| 02:09 | <pikaren> | Did W3C invent HTML5 |
| 02:10 | <jgraham> | They received it by divine revelation. HTML5 is the one true word of God. |
| 02:21 | <caitp> | one would hope god would do a better job than that |
| 02:23 | <caitp> | an omniscient creator wouldn't care about breaking the web, but would make a whole new web, with pancakes |
| 02:23 | <caitp> | but we don't have an omniscient creator, just millions of fallible ones |
| 02:58 | <MikeSmith> | pikaren: no W3C did not invent HTML5 |
| 02:59 | <MikeSmith> | pikaren: if you want a history of where it came from and when there's some places like https://platform.html5.org/history/ that are useful. All the dates there are links to the dated versions of the spec as it progressed |
| 03:02 | <MikeSmith> | pikaren: and there are other things like https://github.com/whatwg/web-history/blob/master/README.md#a-brief-history-of-the-modern-web-platform and http://www.w3.org/html/wg/wiki/History that go into the wider history around it all |
| 03:05 | <caitp> | there's also the brilliant dramatization of the effort on the history channel |
| 03:05 | <MikeSmith> | haha |
| 03:05 | <MikeSmith> | it's actually a reality show |
| 03:07 | <miketaylr> | which is people reading email |
| 03:07 | <miketaylr> | naked |
| 03:08 | <MikeSmith> | touché |
| 03:08 | <MikeSmith> | yeah, just can't understand why viewers don't find that terrifically appealing |
| 03:09 | <MikeSmith> | maybe we need to change the background music |
| 03:09 | <MikeSmith> | and add some mood lighting |
| 03:10 | <miketaylr> | :)~ |
| 03:29 | <erlehmann> | hey gsnedders remember that 5 years ago you helped me transcribing an interview with moot? |
| 03:29 | <erlehmann> | http://news.dieweltistgarnichtso.net/interviews/moot-4chan.html |
| 03:30 | <erlehmann> | i still have not transcribed the second half of that interview |
| 03:30 | <erlehmann> | btw, what would be the markup for a dialogue? |
| 03:33 | <erlehmann> | for „interview moot“ my site ranks for higher than the AMA |
| 03:33 | <erlehmann> | funny |
| 03:35 | <caitp> | do people actually call him moot when they're talking to him |
| 03:36 | <erlehmann> | i did |
| 03:37 | <erlehmann> | caitp interface-wise, can you tell me what you dislike about http://news.dieweltistgarnichtso.net/ ? :) |
| 03:38 | <caitp> | is this back on the topic earlier about "urls control everything" |
| 03:39 | <erlehmann> | no |
| 03:39 | <erlehmann> | it is about what you expect from an interface |
| 03:39 | <erlehmann> | is this pleasant to read for someone who channels angular? |
| 03:40 | <caitp> | there are some things you could add, like a way to sort the table view, or a way to filter by search terms --- maybe a short synopsis of each item (slightly longer than just the date or the one-word description of what it is) |
| 03:41 | <caitp> | but I dunno, pretty straight forward minimalist thing, it works |
| 03:41 | <erlehmann> | that is exactly what i was looking for, thank you for the feedback |
| 03:41 | <erlehmann> | how did i sort table |
| 03:41 | <erlehmann> | polyfill is not ready yet |
| 03:42 | <erlehmann> | caitp the synopsis item comes up from time to time, i think it might be better to use more descriptive titles |
| 03:43 | <terinjokes> | i call him moot |
| 03:44 | <caitp> | a quick google search for "table sorting" finds http://tablesorter.com/, which claims to be a jquery plugin that you should donate money to |
| 03:45 | <erlehmann> | > jquery |
| 03:45 | <erlehmann> | everything that would not fit on a floppy disc is disqualified |
| 03:46 | <terinjokes> | i believe jQuery minifed and gzipped fits on a floopy |
| 03:48 | <caitp> | about 30kb, based on google's cdn |
| 03:49 | <caitp> | it probably wouldn't be too much trouble to write one from scratch that was a lot tinier, but you'd probably have to do gross things like listen for DOMContentLoaded or other badness |
| 03:50 | <terinjokes> | i just did this for my project |
| 03:50 | <terinjokes> | thus all the complaints in this channel over the last week |
| 03:52 | <erlehmann> | caitp 30kb is still 15 times the size of my biggest page with a table. |
