| 03:55 | <jwalden> | there's no way to -- within a script -- get a reference to its corresponding script element, is there? I can't think of one offhand, but I don't know that there isn't one |
| 04:19 | <Hixie> | jwalden: if it's a blocking script run from the parser in a document that isn't being silly with its parsing environment, you can just get the last <script> element of the document |
| 04:20 | <Hixie> | jwalden: failing that, you can set a capture listener for afterscriptexecute during the script and then do whatever it is you want to do when it fires |
| 04:20 | <jwalden> | hmm, I didn't know about the latter |
| 04:20 | <jwalden> | the former I had thought of as a worst-case hack, definitely :-) |
| 04:20 | <Hixie> | might not work reliably everywhere, it's relatively new |
| 04:21 | <Hixie> | if it's an external script, you can do it using the 'load' event instead of 'afterscriptexecute' |
| 04:22 | <Hixie> | if it's an internal script then you can do that too except the event will be queued and i'm not 100% sure you can't have another load fire first |
| 04:22 | <Hixie> | depending on what's going on |
| 05:26 | <zcorpan> | MikeSmith: is it possible to CC a mailing list automatically for new bugs filed that have a particular string in the summary, or when a bug changes summary and now has that string? |
| 05:27 | <zcorpan> | MikeSmith: concretely, public-texttracks and "<track>" |
| 05:27 | <MikeSmith> | not possible with bugzilla as far as I know |
| 05:27 | MikeSmith | checks |
| 05:30 | <MikeSmith> | zcorpan: yeah, no way to do if from bugzilla config |
| 05:31 | zcorpan | ponders about possible ways to do it anyway |
| 05:34 | <zcorpan> | there's a feed: https://www.w3.org/Bugs/Public/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&component=HTML5%20spec%20%28editor%3A%20Ian%20Hickson%29&component=other%20Hixie%20drafts%20%28editor%3A%20Ian%20Hickson%29&product=HTML%20WG&query_format=advanced&short_desc=%3Ctrack%3E&short_desc_type=allwordssubstr&title=Bug%20List:%20track&ctype=atom |
| 05:34 | <zcorpan> | can something hook into that and send email about new items? |
| 05:36 | <MikeSmith> | wow cool that link crashed my browser |
| 05:37 | <MikeSmith> | sweet it just did it again |
| 05:38 | <zcorpan> | nice |
| 05:42 | <MikeSmith> | zcorpan: anyway, yeah, I guess something could hook into that |
| 08:10 | zcorpan | wonders what happens with <img srcset="foo.jpg 2x" style="image-resolution:2dppx"> |
| 08:18 | <odinho> | zcorpan: Hmm. I would guess nothing. Because image-resolution would probably be at that already anyway? At least how I understand it. |
| 08:19 | <odinho> | <img srcset="foo.jpg 2x" style="image-resolution:1dppx"> is IMHO more of a braintwister, -- under the same "I haven't looked up how image-resolution is supposed to work" caveat. |
| 08:21 | <odinho> | MikeSmith: How much work is it approving PHP files for w3c-test.org? -- I keep moving stuff, and it's kinda stupid, I have a "staging" directory now and moved some PHP there. If it's approved, I will move it once more to /approved/ |
| 08:24 | zcorpan | doesn't like the fact that tests move |
| 08:25 | <odinho> | zcorpan: Yeah, I don't really like it either. -- But I don't like them being in submitted/<vendor>/whatever either. And being of different quality and make. |
| 08:27 | <Ms2ger> | And there being too few, and nobody reviewing them ;) |
| 08:28 | <annevk> | Ms2ger: so should I remove that third point then? |
| 08:29 | <Ms2ger> | Yeah |
| 08:29 | <Ms2ger> | And maybe add a note along the lines of "This implies that parent is a Document, DocumentFragment, or Element node." |
| 08:30 | <annevk> | so why did we end up with that step? |
| 08:30 | <annevk> | because of Attr? |
| 08:30 | <zcorpan> | should we change the procedure with submitted tests to be closer to various people's liking? |
| 08:30 | <Ms2ger> | Hmm, that seems possible |
| 08:32 | <annevk> | but even if Attr remains a node it will never have children |
| 08:32 | <annevk> | so we can remove that |
| 08:33 | <zcorpan> | it'd be nice to have a review system (something like what we use at opera, called "critic") for tests, and you can look up the review for a test and see the history |
| 08:33 | <zcorpan> | and not have "accepted" at all, since specs change over time anyway |
| 08:33 | <jgraham> | zcorpan: Spent ages investigating review systems, turns out there aren't any good ones for hg |
| 08:34 | <zcorpan> | can we move to git? :-) |
| 08:34 | <Ms2ger> | No |
| 08:34 | <odinho> | Hmm. Another problem is that some of the tests are kinda OCD. |
| 08:34 | <odinho> | :( |
| 08:34 | <annevk> | if git has all the tooling... |
| 08:34 | <annevk> | git also has the community |
| 08:34 | <zcorpan> | Ms2ger: why? |
