| 01:51 | <ian128K> | Just wanted to put in my 2¢ regarding responsive images: I much prefer the proposed <picture> element over adding a "set" attribute to the existing <img> element. |
| 02:14 | <zewt> | (clearly what we need are: more unsubstantiated opinions) |
| 04:32 | <Hixie> | i have a utf-8 validator function (given input bytes, give a thumbs-up or thumbs-down on whether it's valid utf-8) |
| 04:32 | <Hixie> | anyone got any good test data for that? |
| 05:22 | <asmodai> | Hixie: http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt |
| 05:23 | <asmodai> | http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-demo.txt |
| 06:26 | <zcorpan> | asmodai: nice tests |
| 06:59 | <zcorpan> | TabAtkins: http://www.zurb.com/playground/responsive-tables |
| 07:17 | <Hixie> | asmodai: yeah, i know about those, but they don't list the utf-8 bytes nor indicate valid vs not-valid so they're hard to use |
| 07:17 | <asmodai> | zcorpan: Yeah, useful at least |
| 07:17 | <asmodai> | Hixie: So what exactly are specifically looking for? |
| 07:17 | <asmodai> | Hixie: Because I am sure some of my acquaintances will have these tests lying around |
| 07:17 | <Hixie> | something like "0xC0 0x01 INVALID" "0x01 0x01 VALID" |
| 07:18 | <asmodai> | ah |
| 07:18 | <asmodai> | ok, let me dig around for you |
| 07:18 | <Hixie> | or "0x01 0x01 => U+00001 U+00001" "0xC0 0x01 => U+0FFFD U+00001" |
| 07:19 | <Hixie> | (since i can derive the former from the latter) |
| 07:19 | <Hixie> | (by looking for FFFD in the output side) |
| 07:31 | <matjas> | http://mathias.html5.org/tests/named-character-references/ Opera and Firefox pass all tests, WebKit nightly fails one, IE10pre fails 3. IE9 fails 1869 of them. |
| 07:36 | <smaug____> | I wonder why "If current node is a pre, textarea, or listing element, and the first child node of the element, if any, is a Text node whose character data has as its first character a U+000A LINE FEED (LF) character, then append a U+000A LINE FEED (LF) character." is needed |
| 07:40 | <Ms2ger> | Because otherwise the one LF is discarded when reparsing |
| 07:41 | <smaug____> | ahaa |
| 07:58 | <smaug____> | hmm, does innerHTML work differently for data documents |
| 07:59 | <smaug____> | seems like |
| 08:00 | <smaug____> | I doubt that is right |
| 08:02 | <zcorpan> | smaug____: data documents? |
| 08:02 | <smaug____> | zcorpan: I mean something like createHTMLDocument() |
| 08:02 | <zcorpan> | how is it different? |
| 08:03 | <smaug____> | "or if the parent of current node is noscript element and scripting is enabled for the node, then append the value of current node's data IDL attribute literally." |
| 08:04 | <smaug____> | I don't see what defines that scripting is enabled for a node in a data document |
| 08:05 | <smaug____> | yes, "(A Document created using an API such as createDocument() has no browsing context." |
| 08:05 | <smaug____> | so, different serialization |
| 08:06 | <annevk> | how can scripting be disabled if you're doing scripting? |
| 08:06 | <zcorpan> | but the spec doesn't say that scripting should be disabled for such documents, does it? |
| 08:06 | <zcorpan> | (scripts don't run in documents without a browsing context, but that's different) |
| 08:06 | <smaug____> | "Scripting is enabled for a node if the Document object of the node (the node itself, if it is itself a Document object) has an associated browsing context, " |
| 08:07 | <zcorpan> | aha |
| 08:08 | <zcorpan> | yeah that seems like a bug then |
| 08:08 | <smaug____> | and I was convinced that spec's html-fragment algorithm might be bug-free... |
| 08:09 | <smaug____> | but good to go through |
| 08:09 | <zcorpan> | who convinced you? there are open bugs on it :-) |
| 08:10 | <smaug____> | well, perhaps no one said bug-free |
| 08:10 | <smaug____> | but not full of bugs |
| 08:11 | <zcorpan> | i guess that's about right |
| 08:14 | <zcorpan> | would be nice if the spec had annotations about open bugs |
| 08:14 | <zcorpan> | Hixie: ^ |
| 08:16 | <smaug____> | that would be awesome |
| 08:19 | <rniwa> | smaug____: hi smaug____! |
| 08:20 | <rniwa> | smaug____: slowing improving webkit's fragment parsing APIs :D https://bugs.webkit.org/show_bug.cgi?id=87454 |
