in Safari, Chrome, and Firefox, it would end up taking away from another feature |
| 15:58 | <Hixie> | Asterfield: i'm saying that without vendor buy-in, it's pointless. If the vendors can be convinced that it's worth it, then I'd be happy to spec it or work with someone to spec it |
| 15:58 | <Asterfield> | Hixie: So where does vendor buy in come from? Do I google around for the emails of all the important Mozilla people? :P |
| 15:59 | <Asterfield> | Make a petition perhaps? |
| 15:59 | <Hixie> | i've been at this for about 15 years and i've still no idea how to convince people :-( |
| 16:00 | <Asterfield> | :P Aha, awesome. Do you know where I might be able to start? |
| 16:00 | <Hixie> | generally, posting in the project mailing lists is a good place to start |
| 16:00 | <Hixie> | but you have to be careful |
| 16:00 | <Hixie> | most people who do that manage to flip their bozo bit really early |
| 16:00 | <Asterfield> | How would you recommend avoiding this flipping of the bozo bit? |
| 16:03 | <Asterfield> | Or at least explain what constitutes flipping the bozo bit :P |
| 16:03 | <Hixie> | http://en.wikipedia.org/wiki/Bozo_bit :-) |
| 16:04 | <Hixie> | a lot of people come in with all these high-minded ideas and little patience for the existing engineers' opinions |
| 16:05 | <Asterfield> | Hixie: Hm. I shall be careful then. Thanks for your help, Hixie :) |
| 16:05 | <mounir> | annevk: http://www.w3.org/TR/page-visibility/#now-visible-algorithm < would that be find to hook at the end of that? |
| 16:05 | <Hixie> | Asterfield: good luck |
| 16:05 | <mounir> | annevk: or would that still be monkey patching to you? |
| 16:05 | <caitp> | people who you consider to have their bozo bit set, are people who consider you to have your bozo bit set |
| 16:06 | <caitp> | it's not really a one-sided thing |
| 16:07 | <Asterfield> | Oh, while I'm at it. Do event models fall under the DOM spec? |
| 16:08 | <Asterfield> | It seems they do |
| 16:10 | <Hixie> | caitp: that is often not the case |
| 16:10 | <Hixie> | caitp: there's plenty of times where someone will keep trying to convince someone of something long after that other person has stopped listening |
| 16:10 | <Hixie> | Asterfield: "event models"? |
| 16:10 | <Hixie> | Asterfield: they belong in the spec of whatever they're modeling |
| 16:11 | <caitp> | I think you'll find that people will tend to ignore your opinion if they believe you've thoroughly missed the point regarding something, or just fundamentally disagree with your perspective. the reason they'll keep yammering at you after you've stopped listening, is because they consider you to be in a position of authority or influence |
| 16:11 | <caitp> | it's just basic human dynamics, they might think they need to convince you even they think you are wrong |
| 16:11 | <caitp> | or especially if |
| 16:17 | <caitp> | it's like the intent-to-remove showModalDialog thread, even though people have very good reasons for wanting to get rid of it, you'll still see people trying to convince people with influence not to remove it, because to them, the reasons for removing it are not worthwhile --- thy aren't stupid just because they don't align their opinions with the opinions of implementors and maintainers, they just put importance on d |
| 16:17 | <caitp> | ifferent things which are more pertinent to their own needs |
| 16:27 | <Hixie> | caitp: yes, it absolutely is because they want to convince you. That's my point. Asterfield has to be careful not to have the people he wants to convince flip the bozo bit on him. I don't think in that context it's usual for the bozo bit flipping to be symmetric. |
| 16:28 | <annevk> | mounir: yes, because it is not apparent from the algorithm it can be extended |
| 16:28 | <annevk> | mounir: it's already better than what you have now, but you need to get page visibility updated |
| 16:28 | <annevk> | mounir: it seems it might need to be updated anyway to also make use of the animation frame task... |