| 08:34 | <jgraham> | Actually moving to git would help a lot with the review problem even without open Opera-Critic |
| 08:34 | <odinho> | Some tests are more equal than others, to use Animal Farm terms :] |
| 08:35 | <jgraham> | Since there are a non-zero number os possible review tools |
| 08:35 | <jgraham> | notably gerrit, but - at a push - just github |
| 08:36 | <jgraham> | (of course pliniss is developing some custom thing based around the CSS WG workflow) |
| 08:37 | <annevk> | zcorpan: image-resolution seems like something we should not implement... |
| 08:37 | <zcorpan> | annevk: why not? |
| 08:39 | <AryehGregor> | jgraham, shepherd supports test review, and has for ages. |
| 08:39 | <AryehGregor> | I doubt it's comparable to gerrit, though. |
| 08:40 | <annevk> | zcorpan: does not seem like a good idea to complicate image math even further |
| 08:40 | <jgraham> | AryehGregor: Yes, I think I said that |
| 08:41 | AryehGregor | was replying to "(of course pliniss is developing some custom thing based around the CSS WG workflow)", which was phrased in the present participle and therefore implied it was still under development |
| 08:41 | <jgraham> | Or at least that it was being developed. Last I heard it didn't really support > 1 test per file |
| 08:41 | <AryehGregor> | No, it doesn't. |
| 08:41 | <AryehGregor> | AFAIK. |
| 08:41 | <jgraham> | Well then it is either "still being developed" or "useless" |
| 08:41 | <jgraham> | I was giving it the benefit of the doubt |
| 08:41 | <AryehGregor> | It's really meant originally for manual CSS tests, which is logically going to be one test per file. |
| 08:42 | <jgraham> | Yes |
| 08:42 | <AryehGregor> | It has some support for reftests too, although inadequate. |
| 08:42 | <AryehGregor> | I don't think it has reasonable support for JS tests. |
| 08:42 | <jgraham> | I just think tests should be reviewed like code |
| 08:42 | <AryehGregor> | I completely agree for JS tests. |
| 08:42 | <jgraham> | A sequence of commits |
| 08:42 | <AryehGregor> | Since, well, they are code. :) |
| 08:42 | <jgraham> | Yeah, but not everyone thinks that |
| 08:43 | <AryehGregor> | Reftests and such make sense to review as files, since they tend to be very short and simple and logically delineated as-is. |
| 08:43 | <AryehGregor> | I don't think reviewing them as commits makes as much sense. |
| 08:44 | <jgraham> | I think having one system for everything makes a lot of sense |
| 08:44 | <jgraham> | and using a commit-based review system at Opera for tests is working fine |
| 08:45 | <jgraham> | (although I have personally never used it for reviewing reftests, I have reviewed manual tests) |
| 08:51 | <zcorpan> | AryehGregor: i don't see much problem with reviewing reftests as commits. of course you'd want to run the test as well, but that applies to other tests equally i think |
| 08:56 | <jgraham> | So if I have data:text/html,<body onload="location='data:text/html,1'" onunload="'location=data:text/html,2'">3 what should I end up with per spec? |
| 08:56 | <Ms2ger> | COME_ON_ERR |
| 08:56 | <jgraham> | Browsers seem to give 1) which makes sense if the unload navigation is cancelled, but afaict that would happen in step 4) which only applies to different-origin navigations |
| 08:58 | <zcorpan> | jgraham: material for #htmlpubquiz :-P |
| 09:03 | <jgraham> | zcorpan: Not useful :p |
| 09:05 | <zcorpan> | jgraham: you could use the majority answer as the correct answer on the basis that it's what people expect to be right :-) |
| 09:06 | <zcorpan> | jgraham: is a data: navigation considered to be a same-origin navigation? |
| 09:13 | <odinho> | zcorpan: lol, using majority - that's a sneaky method. |
| 09:13 | Ms2ger | sighs |
| 09:14 | <odinho> | BTW, I'm no big fan of the long names. Anyone care to think about something better than navigator.persistentStorageSettings.queryUsageAndQuota or navigator.storageInfo.queryPersistentUsageAndQuota? http://lists.w3.org/Archives/Public/public-webapps/2012AprJun/1044.html |
| 09:14 | <Ms2ger> | And here I thought that Node.replaceChild would be interoperable |
| 09:14 | <Ms2ger> | Should have known better |
| 09:15 | <zcorpan> | odinho: where's the spec for those? |
| 09:16 | <odinho> | zcorpan: Seems to be http://dvcs.w3.org/hg/quota/raw-file/tip/Overview.html |
| 09:16 | <zcorpan> | odinho: although maybe it's not so reliable since people may be answering the answer they think is *least* reasonable hoping that it's the correct answer (since it probably has been in the past) |
| 09:17 | <odinho> | zcorpan: Haha |
| 09:17 | <odinho> | zcorpan: So, phrase the question: "What would be most reasonable? <example>" |
| 09:18 | <matjas> | http://twitter.com/mattur/status/209574686739275776 lolol |
| 09:18 | <Ms2ger> | "Cannot convert 'document.body' to object" |