| 08:20 | <rniwa> | smaug____: please file bugs & cc me if you find more bugs w.r.t. createContextualFragment, etc... |
| 08:21 | <rniwa> | oops I misread the long |
| 08:21 | <rniwa> | matjas: ^ what I just said to smaug____ |
| 08:21 | <smaug____> | ok |
| 08:21 | <smaug____> | uh |
| 08:22 | <smaug____> | I'll file some spec bugs |
| 08:22 | <rniwa> | smaug____: yeah, it'll be nice if Ms2ger's spec becomes more mature |
| 08:22 | <matjas> | zcorpan: +∞ |
| 08:22 | <rniwa> | smaug____: there are a lot of vaguely defined stuff in the spec at the moment :\ |
| 08:22 | <smaug____> | rniwa: this is in HTML spec |
| 08:22 | <smaug____> | http://www.whatwg.org/specs/web-apps/current-work/multipage/the-end.html#html-fragment-serialization-algorithm |
| 08:23 | <rniwa> | smaug____: oh, sure browser context, etc... stuff is in, yes |
| 08:23 | <smaug____> | looks like noscript handling is all wrong there |
| 08:23 | <rniwa> | smaug____: but that alone doesn't do much |
| 08:23 | <rniwa> | smaug____: because it doesn't define APIs for them |
| 08:23 | <rniwa> | smaug____: oh yeah? |
| 08:23 | <smaug____> | spec defines something that browsers don't do |
| 08:23 | <rniwa> | :( |
| 08:24 | <rniwa> | scumbag spec monsters. |
| 08:25 | <rniwa> | smaug____: one of these days, i'll go through that part of the spec and give a more comprehensive feedback. |
| 08:28 | <rniwa> | smaug____: btw, https://plus.google.com/u/0/105748986001435560355/posts/63GRsrNgGK3 |
| 08:29 | <rniwa> | smaug____: i made a sample code for obtaining mutation records between events |
| 08:29 | <rniwa> | the specific use case i was thinking of was for execCommands since those will how fire input event. |
| 08:31 | <smaug____> | rniwa: your createMutationObserver doesn't probably work in Nightly |
| 08:32 | <smaug____> | which has MutationObserver |
| 08:32 | <rniwa> | smaug____: yeah, saw just your prefix removal. |
| 08:32 | <rniwa> | smaug____: i'll find sometime to update it this weekend |
| 08:33 | <smaug____> | rniwa: hey, btw, since I haven't managed to figure out how to update Chromium, does the latest build handle documentFragments correctlyl |
| 08:33 | <smaug____> | correctly |
| 08:33 | <smaug____> | I mean DOM changes to documentFragment |
| 08:35 | <rniwa> | smaug____: what do you mean by "correctly"? |
| 08:36 | <smaug____> | rniwa: I mean, I couldn't get any mutation records |
| 08:36 | <rniwa> | smaug____: oh i see. |
| 08:36 | <smaug____> | when observing documentfragment |
| 08:36 | <rniwa> | smaug____: it should be fixed. |
| 08:36 | <smaug____> | and adding new child nodes |
| 08:36 | <smaug____> | ok |
| 08:37 | <rniwa> | smaug____: http://www.chromium.org/getting-involved/download-chromium |
| 08:37 | smaug____ | really needs to figure out how to get up-to-date C |
| 08:37 | <smaug____> | hey, that is what I need :) |
| 08:37 | <smaug____> | rniwa: thanks! |
| 08:37 | <rniwa> | smaug____: this is the latest trunk build |
| 08:37 | <rniwa> | of chromium |
| 08:37 | <rniwa> | smaug____: but which means it won't self-update :( |
| 08:38 | <smaug____> | I don't care about that |
| 08:38 | <smaug____> | but I couldn't find that link to .zip file |
| 08:38 | <rniwa> | smaug____: if you want chromium equivalent of aurora, then go to http://www.chromium.org/getting-involved/dev-channel |
| 08:38 | <rniwa> | smaug____: http://download-chromium.appspot.com/ ? |
| 08:38 | <smaug____> | yes |
| 08:38 | <rniwa> | smaug____: no link? |
| 08:39 | <smaug____> | er, I mean, there was .zip file somewhere |
| 08:39 | <rniwa> | smaug____: oh |
| 08:39 | <smaug____> | but couldn't find the page |
| 08:39 | <rniwa> | :\ |
| 08:39 | <rniwa> | smaug____: which OS do you use? |
| 08:39 | <smaug____> | linux |
| 08:39 | <smaug____> | fedora |
| 08:39 | <rniwa> | smaug____: x64? |
| 08:39 | <smaug____> | yup |
| 08:39 | <rniwa> | smaug____: http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Linux_x64/ |
| 08:40 | <rniwa> | smaug____: (warning takes forever to load) |
| 08:40 | <rniwa> | smaug____: http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Linux_x64/139006/ is better :D |
| 08:41 | <smaug____> | (all the browsers are getting so fat nowadays) |
| 08:43 | <rniwa> | smaug____: yeah i know :\ |
| 08:43 | <rniwa> | smaug____: we need to get rid of all the features |