| 03:52 | <erlehmann> | ; curl -IsH 'Accept-Encoding: gzip,deflate' http://news.dieweltistgarnichtso.net/diary/index.html | grep '^Content-Length' |
| 03:52 | <erlehmann> | Content-Length: 2035 |
| 03:53 | <erlehmann> | i am not willing to bear the performance implications of bloating my content that much |
| 03:55 | <caitp> | remember that content-length isn't the only thing going over the wire |
| 03:55 | <erlehmann> | yeah, but what else should there be to bloat? |
| 03:56 | <erlehmann> | hahaha |
| 03:56 | <erlehmann> | chromium finds only one fault |
| 03:56 | <erlehmann> | unused css rules |
| 03:57 | <caitp> | socket acknowledgement, tls negotiation, etc |
| 03:57 | <erlehmann> | no shit sherlock, i don't have every element on my page |
| 03:57 | <erlehmann> | also chromium will probably always say that :target and :hover and :active are unneeded |
| 03:57 | <erlehmann> | no TLS here, sorry |
| 03:57 | <caitp> | just throwing out examples :P |
| 07:08 | <karlcow> | annevk: is there a banana bug in bugzilla? |
| 07:29 | <zcorpan> | erlehmann: what, no tls? heresy! |
| 08:30 | <annevk> | karlcow: you mean re that WHATWG thread? |
| 08:30 | <annevk> | karlcow: yes |
| 08:30 | <karlcow> | yup |
| 09:22 | <Domenic> | Hixie: FWIW in my custom elements I don't do the silly per-element <style> thing. I just use a stylesheet with custom rules |
| 09:22 | <Domenic> | For scoping styles via shadow DOM though you do need to insert them in the shadow root I guess |
| 09:23 | <Domenic> | (Also I am in Europe on vacation \o/) |
| 09:25 | <The-Compiler> | Is there some accepted syntax I could use in a browser to say "load foo but without proxy" in an URL? |
| 09:26 | <Domenic> | Hixie: arrow functions don't have a [[Construct]]. |
| 09:28 | <annevk> | The-Compiler: afaik, no |
| 09:32 | <The-Compiler> | also, why does "//whatwg.org" work as URL in browsers? Is that a valid URL with no scheme? |
| 09:35 | <annevk> | The-Compiler: work where? |
| 09:35 | <annevk> | The-Compiler: my address bar takes me to file:////whatwg.org ... |
| 09:36 | <The-Compiler> | annevk: I tried Chrome and Firefox |
| 09:36 | <annevk> | The-Compiler: if you're just trying the address bar, it's up to the browser really, as that's just a UI field |
| 09:37 | <annevk> | The-Compiler: if you're talking about <a>, we have a say in that, and it might work there depending on the base URL, as it's a scheme-relative URL |
| 09:37 | <The-Compiler> | annevk: ah - I was talking about the address bar, sorry |
| 09:38 | <The-Compiler> | I'm writing my own browser (based on QtWebKit), and would find such a "use no proxy" syntax useful, so I wondered if //URL would be a sane syntax for that :) |
| 09:39 | <annevk> | I see, it's not |
| 09:39 | <The-Compiler> | (got the idea when I watched a co-worker using that for a machine in the LAN, probably used to it by Windows/Samba) |
| 09:52 | <Domenic> | annevk: after sleeping on it I think I'll just send an email to the list saying "here is the version with overloading; seem OK?" |
| 09:54 | <annevk> | Domenic: yeah, but it sounds like you should be seeing a museum rather than your outbox |
| 09:55 | <Domenic> | annevk: heh. i'm at a science-programming thing for the first few days and this is downtime. |
| 09:55 | <annevk> | https://tools.ietf.org/html/draft-reschke-objsec Julian Reschke has been coopted by the GSMA? |
| 09:57 | <annevk> | The same crowd that ads unique tracking headers, advertisements, and other crap over insecure connections, would really like to continue doing that... As if they are unclear on the fact that this is part of why we move to HTTPS. |
| 10:01 | <erlehmann> | zcorpan what are you referring to? |
| 10:02 | <zcorpan> | erlehmann: http://krijnhoetmer.nl/irc-logs/whatwg/20141105#l-163 |
| 10:06 | <erlehmann> | ok |
| 10:10 | <Domenic> | annevk: no twitter account for fetch? |
| 10:12 | <annevk> | Domenic: I was thinking of morphing xhrstandard at some point |
| 10:12 | <Domenic> | hmm |
| 11:50 | <annevk> | Domenic: maybe I'll add one... I added @notifyapi for now |
| 11:50 | <Domenic> | annevk: any rhyme or reason on -standard vs. -api? |