| 16:29 | <mounir> | hmm page visibility would be updated at that point, right? |
| 16:30 | <caitp> | that's not what I'm saying --- on both sides, you're ignoring the other sides input. I mean yeah, they aren't "totally" ignoring the other sides input, since they allow people to use showModalDialog for another year with special flags, but they're basically still stuck behind "this is a really bad API, a really bad user experience, and it's just generally bad --- users don't seem to agree that it's a really bad API, a |
| 16:30 | <caitp> | nd users aren't going to care how complicated it is, so obviously they're going to ignore that opinion and believe that it's boneheaded to remove it |
| 16:31 | <caitp> | like, there's no real way around it, they're going to feel like their needs are being ignored, and developers are going to worry about maintainability more than meeting their needs, which they deem to be invalid |
| 16:31 | <caitp> | so both sides are essentially getting it wrong, despite care |
| 16:31 | <caitp> | it's just the way people are |
| 16:32 | <caitp> | but I dunno why I'm talking about it, it's just one of those fascinating things |
| 16:32 | caitp | moves on to other endeavours |
| 16:35 | <annevk> | mounir: I guess, but this is the W3C, you never know |
| 16:36 | <annevk> | mounir: took them almost a decade to start updating HTML again, or DOM |
| 16:36 | <mounir> | annevk: instead of trolling, wanna have a look at that PR? :) |
| 16:39 | <Hixie> | caitp: flipping the bozo bit means more than just talking at instead of talking with. It means essentially killfiling. |
| 16:39 | <annevk> | mounir: hey you started with "I'm not sure we can change that now we're going to LC" :p |
| 16:39 | <annevk> | done btw, back later |
| 17:29 | <Hixie> | MikeSmith: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=26195 |
| 17:37 | <annevk> | MikeSmith: thanks |
| 17:37 | <annevk> | jgraham: sorry for not remembering |
| 17:37 | <annevk> | jgraham: Linode looks pretty good |
| 17:44 | <TabAtkins_> | roc: L2L is a crazy person and went in my killfile pretty quickly. We should ban him from the w3 lists. |
| 17:49 | <TabAtkins_> | What's a polite way to say "the entire preceding paragraph is nonsense"? |
| 17:52 | <TabAtkins> | Okay, going with "None of what you just said is true". |
| 17:52 | <TabAtkins> | Can't come up with anything better. |
| 17:53 | <TabAtkins> | And don't really have any respect for the person I'm talking to, just trying to avoid looking overtly hostile to the rest of the list. |
| 18:01 | <caitp> | still sounds a bit hostile |
| 18:01 | <TabAtkins> | Sure, it definitely is, but this guy knows he's spewing nonsense, and he always does it. I don't like that sort of shit looking official and correct when it's complete fantasy. |
| 18:02 | <caitp> | I'm not telling you what to do :p but maybe clearly and politely showing why it's nonsense might be more effective and also time consuming |
| 18:03 | <TabAtkins> | The rest of my email outside that sentence was fine. |
| 18:10 | <annevk> | So I just found out via @antimattur that there's a thread going on in public-html about "after 5"... Seems to be mostly about modularizing, which seems weird, since they still don't really write any of the text. |
| 18:12 | <annevk> | Who volunteered again for console.og? |
| 18:12 | <annevk> | console.log even. That someone should read https://www.w3.org/Bugs/Public/show_bug.cgi?id=26182 |
| 18:12 | <annevk> | Domenic: ^ |
| 18:13 | <Domenic> | terinjokes did, haven't heard any updates recently |
| 18:13 | <terinjokes> | Domenic: i filed a bug https://code.google.com/p/chromium/issues/detail?id=415645 |
| 18:14 | <Domenic> | paul_irish ^ |
| 18:14 | <terinjokes> | i already messaged him |
| 18:14 | <terinjokes> | :) |
| 18:14 | <Domenic> | terinjokes we were talking about console spec |
| 18:14 | <terinjokes> | ah |