| 09:18 | <zcorpan> | odinho: but that probably quickly leads to useless bikeshedding and ultimately the Complicator's Gloves |
| 09:19 | <AryehGregor> | jgraham, you probably mean: data:text/html,<body onload="location='data:text/html,1'" onunload="location='data:text/html,2'">3 |
| 09:19 | <AryehGregor> | (note location of ' in onunload attribute) |
| 09:20 | <AryehGregor> | This is just a question of what happens if onunload navigates, right? |
| 09:21 | <AryehGregor> | I'd think it would either fail, or succeed and automatically cancel the pending navigation but not trigger the onunload handler again (obviously). |
| 09:21 | <AryehGregor> | I would sure hope that it doesn't succeed *followed by* the original navigation succeeding. That would be kind of silly. |
| 09:22 | <AryehGregor> | zcorpan, reviewing reftests as commits makes sense, but if you're reviewing *only* reftests (and reftest-like manual tests) I think reviewing per-file probably makes more sense. Which is probably why the CSSWG does it. |
| 09:22 | <AryehGregor> | If you're reviewing both reftests and JS tests, I agree that it makes the most sense to do commit-based review for all of them, but the CSSWG is kind of stuck with its existing system, I guess . . . |
| 09:26 | <matijsb> | good morning, Whatwg! |
| 09:27 | <krijn> | Slijmbal |
| 09:30 | <Ms2ger> | Hah |
| 09:30 | Ms2ger | whacks krijn |
| 09:32 | <MikeSmith> | odinho: it's not much work |
| 09:32 | <MikeSmith> | if you have stuff you want to have approved, let me know |
| 09:33 | <odinho> | MikeSmith: OK webappsec/tests/cors/submitted/opera/staging/resources/cors-makeheader.php << that one. :P |
| 09:34 | <odinho> | If I update it later (same path), is it still approved? |
| 09:34 | <MikeSmith> | no |
| 09:34 | <jgraham> | AryehGregor: Oh, good catch, thanks |
| 09:34 | <jgraham> | Seems to change the behaviour of Opera |
| 09:34 | <Kolombiken> | good morning, Whatwg! |
| 09:38 | <MikeSmith> | odinho: done |
| 09:39 | <Ms2ger> | Fail: assert_throws: function "function() { doc.replaceChild(df, doc.documentElement..." Fail: assert_throws: function "function() { a.replaceChild(a, c); }" threw object "DOMException: HIERARCHY_REQUEST_ERR" that is not a DOMException NotFoundError: property "code" is equal to 3, expected 8 |
| 09:39 | <Ms2ger> | Pass |
| 09:39 | <Ms2ger> | Fail: assert_throws: function "function() { document.replaceChild(doc, document.docu..." threw Element node <html><head><meta charset="utf-8"> <title>Node.replaceChi... that is not a DOMException HierarchyRequestError: property "code" is equal to undefined, expected 3 |
| 09:39 | <Ms2ger> | Fail: assert_throws: function "function() { doc.replaceChild(df, doc.documentElement..." threw Element node <html><head><title>title</title></head><body></body></html> that is not a DOMException HierarchyRequestError: property "code" is equal to undefined, expected 3 |
| 09:40 | <Ms2ger> | 1. Why does Opera fail at copying stuff in columns |
| 09:40 | <Ms2ger> | 2. Why would it claim to throw an Element as an exception? |
| 09:40 | <zcorpan> | maybe it does? |
| 09:42 | <Ms2ger> | It does appear to |
| 09:42 | <Ms2ger> | Fascinating |
| 09:43 | <Ms2ger> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1559 |
| 09:46 | <AryehGregor> | nsCOMPtr<nsIContentIterator> iter = do_CreateInstance("@mozilla.org/content/post-content-iterator;1";, &rv); |
| 09:46 | <AryehGregor> | How could that possibly fail? |
| 09:46 | <Ms2ger> | Wrong window? :) |
| 09:47 | <AryehGregor> | Argh! |
| 09:47 | <AryehGregor> | I get confused when the last person talking is someone like you. |
| 09:48 | <odinho> | lol, someone like you. Sounds ... a bit accusing. :P |
| 09:49 | <matjas> | yeah, why can’t Ms2ger just SHUT UP |
| 09:51 | <jgraham> | I think he means a dirty, good for nothing, Gecko hacker |
| 09:51 | <jgraham> | (uh, does that need a smilie?) |
| 09:52 | <jgraham> | (just in case: ;) |
| 09:52 | <Ms2ger> | ) |
| 09:53 | <jgraham> | hsivonen: I'm not sure I understand the point you were trying to make in bug 17231. |
| 09:53 | <Ms2ger> | "There�s been a big spat over responsive images but WHATWG is unwilling to reverse its decision" |
| 09:54 | <Ms2ger> | Trolling much, leaverou? |
| 10:15 | Ms2ger | grumbles at Opera |
| 10:17 | <annevk> | "Full Screen API" spelling is hard I guess... |
| 10:17 | <annevk> | "The Full Screen API is a joint effort between the WebApps and CSS Working Groups." comedy gold |
| 10:18 | <annevk> | but overall it's super that someone is writing this up |
| 10:28 | <odinho> | Ms2ger: Yeah, storm in a tea cup. |
| 11:02 | Ms2ger | is glad that Opera manages to avoid creating cycles, at least... |
| 11:08 | <zcorpan> | Hixie: is it intentional that cues can have background images? |
