| 03:40 | <zewt> | ... amazon.co.jp uses doublewidth backslashes in JS code for string escapes? gross |
| 07:48 | <rabbi1> | using this innerHTML, innerText, innerShit, innerDick are these supported by FF, and Chrome, ? |
| 09:58 | <Ms2ger> | OH: "I do not know how booleans work tralala" |
| 10:02 | annevk | gets curious about context |
| 10:09 | <Ms2ger> | http://dvcs.w3.org/hg/webapps/rev/7d1f6ad8a1fb |
| 10:09 | <annevk> | ouch |
| 10:09 | <annevk> | :) |
| 10:11 | <Ms2ger> | :) |
| 10:12 | <tmzt> | it was the beginning of time... |
| 10:12 | <tmzt> | decemember 31, 1969 to be precise |
| 10:13 | <Ms2ger> | Oh man |
| 10:13 | <Ms2ger> | You made me realize that people in their forties can represent their birth dates as a unix time stamp |
| 10:14 | <Ms2ger> | I feel like living in the future |
| 14:11 | <Lachy> | Hixie, I'm back now. I've been on holiday for 10 days, just got home. Ping me again later re :scope, <style scoped>, etc. |
| 14:14 | <Lachy> | Hixie, I'll catch up on the email you wrote about it a bit later. |
| 16:14 | <remysharp> | am I right in thinking that web workers and shared workers should be able to access IndexedDB methods? |
| 16:18 | <beverloo> | IndexedDB should make its features available in the WorkerUtils interface, which is part of a worker's global scope |
| 16:19 | <remysharp> | cheers. |
| 16:19 | <remysharp> | but sessionStorage (and local) isn't available, right? |
| 16:21 | <beverloo> | According to http://dev.w3.org/html5/webstorage/#the-localstorage-attribute it extends the Window interface, so no |
| 16:22 | <beverloo> | which would make sense as (at least shared) workers don't have a domain scope |
| 16:24 | <remysharp> | that's what I thought - cheers for the tips beverloo |
| 16:24 | <zewt> | run! |
| 17:01 | <Philip`> | Ms2ger: Ping |
| 17:01 | <Ms2ger> | Hi |
| 17:01 | <Philip`> | Ms2ger: About general canvas-test-related matters: I'll probably continue to ignore you for the next couple of weeks (got a PhD to finish) but should have some time after that, so please poke me then if I still forget :-) |
| 17:02 | <Ms2ger> | Will do :) |
| 17:08 | <zewt> | it's bad that it's getting refreshing when pages *don't* have kludgy GUI text editors on pages--where I can undo and actually *not* have the cursor madly jump to some random place in the document |
| 18:17 | <Ms2ger> | Anybody who happens to have outerHTML tests lying around? |
| 18:38 | <jarek> | Hi |
| 18:38 | <jarek> | I remember reading somewhere that SVG filters are going to be replaced with CSS |
| 18:39 | <jarek> | I mean that filters would be no longer defined with XML |
| 18:39 | <jarek> | is this proposal still actual? |
| 18:43 | <jarek> | Microsoft initially said that it is not planning to support SVG fillters (+ SMIL and SVG Fonts) because they thought that the standard is going to change soon |
| 18:44 | <jarek> | but they added SVG filters in latest IE10 preview |
| 21:00 | <hsivonen> | is there already a wiki page for listing stuff that's going in Acid4? |
| 21:00 | <hsivonen> | or should it be Acid Living Test? |
| 21:00 | <Ms2ger> | Weren't we giving up on acid tests? |
| 21:01 | <hsivonen> | Ms2ger: maybe. I don't know. |
| 21:01 | Ms2ger | prefers conformance test suites |
| 21:02 | <Ms2ger> | I'd also prefer if someone else wrote outerHTML tests for me :) |
| 21:03 | <AryehGregor> | We should do proper conformance tests. |
| 21:03 | <AryehGregor> | Those are actually useful on getting browsers to converge on real standards, not just a handful of random unrelated details. |
| 21:04 | <AryehGregor> | Passing Acid3 is almost meaningless. You can support tiny subsets of the relevant features and get a full score. |
| 21:05 | <Ms2ger> | Well... |
| 21:05 | <hsivonen> | yay wikipedia: http://en.wikipedia.org/w/index.php?title=Acid3&diff=prev&oldid=451178499 |
| 21:05 | <Ms2ger> | Firefox passes acid3 now, so it must be a good test :) |
| 21:05 | <hsivonen> | Ms2ger: are you still planning on implementing outerHTML? |