| 11:50 | <Domenic> | annevk: anyway if you add one before 15:00 UTC I'll update my scheduled tweet to say @fetchstandard instead of "Fetch Standard" :P |
| 11:51 | <annevk> | Domenic: it's mostly being constrained by existing Twitter usernames |
| 11:52 | <annevk> | Domenic: scheduled tweets? Are you into social marketing or something? :p |
| 11:52 | <Domenic> | I don't want to abandon my US audience :P |
| 12:00 | <annevk> | Domenic: added account |
| 12:05 | <annevk> | JavaScript, Books, Figures, and Quirks Mode lack accounts |
| 12:06 | <annevk> | And I guess some of the repositories do |
| 12:10 | <Domenic> | I like this http://static.manuel-strehl.de/StillImage/htmlnext_logo2.svg |
| 12:17 | <Domenic> | annevk: why can TextEncoder encode as UTF-16? Don't we only like UTF-8? |
| 12:44 | <annevk> | Domenic: I think there was some argument that utf-16 was needed |
| 12:44 | <annevk> | Domenic: and that at the time we were less concerned about utf-16 because we didn't know much |
| 12:48 | <espadrine_> | rubys: thanks a lot for this! http://intertwingly.net/projects/pegurl/liveview.html |
| 12:48 | <rubys> | yw |
| 12:49 | <espadrine_> | Does red indicate something that the browsers do wrong, assuming they intend to implement this? |
| 12:49 | <rubys> | If you assume that, then yes. :-) |
| 12:49 | <espadrine_> | I'm surprised by the canonicalization of href, could it break things? |
| 12:50 | <rubys> | some browsers already do canonicalize; given the diversity of implementations breakage may be limited |
| 12:51 | <The-Compiler> | rubys: oooooh! That's an useful tool! :) |
| 12:51 | <rubys> | an overview of test results: http://intertwingly.net/projects/pegurl/urltest-results/ |
| 12:51 | <espadrine_> | ah, you're right |
| 13:13 | <gsnedders> | erlehmann: no :) |
| 13:25 | <lolmaus> | TabAtkins: hi! You around? Wanna ask you resolve my spec confusion. |
| 13:30 | <zcorpan> | mathiasbynens: https://mathiasbynens.be/notes/css-escapes idents can start with two dashes now |
| 13:58 | <mathiasbynens> | zcorpan: thanks. i see that 0xA0 is no longer allowed too – nice |
| 14:03 | <JakeA> | annevk: How have you speced Headers to be an iterator? Just the last sentence of https://fetch.spec.whatwg.org/#headers? |
| 14:03 | <JakeA> | I guess there's no IDL for it? |
| 14:03 | <annevk> | JakeA: iterable<ByteString, ByteString> |
| 14:03 | <JakeA> | omg how did I miss that |
| 14:03 | <annevk> | JakeA: plus that text, yes |
| 14:04 | <JakeA> | ta |
| 14:15 | <zcorpan> | annevk: what does IE do for \#\u03B2 ? it seems blink doesn't escape |
| 14:16 | <zcorpan> | rubys: ^ |
| 14:17 | <rubys> | zcorpan: check for yourself: http://intertwingly.net/projects/pegurl/urltest-results/ ;-) |
| 14:17 | <rubys> | direct link: http://intertwingly.net/projects/pegurl/urltest-results/f0b7d5c4b4 |
| 14:17 | <annevk> | zcorpan: I broke my IE setup |
| 14:17 | <annevk> | zcorpan: Safari escapes |
| 14:17 | <annevk> | zcorpan: I tried to set up remote.modern.ie, but it doesn't seem to work on Mac |
| 14:18 | <MikeSmith> | it does |
| 14:18 | <rubys> | annevk: I highly recommend https://www.modern.ie/en-us/virtualization-tools#downloads |
| 14:18 | <rubys> | I run modern.ie on a mac |
| 14:19 | <MikeSmith> | rubys: I think he's asking about the remote thing that was just released within the last couple days |
| 14:19 | <MikeSmith> | which doesn't require you to run virtualbox or install 10GB VM images and such |
| 14:19 | <annevk> | Oh, I got access now, but there's too many users |
| 14:19 | <MikeSmith> | ah yeah |
| 14:19 | <rubys> | oh. I wasn't aware that there was a remote thing. I much prefer installing things locally myself. |
| 14:19 | <MikeSmith> | they seem to be having a problem handling the load |
| 14:20 | <MikeSmith> | rubys: this remote thing is pretty cool |
| 14:20 | <annevk> | Yeah I would install modern.ie but I need to upgrade VMWare and I've no idea who to contact about a new license at Mozilla |
| 14:20 | <MikeSmith> | rubys: I think you'll be impressed |
| 14:20 | <annevk> | I should probably stop being lazy and figure it out |
| 14:21 | <rubys> | I run VirtualBox. It is free. |
