| 00:58 | <boblet> | hey all, I’m under the impression that the hr element shouldn’t be used to indicate a break between sections, but the description and paragraph content model explanation don’t really say this explicitly |
| 00:59 | <boblet> | am I mistaken? |
| 01:01 | <boblet> | will check logs for reply… TIA |
| 01:51 | <werdnativ> | Is HTML5 practically set in stone now, or is there still some possibility to consider new elements? |
| 01:55 | <erlehmann> | werdnativ, there is a process for that. http://wiki.whatwg.org/wiki/Feature_Proposals#Guidelines |
| 01:56 | <werdnativ> | cool, thanks. so it's still open for consideration? |
| 02:20 | <wolf2k_ubuntu> | Evening. So far my XHTML5 SVG web app is coming along. I've run into an issue, however. Due to the nature of the app, I have to be able to insert SVG elements dynamically. Without some sort of z-index, all new elements are overlayed on top: that's not what I'm after. Based on my own testing, I don't think SVG elements respect HTML's z-index CSS attribute. Is there an alternative? |
| 02:24 | <wolf2k_ubuntu> | ...nevermind. I got the answer. Painters model. |
| 02:52 | <othermaciej> | wolf2k_ubuntu: I am surprised that SVG elements don't respect z-index - I believe they should |
| 02:53 | <wolf2k_ubuntu> | othermaciej: Try talking to them for version 1.2, whenever that comes |
| 02:53 | <wolf2k_ubuntu> | Either way, you've got a new FAQ now. Even if it's not so F |
| 02:54 | <othermaciej> | by "should" I mean I expect that to work now - do you have a test case to show that it doesn't work? |
| 02:54 | <wolf2k_ubuntu> | othermaciej: Unless you plan on visiting my house and seeing my development area this instant, don't think I can easily show it |
| 02:54 | <wolf2k_ubuntu> | It's a combination of XHTML, SVG, and JS |
| 02:56 | <wolf2k_ubuntu> | This app will be adding and removing elements dynamically. Now we've hit the point where I basically have to keep track of the nodes. |
| 02:56 | <roc> | according to the spec, z-index should not work on SVG elements (except possibly a positioned <svg> element with non-SVG parent) |
| 02:56 | <roc> | everyone agrees we need to change SVG to support some form of z-index |
| 02:57 | <wolf2k_ubuntu> | I can work around it. jQuery will be a big help. |
| 02:57 | <roc> | jwatt and I have argued pretty hard that we should support CSS z-index directly, with stacking contexts etc |
| 02:57 | <wolf2k_ubuntu> | But if you need someone else to sign your petition, you've got me |
| 02:58 | <wolf2k_ubuntu> | I had a ball of a time getting a mouseover/mousemove shadow rectangle to display over the SVG due to lack of z-index |
| 02:58 | <roc> | other people have argued that CSS z-index stacking contexts are not necessary for SVG |
| 02:58 | <wolf2k_ubuntu> | I personally think it should be available. |
| 03:01 | <othermaciej> | it looks like WebKit ignores z-index on SVG elements |
| 03:01 | <othermaciej> | seems bogus to me |
| 03:01 | <wolf2k_ubuntu> | othermaciej: Firefox does it too |
| 03:01 | <othermaciej> | roc: you could just say that nothing besides the SVG root element creates a stacking context |
| 03:02 | <othermaciej> | roc: that's assuming CSS position does not apply to SVG elements |
| 03:02 | <roc> | Then SVG elements with z-index not 'auto' would be positioned but not create a stacking context for their descendants |
| 03:02 | <roc> | which would be weird |
| 03:02 | <roc> | I mean positioned in z-order |
| 03:03 | <roc> | Webkit and Firefox are just following the current spec |
| 03:03 | <othermaciej> | is there any way to escape your ancestor's stacking context in CSS? |
| 03:03 | <roc> | one reason to want stacking contexts for SVG is that you want 'opacity' and 'filter' and other effects to induce a stacking context |
| 03:03 | <roc> | no |
| 03:04 | <roc> | otherwise you get weirdness like a blur filter applied to an element subtree, but the element's children are not all part of the same rendering group because there's another non-blurred element in between |
| 03:05 | <roc> | and then it becomes really unclear what the rendered results should be, or how to produce them |
| 03:06 | <othermaciej> | I see |
| 03:06 | <othermaciej> | I think most SVG elements either have their own rendering, or contain other elements that may have rendering, but not both |
| 03:08 | <roc> | that's true, but doesn't help us here, I think |
| 03:45 | <othermaciej> | roc: the reason I mention it is that this may seem like a reason to not support stacking contexts, although you could argue the opposite on the same basis I think |
| 03:46 | <roc> | it may *seem* like a reason but I don't think it actually is :-) |
| 04:10 | <JonathanNeal> | I made an HTML5 Pangram http://sandbox.thewikies.com/html5-pangram/ |
| 04:23 | <wolf2k_ubuntu> | JonathanNeal: I'm guessing pangrams are supposed to show off everything about a protocol? |
| 04:23 | <JonathanNeal> | wolf2k_ubuntu, yes |
| 04:24 | <wolf2k_ubuntu> | ...maybe it was just me, but yours felt a little disorganized |
| 04:24 | <JonathanNeal> | With as few extra properties on elements as possible to validate. |
| 04:26 | <wolf2k_ubuntu> | ...you know, it feel funny. The closer I get done with this HTML5 SVG app, the more I'm thinking that it's a good thing I didn't try to use GWT or whatever Google calls their Java -> Javascript thing. |
| 07:22 | <JonathanNeal> | http://sandbox.thewikies.com/html5-pangram/ I noticed that http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html doesn't have any rendering for video / audio / canvas |
| 07:27 | <hsivonen> | does anyone have data on what maximum values different browser engines place on attribute value length? |
| 07:29 | <othermaciej> | it looks like WebKit will accept arbitrarily long attribute names (until you run out of RAM or characters in your stream) |
| 07:29 | <othermaciej> | checking values now |
| 07:30 | <othermaciej> | wait, I'm wrong |
| 07:30 | <othermaciej> | 1024 is the limit for attribute names |
| 07:30 | <othermaciej> | now checking vlues |
| 07:31 | <othermaciej> | attribute values can be as long as you can fit in the source file |
| 07:31 | <othermaciej> | at least as far as parsing goes |
| 07:31 | <hsivonen> | ouch |
| 07:32 | <othermaciej> | if added via the DOM, either name or value can be as long as you can fit in a String |
| 07:32 | <hsivonen> | so you can OOM with <a a="aaaaaa..... |
| 07:32 | <othermaciej> | it won't crash, but it can use up a lot of memory before an allocation fails |
| 07:32 | <othermaciej> | that's assuming your source was long enough for that in the first place |
| 07:33 | <hsivonen> | well, the HTML5 specs is probably larger than any page Schwab's banking UI |
| 07:33 | <othermaciej> | and did not itself get too big to fit in RAM as UTF-16 |
| 07:33 | <hsivonen> | so if I made the DoS limit the size of the HTML5 spec, banking sites probably wouldn't break... |
| 07:34 | <hsivonen> | Gecko is getting an infallible malloc, so it might be a problem to have the parser thread request so much memory that infallible malloc has to scramble for more memory on the main thread |
| 07:35 | <hsivonen> | not sure how exactly these things are supposed to work |
| 07:35 | <othermaciej> | in fallible? |
| 07:35 | <hsivonen> | never returns null |
| 07:35 | <othermaciej> | what happens if you exhaust the address space? |
| 07:35 | <Hixie> | does it buy more ram for you? |
| 07:35 | <hsivonen> | I believe it kills the app |
| 07:36 | <othermaciej> | we have a version of malloc that will crash if it runs out of memory but we have a few places that are safe against allocation failure |
| 07:36 | <othermaciej> | and thus do not use that variant |
| 07:36 | <hsivonen> | though IIRC, some design doc says that it's going to ask things to stop when there's little RAM left |