| 11:14 | <matjas> | annevk: yeah, it’s like the WHATWG weekly, but with a broader scope. Typos or not, I dig it. |
| 11:20 | <Ms2ger> | Too few DOM4 tabs open... |
| 11:20 | Ms2ger | opens a 7th |
| 11:57 | <Ms2ger> | zcorpan, do you know if Opera has a bug for throwing nodes as exceptions from replaceChild? |
| 11:58 | <zcorpan> | i don't see one |
| 12:00 | <Ms2ger> | http://w3c-test.org/webapps/DOMCore/tests/submissions/Ms2ger/Node-replaceChild.html if you want more of a test case |
| 12:00 | <zcorpan> | thanks! i'll file a bug |
| 12:01 | <Ms2ger> | Np |
| 12:04 | <annevk> | heh, fun bug |
| 12:05 | <zcorpan> | CORE-46726 |
| 12:05 | <Ms2ger> | Definitely not a failure mode I'd met before |
| 12:05 | <Ms2ger> | zcorpan, thanks |
| 12:08 | <annevk> | Ms2ger: I've seen it before; don't remember where though |
| 12:08 | <Ms2ger> | Probably in Opera :) |
| 12:08 | <annevk> | in part it's because we don't have bindings |
| 12:08 | <annevk> | Ms2ger: well yeah :) |
| 12:09 | <Ms2ger> | You don't have bindings? |
| 12:10 | <Ms2ger> | I'm rather surprised how well it works out, then :) |
| 12:22 | <annevk> | http://vimeo.com/43380467 is pretty cool |
| 12:22 | <annevk> | bit of standards bashing |
| 12:25 | <moo-_-> | annevk: nice :) |
| 12:26 | <Ms2ger> | LaTeX slides? |
| 12:27 | <annevk> | what seems somewhat weird to me is that he tells this at a developer's conference |
| 12:45 | <jgraham> | Ms2ger: Mozilla don't exactaly have a glorious history of making DOM-via-code-generation (i.e. bindings) work in a standards compliant / performant way :) |
| 12:46 | <jgraham> | (or at least, so http://jstenback.wordpress.com/2012/04/07/history-of-mozillas-dom-bindings/ would have you believe) |
| 12:47 | <matjas> | in case any of you know: |
| 12:47 | <matjas> | “SVG images know their aspect ratio; why don't browsers respect it?” — http://stackoverflow.com/q/10874763/96656 |
| 12:49 | <gsnedders> | annevk: XSLT sounds a lot like what he wants from CSS. |
| 12:52 | <annevk> | heh |
| 12:55 | <zcorpan> | matjas: it seems the question boils down to s/browsers/chrome 19/ |
| 12:57 | <annevk> | matjas: added an answer |
| 12:57 | <gsnedders> | zcorpan: Who cares about anything apart from WebKit? |
| 12:58 | <matjas> | annevk: thanks! |
| 12:58 | <matjas> | gsnedders: wait, is there anything else? |
| 12:58 | <matzipan> | gsnedders: sarcasm ? |
| 12:59 | <gsnedders> | matjas: What's this Gecko thing? Sounds scary! |
| 13:00 | <matzipan> | I don't want lizards in my computer |
| 13:29 | <zcorpan> | Hixie: the file-spec-bug thingie in the spec dialog saying that the bug was filed (with a link) often doesn't appear for me (maybe because i switch tabs while it's doing its thing and then decides i don't want to see it anymore, or i dunno) |
| 13:30 | <zcorpan> | Hixie: i think you made it fragile and disappear-for-nothing after i complained that i wanted "enter" to dismiss the dialog instead of filing a new bug |
| 13:30 | <zcorpan> | Hixie: but i'm afraid you didn't fix what i wanted and instead made it worse :-P |
| 13:44 | <Ms2ger> | jgraham, definitely not |
| 13:45 | <Ms2ger> | jgraham, but the places where we have to deal with JS values directly tend to end up being rather horrible code |
| 13:49 | <smaug____> | Ms2ger: you might remember |
| 13:49 | <smaug____> | why would anyone use float in idl |
| 13:49 | <smaug____> | and not double |
| 13:49 | smaug____ | needs to check how those map to JS Number |
| 13:50 | <zcorpan> | because they write IDL targeting the Java binding? |
| 13:50 | <smaug____> | this is new stuff |
| 13:50 | <smaug____> | certainly for web only |
| 13:50 | <smaug____> | Gamepad |
| 13:50 | <smaug____> | is there *any* reason to use float |
| 13:50 | <Ms2ger> | No idea |
| 13:51 | <Ms2ger> | heycam|away, can we just kill float? |
| 13:51 | <smaug____> | and this happens again...start reviewing a patch by filing spec bugs first |
| 13:52 | <zcorpan> | the html spec switched all floats to doubles some time (few years maybe?) ago, iirc |
| 14:03 | <smaug____> | there is no such idl type as string, right? |
| 14:03 | <smaug____> | it is DOMString |
| 14:04 | <Ms2ger> | Right |
| 14:27 | <gsnedders> | Is there some reason we can't just call it "string"? |
| 14:36 | <Ms2ger> | gsnedders, objection, I don't want to change all the specs :) |
| 14:39 | <odinho> | ssh dvcs.w3.org -e sed -Ri 's/DOMString/string/' * |
| 14:39 | <odinho> | :P |
| 15:15 | <gsnedders> | Can anyone, from spec, justify what Object.getPrototypeOf(crossOriginWindow) should do? |
| 15:18 | <Ms2ger> | Does the spec for Object.getPrototypeOf no longer deny the existence of crossOriginWindow? |
| 15:28 | <gsnedders> | Ms2ger: The spec for Object.getPrototypeOf says "return [[Prototype]]", pretty much] |
| 15:30 | gsnedders | checks ES6 |
| 15:35 | <boaz> | good morning, Whatwg! |