| 08:43 | <zcorpan> | first, get rid of this thing "HTML5" |
| 08:43 | <smaug____> | HTML2 was enough |
| 08:44 | <smaug____> | hmm, it had JS |
| 08:44 | <smaug____> | HTML1 |
| 08:44 | <smaug____> | brb |
| 08:44 | <zcorpan> | responsive images? how about no images?? |
| 08:45 | <annevk> | shit would be fast |
| 08:45 | <rniwa> | zcorpan: THAT SOLVES EVRYTHING :D |
| 08:45 | <rniwa> | zcorpan: welcome to my website: https://rniwa.com/ |
| 08:45 | <zcorpan> | <!DOCTYPE html dir="ltr" lang="en-US"> |
| 08:45 | <rniwa> | (except i still use images for math eqn :( ) |
| 08:46 | <rniwa> | zcorpan: huh, that's weird. |
| 08:46 | <rniwa> | zcorpan: it seems like my DOCTYPE and html element got smashed together :\ |
| 08:46 | <zcorpan> | nice :) |
| 08:48 | rniwa | fixes the stupid bug |
| 08:48 | <rniwa> | fixed. |
| 08:49 | <rniwa> | zcorpan: anyway, no image is the future. |
| 08:49 | <rniwa> | zcorpan: with inline svg and all that jazz, i don't know why we need images. did you also see github's latest ui change to use custom fonts for icons? |
| 08:50 | <rniwa> | zcorpan: https://github.com/blog/1135-the-making-of-octicons ? |
| 09:45 | <david_carlisle> | rniwa: There is an alternative to images for math you know... |
| 09:45 | <Ms2ger> | No way! |
| 09:47 | <jgraham> | Pencil + paper + the postal service? |
| 09:52 | <zcorpan> | use a custom font: one glyph for each formula you want to use |
| 09:54 | <jgraham> | Ah, the github solution |
| 10:01 | <[tm]> | david_carlisle: I removed the openmath stuff from the schema in my workspace but doing that makes your doc invalid |
| 10:10 | <david_carlisle> | That's OK I was waiting to see exactly what went and what you're allowing and then I shall squirrel away the openmath in some plausibly valid way:-) |
| 10:10 | <david_carlisle> | better do that now.... |
| 10:12 | <david_carlisle> | [tm]: have you pushed that out yet? is saying my existing file is valid? |
| 10:13 | <david_carlisle> | [tm]: meant to say http://validator.w3.org/nu says my existing file is valid |
| 10:14 | <david_carlisle> | sigh you did say "in my workspace " which might have been a clue it wasn't the public copy, if I could read:-) |
| 10:24 | <david_carlisle> | [tm]: Can you try now, I just updated mathml.html |
| 10:25 | <smaug____> | no, I don't understand the newline insertion |
| 10:25 | <smaug____> | hsivonen: ping |
| 10:26 | <jgraham> | smaug____: What's the question? |
| 10:26 | <smaug____> | in http://www.whatwg.org/specs/web-apps/current-work/multipage/the-end.html#html-fragment-serialization-algorithm |
| 10:26 | <smaug____> | why we want extra newline |
| 10:26 | <smaug____> | for pre and textare and listing |
| 10:26 | <smaug____> | textarea even |
| 10:26 | <smaug____> | or how should it work |
| 10:28 | <Ms2ger> | So if you parse <pre>\n\n |
| 10:28 | <Ms2ger> | You get a pre element with a text child whose data starts with one \n |
| 10:29 | <Ms2ger> | And in that case, you want to serialize two \ns |
| 10:30 | <smaug____> | if I do document.body.appendChild(document.createElement("textarea")); document.body.firstChild.appendChild(document.createTextNode("\nhello")); document.body.innerHTML |
| 10:30 | <smaug____> | what should be the result? |
| 10:31 | <Ms2ger> | <textarea>\n\nhello</textarea> |
| 10:31 | <Ms2ger> | No? |
| 10:31 | <smaug____> | it should, but that is not what browser seem to do |
| 10:31 | <smaug____> | I mean, per spec it should |
| 10:32 | <Ms2ger> | That doesn't surprise me all that much :) |
| 10:34 | <jgraham> | Anyone got any opinion on whether javascript URIs should be dereferenced in a sync way or async way? |
| 10:35 | <jgraham> | The spec seems to say async like other URIs |
| 10:35 | <jgraham> | and gecko seems to mostly do that |
| 10:35 | <jgraham> | But WebKit seems to make it sync in at least some cases |
| 10:35 | <jgraham> | Sync has the advantage of being less racy |
| 10:36 | <jgraham> | But the disadvantage of being more magic |
| 10:36 | <smaug____> | wasn't javascript: sync in gecko and then changed to async |
| 10:36 | <smaug____> | bz would know that |
| 10:37 | <jgraham> | Maybe? |
| 10:37 | smaug____ | checks the blame |
| 10:37 | <jgraham> | Inconveniently, bz isn't here |
| 10:37 | <smaug____> | it would be quite early for bz |