| 07:36 | <othermaciej> | (theory being that clean crash beats a buffer overrun) |
| 07:37 | <hsivonen> | I wonder if IE has a hard limit on attribute value length. |
| 07:37 | <hsivonen> | maybe I should just bite the bullet and test |
| 07:37 | <othermaciej> | we did find that for attribute *names*, 1024 is a sufficient practical limit |
| 07:37 | <othermaciej> | not sure it would fly for values |
| 07:38 | <Philip`> | Does it say in the IE interoperability reports that were released recently? |
| 07:38 | Philip` | vaguely remembers them mentioning limits, compared to the limits HTML4 mentions, or something |
| 07:38 | <othermaciej> | tag names also have a 1024 limit |
| 07:38 | <othermaciej> | (in the HTML parser only) |
| 07:39 | <hsivonen> | my plan is to change the tokenizer to use the same buffer for both names and values |
| 07:39 | <Philip`> | I've seen pages with tens of kilobytes of text in an <img alt> attribute |
| 07:39 | <hsivonen> | because you can't be tokenizing an attribute name and value at the same time |
| 07:39 | <Hixie> | meta content (with name=keywords) also often has huge values |
| 07:39 | <Hixie> | sometimes it is missing quotes and you end up with bazillions of attributes |
| 07:39 | <hsivonen> | the only case where the tokenizer needs two accumulation buffers is when a character reference is being tokenized inside an attribute value |
| 07:39 | <othermaciej> | we use a variable-size buffer with a small inline capacity for attribute values |
| 07:40 | <othermaciej> | i.e. in the common case it doesn't malloc, just uses a fixed-size buffer, but will malloc on overflow |
| 07:40 | <hsivonen> | but in that case, the other buffer can't grow longer that the longest named character name |
| 07:40 | <hsivonen> | Hixie: truncating keywords isn't fatal |
| 07:41 | <hsivonen> | Hixie: truncating huge attributes on Schwab's netbank apparently is |
| 07:41 | <Hixie> | <input type=hidden value="">? |
| 07:42 | <hsivonen> | Hixie: not sure. I don't have login credentials. |
| 07:43 | <othermaciej> | Hixie: you're the only one who writes Change Proposal summaries that I can use as-is on the status page |
| 07:44 | <Hixie> | i do my best :-) |
| 07:45 | <hsivonen> | I think http://www.w3.org/Bugs/Public/show_bug.cgi?id=7670#c43 is a mockery of RDFa in HTML being a deliverable of the HTML WG |
| 07:45 | <hsivonen> | why bother pretending that it's contents aren't delivered by the RDFa WG? |
| 07:46 | <hsivonen> | s/it's/its/ |
| 07:47 | <hsivonen> | scary. I do the "it's" thing now. |
| 08:02 | <hsivonen> | click and hold Dock icon for Exposé has to be the worst UI feature of Snow Leopard |
| 08:02 | <hsivonen> | I have never wanted to active it, but I accidentally activate it on a daily basis |
| 08:07 | <othermaciej> | hsivonen: I don't think accidental click-and-hold is a common user behavior |
| 08:19 | <jgraham> | I have done that a few times too |
| 08:20 | <jgraham> | Although plural of anecdotes and all that |
| 08:20 | <hsivonen> | othermaciej: I use a Wacom tablet--not a mouse |
| 08:20 | <othermaciej> | does that make it more likely? |
| 08:20 | <hsivonen> | dunno |
| 08:21 | Philip` | sometimes half-clicks on something before he's quite decided whether he meant to click on it or not, and then holds the mouse down while he's thinking |
| 08:27 | <asmodai> | http://www.smashingmagazine.com/2010/03/15/showcase-of-web-design-in-china-from-imitation-to-innovation-and-user-centered-design/ <-- I knew before how click intense Chinese websites are, but this really confirms it. |
| 08:29 | <othermaciej> | what parts of the article show the click-heaviness? |
| 08:29 | othermaciej | is just skimming |
| 08:31 | <annevk> | bah, got disconnected |
| 08:40 | annevk | wonders in what situation you would use hidden= on <td> |
| 08:44 | <lazni> | empty cell? |
| 08:45 | <Philip`> | It'd be more useful to support <col hidden> |
| 08:46 | <annevk> | again, would that usage match the semantics of hidden? |
| 08:50 | <Philip`> | <table><col><col id=rc hidden><tr><th>Test case input<th>Result<tr><td>1+1<td id=r0></table> <script>r0.innerHTML = '11'; rc.hidden = false;</script> |
| 08:51 | <zcorpan> | hmm, regarding allow-top-navigation, don't ads usually open links as target=_blank anyway? |
| 08:51 | <Philip`> | (Don't want to make the results visible until they're all filled in) |
| 08:51 | <zcorpan> | Philip`: does visibility:collapse work on colgroup and col? |
| 08:52 | <Philip`> | zcorpan: Don't ask me |
| 08:52 | <Philip`> | If it doesn't, HTML shouldn't be restricted by current bugs in CSS anyway |
| 08:53 | <zcorpan> | it seems to work |
| 09:04 | <annevk> | any chance btw we can get a multipage version of complete.html? |
| 09:04 | <annevk> | would be especially useful for WebSocket |
| 09:04 | <Hixie> | just get a better browser :-P |
| 09:05 | <Hixie> | seriously though, complete.html works fine in chrome on mac |
| 09:05 | <zcorpan> | still annoying for page search |
| 09:05 | <annevk> | on Ubuntu too, but loading a smaller file is still nice |
| 09:06 | <JonathanNeal> | I tried to translate the x h1 example from http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#margins-and-padding into actual css |
| 09:07 | <zcorpan> | maybe we should ditch the "html5" version and go back to only web apps 1.0 |
| 09:07 | <JonathanNeal> | in LESS it is 20 lines, in actual css it is 1383 lines. |
| 09:08 | <roc> | the comment about "better browser" is slightly annoying given http://weblogs.mozillazine.org/bz/archives/020267.html |
| 09:10 | <JonathanNeal> | http://pastie.org/884348 written in LESS css, http://pastie.org/884350 written in normal css |
| 09:12 | <zcorpan> | JonathanNeal: it seems that's not correct |
| 09:13 | <JonathanNeal> | zcorpan, like I'm wrong, or it's just suprising? |
| 09:13 | <zcorpan> | JonathanNeal: you want article article article article article h1, not article h1 article h1 article h1 article h1 article h1, |
| 09:14 | <JonathanNeal> | Sure, I'll update the examples, they'll still be about 70 fold different. |
| 09:15 | zcorpan | doesn't follow |
| 09:16 | <zcorpan> | JonathanNeal: the LESS version should probably be something more like article, aside, etc { h1 { ... } article, aside, etc { h1 { ... if i understand the LESS syntax correctly |
| 09:18 | <JonathanNeal> | zcorpan, here you go http://pastie.org/884363 and http://pastie.org/884364 |
| 09:19 | <zcorpan> | JonathanNeal: yep, that looks better |
| 09:20 | <JonathanNeal> | So, I'll have to use some kind of class on all sectioning content. |
| 09:23 | <annevk> | roc, maybe we should all start cheating |
| 09:24 | <roc> | 'cheating' is a stronger word than I would use |
| 09:24 | <roc> | browsers make different performance vs correctness tradeoffs :-) |
| 09:26 | <zcorpan> | i really don't understand why anyone would want to use <img role="presentation" alt="" src="img" > or <img role="presentation" src="img" > instead of <img alt="" src="img" > |
| 09:27 | <lazni> | http://3273372964/en/weblog?weblogid=208188044 |
| 09:27 | <Philip`> | annevk: Feel free to update http://code.google.com/p/html5/source/browse/trunk/spec-splitter/spec-splitter.py to work on it :-) |
| 09:27 | <Philip`> | (Maybe the only difference is it should have a different list of sections to split?) |
| 09:28 | <roc> | I shouldn't say this, but "figuring out how to fix selectors without regressing performance" reminds me of OJ searching for the real killers |
| 09:28 | <Philip`> | and then either run it yourself, or ask Hixie to change his server stuff (and me to change mine a bit) |
| 09:28 | <annevk> | Philip`, I suppose, how do you deal with W3C HTML5 vs WHATWG HTML5? |
| 09:30 | <Philip`> | annevk: I don't |