| 14:21 | <annevk> | I used to have that, but something was off |
| 14:21 | <annevk> | Guess I could try it again |
| 14:22 | <MikeSmith> | I have remote.modern.ie running in my Android phone as well. It's pretty cool to be able to, e.g., check the result for a test case in IE from and Android phone |
| 14:23 | <zcorpan> | rubys: thx. seems like we could go either way on escaping the fragment or not |
| 14:23 | <rubys> | hm. modern.ie doesn't have a remote app for ubuntu. I am NOT impressed. |
| 14:24 | <zcorpan> | .hash has clear majority among browsers to not escape |
| 14:24 | <MikeSmith> | rubys: yeah nothing for linux so far. Seems odd that they have it working on android but not on linux |
| 14:25 | <rubys> | zcorpan: s/not escape/escape less/ perhaps? |
| 14:25 | <rubys> | I'd think that there are still some characters that would need to be escaped. %20, for example. |
| 14:25 | <MikeSmith> | but at the same time I can admire them for not blocking the release to the rest of us just it get it working for y'all that choose to torture yourselves by running desktop linux |
| 14:26 | <zcorpan> | rubys: yeah i mean this particular character. don't have the entire picture |
| 14:26 | rubys | enjoys seeing those that run Mac have troubles running open source software after nearly every upgrade |
| 14:26 | <MikeSmith> | heh |
| 14:26 | rubys | only has pity for those that run Windows. |
| 14:27 | <MikeSmith> | fair enough :-) |
| 14:29 | <annevk> | zcorpan: yeah, the URL fragment stuff is annoying inconsistent |
| 14:29 | <rubys> | should the consensus turn out to be different than what the spec currently says, it would not be a big change. |
| 14:30 | <rubys> | meanwhile, the expected test results and the spec disagree. :-/ |
| 14:31 | <mathiasbynens> | zcorpan: updated https://mathiasbynens.be/notes/css-escapes and https://mothereff.in/css-escapes to refer to the latest draft spec where applicable |
| 14:32 | <mathiasbynens> | thanks for the heads up |
| 14:32 | <zcorpan> | mathiasbynens++ |
| 14:32 | <annevk> | rubys: I think for now we should patch the test to match the spec |
| 14:32 | <rubys> | wfm |
| 14:32 | <rubys> | meanwhile, I have a patches file in my branch |
| 14:33 | <annevk> | I'm happy to review a PR to web-platform-tests |
| 14:33 | <rubys> | https://github.com/rubys/url/blob/peg.js/reference-implementation/test/patchtestdata.txt |
| 14:33 | <rubys> | I'll look at making a PR in a bit. |
| 14:42 | <zcorpan> | annevk: rubys: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3289 |
| 14:44 | <rubys> | zcorpan: very interesting. |
| 14:45 | <zcorpan> | hmm. checking safari suggests there's a bug in the script |
| 14:45 | <annevk> | zcorpan: file a bug maybe |
| 14:46 | <rubys> | I'm surprised that Chrome doesn't escape U+0025 (percent sign) |
| 14:46 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3290 |
| 14:46 | <zcorpan> | annevk: ok |
| 14:47 | <rubys> | firefox either. weird. |
| 14:49 | <zcorpan> | annevk: your file bug link broke the bug filer script |
| 14:49 | <zcorpan> | i guess i can fix the script |
| 14:50 | <annevk> | zcorpan: if you look at script.src it should be resolved |
| 14:51 | <zcorpan> | annevk: var link = document.querySelector('a[href^="https://www.w3.org/Bugs/Public/enter_bug.cgi?"]';); |
| 14:52 | <annevk> | oh that link |
| 14:52 | <annevk> | CSS not being able to look at resolved URLs anno 2014 is a bit of joke too |
| 14:52 | <Ms2ger> | I fixed that somewhere |
| 14:52 | <annevk> | yeah for one out of 10 specs |
| 14:53 | <annevk> | zcorpan: I can fix the link in URL and such |
| 14:53 | <Ms2ger> | Damned boilerplate :) |
| 14:53 | <zcorpan> | annevk: i'm changing file-bug.js to be more flexible |
| 14:53 | <annevk> | ok |
| 15:04 | <zcorpan> | annevk: rubys: https://www.w3.org/Bugs/Public/show_bug.cgi?id=27252 |
| 15:05 | <zcorpan> | maybe testing 0..0xffff for each part of a url should be in web-platform-tests |
| 15:06 | rubys | notes that the tests in the range of d800 through dfff are suspect as these represent half of surrogate pairs |