| 10:38 | <jgraham> | Or quite late :p |
| 10:38 | <jgraham> | In any case he doesn't use #whatwg |
| 10:38 | <smaug____> | jgraham: https://bugzilla.mozilla.org/show_bug.cgi?id=351633 |
| 10:40 | <jgraham> | smaug____: Thanks |
| 10:52 | <AryehGregor> | "The font element must override the color of any text decoration that spans the text of the element to the used value of the element's 'color' property. Note: This applies in all modes." zcorpan, if it applies in all modes, then why is it in a spec entitled "Quirks Mode"? |
| 10:52 | <Ms2ger> | At least Gecko still has it in quirks only |
| 10:54 | <zcorpan> | AryehGregor: it was quirks-only in the spec first but i changed it |
| 10:55 | <zcorpan> | AryehGregor: ideally all requirements in the quirks mode spec are moved to the "main" specs that define the features involved |
| 10:55 | jgraham | hopes that note isn't normative |
| 10:55 | <AryehGregor> | I'm against any legacy HTML feature not aligning exactly with an equivalent CSS feature if we can possibly avoid it, so I vote that it be quirks-only. Does it really cause compat problems if it's not? |
| 10:55 | <AryehGregor> | jgraham, no -- every other thing in the spec says "In quirks mode, . . ." |
| 10:56 | <AryehGregor> | So the note is redundant. |
| 10:56 | <jgraham> | Good :) |
| 10:56 | jgraham | would be very surprised if zcorpan had not got that right |
| 10:56 | <Ms2ger> | I object to it being quirks-only :) |
| 10:59 | <zcorpan> | AryehGregor: it applies in all modes in webkit, which suggests we can get away with it not being quirks-only :-) |
| 10:59 | <AryehGregor> | Ah, I see -- having legacy features not match CSS is bad, but having differences between quirks and standards is worse. |
| 10:59 | <AryehGregor> | Okay, I changed my mind, I'm in favor of it being true in all modes. |
| 11:00 | <zcorpan> | it's a trade-off, of course. depends on how ugly the quirk is |
| 11:00 | <zcorpan> | for instance mozilla want window.foo to be limited to quirks mode |
| 11:01 | <Ms2ger> | Pff |
| 11:01 | <zcorpan> | or some mozilla people |
| 11:01 | <zcorpan> | i know y'all aren't one entity |
| 11:01 | <Ms2ger> | We can want all we want, but we gave up |
| 11:01 | <Ms2ger> | Thank you, Google! |
| 11:07 | <jgraham> | So is the whole "Mozilla are individuals" thing supposed to make me imagine that Mozilla are starfleet and everyone else are the Borg? Should I imagine that Mozilla HQ is like the enterprise with Gary Kovacs as Picard and Brendan Eich as Geordi Laforge? When your senior crew, I mean staff, leave for Facebook do you talk about them being "assimilated"? |
| 11:07 | <Ms2ger> | You know what our meeting rooms are called, right? |
| 11:07 | <jgraham> | And, most importantly in this scenario, who is Wesley Crusher? |
| 11:08 | <jgraham> | Because they deserve to be punched |
| 11:08 | <jgraham> | Ms2ger: I assume given context that they must be star trek based, although iirc the smaller ones are internet memes |
| 11:09 | <Ms2ger> | http://25.media.tumblr.com/tumblr_m4gatutgsA1rrf1eeo1_400.jpg |
| 11:11 | <jgraham> | So basically, I *am* supposed to have this picture |
| 11:11 | <Ms2ger> | No comment |
| 11:11 | jgraham | assumes Ms2ger is a random redshirt |
| 11:15 | <AryehGregor> | Does anyone here have ready access to Word? If so, could you do a quick test for me? |
| 11:15 | <AryehGregor> | 1) Type "foo\nbar\nbaz". 2) Select all and make bold. 3) Delete the three letters "bar", so the middle line is empty. 4) Move the cursor someplace else, then move it back to the empty middle line. 5) Type "quz". 6) Is "quz" bold or not? |
| 11:15 | <AryehGregor> | (also tell me what version) |
| 11:16 | <AryehGregor> | (my wife has Word on her laptop, but I can't figure out how to turn the laptop on . . .) |
| 11:30 | <jgraham> | AryehGregor: Look for a button on the outer casing? |
| 11:30 | <AryehGregor> | jgraham, I see a power button! But pushing it doesn't seem to do anything . . . |
| 11:30 | <AryehGregor> | Thanks for the tip, though! I appreciate it. |
| 11:31 | <Philip`> | Is it plugged in? |
| 11:31 | <AryehGregor> | I mean, my grandfather would reportedly call my father sometimes to get a step-by-step walkthrough on how to turn his computer on. |