| 09:30 | <Philip`> | except to the extent of adding a --w3c option that twiddles a few things |
| 09:30 | <Philip`> | MikeSmith set it up to run on the W3C copy, I believe |
| 09:31 | <annevk> | I wonder if I can just add the new section names without having the HTML5 copy left broken |
| 09:33 | <annevk> | oh, it already splits on all <h2> elements |
| 09:33 | <annevk> | that sounds perfect |
| 09:33 | <Philip`> | The split_exceptions list is just a set of (approximately) childs of <body> that it'll split on, when the <h2>s are too big |
| 09:33 | <Philip`> | and it'll ignore anything that's in that set and doesn't exist in the document |
| 09:34 | <annevk> | oh, maybe not |
| 09:34 | <annevk> | cool, so we could add sections there for eventsource, websocket |
| 09:36 | <Philip`> | That should work |
| 09:37 | Philip` | tried to choose splits so each page was 100-200KB |
| 09:37 | <Philip`> | Oh, the source code already says that |
| 10:05 | <Philip`> | Does anyone want a free 1-year domain registration at gandi.net? |
| 10:05 | Philip` | has a couple of promo codes that expire at the end of the month |
| 10:09 | <hsivonen> | Philip`: for new domains only, I assume? |
| 10:09 | <Philip`> | hsivonen: Yes |
| 10:09 | <Philip`> | (https://ten.gandi.net/faq) |
| 10:12 | <hsivonen> | Philip`: then, I guess I don't have a use for it |
| 10:13 | <annevk> | registering at a new host is not worth the hassle either for something that costs EUR 5-10 |
| 10:15 | <Dashiva> | "Currently <p /> is forbidden - validator.nu doesn't even permit it in polyglot mode, it seems. Yet, it has been permitted for 10 years. Or more: It is even permitted in HTML4 documents." |
| 10:15 | <Dashiva> | What black magic is this? |
| 10:16 | <Philip`> | Isn't that just <p>> ? |
| 10:17 | <Dashiva> | That's what I thought |
| 10:17 | <hsivonen> | Dashiva: where's the quete from. It's wrong. |
| 10:17 | <Dashiva> | http://lists.w3.org/Archives/Public/public-html/2010Mar/0601.html |
| 10:17 | <Dashiva> | Leif seems to claim <p /> is identical to <p></p> |
| 10:17 | <Philip`> | validator.w3.org even gives a warning ("NET-enabling start-tag requires SHORTTAG YES") |
| 10:18 | <hsivonen> | Dashiva: incorrect |
| 10:18 | <hsivonen> | I need to remove the "polyglot" mode |
| 10:18 | <hsivonen> | since it hasn't gone anywhere after the first features |
| 10:18 | <Philip`> | validator.w3.org "show outline" of <h1><p/>foo shows how it results in a literal ">" |
| 10:19 | <hsivonen> | I don't know who dropped the ball. Me regarding repo access? Sam regarding impl? tantek and zeldman regarding requirements? |
| 10:19 | <jgraham> | hsivonen: Did you move to hg yet? |
| 10:20 | <jgraham> | If so it seems like repo access shouldn't be a big issue |
| 10:21 | <hsivonen> | jgraham: the parser is in hg, but Sam doesn't have access, because the central repo is under Mozilla committer rules |
| 10:21 | <hsivonen> | the rest is still in svn and Sam has access |
| 10:21 | <hsivonen> | s/access/write access/ |
| 10:33 | <othermaciej> | hsivonen: you could ask Sam if he has any intent to continue the work |
| 10:33 | <annevk> | http://www.w3.org/QA/2010/03/update_on_html_5_document_lice.html -- hmm, it almost sounds like they want to invent yet another license |
| 10:34 | <othermaciej> | annevk: but see Ian Jacobs' comment at the end |
| 10:36 | <annevk> | I wonder how that will work out |
| 10:36 | <annevk> | It's annoying to have see the confidential information as well... |
| 10:41 | <othermaciej> | you mean the AC poll? or is there other confidential informaiton? |
| 10:44 | <annevk> | yeah |
| 10:44 | <annevk> | there's some presentation as well |
| 10:44 | <othermaciej> | I was surprised at Opera's vote, but then again, I was also surprised at Apple's vote |
| 10:45 | <annevk> | tell me about it |
| 10:45 | <Philip`> | "Having the specification available in two places is suboptimal, and we'll see whether we can improve the situation." - so they want to restrict the licensing of the W3C copy of the spec, and also get rid of the freely licensed copy? |
| 10:45 | <annevk> | oh, the slides are public |
| 10:46 | <annevk> | Philip`, the W3C copy is already restricted |
| 10:47 | <annevk> | a colleague of mine argued the WHATWG license was not adequate btw, but I've no idea whether that is actually true |
| 10:47 | <jgraham> | I note that "a license that is familiar to the free software community" would include, for example "all rights reserved" |
| 10:47 | <annevk> | though it would have been nicer if it just said MIT |
| 10:47 | <annevk> | jgraham, :p |
| 10:49 | <Philip`> | s/to restrict/to continue restricting/ |
| 10:50 | <jgraham> | annevk: Well it is kind of important; you can't have a license that is actually a free software license and prevent forking. But "familiar to" doesn't mean "compatible with" |
| 10:50 | <annevk> | jgraham, if it's important you should add a comment |
| 10:51 | annevk | added a comment asking about non-HTML5 specifications |
| 10:51 | annevk | wonders if he should publish CORS/CSSOM/etc. under some different venue with another license |
| 10:51 | <othermaciej> | or XHR for that matter |
| 10:51 | <annevk> | do I just give up my copyright if I commit to dev.w3.org? |
| 10:52 | annevk | wonders how that works |
| 10:53 | <Dashiva> | No |
| 10:54 | Philip` | has a vague recollection that you remain the copyright owner of anything you produce (except work-for-hire etc) unless you explicitly sign an agreement to transfer ownership |
| 10:55 | <annevk> | cool, I haven't done that |
| 10:55 | <annevk> | how does it work if HTML WG agreed upon text gets included in the WHATWG copy? |
| 10:56 | <annevk> | do change proposal people give away their copyright? |
| 10:57 | <Philip`> | I think it works on the basis that the HTML WG only agrees on text that is low quality and needs to be rewritten by Hixie anyway |
| 10:57 | <Dashiva> | Indeed |
| 10:57 | <Philip`> | and also on the basis that nobody cares that much about the precise correct application of copyright laws |
| 10:58 | <othermaciej> | I think W3C participation agreement might include something about copyright |
| 10:59 | <annevk> | yeah, was afraid of that |
| 10:59 | <Dashiva> | What if a non-wg member posts to public-html? |
| 11:00 | <othermaciej> | I don't know if it implies an exclusive copyright assignment or non-exclusive |
| 11:01 | <othermaciej> | http://www.w3.org/2009/12/Member-Agreement |
| 11:01 | <Dashiva> | The slides from yesterday said that contributors have the right to publish their own contributions, but not others', which seems to imply non-exclusive |
| 11:01 | <othermaciej> | (not sure if that is the latest version and not sure whether it also applies to standards documents |
| 11:01 | <workmad3> | it's probably something along the lines of contributors grant a non-exclusive, perpetual licence to w3c/whatwg |
| 11:01 | <Philip`> | http://www.w3.org/Consortium/Legal/2007/06-invited-expert.html - "The Invited Expert hereby grants to the W3C a perpetual, nonexclusive, royalty-free, world-wide right and license under any Invited Expert's copyrights on his or her contributions, to copy, publish and distribute the contribution under the W3C document licenses." |
| 11:01 | <othermaciej> | so you could publish under Opera's copyright |
| 11:02 | <othermaciej> | per the member agreement |
| 11:02 | <Philip`> | "The Invited Expert agrees to refrain from creating derivative works that include the Invited Expert's contributions when those derivative works are likely to cause confusion about the status of the W3C work or create risks of non-interoperability with a W3C Recommendation. «Branching» is one example of a non-permissible derivative work." |
| 11:02 | <othermaciej> | that doesn't seem to be in the Member agreement |