| 15:37 | <Ms2ger> | Good morning, boaz |
| 15:40 | <gsnedders> | I think per spec you should return the object and then that should be poisoned. |
| 15:40 | <gsnedders> | That isn't what ourselves or WebKit do. |
| 15:40 | <gsnedders> | Dunno about IE. |
| 15:50 | <gsnedders> | So much variation in behaviour. |
| 15:51 | <dglazkov> | good morning, Whatwg! |
| 15:55 | <Ms2ger> | dglazkov, today, you are not alone |
| 15:56 | <dglazkov> | yaaaay! |
| 16:07 | <timeless> | matjas: that /timeless/ stuff is confusing :( |
| 16:08 | <timeless> | odinho: i hope you realize that wouldn't work right at all :) |
| 16:45 | <jgraham> | Must. Not. Use. Sarcasm. On. Mailing. List. |
| 16:46 | <WeirdAl> | that's right. IRC is much, much more forgiving. |
| 16:46 | <WeirdAl> | (mainly because we're much, much more forgetting) |
| 16:46 | <jgraham> | But seriously, wtf at the coremob people saying "we're having a meeting, there's no agenda, decisions are binding, if you can't come provide input in advance" |
| 16:47 | <WeirdAl> | coremob? |
| 16:47 | <jgraham> | I mean apart from the fact that they are requiring claivoyance from non-participants, who the fuck still thinks that F2F meetings should have binding decisions |
| 16:48 | <Ms2ger> | The WG |
| 16:49 | <jgraham> | After all, there's no better way to make a decision than get a small subset of interested parties into a time-constrained situation where they can't digest all the arguments properly and can't change their position on something controversial without losing face. |
| 16:49 | <Hixie> | zcorpan: patches welcome |
| 16:50 | <Hixie> | jgraham: tell them that's not how it's going to happen (assuming you're the editor... if you're not, not much you can do) |
| 16:50 | <timeless> | Ms2ger: the WG doesn't technically exist yet iirc |
| 16:50 | <timeless> | or does it |
| 16:50 | timeless | can't remember which WGs have had their charters' approved |
| 16:50 | <Ms2ger> | timeless, I meant CSS |
| 16:50 | <Hixie> | short of not implementing it |
| 16:51 | <jgraham> | Hixie: I'm not and there's not. Which is why I have resorted to ineffectual whining here |
| 16:51 | <Hixie> | heh |
| 16:51 | <Hixie> | which wg is this? |
| 16:51 | <timeless> | coremob |
| 16:51 | <Hixie> | what is "coremob" |
| 16:51 | <timeless> | it's totally unclear what it will do anyway |
| 16:51 | <Hixie> | sounds like a gang or something |
| 16:51 | <jgraham> | timeless: Maybe that's what they will decide |
| 16:51 | <timeless> | i think the best possible outcome is that it writes test cases for other wg's specs |
| 16:52 | <Ms2ger> | "Anti-device-independent-web" |
| 16:52 | <timeless> | jgraham: dunno, i haven't gotten around to asking my manager if i should go |
| 16:52 | <timeless> | at one point i assumed i would go |
| 16:52 | <jgraham> | Ms2ger: I think that mnight be unfair |
| 16:52 | <Hixie> | oh, a mobile-web wg? in that case it's great news that it's dysfunctional :-) |
| 16:53 | <timeless> | Ms2ger: don't forget the sysapps wg (charter not finalized) |
| 16:53 | <jgraham> | Hixie: I like to think of it as "an interoperability-increaing oppertunity that just happens to be riding on the fashion for the word 'mobile'" |
| 16:53 | <timeless> | if you want the anti-device-independent-web, i think sysapps is a better place for you :) |
| 16:53 | <timeless> | coremob is more like "let's standardize webkit" |
| 16:53 | WeirdAl | waves to his old friend |
| 16:54 | <Ms2ger> | timeless, well, that or my friends at Mozilla |
| 16:54 | <timeless> | hi |
| 16:54 | <jgraham> | Yeah, "let's standardise WebKit" is certianly the dark side of the path it could take |
| 16:54 | <timeless> | technically the path is more like |
| 16:54 | <timeless> | "let's write tests for webkit specific features" |
| 16:54 | <timeless> | ... as an end result, everyone will have to implement them to pass the tests |
| 16:54 | <timeless> | it isn't technically "let's standardize webkit" |
| 16:55 | <timeless> | but the distinction isn't significant |
| 16:55 | <Hixie> | how frightening |
| 16:55 | <timeless> | Hixie: i'm not sure that coremob is the scariest nascent group out there |
| 16:55 | <timeless> | sysapps could be significantly scarier |
| 16:56 | <timeless> | it's basically asking most UAs to implement a non-web security model |
| 16:56 | <timeless> | and note, it |
| 16:56 | <timeless> | 's possible for coremob to be useful and helpful to the web |
| 16:56 | <timeless> | ... IFF it spends its time writing missing test cases for other w3 specs |
| 16:57 | <Hixie> | does sysapps have any UA vendors in it? |
| 16:57 | <jgraham> | I'm pretty sure that the coremob stuff will end up setting some aspirational goals for UAs like "implement spec X before spec Y" and, perhaps, might get some more people writing tests |