| 15:22 | <karanlyons> | Does XHR support accessing the response attribute before the request is complete like you can with responseText, and if not are there plans to include this support? |
| 15:23 | <karanlyons> | I know there are moz prefixes for responseType = chunked-[blob|text|arraybuffer], which gives access to response inside of progress events. |
| 15:23 | <caitp> | you should be able to get headers as soon as headers are available, I think |
| 15:24 | <caitp> | oh |
| 15:24 | <caitp> | response attribute |
| 15:24 | <caitp> | derp |
| 15:24 | <karanlyons> | Yep. |
| 15:24 | <karanlyons> | Headers I'm not sure about, but I'll just fire off a separate HEAD request if needed. |
| 15:25 | <Domenic> | karanlyons: the current plan is to make that available via the new fetch API by exposing its body stream |
| 15:27 | <karanlyons> | Is that API implemented in any fashion currently? |
| 15:28 | <Domenic> | nope, in progress though |
| 15:29 | <karanlyons> | So currently my only option would be many requests with Ranges headers, I think. |
| 15:30 | <karanlyons> | Or pretending it's text. |
| 15:31 | <Domenic> | yeah, that's what people generally do, which is why we're fixing it :) |
| 15:32 | <karanlyons> | Is the fetch API meant to replace XHR, and if not will XHR get the same support? (Apologies for all the questions.) |
| 15:34 | <Ms2ger> | XHR will stay around |
| 15:34 | <Domenic> | But Fetch will be better in most every way; in developers minds it should be a replacement |
| 15:35 | <karanlyons> | But just in case there are discovered shortcomings in the Fetch API, it may be wise to add that same support to XHR. |
| 15:35 | <Ms2ger> | Domenic, we'll see if that ends up being true :) |
| 15:35 | <Domenic> | karanlyons: well, good luck convincing browser vendors to do twice the work :) |
| 15:36 | <karanlyons> | Domenic: Yeah :( |
| 15:40 | <caitp> | isn't it hard when you just have tons of APIs that are just complete crap, and you can't fix them? |
| 15:40 | <karanlyons> | I'm trying to fix buffering in HTML5MediaElement, so tell me about it. |
| 15:58 | <annevk_> | karanlyons: why would it be wise to add them to XMLHttpRequest? |
| 15:58 | <annevk> | karanlyons: please don't hold back on questions btw, we're all here to learn (and have fun) |
| 15:58 | <karanlyons> | annevk: Just in case there's some shortcoming of the Fetch API that XHR would be able to address. |
| 15:59 | <karanlyons> | Like right now I'm swimming in issues due to non-overlapping shortcomings of HTMLMediaElement and MSE. |
| 15:59 | <annevk> | We'll try to avoid that, but if you find anything do say so |
| 16:00 | <annevk> | XMLHttpRequest's responseURL was backported, in a way; I wasn't really planning on doing anything beyond that, but it's not a principled thing |
| 16:01 | <karanlyons> | Ideally backporting won't be needed, but it's one of those things where you can never be too sure. |
| 16:05 | <annevk> | I'm not envying you, pioneering this stuff is prolly painful as hell |
| 16:07 | <karanlyons> | Yeah, right now I'm trying to turn a streamed file from XHR as 'text\/plain; charset=x-user-defined' back into a proper arraybuffer, and I'm fighting back tears. |
| 16:15 | <annevk> | karanlyons: with XMLHttpRequest you should be able to use arraybuffer at least |
| 16:15 | <karanlyons> | Can't stream an arraybuffer. |
| 16:15 | <karanlyons> | Well, can't with XHR. |
| 16:15 | <karanlyons> | The only type I can access mid download is text. |
| 16:16 | <karanlyons> | (I could use Ranges headers, but I'd like to do it without opening many small requests to the server) |
| 16:16 | <annevk> | karanlyons: oh man, that people rely on that... |
| 16:16 | <karanlyons> | There's no other option, sadly. |
| 16:16 | <annevk> | karanlyons: many small requests with HTTP/2 seems better, not sure if you can rely on that |
| 16:16 | <karanlyons> | annevk: Nor can I even rely on the server accepting range requests. |
| 16:24 | <karanlyons> | Holy crap, I think this works. |
| 16:40 | <karanlyons> | annevk: How often does XHR fire a progress event? Is it up to the browser? |
| 16:41 | <gsnedders> | karanlyons: "about every 50ms or for every byte received, whichever is least frequent" |