| 11:03 | <annevk> | I wonder if "c. The Member shall have access to software and documentation produced by the Consortium prior to release of such software and documentation to non-members, as set forth in Appendix 1." is still true with all the public drafts |
| 11:04 | <othermaciej> | I wonder if it is in the Process Document |
| 11:05 | <othermaciej> | the member agreement I linked is marked *** DRAFT *** but also linked as the latest |
| 11:05 | <Philip`> | Is something like http://philip.html5.org/tests/canvas/suite/tests/spec.html okay? It's a derivative work (of the WHATWG spec) that contains some of my contributions that were posted to public-html and were integrated into the W3C spec |
| 11:06 | <othermaciej> | you participate in public-html as an invited expert? |
| 11:06 | <Philip`> | and it's based on a 'branch' of the W3C specs (the HTML5 and 2d-context specs joined together with some extra text) |
| 11:06 | <othermaciej> | er in the HTML WG |
| 11:07 | <Philip`> | and it might confuse people into thinking it's related to the W3C spec |
| 11:07 | <Philip`> | othermaciej: Public invited expert, which I think involves the same agreement |
| 11:09 | <othermaciej> | I think it would take a lawyer to answer |
| 11:09 | <Philip`> | I think I won't bother asking one :-) |
| 11:09 | <othermaciej> | I don't think you made any contributions that would require a copyright grant |
| 11:09 | <othermaciej> | but I don't know if that affects the second clause |
| 11:10 | <othermaciej> | clause 2.4 of the Invited Expert agreement is interesting |
| 11:10 | <Philip`> | It just seems strange that it imposes more restrictions than the W3C document license itself |
| 11:11 | <othermaciej> | does the document license allow derivative works? |
| 11:12 | <Philip`> | The document license doesn't (can't) prevent works that are derivative of non-W3C works (e.g. the WHATWG spec, in my case) |
| 11:13 | <Philip`> | (because the license can only restrict the copying of documents which are published under that license) |
| 11:16 | <othermaciej> | ah, right |
| 11:16 | <othermaciej> | copyright can't prevent you from making a derivative work of something else, and most particularly of something on which you hold copyright |
| 11:19 | <Philip`> | The IE agreement sounds like it does try to prevent that, though |
| 11:19 | <othermaciej> | yes it does |
| 11:20 | <othermaciej> | I wonder why they make the assignment non-exclusive if they then try to restrict your use of your contributed material |
| 11:27 | <Philip`> | I'd guess because exclusive assignment is much more complex to do |
| 11:27 | <Philip`> | http://www.copyright.gov/title17/92chap2.html#204 - "A transfer of copyright ownership, other than by operation of law, is not valid unless an instrument of conveyance, or a note or memorandum of the transfer, is in writing and signed by the owner of the rights conveyed or such owner's duly authorized agent." etc |
| 11:28 | <Philip`> | ("A “transfer of copyright ownership” is an assignment, mortgage, exclusive license, [...] but not including a nonexclusive license.") |
| 11:29 | <Philip`> | and I never had to write or sign anything to become a PIE |
| 11:30 | <othermaciej> | I see |
| 11:30 | <Philip`> | and I guess other sorts of contracts and agreements don't necessarily have to be in writing so it's okay to just do those over the internet and try to restrict people that way |
| 11:30 | <othermaciej> | FSF makes people sign things even for non-exclusive copyright assignment |
| 11:30 | <othermaciej> | (or maybe they do exclusive assignment and license your contribution back to you, I can't remember) |
| 11:31 | <hsivonen> | IIRC, FSF wants assignment and licenses stuff back |
| 11:31 | <hsivonen> | OpenOffice.org wants join copyright holdership |
| 11:31 | <hsivonen> | *joint |
| 11:32 | <hsivonen> | as I understand it, in the U.S. you need to have at least joint ownership to have standing to enforce a copyright |
| 11:36 | <othermaciej> | does Mozilla require any copyright assignment or joint ownership? |
| 11:36 | <othermaciej> | (neither Apple nor the WebKit project do, for WebKit) |
| 11:36 | <hsivonen> | IIRC, Mozilla Foundation requires neither. |
| 11:37 | <othermaciej> | is work done by MoCo employees copyright MoCo or copyright MoFo? |
| 11:37 | <hsivonen> | othermaciej: MoFo |
| 11:37 | <gsnedders> | othermaciej: Who should be CC'd on a JS parser bug in JSC? |
| 11:37 | <othermaciej> | gsnedders: oliver⊙ac, mjs⊙ac, ggaren⊙ac |
| 11:37 | <othermaciej> | (I'm not strictly necessary but I am curious) |
| 11:38 | gsnedders | thought olliej did more of the JIT stuff and not so much parser stuff |
| 11:39 | <othermaciej> | there are also some uni-szeged people who know the parser code but I can't recall names offhand |
| 11:39 | <othermaciej> | perhaps also some nokia people |
| 11:39 | <othermaciej> | olliej is fully qualified to hack any part of JSC |
| 11:41 | <gsnedders> | Shows my memory :) |
| 11:41 | <othermaciej> | now I wonder who has their svn username in the most distinct JSC source files |
| 11:41 | <othermaciej> | but I think I am too lazy to make a script to find out |
| 11:53 | <hsivonen> | aargh. looks like USB on my Linux box died and I don't have sshd up |
| 11:53 | <hsivonen> | one should always have sshd up |
| 12:09 | <hsivonen> | well, that wasn't nice. I did a hardware shutdown and reboot and grub wouldn't find my boot volume. |
| 12:09 | <hsivonen> | I was rather worried until I decided to reboot again |
| 12:09 | <hsivonen> | and then grub found the boot volume |
| 12:10 | <hsivonen> | things not going great today with computers |
| 12:11 | <hsivonen> | earlier today, my Mac was unusable for 3 hours, because Parallels decided to update Parallels Tools on a Windows VM and I was crazy enough to put other load on the machine at the same time (as opposed to just waiting for Parallels rendering the computer inoperable for half an hour or so) |
| 12:20 | <gsnedders> | Anyone with Minefield handy? |
| 12:20 | <jgraham> | Sure |
| 12:20 | hsivonen | has |
| 12:20 | <gsnedders> | Does <http://stuff.gsnedders.com/es/getterparsing.html>; fail for the last four tests? |
| 12:21 | <hsivonen> | yes |
| 12:21 | <jgraham> | yes |
| 12:21 | <jgraham> | Also in Chrome |
| 12:21 | <gsnedders> | OK |
| 12:21 | <gsnedders> | I know about other browsers |
| 12:21 | <jgraham> | Yay for interoperability |
| 12:22 | gsnedders | wonders why he doesn't have up-to-date Minefield :\ |
| 12:22 | <gsnedders> | (Like, the latest Fx nightly I have is 3.6 alpha2-pre) |
| 12:22 | <gsnedders> | (Which isn't even Minefield) |
| 12:22 | <annevk> | can you bikeshed just by yourself? |
| 12:22 | <jgraham> | You don't have the Mozilla PPA? |
| 12:23 | <gsnedders> | Uh, yeah. I could've just done it over remote desktop. |
| 12:23 | <gsnedders> | That's true |
| 12:23 | <jgraham> | annevk: Isn't that just called being indecisive? |
| 12:23 | <jgraham> | In which case: ask gsnedders :) |
| 12:23 | gsnedders | slaps jgraham |
| 12:24 | <annevk> | hehe |
| 12:24 | <jgraham> | Sadly I'm out of reach, so I imagine that without a face to absorb the momentum, your hand just continued round and you actually slapped yourself on the other side of your head |
| 13:26 | <AryehGregor> | <roc> the comment about "better browser" is slightly annoying given http://weblogs.mozillazine.org/bz/archives/020267.html <-- Incorrect but fast can legitimately be called "better" than correct but slow. Browsers are meant to create a good user experience, not conform to standards at all costs. Realistically, IMO, refraining from fully implementing features until you can do so with no performance regression does make for a better browser f |