| 16:57 | <TabAtkins> | If so, they're not doing their jobs... |
| 16:57 | <jgraham> | sysapps is Mozilla driven, I think |
| 16:57 | <timeless> | Hixie: oddly, yes |
| 16:57 | <Hixie> | christ |
| 16:57 | <timeless> | it has mozilla and chrome at least |
| 16:57 | <timeless> | that's what makes it scary |
| 16:58 | timeless | wonders how Hixie has missed it |
| 16:58 | <Hixie> | how would i have known about it? :-) |
| 16:58 | <timeless> | (well, ok, you were being productive instead of spending your time worrying about uncertain doom) |
| 16:58 | <TabAtkins> | timeless: Do you know if it's actually Chrome, or just Google that's on it? |
| 16:59 | <timeless> | TabAtkins: it isn't the Android-Browser if that's what you're asking |
| 16:59 | <Ms2ger> | timeless, productive or on holiday ;) |
| 16:59 | <timeless> | it seems to be either Chrome-OS or Chrome-Android |
| 17:00 | <timeless> | Hixie: http://www.w3.org/2012/05/sysapps-wg-charter.html |
| 17:00 | <timeless> | i'm not sure if that's current |
| 17:00 | <timeless> | they revised it recently |
| 17:00 | <Hixie> | oh |
| 17:00 | <Hixie> | well |
| 17:00 | <Hixie> | that's doomed |
| 17:00 | <timeless> | TabAtkins: and it seems that certain other webkit groups don't object to adjusting the security model of webkit |
| 17:01 | <timeless> | Hixie: i wish i had your confidence in that |
| 17:01 | <Hixie> | trying to define a spec that competes with native OS apps is a non-starter |
| 17:01 | timeless | is actually more afraid of B2G than Chrome-OS/Chrome-Android |
| 17:01 | <Ms2ger> | Hixie, did you hear about B2G? |
| 17:01 | <Hixie> | b2g is fine so long as it's proprietary |
| 17:01 | <timeless> | Hixie: native-os api's are more or less going out the window |
| 17:02 | <timeless> | Hixie: B2G is one of the groups trying to drive sys-apps |
| 17:02 | <Hixie> | it takes all the worst thins of apps (installation, security nightmare, etc), and combines them with the worst things of the web (slow development, crappy APIs, etc) |
| 17:02 | <timeless> | they don't want to "be proprietary" |
| 17:02 | <timeless> | because they want to borrow apps written for other "devices" |
| 17:02 | <timeless> | note: i'm not endorsing this stuff, |
| 17:03 | <Hixie> | a non-proprietary platform will be unable to compete with a proprietary platform so long as it tries to play the same game |
| 17:03 | <timeless> | i'm just more afraid of it being partially successful than you are confident in it not being so |
| 17:03 | <timeless> | Hixie: btw, would you classify Android as proprietary? |
| 17:03 | <Hixie> | i should say, i don't think they'll fail to write specs and waste lots of engineer time implementing it |
| 17:03 | timeless | really hasn't considered that question in a while |
| 17:03 | <Hixie> | i just think it'll fail in the market |
| 17:03 | <Hixie> | timeless: yeah. i should really say one-vendor-driven, not proprietary. the word is a bit ambiguous in cases like this. |
| 17:04 | <timeless> | Hixie: right |
| 17:04 | <timeless> | so the question is, if it's 2 vendor driven instead of 1 vendor driven |
| 17:04 | <timeless> | does that make things better or worse? |
| 17:04 | <timeless> | (anyone here remember MeeGo?) |
| 17:04 | <Hixie> | it makes them slower |
| 17:04 | <timeless> | (that was 2 vendor driven) |
| 17:04 | <Hixie> | which makes it impossible for it to compete with native OSes |
| 17:04 | <Hixie> | at least while they're playing the same game |
| 17:05 | <gsnedders> | Hixie: Surely the point is to be cross-platform, so complete in the sense that Qt does? |
| 17:05 | <gsnedders> | Which, sure, isn't exactly rivaling native APIs, but still is significant |
| 17:06 | <Hixie> | Qt is single-vendor, and i would still not characterise it as a success in this space |
| 17:08 | <timeless> | Qt is an odd duck |
| 17:08 | <timeless> | there was a time when it made a lot of sense and could have done well |
| 17:08 | <timeless> | but at this point it's implementing stuff over a decade late |
| 17:08 | <timeless> | (see VB/Delphi) |
| 17:09 | <Hixie> | ahh, delphi |
| 17:09 | Hixie | has recently taken to writing all his server-side code in freepascal, which is a free implementation of a superset of delphi |
| 17:09 | <Hixie> | so nice |
| 17:10 | <TabAtkins> | So weird. |
| 17:10 | <Hixie> | there's only really two things about it that i would say are weird, which is really orders of magnitude better than most other platforms |
| 17:11 | <Hixie> | (a) strings are 1-indexed for historical reasons :-/ |
| 17:11 | <Hixie> | and (b) the support for generics is a bit wonky in its current implementation, e.g. it fails when combined with structured types with overloaded operators |
| 17:11 | <Hixie> | (i've run into (b) more often than i'd like) |