| 16:41 | <karanlyons> | Awesome. |
| 16:42 | <gsnedders> | karanlyons: historically it's been all over the place |
| 16:42 | <karanlyons> | Oh. Not so awesome. |
| 16:42 | <karanlyons> | So I may be better off with an interval. |
| 16:43 | <karanlyons> | Not so nice for the user, though. |
| 16:49 | <MikeSmith> | TabAtkins: dunno if I'm supposed to ask for permission to add somebody to https://github.com/orgs/w3c/teams/csswg-reviewers but I wanted to give you a heads-up at least that I just added a guy from Igalia there because he's in the midsts of submitting some issues and wants to label them but can't without having perms |
| 17:02 | <rego> | MikeSmith: TabAtkins: just in case, I'm the one working on that, Tab already knows us from the www-style mailing list :-) |
| 17:13 | <annevk> | So I'm running IE11 now in Virtualbox and the TLS UI in the address bar... |
| 17:13 | <annevk> | How can this be so bad? |
| 17:23 | <Domenic> | annevk: your complains being? seems ... okay-ish... the favicon seems likely to be exploitable |
| 17:24 | <Domenic> | except ... they put it on top of a document icon? |
| 17:24 | <annevk> | Domenic: the favicon is exploitable, the lock icon is far away from the domain |
| 17:24 | <annevk> | Domenic: insecure pages look as secure as secure pages if you don't look carefully |
| 17:24 | <Domenic> | i think a lock favicon would show up on top of a "document" icon so slightly less exploitable |
| 17:24 | <MikeSmith> | rego: ah yeah I should have figured TabAtkins and you already were in contact. Anyway, no worries |
| 17:25 | <annevk> | Domenic: what is a document icon? |
| 17:25 | <rego> | yeah no problem, I didn't know the procedure, so I was just asking here and there :) |
| 17:25 | <karanlyons> | I don't see this document icon |
| 17:25 | <Domenic> | annevk: looks like a piece of paper |
| 17:25 | <annevk> | Domenic: both w3.org and annevankesteren.nl look identical, despite the former not using TLS |
| 17:25 | <karanlyons> | Just looking at a screenshot, but that lock is gray and easy to miss. |
| 17:26 | <karanlyons> | Safari's lock is gray too, but it's at least right next to the domain. |
| 17:27 | <Domenic> | Firefox's lock is gray |
| 17:27 | <Domenic> | but they distinguish insecure sites by giving them a "globe" |
| 17:27 | <karanlyons> | The big difference is IE11's lock looks to be sandwiched between reload icons, etc., so it can be easily overlooked. |
| 17:29 | <MikeSmith> | annevk: on that note I'm really surprised the Firefox Mobile devs have not fixing that seriously problem that allows but the URL and lock icon to be spoofed in the address bar. the last time I looked at that bug, they were still saying that didn't think it was a real problem. https://bugzilla.mozilla.org/show_bug.cgi?id=605206 |
| 17:29 | <MikeSmith> | https://gist.github.com/sideshowbarker/8284404#file-phish-png |
| 17:30 | <MikeSmith> | source of which is http://people.w3.org/mike/phish/ |
| 17:30 | <annevk> | MikeSmith: the response you got there is still baffling today |
| 17:30 | <karanlyons> | Oof. |
| 17:31 | <MikeSmith> | even more baffling that although dveditz is Cc'ed there, he hasn't chimed in on it |
| 18:12 | <annevk> | So I remember why I had VMWare now |
| 18:12 | <annevk> | The integration between host and guest is lacking in Virtualbox, at least by default |
| 18:12 | <annevk> | Even after enabling bidirectional copy and paste it isn't really working |
| 18:13 | <karanlyons> | annevk: What's your host OS? |
| 18:13 | <annevk> | Mac |
| 18:13 | <karanlyons> | Parallels on OS X is pretty good. |
| 18:13 | <karanlyons> | I still end up dual booting for most things, but for quick checks it works. |
| 18:14 | <annevk> | VMWare was too until it broke apart |
| 18:14 | <annevk> | I wouldn't really consider dual boot |
| 18:14 | <annevk> | I just need to check things in IE |
| 18:14 | <annevk> | And perhaps other browsers, on Windows |
| 18:14 | <karanlyons> | What's kinda neat about Parallels is that it uses my Bootcamp partition as the image. |
| 18:55 | <Hixie> | annevk: if someone wanted to invent a URL scheme that was hierarchical but didn't use IP addresses or domain names, what would you do? |