| 11:31 | <AryehGregor> | No reason for you to think I'm any different. |
| 11:31 | <Philip`> | Could you ask your wife? |
| 11:31 | <AryehGregor> | I will when she gets back. |
| 11:32 | <AryehGregor> | However, I was hoping that maybe someone here had a copy of Word handy, which might be faster. |
| 11:33 | Philip` | doesn't :-( |
| 11:34 | <Philip`> | (It's strange being in a community where most people don't have access to probably the second most widely installed piece of software in the world) |
| 11:34 | <jtcranmer> | most people I know don't have it installed |
| 11:34 | <asmodai> | What about Word? |
| 11:34 | <asmodai> | I have 2010 installed |
| 11:35 | <AryehGregor> | [120525 14:16:48] <AryehGregor> Does anyone here have ready access to Word? If so, could you do a quick test for me? |
| 11:35 | <AryehGregor> | [120525 14:17:16] <AryehGregor> 1) Type "foo\nbar\nbaz". 2) Select all and make bold. 3) Delete the three letters "bar", so the middle line is empty. 4) Move the cursor someplace else, then move it back to the empty middle line. 5) Type "quz". 6) Is "quz" bold or not? |
| 11:35 | <AryehGregor> | [120525 14:17:33] <AryehGregor> (also tell me what version) |
| 11:35 | <AryehGregor> | asmodai, thanks. :) |
| 11:35 | <asmodai> | yes, bold |
| 11:37 | <asmodai> | on 2010 |
| 11:37 | <asmodai> | I think the begin and end mark for bold are at f and z, respectively, so deleting the middle text doesn't change that |
| 11:40 | <AryehGregor> | Right. |
| 11:43 | <AryehGregor> | asmodai, https://bugs.webkit.org/show_bug.cgi?id=81656#c10 |
| 11:43 | <asmodai> | Awesome |
| 12:19 | <annevk> | Ms2ger: is there a way for Anolis to recognize [VERSION] and such but not complain about references being in a single section? |
| 12:19 | <annevk> | Ms2ger: also, having a flag for First Public Working Draft would be nice |
| 12:19 | <annevk> | although maybe it's not worth optimizing for |
| 12:27 | <gsnedders> | annevk: What's different for FPWD? |
| 12:30 | <annevk> | gsnedders: SotD contains "First Public Working Draft" |
| 12:30 | <annevk> | (yes it is insane) |
| 12:41 | <david_carlisle> | fpwd don't have to have a "previous version" (obvious but true:-) |
| 12:51 | <odinho> | gsnedders: Oh, that's sooo nice. http://ecma-international.org/ecma-262/5.1/ Didn't see it till now. |
| 12:52 | <annevk> | first have to see what they're releasing ES6 in |
| 12:52 | <annevk> | it seems like this was a transformation done by someone, similarly to http://es5.github.com/ |
| 12:52 | <annevk> | not done by the editors of the spec... |
| 12:53 | <gsnedders> | annevk: That's what was said in the email to es-discuss |
| 12:53 | <gsnedders> | The *big* deal here is that ECMA agreed to publish it. |
| 12:54 | <odinho> | It's on their site. |
| 12:54 | <gsnedders> | Still, in the event of any dispute, the PDF version is normative, which makes sense. |
| 12:54 | <odinho> | one small step ... etc |
| 12:54 | <gsnedders> | (Reviewing a the transformation to HTML would be a lot of work to ensure there are no differences) |
| 12:54 | <Ms2ger> | annevk, doesn't --w3c-compat-substitutions get you that? |
| 12:57 | <annevk> | Ms2ger: ah good to know |
| 13:01 | <matjas> | annevk: it’s https://github.com/jorendorff/es-spec-html and it looks like this will be used for ES6 as well |
| 13:01 | <matjas> | i actually configured the ecma-international.org server the other day — they were sending the SVG figure as text/plain which broke things |
| 13:03 | <gsnedders> | The interesting question is whether anyone can make a stylesheet for the HTML copy that meets the publishing format rules, of cover pages and such like. |
| 13:04 | <Ms2ger> | Well, the more interesting question is whether ECMA will ever make those rules available |
| 13:04 | Ms2ger | gets his towel out |
| 13:05 | <gsnedders> | But that'd allow non-members to participate! |
| 13:05 | <gsnedders> | And how are we meant to stay financially viable unless members get something for their money!? |
| 13:06 | <Ms2ger> | In the case of ECMA? |
| 13:06 | <Ms2ger> | I suspect "not" |
| 13:06 | <Philip`> | gsnedders: Make everything free and open, but encourage bribery |
| 13:07 | <gsnedders> | Ms2ger: Yeah, in the case of Ecma. |