| 21:05 | <gsnedders> | AryehGregor: And people claim that smaller browsers did exactly that (like Opera and SVG fonts, which we supported for years before Acid3…). :\ |
| 21:06 | <Ms2ger> | I spent some time on it today, actually |
| 21:06 | <hsivonen> | Ms2ger: cool. any guestimate of completion time frame? |
| 21:06 | <AryehGregor> | gsnedders, I thought there were some cases where certain browsers did actually add bogus special cases to pass Acid3. |
| 21:06 | <Ms2ger> | AryehGregor, sure, webkit does |
| 21:06 | <AryehGregor> | I was going to say, I had heard that. |
| 21:06 | <Ms2ger> | But that's the way they roll |
| 21:06 | <Ms2ger> | s/does/did/, I should say |
| 21:07 | <Ms2ger> | hsivonen, I'll try to finish it this week, so expect a lot of stupid questions coming your way :) |
| 21:07 | <hsivonen> | Ms2ger: ok. great. |
| 21:07 | <gsnedders> | AryehGregor: It's one thing something getting bad publicity for doing so, it's another thing getting bad publicity when it's untrue. :\ |
| 21:08 | <hsivonen> | AryehGregor: was this the font anti-aliasing thing? |
| 21:08 | <hsivonen> | didn't that get fixed in the test rather soon? |
| 21:08 | <AryehGregor> | gsnedders, well, it's Opera's fault for being closed-source. You force people to make up non-falsifiable malicious rumors instead of demonstrably true malicious rumors. |
| 21:08 | <AryehGregor> | hsivonen, I vaguely remember it as being about that, yes. |
| 21:09 | <gsnedders> | AryehGregor: How so? We shipped browsers for years that disprove that theory. |
| 21:09 | <gsnedders> | AryehGregor: Shipping a browser a number of years before Acid3 was published with SVG fonts support would appear to disprove that. |
| 21:10 | <hsivonen> | for some levels of "support" |
| 21:10 | <Ms2ger> | Were the changes to acid3 announced somewhere, btw? |
| 21:10 | <hsivonen> | Ms2ger: G+ |
| 21:10 | <AryehGregor> | gsnedders, you can't prove it. Maybe you secretly changed the old released versions to add the feature so that it looked like you always supported it. The only way to disprove it would be to have the whole version history in a DVCS that uses cryptographic hashes for revision id's so it's impossible to rewrite the history without everyone noticing. |
| 21:10 | <hsivonen> | Ms2ger: https://plus.google.com/u/0/107429617152575897589/posts/JdHnqpuUER4 |
| 21:11 | <Ms2ger> | Meh, G+ |
| 21:12 | <Philip`> | AryehGregor: How is that any different to people having downloaded binaries in the past and verifying that the hashes are the same as the current provided binaries? |
| 21:13 | <AryehGregor> | Philip`, how many people keep around old binaries, and of those, how many are actually going to do the comparison? With a DVCS, anyone with a checkout automatically verifies the history every time they do a fetch. |
| 21:13 | <AryehGregor> | Also, anything closed-source is automatically evil and wrong, regardless of facts or other such distractions. |
| 21:14 | <hsivonen> | AryehGregor: assuming your DVCS client hasn't been compromised |
| 21:14 | AryehGregor | goes back to writing an e-mail in Gmail using Chrome |
| 21:14 | <AryehGregor> | hsivonen, yes, well, come on. Let's not be paranoid here. |
| 21:15 | <gsnedders> | AryehGregor: People don't need that level of guarantee. Plenty of people believed CVS servers's dates, yet those are trivially changed. |
| 21:15 | <AryehGregor> | gsnedders, those people are deluded victims of pre-DVCS technology. Any centralized version control system is also evil and wrong. |
| 21:16 | <gsnedders> | AryehGregor: Also, git commit --date allows you to change the authored date of a commit. |
| 21:16 | <AryehGregor> | Also KDE, emacs, and Microsoft. |
| 21:16 | <AryehGregor> | gsnedders, it will still look like a descendant of earlier commits, so you're not going to fool anyone. |
| 21:17 | hsivonen | mumbles about WebKit using SVN |
| 21:18 | <gsnedders> | AryehGregor: And if someone like bz seriously thinks we doctor our history, then that's fucking sad. |