| 18:57 | <karanlyons> | > as a delimiter? |
| 18:57 | <annevk> | Hixie: I would try to figure out what comes out of fixing https://www.w3.org/Bugs/Public/show_bug.cgi?id=27233 I suppose |
| 18:58 | <Hixie> | annevk: like, suppose i had a url scheme to select tracks in my train layout. track://countryname/stationname/stationsection/tracknumber |
| 18:58 | <Hixie> | annevk: i'm in the context of track 2, i want to say "go see track 3, whatever station this is", so i want to say ../3 |
| 18:58 | <Hixie> | or i want to say "go to the freight section of this station, track 1", so i say "../freight/1" |
| 18:58 | <annevk> | Hixie: if you didn't want countryname to be a host, I suspect track:///countryname/... |
| 18:59 | <Hixie> | right now this is non-conforming because a country name is not a host |
| 18:59 | <Hixie> | ah, interesting |
| 18:59 | <annevk> | Hixie: because I think we could make the parsing rules for these new URLs strict enough that three slashes meant the host was omitted |
| 18:59 | <annevk> | Hixie: we can't do that for http / ws, there it just means you made a typo |
| 19:00 | <Hixie> | interesting |
| 19:00 | <Hixie> | how about if you want to have a host-like concept, separate from the path, but where it's not a domain/ip? |
| 19:01 | <Hixie> | e.g. suppose you want train://engineid/... |
| 19:01 | <Hixie> | e.g. because you want to use the urls to create a security boundary like we do with web origins |
| 19:01 | <Hixie> | where the scheme and host is used, but not the path |
| 19:02 | <TabAtkins> | The meaning of a "domain" is up to the scheme to decide, no? |
| 19:02 | <Hixie> | not per the url spec |
| 19:03 | <annevk> | Hixie: I suspect you'd have to stay within the parsing space of domain names, but perhaps treat them as something else |
| 19:03 | <annevk> | Hixie: I'm not sure I'd want to throw away normal host parsing rules even for unknown schemes, seems like too big of a footgun, but who knows |
| 19:04 | <Hixie> | should i file a bug on saying that schemes can use a non-host/ip value for the host component? |
| 19:05 | <JakeA> | TabAtkins: so, I was playing around with svg+viewBox+foreignObject as a way of making html that scales the way svg/img does. Chrome's a bit broken there. Is it worth fixing or is there a saner way to do what I'm trying to do on the horizon? |
| 19:05 | <JakeA> | Like viewBox on an element (which maybe could be a path towards responsive elements) |
| 19:05 | <TabAtkins> | Dunno what's wrong, so I dunno what needs fixing. |
| 19:06 | <Hixie> | annevk: (this isn't academic, fwiw. chrome-extension:// and android.resource:// both do this today.) |
| 19:07 | <JakeA> | TabAtkins: We balls up the scaling of gpu-layered content. But still, it seems a pretty round-about way of achieving what I'm trying to achieve (html with a viewbox), so thought there might be something in the works that doesn't involve SVG |
| 19:08 | <TabAtkins> | viewBox is just a way of applying an internal transform to the content. You can do it yourself. |
| 19:10 | <TabAtkins> | to reproduce <svg width=100 height=100 viewBox="-10 -10 220 220">, just do... |
| 19:10 | <TabAtkins> | (gimme a sec) |
| 19:10 | <TabAtkins> | <div width=100 height=100><div width=220 height=200 transform="scale(.5) translate(-10px, -10px)"></div></div> |
| 19:10 | <JakeA> | You'd have to add a resize listener + mutation observers, and I still don't know if that captures all the times you'd need to recalculate the transform |
| 19:10 | <TabAtkins> | (I think) |
| 19:12 | <Hixie> | annevk: what's the state with https://www.w3.org/Bugs/Public/show_bug.cgi?id=23250 ? |
| 19:12 | <JakeA> | TabAtkins: it's easy if the container is fixed, but not if it's dynamic, eg https://jsbin.com/nafawe/quiet |
| 19:13 | <TabAtkins> | Oh, sure. |
| 19:14 | <TabAtkins> | Yeah, don't think there's any way to do a dynamic one in HTML without manual handling of resizes. |
| 19:14 | <JakeA> | That'd be a lot easier with some kind of layout observer |
| 19:14 | <Hixie> | why do mutation observers have a .observe() method instead of just being configured in the constructor? |