| 13:26 | <AryehGregor> | rom a user standpoint. As Dave Hyatt points out on that post, it's just not fully implemented -- should they have not implemented it at all until they could get all corner cases right? |
| 13:29 | <annevk> | AryehGregor, I think the point is that it is is probably not possible to get it right without regressing performance |
| 13:29 | <annevk> | arguably we should nuke some selectors |
| 13:30 | <AryehGregor> | Interesting. |
| 13:30 | <AryehGregor> | Well, UAs often just refuse to implement features that would cost performance, right? |
| 13:31 | <AryehGregor> | That's not unique to WebKit. |
| 13:32 | <annevk> | yes, but that's not what happened here |
| 13:32 | <hsivonen> | if the selector features in question have been permanently vendor vetoed, the selector spec should change, IMO |
| 13:32 | <AryehGregor> | Looks like what happened here to me. |
| 13:32 | <AryehGregor> | "Boris, I'm just saying that I never looked at benchmarks or cared about benchmarks when adding this code. I held off on generalizing the code because I needed to think about a way to do it that would still be performant." |
| 13:33 | <AryehGregor> | = "won't implement it unless it can perform well" |
| 13:33 | <AryehGregor> | + "probably not possible to get it right without regressing performance" => "won't implement", simple syllogism. |
| 13:34 | <hsivonen> | I suggest Hixie stops using the feature that has been vendor vetoed, then |
| 13:34 | <annevk> | Selectors is also in CR... |
| 13:35 | <annevk> | there'd be quite some outcry if we killed nth-child and friends |
| 13:36 | <AryehGregor> | Is Hixie actually using a CSS feature that kills performance in every browser that implements it? That seems pointless. I thought Firefox's painful handling of the HTML5 spec was mainly due to some JS thing and was fixed in 3.7. |
| 13:40 | <ment> | AryehGregor: the problem with css and js is, that you have to maintain correct styling during execution of the script |
| 13:42 | <AryehGregor> | What do you mean? You don't have to rerender styles synchronously during script execution, right? If I do foo.width = 100; foo.height = 100; then foo is never actually rendered with width 100 but not height 100, AFAIK. |
| 13:42 | <gsnedders> | AryehGregor: No, but if you reflow during script execution perf can be bad (as is the case with HTML5 and Opera, which reflows on a timeout during script execution) |
| 13:43 | <AryehGregor> | Mm. |
| 13:43 | <ment> | AryehGregor: yes, you could postpone rematching. but in some cases you just have to rematch the whole tree after insertion of one node in the right place |
| 13:44 | <AryehGregor> | Interesting, although this is somewhat over my head. |
| 13:44 | <ment> | during the time i've been writing css matcher i've wondered why the selector language wasn't designed with some theory backing it up |
| 13:46 | <gsnedders> | AryehGregor: Otherwise, only trying to access the styles through CSSOM causes a reflow during script execution |
| 13:54 | <hsivonen> | annevk: in that case, I think Hixie should make the spec obviously misrender if the selectors aren't implemented right :-) |
| 13:55 | <hsivonen> | misrender in an obvious way, that is |
| 13:56 | <hsivonen> | applying the strategy of making stuff look ugly in WebKit to pressure WebKit devs to fix WebKit isn't working on my personal site, but doing it with the spec might work |
| 13:57 | <othermaciej> | hsivonen: I didn't know your site looked any different in WebKit-based browsers than it was supposed to |
| 13:57 | <othermaciej> | what's the bug? |
| 13:58 | <othermaciej> | other than the Firefox ad |
| 13:58 | <othermaciej> | but I assume that is by design |
| 13:58 | <hsivonen> | othermaciej: bug 1: text doesn't use a local font while fonts are downloading |
| 13:58 | <othermaciej> | ah, that I've noticed |
| 13:58 | <hsivonen> | othermaciej: bug 2: the double-struck rounded corners are ugly |
| 13:59 | <othermaciej> | that I didn't notice, but I see your point |
| 14:00 | <hsivonen> | which reminds me that I should add support for -moz- transitions |
| 14:00 | hsivonen | doesn't like vendor prefixes |
| 14:00 | <othermaciej> | your site tells me to upgrade to Firefox 3.5 when I visit with Firefox 3.6 |
| 14:01 | <hsivonen> | othermaciej: yeah, the back end for that ad in FAIL |
| 14:01 | <hsivonen> | I guess I should remove the ad |
| 14:01 | <hsivonen> | s/in/is/ |
| 14:02 | <othermaciej> | hmm |
| 14:02 | <othermaciej> | I wonder if we are even using the right font for your site at all |
| 14:03 | <othermaciej> | the font it ends up with in Safari seems to be the one that flashes temporarily in Firefox |
| 14:03 | <AryehGregor> | I also noticed on IE6 that it shows the "You're settling for good" message for a second before switching to the "You don't use VHS" one. |
| 14:05 | <hsivonen> | othermaciej: because Apple doesn't allow Mac OS X to be virtualized and because Safari detects uninstalled local fonts(!), it's too hard for me to test if the page looks right in Safari on a Mac that doesn't already have the fonts installed locally |
| 14:06 | <hsivonen> | maybe I should test on my mother's Mac when I visit my parents |
| 14:08 | <jgraham> | ment: Any theory you hand in mind? |
| 14:09 | <hsivonen> | I also added -o- and -ms- transitions completely defeating the point of vendor prefixes |
| 14:10 | <MikeSmith> | http://twitter.com/futomi/status/10978894529 |
| 14:10 | <MikeSmith> | "#html5 spec says video.width must be DOMString while canvas.width must be unsigned long. Why inconsequence?" |
| 14:10 | <MikeSmith> | of course he means, "Why the inconsistency?" |
| 14:10 | <MikeSmith> | but anybody know the answer to that? |
| 14:15 | <hsivonen> | can someone who understands CSS Transitions explain to me why my link styling transitions work on unvisited links but not on visited links? |
| 14:16 | <hsivonen> | also, what's the deal with Opera supporting border-radius but not outline-radius? |
| 14:17 | <annevk> | outline-radius is not an official property |
| 14:17 | <hsivonen> | furthermore, Opera mistransitions the background color of my links |
| 14:17 | <annevk> | bugs.opera.com/wizard |
| 14:18 | <hsivonen> | hmm. the visited links do transition in WebKit |
| 14:18 | <hsivonen> | I guess I got one Gecko bug report and one Presto bug report out of this |
| 14:18 | <karlcow> | MikeSmith: it seems there are two attributes on video |
| 14:19 | <karlcow> | readonly attribute unsigned long videoWidth; |
| 14:19 | <karlcow> | attribute DOMString width; |
| 14:20 | <karlcow> | oh and img |
| 14:20 | <karlcow> | attribute unsigned long width; |
| 14:20 | <karlcow> | readonly attribute unsigned long naturalWidth; |
| 14:21 | MikeSmith | nods |
| 14:21 | <karlcow> | a videoWidth is not a naturalWidth ? |
| 14:21 | <MikeSmith> | dunno |
| 14:21 | <MikeSmith> | I guess naturalWidth must be the intrinsic width |
| 14:21 | <MikeSmith> | and the other one is the displayed width |
| 14:22 | <annevk> | nope, both are intrinsic |
| 14:22 | <MikeSmith> | oh, ok |
| 14:22 | <karlcow> | nativeWidth have been exterminated I guess |
| 14:22 | <karlcow> | :p |
| 14:22 | <annevk> | i complained about this once, but didn't fight it through |
| 14:22 | <MikeSmith> | anyway, it seems futomi's question is, why is one width value a string datatype and another is an unsigned long |
| 14:23 | <annevk> | naturalWidth was already there, and videoWidth was implemented when we found that out |
| 14:23 | <annevk> | maybe someone should complain again |
| 14:23 | <annevk> | MikeSmith, dunno really |
| 14:24 | <karlcow> | object has no intrinsic width, just attribute DOMString width; |
| 14:24 | <annevk> | i guess to be the same as <iframe> and <object> |
| 14:24 | <annevk> | <img>.width is quite special |
| 14:30 | <karlcow> | % grep -i width html5.txt | grep attribute | grep long |