| 13:07 | <gsnedders> | The only thing you get, in the case of ECMAScript, is to get to go to F2F meetings. That is literally the only gain of being an ECMA member. |
| 13:07 | <jgraham> | And you can submit tests |
| 13:08 | <gsnedders> | Though Mozilla people have in principle agreed to submit third-party tests. |
| 13:08 | <gsnedders> | Provided they meet the licensing requirements. |
| 13:08 | <gsnedders> | So even that gain is questionable. |
| 14:35 | <odinho> | Who did/knows IDLharness? |
| 14:35 | <jgraham> | AryehGregor |
| 14:36 | <odinho> | Trying to make a interface test for IDB, but it's really not all roses. |
| 14:36 | <odinho> | AryehGregor: // XXX: [17:25:40.864] uncaught exception: Window implements IDBEnvironment, but one is undefined or not an interface |
| 14:36 | <odinho> | Window implements IDBEnvironment; |
| 14:37 | <odinho> | It doesn't like that. And I have an [NoInterfaceObject] interface IDBEnvironment { readonly attribute IDBFactory indexedDB; }; |
| 14:37 | <odinho> | Tried removing [NoInterfaceObject] and other stuff, but I don't really know where to start at all. |
| 14:38 | <AryehGregor> | odinho, I'm about to run now -- could you send an e-mail? I should be able to look at it Monday (UTC+0300). At a glance, I'd guess the problem is that "Window" isn't defined -- you have to include the declarations of any interfaces you depend on too. |
| 14:38 | <AryehGregor> | You can include them in a fashion that they don't get tested themselves. |
| 14:39 | <odinho> | AryehGregor: Ah, I can look at that. It's a way I haven't explored, it might fix it, yea :] |
| 14:39 | <odinho> | AryehGregor: Thanks. |
| 14:40 | <AryehGregor> | odinho, look here for inspiration: http://dvcs.w3.org/hg/html/file/0bc4307c200a/tests/submission/AryehGregor/interfaces.html Note use of "add_untested_idls" -- this is to avoid double-testing the DOM IDLs, which the HTML IDLs depend on. |
| 14:45 | <odinho> | AryehGregor: Wohoo! :D |
| 16:30 | <TabAtkins> | zcorpan: Yeah, responsive tables of that variety have been seriously discussed before, but didn't have sufficient momentum to get one of us to work on them. |
| 16:30 | <TabAtkins> | zcorpan: But the idea of pinning rows/columns is very cool and obviously useful. Mobile presentation of tables just drops the threshold for how large of a table it's useful on. |
| 16:36 | <Hixie> | zcorpan: patches welcome. :-) |
| 16:37 | <tantek> | welcome back Hixie |
| 16:37 | <Hixie> | still on vacation til tuesday |
| 16:37 | <Hixie> | but back in the area and on the net |
| 16:37 | <tantek> | good for you for taking vacation |
| 16:37 | <Ms2ger> | And bad for us :) |
| 16:55 | <Ms2ger> | Anyone around here who cares about <script type>? |
| 16:59 | <TabAtkins> | "cares" in what sense? |
| 16:59 | <Ms2ger> | About the supported values |
| 17:00 | <TabAtkins> | Nope. |
| 17:13 | <Philip`> | Is there any good reason why someone would send dozens of packets per second to a server's HTTPS port, for many hours, when that server doesn't have anything responding to that port? |
| 18:38 | <TabAtkins> | Ms2ger: D'oh, I had a good snarky comment for that bug. Damn your quick AI reflexes! |
| 18:38 | <Ms2ger> | :) |
| 18:42 | <Hixie> | which bug? :-) |
| 18:42 | <TabAtkins> | A webapps bug. |
| 18:42 | <TabAtkins> | Entitled: "Looking for the music I downloaded" or similar. |
| 18:42 | <Hixie> | aah |
| 18:42 | <annevk> | Ms2ger: is bz saying Gecko wants to implement the spec or something else? |
| 18:42 | <Ms2ger> | The spec |
| 18:43 | <annevk> | k |
| 18:43 | <Ms2ger> | But since nobody else does, we'd like to hear if anybody objects |
| 19:06 | <Ms2ger> | Hmm, wasn't [ArrayClass] interface NodeList not going to fly? |
| 19:06 | <annevk> | I believe WebKit was trying |
| 19:07 | <Ms2ger> | Also |
| 19:07 | <Ms2ger> | http://www.whatwg.org/specs/web-apps/current-work/multipage/common-dom-interfaces.html#htmlpropertiescollection-0 |
| 19:07 | <annevk> | and I believe they did run into issues, but not so much that people were admitting defeat... |
| 19:08 | <annevk> | Ms2ger: ? |
| 19:08 | <Ms2ger> | Would you consider the "represents" in the first paragraph there as overriding http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#represented-by-the-collection |
| 19:08 | <Ms2ger> | ? |