| 18:14 | <terinjokes> | i created a github repo, but haven't actually pushed anything I've done up yet |
| 18:17 | <TabAtkins> | terinjokes: Get it up! Develop in the open. ^_^ |
| 18:18 | <terinjokes> | two seconds |
| 18:18 | <terinjokes> | :) |
| 18:21 | <terinjokes> | https://github.com/terinjokes/console-spec/blob/master/index.bs |
| 18:21 | <terinjokes> | just a basic outline so far :'( |
| 18:22 | <terinjokes> | TabAtkins: i couldn't get links in the IDL working with <dfn>s that weren't part of a <dd>, not sure what I'm missing |
| 18:23 | <TabAtkins> | terinjokes: The one <dfn> I see has a capitalization error - you need to say for="Console". |
| 18:23 | <terinjokes> | oh, i didn't know it was case sensative |
| 18:23 | <TabAtkins> | (Bikeshed's handling of capitalization is still kinda fucked up; I need to spend time making it sane. The code I wrote there predates IDL definitions, where capitalization matters.) |
| 18:24 | <terinjokes> | yep, that did it |
| 18:24 | <TabAtkins> | terinjokes: Assume it's only sensitive for IDL things (after all, `var x = console;` and `var x = Console;` are two very different statements). |
| 18:24 | <TabAtkins> | terinjokes: Even if my capitalization handling is messed up for now, it'll be fixed later, so might as well write it correctly. |
| 18:24 | <terinjokes> | noted |
| 18:25 | <TabAtkins> | (Other things will remain case-insensitive, so you can, for example, capitalize a term that appears as the first word of a sentence and still have it link up.) |
| 18:25 | <TabAtkins> | terinjokes: That said, I think the interface really is spelled "console". ^_^ |
| 18:25 | <terinjokes> | also should probably create a real logo… i'm sure that kitten, while cute, has a copyright |
| 18:25 | <TabAtkins> | hahahahaha |
| 18:25 | <TabAtkins> | I didn't even see that. |
| 18:26 | <terinjokes> | so in this case there's a one-to-one interface name to what the variable exposed is called? |
| 18:27 | <TabAtkins> | Yeah, that's how WebIDL works. |
| 18:27 | <Domenic> | Unsure |
| 18:27 | <Domenic> | Is console an instance of Console |
| 18:27 | <TabAtkins> | Console doesn't exist. |
| 18:27 | <Domenic> | or is console a single interface with lots of static methods |
| 18:28 | <terinjokes> | Domenic: sounds like an openable issue |
| 18:28 | <TabAtkins> | So etiher it's a [NoInterfaceObject], and you're doing something complicated, or it's just spelled "console". |
| 18:28 | <Domenic> | I think in browsers it is an instance of a Console |
| 18:28 | <TabAtkins> | Whoops, you're right. |
| 18:28 | <terinjokes> | yes, there's a Console constructor |
| 18:28 | <TabAtkins> | Console is the interface name. |
| 18:29 | <TabAtkins> | terinjokes: What browser? |
| 18:29 | <TabAtkins> | Chrome doesn't expose Console on the window. |
| 18:29 | <Domenic> | If we can convince people to use static methods then you get the benefit of them being "bound" (so you can do .forEach(console.log) |
| 18:29 | <terinjokes> | console.constructor exists and is labeled "Console" |
| 18:29 | <Domenic> | but no browser implements that sooooo |
| 18:29 | <terinjokes> | but not callable |
| 18:29 | <terinjokes> | well, callable, but throws |
| 18:29 | <TabAtkins> | terinjokes: Ah, right. |
| 18:30 | <TabAtkins> | terinjokes: So, what you wanna do is name the interface "Console", but put "[NoInterfaceObject]" on the line before it. |
| 18:30 | <TabAtkins> | That means it doesn't show up as a property on window. |
| 18:31 | <Domenic> | You'll also need some kind of patch onto window to expose .console. I think partial interfaces is how that is done? |
| 18:31 | <TabAtkins> | Yup, |
| 18:31 | <Domenic> | Also I think workers too these days |
| 18:31 | <TabAtkins> | partial interface Window { attribute Console console; } |
| 18:32 | <Domenic> | DID YOU KNOW? window.console can be overwritten (in Firefox at least) |
| 18:32 | <TabAtkins> | And the same on WorkerGlobalScope. |
| 18:32 | <TabAtkins> | Domenic: Makes sense. |