| 14:30 | <karlcow> | attribute unsigned long width <#dom-img-width>; |
| 14:30 | <karlcow> | readonly attribute unsigned long naturalWidth <#dom-img-naturalwidth>; |
| 14:30 | <karlcow> | readonly attribute unsigned long videoWidth <#dom-video-videowidth>; |
| 14:30 | <karlcow> | attribute unsigned long width <#dom-canvas-width>; |
| 14:30 | <karlcow> | attribute unsigned long width <#dom-pre-width>; |
| 14:31 | <karlcow> | % grep -i width html5.txt | grep attribute | grep DOMString |
| 14:31 | <karlcow> | attribute DOMString width <#dom-dim-width>; |
| 14:31 | <karlcow> | attribute DOMString width <#dom-applet-width>; |
| 14:31 | <karlcow> | attribute DOMString width <#dom-marquee-width>; |
| 14:31 | <karlcow> | attribute DOMString marginWidth <#dom-frame-marginwidth>; |
| 14:31 | <karlcow> | attribute DOMString width <#dom-col-width>; |
| 14:31 | <karlcow> | attribute DOMString width <#dom-hr-width>; |
| 14:31 | <karlcow> | attribute DOMString marginWidth <#dom-iframe-marginwidth>; |
| 14:31 | <karlcow> | attribute DOMString width <#dom-table-width>; |
| 14:31 | <karlcow> | attribute DOMString width <#dom-tdth-width>; |
| 14:32 | <annevk> | oh, and <canvas> is different because width/height are just different there, they define the grid, not the dimensions |
| 14:36 | <Philip`> | <iframe width="100%"> |
| 14:36 | <Philip`> | (I think that works?) |
| 14:36 | <Philip`> | so it's not just an integer |
| 14:37 | <Philip`> | but canvas dimensions are the number of pixels in the coordinate space, so they must be a fixed number |
| 14:37 | <karlcow> | can video be 100% too |
| 14:37 | Philip` | shrugs |
| 14:37 | <annevk> | nope |
| 14:38 | <annevk> | oh yes |
| 14:39 | <annevk> | it works for both, but is also invalid |
| 14:39 | <hsivonen> | HTMLDocument has 3 charset-related properties. Seriously? |
| 14:40 | <MikeSmith> | hsivonen: btw, http://html5.org/tools/web-apps-tracker?from=4861&to=4862 adds a new "allow-top-navigation" enumerated value to the allowed values for iframe/@sandbox |
| 14:41 | <MikeSmith> | which I think means instead of 3! combinations of the allowed values for that in the schema |
| 14:41 | <MikeSmith> | we now have 4! |
| 14:41 | <MikeSmith> | 4 factorial |
| 14:41 | <hsivonen> | combinations FTW! |
| 14:41 | <MikeSmith> | heh |
| 14:42 | <MikeSmith> | is there any alternative to using an RNG list structure for this? |
| 14:42 | <MikeSmith> | I guess there's probably not |
| 14:42 | <hsivonen> | MikeSmith: a datatype in Java |
| 14:42 | <Philip`> | Isn't it 4!+3!+2!+1!+0!? |
| 14:42 | <MikeSmith> | Philip`: is it? |
| 14:42 | <Philip`> | Any permutation of 4 values + any permutation of 3 values etc |
| 14:43 | <MikeSmith> | Philip`: what we have in the schema now is just 6 combinations |
| 14:43 | <hsivonen> | fortunately, you don't need to list all possibilities, because you can use the ? quantifier inside a list {} |
| 14:43 | hsivonen | wishes list {} permitted interleave |
| 14:45 | <MikeSmith> | hsivonen: do you think this merits creating a new dataype? just adding all the combinations in the schema is easy enough to do, just ugly |
| 14:45 | <Philip`> | hsivonen: Ah, right |
| 14:47 | <hsivonen> | MikeSmith: it doesn't merit creating a datatype at this point of spec development if adding the combos is easy enough |
| 14:47 | <hsivonen> | Hixie: it would be nice to have the new Gecko textarea resizer in Live DOM |
| 14:51 | <jgraham> | hsivonen: (why) Doesn't it work automatically? |
| 15:01 | <hsivonen> | jgraham: it seems it doesn't work automatically. dunno why |
| 15:01 | <hsivonen> | twitter seems to use it already |
| 15:02 | hsivonen | discovers that Opera doesn't implement document.charset and document.characterSet |
| 15:03 | <hsivonen> | and WebKit doesn't appear to support reloading document.open()ed docs from location.reload() |
| 15:08 | <hsivonen> | whoa! is WebKit the only browser that supports document.charset? |
| 15:08 | <hsivonen> | why do we have it in the spec if Gecko, IE and Opera don't have it? |
| 15:10 | <hsivonen> | I'm confused. Live DOM w("Foo") not working in IE8 |
| 15:11 | <jgraham> | hsivonen: It never has |
| 15:11 | <annevk> | hsivonen, maybe Hixie didn't check enough of the other impls? |
| 15:11 | <hsivonen> | jgraham: ooh! suprising |
| 15:11 | <hsivonen> | jgraham: it works occasionally |
| 15:12 | <hsivonen> | oh. IE has .charset after all |
| 15:12 | <jgraham> | Yeah, I guess I should have inserted the word "reliably" in there somewhere |
| 15:12 | <hsivonen> | and for document.open()ed docs it's "unicode" |
| 15:12 | <jgraham> | (it is also differently broken in Chrome) |
| 15:13 | <hsivonen> | browsers shouldn't be so broken that you can't black box poke at them |
| 15:16 | <hsivonen> | anyway, http://software.hixie.ch/utilities/js/live-dom-viewer/saved/420 is what I have so far |
| 15:16 | <hsivonen> | it didn't work the way I expected in the old parser in Gecko |
| 15:16 | <hsivonen> | I wonder why |
| 15:17 | <hsivonen> | (I expected Gecko to reparse the written content upon .close()) |
| 15:17 | <hsivonen> | (so that the written script would run thrice) |
| 15:21 | <TabAtkins> | JonathanNeal: You win. My response was *not* sent to the room, so I can't find it in the logs. ;_; |
| 15:22 | <TabAtkins> | The gist of it was that, sure, a collection of article snippets might be an <article>. No reason why it can't be. That seems like something that could legitimate be an independent chunk of content. |
| 15:30 | <JonathanNeal> | TabAtkins, and rotating banners then could be as well, s'pose? |
| 15:31 | <TabAtkins> | Apply the rule of thumb. Would you want to look at it fullscreen, or as an independent page? |
| 15:31 | <JonathanNeal> | What would the difference be? |
| 15:32 | <JonathanNeal> | When I look at it full screen, is it still interactive (this would be necessary in the article snippits as a collective article viewed fullscreen) |
| 15:32 | <TabAtkins> | Sorry, that was an or choice. Just two variants on the question. |
| 15:32 | <TabAtkins> | s/was/wasn't/ |
| 15:32 | <TabAtkins> | argh |
| 15:32 | <JonathanNeal> | Got it. |
| 15:33 | zcorpan | wonders where that rule of thumb comes from |
| 15:33 | <TabAtkins> | It comes from my own reading of the spec, and reasoning about what the difference should be between an <article> and <section> |
| 15:33 | <JonathanNeal> | So section almost acts like a more generic version of article, in that it doesn't specify that the content is wholey complete on its own, but that it is an independent entity in some form? |
| 15:33 | <TabAtkins> | Also what can be easily and quickly applied. |
| 15:33 | <TabAtkins> | JonathanNeal: Yeah, in my head at least <section> is the most generic sectioning element, and all the others (<article>, <aside>, <nav>, etc.) are specializations on it. |
| 15:34 | <JonathanNeal> | Did you catch my attempt at realizing http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#margins-and-padding last night? |
| 15:34 | <TabAtkins> | Nope. |
| 15:35 | <zcorpan> | i'd want to look at <video>s fullscreen, but wouldn't in general wrap <video>s in <article>s |
| 15:35 | <TabAtkins> | zcorpan: Sure, but <video> has its fullscreen-ness built in. |
| 15:35 | <JonathanNeal> | The section of x h1, x x h1 etc. |
| 15:35 | <TabAtkins> | JonathanNeal: Oh gods. |
| 15:35 | <JonathanNeal> | I tried to realize what that css would look like. |
| 15:36 | <JonathanNeal> | http://pastie.org/884363 in less css it is about 30 lines of code. |
| 15:36 | <TabAtkins> | Ideally, it would look like :section(1) h1, :section(2) h1, or similar. |
| 15:36 | <JonathanNeal> | http://pastie.org/884364 in actual css it is about 1383 lines of code |
| 15:37 | <TabAtkins> | Failing that, :any(article, aside, nav, section) h1, :any(article, aside, nav, section) :any(article, aside, nav, section) h1, etc. |