| 19:09 | <annevk> | Ms2ger: no, but it's confusing |
| 19:09 | <annevk> | Ms2ger: represents in DOM applies to collection; in HTML it applies to interface... |
| 19:09 | <Ms2ger> | Yeah, I thought so |
| 19:09 | <Ms2ger> | It confused bz as well |
| 19:27 | <Ms2ger> | ojan_away, s/you're feeling/your feeling/ :) |
| 19:27 | <annevk> | btw |
| 19:27 | <annevk> | did bz test |
| 19:27 | <annevk> | language=javascript1.0 |
| 19:28 | <Ms2ger> | We're not touching language right now |
| 19:28 | <Ms2ger> | Does anybody forward language to type like the spec claims? |
| 19:28 | <annevk> | the whole point of the list in HTML is that it's the same for type and language |
| 19:28 | <annevk> | I think we might |
| 19:28 | <annevk> | not sure though |
| 19:30 | <Ms2ger> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1541 suggests Opera doesn't |
| 21:05 | <zewt> | nothing speaks of competence in web design as much as phone number entry boxes that swallow keystrokes, breaking browser hotkeys |
| 21:08 | <Hixie> | am i missing something here? UTF-8 byte 0xED followed by two continuation bytes 0x80 0x80 is equivalent to U+D000 right? |
| 21:10 | <gsnedders> | That's non-shortest form, no? |
| 21:10 | <Hixie> | no? |
| 21:11 | <gsnedders> | Yeah, that is U+D000, except 0xED 0x80 0x80 is not a valid UTF-8 string. |
| 21:11 | <Hixie> | how can it be anything but shortest form if all the continuation bytes are 0x80 and the first byte has its second nibble set to anything but 0 |
| 21:11 | <gsnedders> | What, that's wrong. |
| 21:11 | <gsnedders> | Ignore me, ignore me! |
| 21:11 | <Hixie> | ok |
| 21:12 | <Hixie> | phew |
| 21:12 | <gsnedders> | I'm not even going to try and answer that. |
| 21:12 | <TabAtkins> | Yes, it's U+d000 |
| 21:12 | <Hixie> | i'm confused enough already as it is :-P |
| 21:12 | <gsnedders> | I seem to just be speaking nonsense so answering a harder question will likely result in more of that. |
| 21:14 | <TabAtkins> | Hixie: Why do you ask? |
| 21:21 | <jgraham> | zewt: Half the fucking web thinks it's OK for "/" to be used for "focus site search box" |
| 21:28 | <zewt> | jgraham: this is worse--probably cancelling onkeydown if it's not 0-9 |
| 21:28 | <zewt> | which means every browser hotkey stops working (why browsers allow that I can't fathom) |
| 21:29 | <zewt> | not to mention pasting, which is ... sort of basic for an "enter bank phone number" field |
| 21:37 | <jamesr_> | man i really wish you could declaratively say which keys you care about when registering a JS key listener |
| 21:38 | <jamesr_> | so the browser could avoid having to bounce through the page and hope that it didn't fuck up the preventDefault() |
| 21:38 | <TabAtkins> | a;lksdjf;al WANT IT SO BAD |
| 21:38 | <jamesr_> | window.addEventListener("keydown", flobberize(), false, ['/']); ? |
| 21:38 | <jamesr_> | s/flobberize()/flobberize/ |
| 21:38 | <TabAtkins> | Hey, you don't know. Maybe flobberize() returns a function. |
| 21:41 | <jamesr_> | i want it for all event types |
| 21:41 | <jamesr_> | like mouse listeners should say that they only care about right click or whatever |
| 21:41 | <zewt> | jamesr_: well, it doesn't really fit the dom event model, but could have element.addKeyListener(func, keySpecifier, capture) that's functionally equivalent to an event listener with a key check (except easier for the browser to optimize) |
| 21:42 | <jamesr_> | yeah, the model is you fire everything and the handler figures out if it cares or not |
| 21:42 | <jamesr_> | which means if we want to optimize it in the browser we have to do it on an event type level |
| 21:42 | <zewt> | in a sense make it so you can attach a sub-event-level filter to each event handler |
| 21:42 | <jamesr_> | i.e. define a new set of events for each interaction |
| 21:42 | <jamesr_> | even applying an after-the-fact filter might work |
| 21:43 | <zewt> | elem.addEventListener("keydown", func, false, {filter: {keyCode: someKindOfFilterLanguage}}) |
| 21:44 | <zewt> | personally i think it's inefficient but mostly only academically, so I'm not too bothered by it |
| 21:45 | <jamesr_> | it's not quite theoretical for things like pagedown |
| 21:45 | <jamesr_> | we would really like to scroll the page when you hit page down without having to wait for the thread running JS/DOM/etc |