| 21:18 | <AryehGregor> | hsivonen, Chromium has an official git mirror, though. |
| 21:18 | <AryehGregor> | And that includes WebKit somewhere. |
| 21:18 | <AryehGregor> | gsnedders, I don't know how the word "seriously" found its way into this conversation. |
| 21:18 | <AryehGregor> | It must be lost. |
| 21:19 | <Philip`> | What if someone changes the release tags in the DVCS to point to a falsified backdated revision? |
| 21:20 | <gsnedders> | AryehGregor: bz amongst others have repeatedly claimed that we support SVG fonts because of Acid3 and just enough to pass it. It has been pointed out to him that Opera 8 (which predates Acid *2*) supports as much as the latest Opera release, yet he still claims it. |
| 21:20 | <AryehGregor> | I dunno anything about what bz thinks. |
| 21:20 | <AryehGregor> | I was actually thinking of the WebKit font thing, which I vaguely recalled someone mentioning. |
| 21:21 | <AryehGregor> | I think I remember dbaron or someone complaining about WebKit cheating on something for performance and then admitting that Opera was often faster than Gecko but he never caught it cheating yet. |
| 21:21 | <gsnedders> | AryehGregor: Yeah, which they did genuinely temporarily special-cased. |
| 21:21 | <gsnedders> | AryehGregor: and then dhyatt and Hixie came to an agreement to remove that |
| 21:21 | <AryehGregor> | (although it emerged that dhyatt pointed out WebKit wasn't cheating, it just implemented the feature partially to avoid a performance regression) |
| 21:22 | <AryehGregor> | Philip`, why does that matter? You can easily inspect the revision and detect the falsification. |
| 21:23 | <gsnedders> | The only thing I think we only partially fixed for Acid3 was null-terminated strings in DOM, and the chance of us actually fixing that all in one go was always minute, because it means changing basically the entirity of DOM. |
| 21:24 | <AryehGregor> | Well, I plan to make sure that every single DOM-related test that I write tests for proper handling of nulls in strings. |
| 21:25 | <AryehGregor> | Actually, maybe I should update my DOM Range tests to do that. |
| 21:25 | <AryehGregor> | My reflection tests already do. |
| 21:25 | <AryehGregor> | Although there are basically no strings in DOM Range, so I don't know what I'd change exactly. |
| 21:25 | <gsnedders> | Probably won't really prioritize us fixing it, though, as long as there are things that actively are known to hurt site-compat. |
| 21:26 | <AryehGregor> | In the long term everyone will want to score 100% on all these tests. |
| 21:26 | <AryehGregor> | I'm not worried about the precise timescales. |
| 21:26 | <AryehGregor> | If there are bigger fish to fry for now, be my guest. |
| 21:38 | <annevk> | Ms2ger, I quoted the post in full on my blog, in case G+ goes away |
| 21:39 | <jgraham> | gsnedders: Wow, holy not worth geting upset about, batman |
| 21:39 | <gsnedders> | jgraham: You think that's me upset? |
| 21:41 | <jgraham> | gsnedders: Way more than it deserves |
| 21:41 | jgraham | expects he accidentially misrepresents the history of Mozilla and Webkit all the time and hopes no one gets upset about it |
| 21:42 | <jgraham> | And Opera too probably |
| 21:43 | <annevk> | I think it's nice that gsnedders cares and 386s the interwebs |
| 21:44 | <annevk> | I for one did not know we have done SVG fonts since Opera 8 |
| 21:44 | <jgraham> | Coincidentially, 8 is the number of people that ever used SVG fonts |
| 21:45 | <jgraham> | ;) |
| 21:47 | <annevk> | I don't really care for the feature, especially if we don't do per-character-animations-and-unicorns-and-double-rainbows, but if we get unwarranted flak for it, might as well set people straight |
| 21:51 | <jgraham> | Well sure, I woudl rather people had an accurate story. But I'd also rather not accuse people of "thinking we doctor our history" when it seems much more plausible that they made an honest mistake on a subject where it's easy to see how one might be mistaken |