| 15:38 | <JonathanNeal> | I'm guessing :section(1) and :any(section) aren't supported in any browser yet? |
| 15:40 | <TabAtkins> | They're not even close to specced yet. |
| 15:40 | <TabAtkins> | :any() at least has wide support, though we haven't started Selectors 4 yet. |
| 15:42 | <annevk> | You need both :section() and :heading() I guess... |
| 15:42 | <annevk> | And maybe ::section if you want to have boxes for implied sections... |
| 15:42 | <annevk> | But getting that implemented and working dynamically might be quite hard |
| 15:42 | <JonathanNeal> | I tried any() and it did not work |
| 15:43 | <TabAtkins> | JonathanNeal: No, it doesn't exist yet. |
| 15:43 | <jgraham> | "wide support" meaning in the CSSWG not in extant UAs |
| 15:43 | <JonathanNeal> | Ha |
| 15:43 | <TabAtkins> | Nod to jgraham. |
| 15:43 | <JonathanNeal> | Be right back. |
| 15:44 | <TabAtkins> | annevk: Yeah, that's the issue. I'd like to create some sort of entity in selectors that can be *either* a pseudoelement *or* an alias to an existing element, so I can transparently use ::section(1) and refer to either the implicit section or an explicit <section>. |
| 15:45 | <annevk> | :::section |
| 15:45 | <annevk> | more colons is better |
| 15:45 | <jgraham> | Selectors gets more and more like Perl every day |
| 15:45 | <annevk> | only in #whatwg |
| 15:46 | <annevk> | well, and TabAtkins' dreams |
| 15:46 | <annevk> | :p |
| 15:46 | <jgraham> | Well it is pretty close already |
| 15:46 | <jgraham> | You get to play "guess what this random bit of punctuation does" |
| 15:47 | <annevk> | especially the attribute selectors are tricky |
| 15:52 | <zcorpan> | i wonder why some people think it's suboptimal to have a spec in two places |
| 15:52 | <annevk> | it defies their sense of logic |
| 15:53 | <annevk> | prolly roughly corresponds to the people that hate this channel |
| 15:53 | <zcorpan> | the relaxng spec is also in two places (with editorial differences) |
| 15:53 | <zcorpan> | i think it's good because sometimes when i need to access the spec, whatwg.org is down or i can't reach it, or w3.org is down |
| 15:56 | <TabAtkins> | annevk: Answer my email I sent you 12 hours ago. |
| 15:56 | <annevk> | I don't have that email? :/ |
| 15:57 | <annevk> | only email I have from you is about serializing Selectors |
| 15:57 | <TabAtkins> | Hrm. I sent it to the address listed on your blog. |
| 15:57 | <TabAtkins> | (Easiest place to find it on my phone.) |
| 15:57 | <TabAtkins> | (Since I don't have internet at the house until tomorrow.) |
| 15:58 | <annevk> | oh, no email there either? |
| 15:58 | <annevk> | also not in spam |
| 15:59 | <annevk> | did you make a typo? |
| 15:59 | <TabAtkins> | Man, I dunno. I"ll just msg you instead. |
| 16:00 | <zcorpan> | http://twitter.com/futomi/statuses/10978894529 |
| 16:00 | <annevk> | we discussed that a little earlier |
| 16:03 | <jgraham> | zcorpan: I assumed it was all about control |
| 16:20 | <JonathanNeal> | back TabAtkins |
| 16:28 | <JonathanNeal> | How come the rendering section of the html5 spec doesn't mention video, audio, canvas being display: block? Are they meant to be something else? |
| 16:29 | <gsnedders> | Wow. I managed to crash my laptop an hour ago. |
| 16:30 | <JonathanNeal> | You needed an hour to discover this? |
| 16:31 | <Dashiva> | That's how responsive it is, I guess |
| 16:32 | <JonathanNeal> | I'm looking over http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html trying to see where I might assume it is display block or inline-block, any guesses? |
| 16:32 | <JonathanNeal> | Facts will work too :) |
| 16:32 | <jgraham> | YOu clearly should be working harder if it took an hour to notice |
| 16:32 | <gsnedders> | jgraham: Hey, it's not my fault I was talking to James! |
| 16:34 | <jgraham> | You weren't talking to me! |
| 16:36 | <gsnedders> | Did you miss me!? |
| 16:37 | <JonathanNeal> | I noticed there are some vendor styles consistant across Opera, Chrome, Safari, and Firefox --- is there a list they are gathering these from? |
| 16:37 | <JonathanNeal> | For instance, canvas, by default is always 300x150 |
| 16:41 | <gsnedders> | JonathanNeal: Reverse engineering more than anything else |
| 16:41 | <JonathanNeal> | Haha |
| 16:41 | <JonathanNeal> | So I'll create a second css sheet which is what the vendors have said. |
| 16:41 | <gsnedders> | Life of a browser QA involves knowing all your competitor's behaviour as well as your own |
| 16:41 | <gsnedders> | :P |
| 16:41 | <gsnedders> | JonathanNeal: See HTML 5 rendering appendix |
| 16:42 | <JonathanNeal> | gsnedders, ^^ that's what I was referencing. |
| 16:43 | <JonathanNeal> | What's weird is that the rendering spec mentions elements that are no in the html5 spec. |
| 16:43 | <JonathanNeal> | afaik |
| 16:43 | <gsnedders> | Indeed it does |
| 16:44 | <JonathanNeal> | That's kinda weird, huh? |
| 16:44 | <zcorpan> | why is it weird? |
| 16:44 | <gsnedders> | Some things you need to render in a certain way for compat even if they have no semantic meaning :P |
| 16:44 | <zcorpan> | (which elements are those, btw?) |
| 16:48 | <zcorpan> | looking at the rendering section, i wonder if we should make <iframe topmargin> same-origin-only, and whether doing so would break any pages |
| 16:48 | <zcorpan> | sorry, marginheight |
| 16:52 | zcorpan | can't find any elements in rendering that's not mentioned elsewhere in the spec |
| 16:53 | <annevk> | zcorpan, e.g. <center> |
| 16:54 | <zcorpan> | annevk: the obsolete section defines center |
| 16:55 | <annevk> | I'm assuming those elements were excluded |
| 16:55 | <beowulf> | ah <center>, old friend |
| 16:55 | <TabAtkins> | zcorpan: Why would it be necessary to make marginheight same-origin-only? You fearing some negative margins used to position the page to a particular spot? |
| 16:56 | <zcorpan> | TabAtkins: ooh, i didn't even think of negative margins |
| 16:57 | <TabAtkins> | I don't see how positive margins could be maliciously used. |
| 16:58 | <zcorpan> | seems negative margins don't work |
| 16:59 | <zcorpan> | the spec says |
| 16:59 | <zcorpan> | "⚠Warning! The above requirements imply that a page can change the margins of another page (including one from another origin) using, for example, an iframe. This is potentially a security risk, as it might in some cases allow an attack to contrive a situation in which a page is rendered not as the author intended, possibly for the purposes of phishing or otherwise misleading the user." |
| 17:10 | <JonathanNeal> | I'm trying to write the rendering spec out as an actual css file. |
| 17:11 | <JonathanNeal> | It's weird how they confirm that form is display: block, but not fieldset |
| 17:12 | <zcorpan> | the rendering section has some requirements that can't be expressed in css |
| 17:12 | <JonathanNeal> | and then legend is confirmed, which is a child of fieldset |
| 17:12 | <JonathanNeal> | So it's best left implied by the vendors? |
| 17:13 | <zcorpan> | "The fieldset element is expected to establish a new block formatting context." |
| 17:14 | <zcorpan> | what would be useful is a testsuite for the rendering section |
| 17:14 | <annevk> | and a spec for form controls |
| 17:22 | <erlehmann> | some guy just realized that you can do <!dOcTyPe HtMl> |
| 17:23 | <erlehmann> | WHAT HAVE YOU DONE IAN HICKSON |
| 17:23 | <othermaciej> | clearly we need to support w4r3z d00d syntax for all tags and attributes |
| 17:24 | <JonathanNeal> | Here you go http://sandbox.thewikies.com/html5-pangram/ |
| 17:25 | <annevk> | erlehmann, in HTML4 you can write <hTmL> too... |
| 17:25 | <Philip`> | I like writing my pages like <!Doctype Html><Html Lang="En"><Head><Title>... because then it looks more like proper English |