| 17:12 | <Hixie> | (you get used to (a) pretty quickly and there's language features to hide it anyway) |
| 17:12 | <timeless> | Hixie: does freepascal have the VCL? |
| 17:12 | <Hixie> | yeah, a reimplementation called the FCL |
| 17:12 | <Hixie> | i don't generally use it |
| 17:12 | <timeless> | wow |
| 17:12 | <Hixie> | since i'm making server-side software, not guis |
| 17:12 | <timeless> | yeah, it only makes sense if you're doing desktop gui |
| 17:12 | <timeless> | right |
| 17:12 | <Hixie> | also has a reimplementation of TurboVision |
| 17:12 | <Hixie> | cross-platform, in both cases |
| 17:12 | <timeless> | i've been using CherryPy to write an HTML app |
| 17:13 | <Ms2ger> | Back in my day, I used Borland Pascal |
| 17:13 | <moo-_-> | Turbo Pascal FTW |
| 17:13 | <timeless> | Ms2ger: Borland Pascal or Turbo Pascal? |
| 17:13 | <timeless> | and have you seen Quick Pascal? |
| 17:13 | <Ms2ger> | Borland, and no |
| 17:13 | <timeless> | http://en.wikipedia.org/wiki/Microsoft_Pascal |
| 17:14 | <timeless> | how about Turbo Basic? :) |
| 17:14 | <timeless> | http://en.wikipedia.org/wiki/Turbo_Basic |
| 17:14 | <moo-_-> | timeless: microsoft had a Quick Basic which was a commercial version of QBasic |
| 17:14 | <Ms2ger> | (The next year my school moved to JS) |
| 17:14 | <moo-_-> | creating .exes |
| 17:14 | <timeless> | moo-_-: yeah, i used it |
| 17:14 | <timeless> | a friend of mine owned it |
| 17:14 | <timeless> | Ms2ger: i used Turbo Pascal in 10th grade for AP CS |
| 17:14 | <moo-_-> | now, one needs to write qbasic interpreter in JS... |
| 17:15 | <timeless> | the next year they switched to C++ iirc |
| 17:15 | <Hixie> | wow, turbo basic was a compiler, not an interpreter |
| 17:16 | <timeless> | i think i used it when it was powerbasic (much later) |
| 17:16 | <timeless> | Hixie: i'm not sure that's a proper interpretation |
| 17:16 | <timeless> | it /had/ a compiler |
| 17:16 | <timeless> | "The integrated development environment could run a BASIC program internally for traditional BASIC debugging (see sample below)," |
| 17:17 | <Hixie> | that's what i meant, yes |
| 17:17 | <Hixie> | sorry |
| 17:17 | <timeless> | sorry to be picky :) |
| 17:17 | <Hixie> | hey if one can't be a pedant in #whatwg, where can one be a pedant |
| 17:18 | <Hixie> | this is like the capital of pedant country |
| 17:18 | <timeless> | oh good, i knew there was a reason i felt so at home here :) |
| 17:18 | <Ms2ger> | !summon MikeSmith |
| 17:20 | <annevk> | Ms2ger: just talk to [tm] |
| 17:21 | <Ms2ger> | [tm], http://dvcs.w3.org/hg/fullscreen/raw-file/tip/Overview.html has issues |
| 17:21 | <annevk> | whoa what happened there? |
| 17:21 | <jernoble> | the mime type probably changed |
| 17:21 | <Ms2ger> | Yeah |
| 17:22 | <annevk> | weird that it would only affect Fullscreen |
| 17:22 | <jernoble> | Content-Type: application/binary |
| 17:22 | <annevk> | oh it doesn't |
| 17:22 | <annevk> | cannot load the URL spec either |
| 17:22 | <annevk> | or DOM |
| 17:22 | <annevk> | grmbl |
| 17:22 | <timeless> | |
| 17:22 | <timeless> | Content-Type: application/binary |
| 17:22 | <annevk> | shepazu: ^^ |
| 17:22 | <TabAtkins> | CSS specs are okay. |
| 17:22 | <timeless> | according to http://web-sniffer.net/ |
| 17:23 | <timeless> | oh |
| 17:23 | <timeless> | that's probably a Hg upgrade |
| 17:23 | <shepazu> | annevk: sorry, what are you pointing at? |
| 17:23 | <timeless> | there's a risk of hg repos being used to run arbitrary code |
| 17:23 | <timeless> | so you have to tell hg's web server you're ok w/ hosting html |
| 17:23 | <shepazu> | I don't know anything about QBasic |
| 17:23 | <timeless> | otherwise... |
| 17:24 | <annevk> | shepazu: loading http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html or http://dvcs.w3.org/hg/fullscreen/raw-file/tip/Overview.html no longer works |
| 17:24 | <shepazu> | wtf? |
| 17:24 | <timeless> | => annevk |
| 17:24 | <[tm]> | Ms2ger: I'm on a smoke break |
| 17:24 | <annevk> | shepazu: MIME type changed on the server |
| 17:24 | <shepazu> | thanks, annevk, I'll pass that on |
| 17:24 | <shepazu> | d'oh. |
| 17:24 | <timeless> | annevk: not technically on the server ^^ |
| 17:24 | <Ms2ger> | [tm], we've got smart people on it, thanks anyway ;) |
| 17:25 | Hixie | tries to debug some random web page and finds himself deep in a maze of twisty <div>s, all alike |
| 17:25 | <annevk> | timeless: pedant |
| 17:25 | <shepazu> | Ms2ger: no, I'm handling it |
| 17:25 | <annevk> | lol |
| 17:25 | <timeless> | http://selenic.com/repo/hg/rev/d06b9c55ddab |
| 17:25 | <timeless> | annevk: it matters because it affects where the fix is |
| 17:25 | <Ms2ger> | shepazu, every now and then, I try to be nice to you ;) |
| 17:25 | <timeless> | shepazu: guessmime is what you'll need in /etc/hgrc or whatever |