| 21:52 | <gsnedders> | jgraham: I don't think that. I use merely using the suggested theory. |
| 21:53 | <jgraham> | Anyway I think the fundamental point is that the way acid 3 worked was a disaster |
| 21:53 | <jgraham> | ACID4 should be "the full HTML5 conformance testsuite. rendered as a giant smiley face, one pixel per test" |
| 21:54 | <zewt> | AFHCTRAAGSFOPPT4 |
| 21:57 | <annevk> | Really? I thought Acid3 was pretty good |
| 22:01 | <Philip`> | Did it lead to improvements in the lives of normal web authors? |
| 22:04 | <annevk> | Finally Selectors work everywhere, Media Queries, CSS3 Colors, ... |
| 22:04 | <annevk> | Yeah, I would think so |
| 22:06 | <jgraham> | A subset of it was probably helpful. But it was too small and too much was only tested in a shallow way |
| 22:06 | <jcranmer> | the problem with Acid3 is when it tests poorly-specified components (e.g., SVG fonts) |
| 22:06 | <annevk> | Too small? It was much bigger than any previous Acid test... |
| 22:06 | <jcranmer> | if the goal is to promote interoperability |
| 22:07 | <jcranmer> | you really need to focus on the edge cases |
| 22:07 | <jgraham> | And much smaller than a good testsuite |
| 22:07 | <annevk> | Yeah, the drive for having a 100 tests should not have been done |
| 22:07 | <jcranmer> | the CSS 2.1 test suite has tens of thousands of tests |
| 22:07 | <annevk> | jgraham, seems you are missing the point of Acid tests |
| 22:08 | <jgraham> | annevk: publicity? |
| 22:08 | <annevk> | Yeah, lots of Microsoft copy pasta in there (CSS 2.1 test suite) |
| 22:08 | <jcranmer> | and is comprehensive enough that to pass it means you probably interoperably implement it with other UAs |
| 22:09 | <jgraham> | The CSS2.1 testsuite is undoubtedly weirdly balanced |
| 22:09 | <jcranmer> | Acid2 was a stress test for layout edge cases |
| 22:09 | <jgraham> | Like I seem to recall a few hundred tests for drawing outlines of different colours |
| 22:09 | <jcranmer> | Acid3 was a grab bag of "implement this feature" and "this browser has this weird bug" |
| 22:10 | <jcranmer> | so passing Acid2 meant fixing a lot of bugs and truly arriving at a state where browsers interoperably laid out a page |
| 22:11 | <jgraham> | Don't forget the the super-weird "performance test" |
| 22:11 | <jcranmer> | passing Acid3 is more like "here's another checkbox of things we implement" |
| 22:11 | <annevk> | jcranmer, I don't think Acid2 and Acid3 were that different actually |
| 22:11 | <annevk> | from a design perspective |
| 22:12 | <annevk> | both were testing annoying bugs and lack of certain features |
| 22:12 | <jcranmer> | well, I can't say that SVG fonts is a terribly useful feature to have |
| 22:12 | <jcranmer> | :-) |
| 22:13 | <annevk> | water under the bridge |
| 22:15 | <jgraham> | Not just SVG fonts |
| 22:15 | <jcranmer> | the big problem is that while the first Acid tests were fairly comprehensive test suites |
| 22:16 | <jgraham> | Any test that has components that amount to little more than feature detection is bad |
| 22:16 | <jcranmer> | Acid3 tested many more things in a more shallow way |
| 22:16 | <moo-_-> | if audio.loop = true is there a way to get number of loop times and does currentTime go back to zero when the loop happens? |
| 22:22 | <annevk> | jcranmer, http://www.w3.org/Style/CSS/Test/CSS1/current/test5526c.htm is a fairly comprehensive test suite? |
| 22:23 | annevk | goes to watch some series instead |
| 22:25 | <jcranmer> | well, CSS 1 not so much |
| 22:29 | <zewt> | i suggest geometric |
| 23:45 | <shepazu> | wow, there's a lot of ignorance around SVG fonts on this channel |
| 23:46 | <shepazu> | actually, they are very useful |
| 23:46 | <shepazu> | they solve several use cases that WOFF fonts and other EOT fonts don't |
| 23:46 | <shepazu> | (and don't solve others as well) |
| 23:47 | <shepazu> | it's sad when a technical solution becomes political fodder like this |