| 18:32 | <TabAtkins> | Most things can? |
| 18:32 | <Domenic> | Really? My experience is WebIDL people love their `readonly` |
| 18:32 | <terinjokes> | and ditto in Chrome |
| 18:33 | <Domenic> | actually attribute Console console isn't quite right in that case, since I can do `window.console = "foo"` and it works with no type-checking errors |
| 18:33 | <terinjokes> | Domenic: i have a whole project that works by overriding lots of window and DOM things |
| 18:33 | <TabAtkins> | Domenic: Hmmmmmmmm |
| 18:33 | <terinjokes> | Domenic: you can do it, but should you be able to? |
| 18:34 | <TabAtkins> | terinjokes: Best to document what browsers currently do (when they agree), and then have issues in the spec for things you think we might want to change. |
| 18:34 | <Domenic> | WORSE! In Chrome it's a data property; in Firefox it's a getter/setter |
| 18:35 | <Domenic> | And the tiebreaker is... IE11 says data property! |
| 18:35 | <Domenic> | Now we're really in trouble |
| 18:35 | <TabAtkins> | Domenic: We love our data properties in Chrome. |
| 18:35 | <Domenic> | TabAtkins: oh right, I forgot about that |
| 18:35 | <smaug____> | hmm |
| 18:36 | <smaug____> | how would one even then explain window.console in webidl |
| 18:36 | <TabAtkins> | Dunno if we ever made the great "put our data properties on the prototype as getters/setters, like WebIDL requires" migration or not. |
| 18:36 | <TabAtkins> | smaug____: as "any", with prose defining its initial value? |
| 18:36 | <Domenic> | TabAtkins: I believe another heroic attempt is being made, or at least was a week or so ago. |
| 18:37 | <smaug____> | TabAtkins: why would 'any' affect anything here |
| 18:37 | <Domenic> | smaug____: if it's a data property, you're screwed, WebIDL can't do it |
| 18:37 | <smaug____> | yeah, I was talking about being data property |
| 18:37 | <TabAtkins> | smaug____: Oh, sorry, thought you were talking about the "can override with whatever I want" part. |
| 18:37 | smaug____ | certainly doesn't want to make a special case here |
| 18:38 | <TabAtkins> | Yeah, no, just define it as normal in WebIDL, and let browsers fix themselves. Firefox is already fine. |
| 18:40 | <terinjokes> | TabAtkins: bikeshed doesn't seem to like "any" attributes |
| 18:40 | <TabAtkins> | terinjokes: Interesting! It's possible that that's not even allowed. |
| 18:41 | <terinjokes> | you can see from the last push |
| 18:45 | <terinjokes> | TabAtkins: i don't think it is allowed |
| 18:48 | <TabAtkins> | terinjokes: I mean that it might not be allowed in the WebIDL grammar. |
| 18:48 | <terinjokes> | yes |
| 18:48 | <TabAtkins> | (Bikeshed's webidl parser is built directly from the grammar in the spec.) |
| 18:48 | <TabAtkins> | Ah, okay. |
| 18:48 | <terinjokes> | can any's contain nullable sequence types? |
| 18:48 | <TabAtkins> | any is any. |
| 18:48 | <TabAtkins> | I guess mark "console" as type "Console" for now, with an issue tracking its mutability. |
| 18:49 | <terinjokes> | cool |
| 18:49 | <terinjokes> | will get to that tonight |
| 19:43 | <TabAtkins> | "URL modifiers" isn't already a claimed term for some part of a url, is it? |
| 19:47 | <TabAtkins> | annevk: Ugh, I forgot everything I talked to you about a few months ago regarding computed values of URL. :( |
| 20:07 | <Domenic> | wat is this http://www.w3.org/Submission/2014/SUBM-first-screen-paint-20140811/ |
| 20:08 | <Hixie> | a blog post, essentially |
| 21:00 | <caitp> | pdr / basically anybody who can make sense of svg 1.1, is https://code.google.com/p/chromium/issues/detail?id=109212 a valid issue, or is firefox doing the wrong thing? |
| 21:00 | <caitp> | it's pretty non-obvious from reading the spec |
| 21:08 | <pdr> | caitp, I think it's valid from the spec's perspective. I am not sure if it will ever be implemented properly though |
| 21:09 | <caitp> | I see, thanks |