| 17:25 | <timeless> | [web] |
| 17:25 | <timeless> | guessmime=True |
| 17:26 | <timeless> | shepazu: ^ adding that section+line should fix it |
| 17:29 | <shepazu> | timeless, annevk, our crack team is dealing with it |
| 17:29 | <timeless> | shepazu: shrug |
| 17:29 | <timeless> | if you pass along what i suggested it'll save them time |
| 17:29 | <shepazu> | so… it might be a while, until they can afford a really primo rock |
| 17:29 | <timeless> | and thanks for pinging someone :) |
| 17:29 | <shepazu> | timeless: I did pass on your suggestion, thanks! :) |
| 17:30 | <timeless> | :) |
| 17:30 | <shepazu> | hopefully that will be quick |
| 17:30 | <annevk> | thank you! |
| 17:32 | <shepazu> | annevk, timeless, should be fine now… we updated to mercurial 1.8.something this morning, this used to be the default behavior, says our crack team |
| 17:32 | <timeless> | correct |
| 17:33 | <timeless> | the behavior was changed because it was deemed insecure for groups that weren't aware of it |
| 17:33 | <annevk> | shepazu: fine it is :) |
| 17:33 | <timeless> | for specific groups that wanted the feature, the flag was added |
| 17:33 | <shepazu> | cool, sorry for the inconvenience, and thanks for the help |
| 17:47 | <Ms2ger> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=17302 |
| 17:58 | <timeless> | cute |
| 17:58 | <Hixie> | TabAtkins: technically we _did_ "merge", twice so far, once with the htmlwg and once with a cg. :-) |
| 18:02 | <timeless> | heh |
| 18:03 | <TabAtkins> | Hixie: Humor cares nothing for your facts. |
| 18:03 | <Hixie> | hah |
| 18:04 | <TabAtkins> | Btw, how's the CG thing working? When can I stop telling people that we care about the HTMLWG just because they offer a patent policy? |
| 18:05 | <Hixie> | i should do a FSA publication soon |
| 18:06 | <timeless> | fsa? |
| 18:06 | <Hixie> | "Final" Specification Agreement |
| 18:07 | <Hixie> | the patent-policy version of CG publication |
| 18:11 | timeless | ponders kanji and friends |
| 18:19 | <Ms2ger> | "My computer is old and slow. I got a message on the internet I should switch to standard HTML version. Please advise." |
| 18:20 | Ms2ger | closes the bug before TabAtkins humours it |
| 18:25 | <TabAtkins> | Why you do this, Ms2ger? Why you hate humor? |
| 18:25 | <Ms2ger> | Yes |
| 18:25 | <Ms2ger> | I prefer humour |
| 18:25 | <TabAtkins> | I can Brit up my humor, give the people a u if they want. |
| 18:26 | <TabAtkins> | Make some jokes about the queen. |
| 18:29 | <Philip`> | TabAtkins: Only if you want to end up in the Tower |
| 18:57 | <Velmont> | MikeSmith: Yo, there? w3c-test.org doesn't seem to have the latest and greatest /resources/. |
| 18:58 | <Velmont> | MikeSmith: Oh fsck. I didn't commit that yet, no. Stupid. Heh. |
| 18:58 | gsnedders | slow-claps Velmont |
| 18:59 | <Velmont> | :-S |
| 18:59 | <Velmont> | gsnedders: Can you review then, since you're here and acting all cool and all? |
| 18:59 | <Velmont> | https://github.com/w3c/testharness.js/pull/1 |
| 19:00 | <gsnedders> | No. |
| 19:00 | <Velmont> | Hmm. Though, Ms2ger didn't seem to have any more comments. So maybe I can just push it in there... Hmmz. |
| 19:00 | <Velmont> | jgraham: <make noise> |
| 19:00 | <gsnedders> | Mainly because I'm about to head out :) |
| 19:01 | <Velmont> | gsnedders: Tss, life, it's all a lie! |
| 19:01 | <gsnedders> | But I can a fulfill my purpose in life, to impregnate a couple of girls otherwise? |
| 19:02 | <Velmont> | gsnedders: lol, that's so nasty. You shouldn't say stuff like that, boy! |
| 19:02 | <Ms2ger> | Velmont, land it |
| 19:02 | <Ms2ger> | Velmont, he can do that, he's Scottish |
| 19:02 | <gsnedders> | Ms2ger: Aye dinnae no wha' yir oon aboot. |
| 19:02 | Velmont | installing Mercurial and hunts for his W3C-password |
| 19:35 | <timeless> | annevk: thanks for complaining about that hg issue |
| 19:35 | <timeless> | that explains why my Chrome had a bunch of "index.html" files to save |
| 20:17 | <Velmont> | 18:09 < timeless> odinho: i hope you realize that wouldn't work right at all :) <--- way to spoil all the fun!!1 |
| 20:17 | <Velmont> | But, yes, I know... It was highly pseudoish code :P (16:41 < odinho> ssh dvcs.w3.org -e sed -Ri 's/DOMString/string/' *) |
| 20:17 | <timeless> | Velmont: specifically dvcs repos can easily be corrupted by people doing stupid things like that |
| 23:54 | <TabAtkins> | If a script runs before you and overrides one of the window builtins, there's no way to get a copy of the original, right? |
| 23:55 | <TabAtkins> | Say the script preempting you is an extension that runs on all pages, so you can't do something like create an iframe and yank it off of there. |
| 23:56 | <jamesr> | there's no general way in the web platform to hack around a browser's extension system |