| 21:45 | <jamesr_> | but if the page listens to key events, we dunno if it's going to preventDefault() the pagedown |
| 21:46 | <jamesr_> | if the page really just wants to hijack '/' - merits of doing that aside - it'd help us out to know that in the browser |
| 21:48 | <rafaelw_> | Ms2ger: question for you about your Document.parse() WebKit patch comments.... |
| 21:48 | <jwalden> | rafaelw_: gone |
| 21:48 | <jwalden> | (nn, to be precise) |
| 21:48 | <rafaelw_> | ah. thanks. |
| 21:49 | <jamesr_> | it's probably just a pipe dream though :( |
| 21:49 | <TabAtkins> | I have no idea why you think it's a pipe dream. |
| 21:49 | <TabAtkins> | People have watned it for years, just no one's cared enough to make it happen. |
| 21:49 | <TabAtkins> | MAKE IT HAPPEN. |
| 21:49 | <zewt> | jamesr_: if you can think of a static filter language that would work... |
| 21:49 | <TabAtkins> | Easy. |
| 21:49 | <zewt> | i mean, i know of some, like mongodb's queries, but it's not the cleanest |
| 21:50 | <jamesr_> | zewt, that's one tricky part of the problem |
| 21:50 | <jamesr_> | list of keycodes? |
| 21:50 | <zewt> | eg. {$or: [{keyCode: 1}, {keyCode: 2}]} |
| 21:50 | <zewt> | i mean, to be generic for any event |
| 21:50 | <jamesr_> | it gets messy quickly |
| 21:51 | <jamesr_> | and someone always goes "can i just pass a function?" then you have to explain how it picks up the global context and the thread implications, then someone wants a JS isolate, then you're in a world of hurt |
| 21:51 | <zewt> | well, mongo's query language is static |
| 21:51 | <zewt> | the above can be optimized (and even if it's not, it's a lot faster to convert it to a native object and just test them all, since you don't have to fire off JS) |
| 21:52 | <TabAtkins> | Yes, take some object with values. Eagerly parse them to dispose of getters. Done. |
| 21:52 | <zewt> | (idl dictionary) |
| 22:16 | <Hixie> | TabAtkins: my utf-8 validating code isn't vorking |
| 22:16 | <TabAtkins> | Hixie: Interesting. Should be trivial! |
| 22:16 | <Hixie> | specifically, it thinks 0xED 0x9F 0xBF is invalid, because it decodes it to a bogus value |
| 22:17 | <Hixie> | my code for decoding sequences starting with 0xED is equivalent to 0xD000 + ((c1-0x80) << 6) + (c2-0x80) |
| 22:17 | <Hixie> | am i missing something? |
| 22:19 | <TabAtkins> | Assuming that c1 and c2 are guaranteed to be in [00,ff], and you're already making sure that the second bit in each is 0, that should be fine. |
| 22:19 | <TabAtkins> | Better would be to & each with 0x00111111. |
| 22:19 | <Hixie> | each one is known to be in the range 0x80 .. 0xBF |
| 22:20 | <TabAtkins> | Just to be safe, try swapping your "-0x80" with "&63". |
| 22:23 | <Hixie> | that (thankfully) made no difference |
| 22:24 | <TabAtkins> | Okay, good. Just sanity-checking. |
| 22:24 | <TabAtkins> | What's the bogus value it's decoding to? |
| 22:24 | <Hixie> | 0xDB5F |
| 22:25 | <Hixie> | oh i see the problem |
| 22:25 | <Hixie> | off by one error |
| 22:25 | <Hixie> | don't mind me |
| 22:25 | <TabAtkins> | uh, okay. |
| 22:26 | <Hixie> | c2 contained c1 and c1 contained c0 |
| 22:26 | <Hixie> | the leading byte |
| 22:27 | <Hixie> | that resulted in disturbingly few of my tests failing |
| 22:27 | <Hixie> | but that's another story |
| 22:27 | <Hixie> | ok, bbiab |
| 22:29 | <TabAtkins> | Oh, duh, I should have noticed that as soon as I looked at it. |
| 22:33 | <rniwa> | hsivonen: yt? |
| 22:34 | <rniwa> | Hixie: do you know how well tested the part it says to throw SYNAX_ERR in http://html5.org/specs/dom-parsing.html#extensions-to-the-range-interface ? |
| 22:34 | <rniwa> | when the element name is html, body, etc...? |
| 22:53 | <rafaelw_> | Hixie; Where are you WRT Document.parse()? |
| 22:54 | <rafaelw_> | I'd really like to get this in. I'm wondering if you still have strong objections or not. |
| 23:00 | <espadrine> | Firefox cannot redefine Element.prototype.scrollIntoView, while webkit can. Firefox's wrong, right? |
| 23:22 | <jwalden> | espadrine: wrong assuming you consider webidl authoritative now, yes |
| 23:22 | <jwalden> | espadrine: DOM bindings are being rewritten now and should end up fixing that |
| 23:22 | <espadrine> | jwalden: oh ok, thanks! |