| 21:12 | <pdr> | caitp, do you have a usecase in mind? Bundling the resource in the svg document itself seems like it would be a better user experience in almost every case, but I may be missing something. |
| 21:14 | <caitp> | it's this weird thing where if you have a <base> tag in the document, FuncIRI stuff seems to get resolved totally wrong |
| 21:14 | <caitp> | so for instance, clip-path="url(#someId)" results in no clip path |
| 21:16 | <caitp> | or rather, they work on the initial url, but if you use history.pushState(), it no longer works |
| 21:19 | <caitp> | actually I'm wrong, it seems to not work even on the initial url, based on the reproduction in this bug |
| 21:21 | <caitp> | awit no, I wasn't wrong |
| 21:24 | <caitp> | i think the issue is actually someone reproducing a bug wrong, it looks like it works well enough for the use cases we want to support |
| 21:31 | <annevk> | TabAtkins_: you keep both the input string and the parsed URL object |
| 21:31 | <annevk> | TabAtkins_: that was the model, iirc |
| 21:31 | <TabAtkins_> | caitp: What pdr said, yes. |
| 21:31 | <annevk> | TabAtkins_: URLs are parsed relative to the base URL of the style sheet |
| 21:31 | <TabAtkins_> | annevk: Okay. Why do you need the input string? |
| 21:31 | <annevk> | TabAtkins_: for the object model |
| 21:32 | <TabAtkins_> | Ah, right, to serialize. |
| 21:32 | <TabAtkins_> | kk |
| 21:43 | <caitp> | oh no, it is still doing something weird, at least in webkit/blink, basically if you pushState() and then cause the SVG to be re-rendered in some fashion, it gets it wrong, weird |
| 21:44 | <pdr> | caitp, can you create a small example? |
| 21:45 | <caitp> | tobias bosch put together this reproduction http://plnkr.co/edit/YgBEuHuTFZXlpe7SLWrc?p=preview which goes into a bit more detail |
| 21:48 | <caitp> | although I don't see what difference setTimeout makes in his example |
| 22:00 | <pdr> | caitp, yeah, I think this is a bug. I'll file. I would recommend trying a different approach though. Might a and a:active work? |
| 22:05 | <caitp> | i'm not sure what you mean by that |
| 22:06 | <pdr> | caitp, it looks like you are trying to use this to change the appearance of svg based on the url without the svg changing. Is that correct? |
| 22:08 | <caitp> | so, recently we added a change to angular.js which requires a <base> element in the document, so that URLs can be resolved consistently between modern and legacy browsers --- someone complained that this broke their SVGs, because their editor adds a clipPath automatically |
| 22:08 | <caitp> | and it was just breaking whenever an address changed |
| 22:08 | <caitp> | I am not sure I understand how a / a:active could be used to resolve it |
| 22:09 | <caitp> | we're just trying to make our customers happy ;-; |
| 22:10 | <pdr> | Ok, I'll continue to file this bug. I would recommend your customer use another approach, but that's their issue. |
| 22:11 | <caitp> | thanks, would you mind CCing me on it? |
| 22:19 | <caitp> | or just paste a link even, that would be great |
| 22:23 | <pdr> | caitp, http://crbug.com/415772 |
| 22:23 | <caitp> | awesome, thanks |
| 23:28 | <zewt> | some day, "save this password?" prompts in browsers won't be totally broken |
| 23:29 | <zewt> | chrome just asked me for about 1.25 seconds and disappeared it before I could click ... again |
| 23:29 | <terinjokes> | that's the first setting I change |
| 23:29 | <terinjokes> | i never want my browser to save passwords |
| 23:31 | <caitp> | bruce schneier wants you to know that password managers will help security, because if you don't use them, you'll just use easy-to-remember+easy-to-guess passwords or write them down in a text file or something instead |
| 23:32 | <terinjokes> | http://www.geek.com/wp-content/uploads/2012/11/91a-DnnN5WL._AA1500_1.jpg |
| 23:44 | <zewt> | things that i'm not going to do: deal with passwords for some dumb little web forum |
| 23:45 | <caitp> | that's a pretty short list, you must be a very open-minded and adventurous person |