| 19:17 | <annevk> | Hixie: you could comment on the aforementioned bug |
| 19:17 | <Hixie> | annevk: i was just suggesting a semantic change, not a parsing change (assuming we're talking about urls still) |
| 19:17 | <annevk> | Hixie: no input from implementers for that bug |
| 19:18 | <Hixie> | what input are you looking for? |
| 19:18 | <annevk> | Hixie: that they're interested in doing it |
| 19:18 | <Hixie> | ah |
| 19:18 | <annevk> | Hixie: new bug for semantic change I guess |
| 19:19 | <Hixie> | comment 5 is smaug____ saying it would be cheap in gecko |
| 19:19 | <Hixie> | fwiw |
| 19:19 | <Hixie> | and comment 1 is him saying it would be useful |
| 19:20 | <annevk> | But IE and Chrome dislike it |
| 19:21 | <Hixie> | IE and Chrome folk gave reasons they disliked it, but those reasons appear to be false |
| 19:21 | <Hixie> | (i don't mean they're lying, i mean they're wrong about what is possible) |
| 19:23 | <smaug____> | if web components actually ends up being a thing, and it uses MutationObserver, I assume there will need to be a way to observe ancestor chain changes |
| 19:23 | <smaug____> | since that is what implementations need to do now in C++ in certain cases |
| 19:24 | <smaug____> | s/it uses/components using web components use/ |
| 19:24 | <smaug____> | Custom element does have http://w3c.github.io/webcomponents/spec/custom/#types-of-callbacks |
| 19:25 | <zcorpan> | smaug____: examples of things that observe ancestor chain changes? |
| 19:25 | <Hixie> | yeah, seems weird that custom elements don't use mutation observers |
| 19:25 | <annevk> | custom elements wants something that's near synchronous |
| 19:25 | <smaug____> | Hixie: well, mutation observer callbacks run at the end of microtask. I think custom element callbacks run earlier |
| 19:26 | <annevk> | but it's less problematic than mutation events I think |
| 19:26 | <smaug____> | zcorpan: well, being in document or not is the most common |
| 19:26 | <smaug____> | zcorpan: iframe as an example |
| 19:26 | <Hixie> | smaug____: mutation observers run whenever you want if you call takeRecords() |
| 19:26 | <smaug____> | sure |
| 19:26 | <zcorpan> | smaug____: ah right |
| 19:27 | <zcorpan> | smaug____: also changing documents |
| 19:27 | <smaug____> | Hixie: but if you don't control the code which is removing element from document |
| 19:27 | <smaug____> | zcorpan: changing document is unbind from a document + bind to a document |
| 19:28 | <Hixie> | smaug____: yeah, true |
| 19:28 | <Hixie> | smaug____: then again, custom elements don't have a way to react to child list changes |
| 19:29 | <zcorpan> | smaug____: in spec terms i think it's "adoption steps are run" i think |
| 19:30 | <smaug____> | zcorpan: in Gecko ShadowRoot implementation itself uses Gecko's internal MutationObserver to observe changes under the host so that it can distribute new elements to right place |
| 19:30 | <smaug____> | (Gecko's internal MutationObserver is sync) |
| 19:31 | <smaug____> | that just as an example where it is not only about document, but about other changes too |
| 19:35 | <rubys> | TabAtkins: is it possible to have boxes within railroad diagrams to hyperlink to other sections of the document? |
| 19:35 | <rubys> | Example: http://intertwingly.net/projects/pegurl/url.html#user-info |
| 19:36 | <rubys> | user and password inside the railroad diagram don't hyperlink, user and password in the prose below do. |
| 19:36 | <TabAtkins> | rubys: Been a request for a while. File an issue on me, I'll figure it out. |
| 19:36 | <rubys> | sweet. |
| 19:46 | <rubys> | https://github.com/tabatkins/bikeshed/issues/270 |
| 19:46 | <TabAtkins> | yup, thanks. |
| 19:46 | <rubys> | https://github.com/rubys/url/issues/2 |
| 19:59 | <karanlyons> | TabAtkins: Bikeshed is a great name. |
| 20:01 | <TabAtkins> | I think so too. ^_^ |
| 20:05 | <karanlyons> | For development purposes is there an easy way to grab a local file and stash it into an arraybuffer? |
| 20:06 | <karanlyons> | Err, wrong channel, sorry. |
| 22:32 | <karanlyons> | It'd be kinda nice if DataViews had a slice similar to ArrayBuffers. |