| 17:26 | <othermaciej> | I write <!DOCTYPE html>, I am not sure what convinced me that is right |
| 17:26 | <Philip`> | Alternatively, I like writing <body><div><span>things</SPAN> <span>like</SPAN> this</DIV></BODY> because it helps to visually distinguish the different types of tag |
| 17:26 | <JonathanNeal> | http://sandbox.thewikies.com/html5-pangram/rendering.whatwg.css --- that's a collection of many of the styles I could actually see recommended by the whatwg |
| 17:26 | <annevk> | othermaciej, that works in XML too |
| 17:26 | <Philip`> | You can make good use of capitalisation |
| 17:26 | <jgraham> | Philip`: You should do <DIV><span></span></DIV> to distinguish block and inline elements |
| 17:26 | <othermaciej> | annevk: maybe I should add xmlns="http://long/thing/i/cant/remember"; to all my documents after all |
| 17:27 | <JonathanNeal> | http://sandbox.thewikies.com/html5-pangram/rendering.vendor.css --- that's a collection of many of the styles I could see the vendors were also following |
| 17:27 | <annevk> | othermaciej, might wanna check with Paul if it's worth brownie points :) |
| 17:28 | annevk | appreciated the beer at TPAC |
| 17:28 | <erlehmann> | JonathanNeal, a style for headings using sections would be nice to |
| 17:28 | <erlehmann> | too |
| 17:29 | zcorpan | remembers that li[type=I] is still broken |
| 17:30 | <JonathanNeal> | erlehmann, does that have to do with decendants? I tried not to include anything I didn't have to that had to do with decendants. |
| 17:30 | <JonathanNeal> | Just to get it done. Are you talking about h1's affected by being in sectioning elements? |
| 17:30 | <JonathanNeal> | "sectioned content" as it is referred to. |
| 17:31 | <erlehmann> | JonathanNeal, exactly. it would be a huge mess. |
| 17:31 | <erlehmann> | maybe i should autogenerate it |
| 17:31 | <erlehmann> | currently i just use h1 to h3, but it grows exponentially |
| 17:32 | <JonathanNeal> | erlehmann, heh I did, it's 1383 lines of code. |
| 17:33 | <JonathanNeal> | Just for four or five levels of h1's. |
| 17:33 | <erlehmann> | wait, wat |
| 17:33 | <erlehmann> | why so much ? |
| 17:33 | <erlehmann> | argl |
| 17:33 | <erlehmann> | you mean 1383 lines of stylesheet ? |
| 17:33 | <JonathanNeal> | Because they are exponetial decendants. |
| 17:34 | <JonathanNeal> | To test every scenario multiples by itself each time you add a new layer. |
| 17:34 | <JonathanNeal> | So article, aside, nav, or section, that's 4, then 16, then 256, then 65536 |
| 17:35 | <erlehmann> | i am shocked and appalled |
| 17:35 | <JonathanNeal> | Don't take me to the bank on that, but when I actually rendered them out they ended up being 1383 lines of css |
| 17:35 | <JonathanNeal> | I compared this to an example of less css, where it was 30 lines. |
| 17:36 | <zcorpan> | be happy there aren't 5 sectioning elements |
| 17:37 | <JonathanNeal> | So, the best way around this would be to give classnames to all the sectioning elements, and for the exact classname I would borrow from the spec ".sectioning-content" |
| 17:37 | <zcorpan> | or 7 heading levels |
| 17:37 | <JonathanNeal> | Attach those to every article, aside, nav, and section element and you'll only need 20 lines of css. |
| 17:38 | <zcorpan> | or use h1-h6 appropriately and you need 0 lines of css |
| 17:39 | <JonathanNeal> | I would label them ".sectioning-content" vs ".sectioning-element" to describe what it contains rather than what it is (like how roles work, example being main the container isn't itself the main content but contains the main content) ... which (tangent) is also why I do not like naming things "container". |
| 17:45 | <MikeSmith> | paul_irish - you're in Boston, right? |
| 17:49 | <mpilgrim> | jgraham: is http://james.html5.org/microdata/ maintained? |
| 17:49 | <jgraham> | mpilgrim: Not really |
| 17:50 | <mpilgrim> | ok |
| 17:50 | <mpilgrim> | are there similar tools elsewhere, besides google's rich snippets testing tool? |
| 17:50 | <jgraham> | foolip has something |
| 17:50 | <Philip`> | http://philip.html5.org/demos/microdata/demo.html is similar, in that it's also unmaintained |
| 17:51 | <Philip`> | http://foolip.org/microdatajs/live/ |
| 17:51 | <Philip`> | That one's better |
| 17:52 | <Philip`> | (Google's rich snippets testing tool is probably a bad way to attempt to determine correct microdata parsing) |
| 17:53 | <mpilgrim> | i agree |
| 17:53 | <mpilgrim> | but i'm going to talk about microdata in "dive into html5" and the hook to get people to care is "google parses it" |
| 17:53 | <jgraham> | Philip`: Did you have a good way to find that or did you have a reference to it stored somewhere? |
| 17:55 | <Philip`> | mpilgrim: I'm not aware of anybody having tested Google's support (other than some trivial things like it parsing <p><p> wrong), so if you try it out then it'd be interesting to know how well it matches the spec |
| 17:55 | <Philip`> | jgraham: I used a search engine |
| 17:55 | <Philip`> | They're quite good at finding things |
| 17:55 | <jgraham> | Not for me, in this case |
| 17:55 | <mpilgrim> | thanks for the links |
| 17:56 | <mpilgrim> | back to bed |
| 17:56 | <mpilgrim> | (sick today) |
| 17:56 | <Philip`> | I searched for "philip microdata" to find my one, and foolip's one came immediately after it |
| 17:57 | <Philip`> | so it found his one even though I *wasn't* searching for it, which is a step up from old-fashioned search engines that only find what you're searching for |
| 18:12 | <JonathanNeal> | href is not a required attribute of base? |
| 18:13 | <JonathanNeal> | It is ... but html5.validator.nu and validator.w3.org let it slide. |
| 18:14 | <JonathanNeal> | It took IE6 for me to find that out. |
| 18:14 | <annevk> | is it? |
| 18:14 | <MikeSmith> | I don't think href is required on base |
| 18:14 | <annevk> | it's not required |
| 18:14 | <JonathanNeal> | "A base element must have either an href attribute, a target attribute, or both." |
| 18:14 | <JonathanNeal> | http://www.whatwg.org/specs/web-apps/current-work/#the-base-element |
| 18:14 | <annevk> | right, so href is not required if target is there |
| 18:15 | <JonathanNeal> | Sure, I wasn't aware --- I was validating with <base /> |
| 18:15 | <MikeSmith> | hmm, this is probably a bug in the validator |
| 18:15 | <MikeSmith> | in the schema |
| 18:16 | <MikeSmith> | yeah |
| 18:16 | MikeSmith | wonders when that spec change was made |
| 18:18 | <MikeSmith> | JonathanNeal: fwiw, after I make the fix for this, for this case, the validator will eventually report, "Element base is missing one or more of the following attributes: href target" |
| 18:18 | <paul_irishhhh> | MikeSmith: yessir i am. |
| 18:19 | <JonathanNeal> | Coolio Mike! I found this out after IE6 crapped out on me, which had nothing to do with that requirement, but that's how I found it (IE6 just doesn't support the self-closing <base />) |
| 18:20 | <MikeSmith> | JonathanNeal: thanks for catching it |
| 18:34 | <annevk> | ooh, voicexml |
| 18:34 | <annevk> | it returns |
| 18:35 | annevk | reads rest of the message |
| 18:35 | <annevk> | ok, maybe not |
| 18:36 | zcorpan | figured out why dailymotion doesn't work in opera |
| 18:38 | annevk | is curious |
| 18:39 | <JonathanNeal> | http://localhost/html5-pangram/ you know those whatwg styles really do conform most of the browsers to look about the same. |
| 18:39 | <JonathanNeal> | You still need to set some styles not mentioned that the vendors are following on video, canvas, etc, but for the most part they're all starting to look the same. |
| 18:39 | <JonathanNeal> | localhost! For shame ... |
| 18:39 | <JonathanNeal> | http://sandbox.thewikies.com/html5-pangram/ |
| 18:40 | <zcorpan> | annevk: we fire canplaythrough before the script registers the listener |
| 18:40 | <zcorpan> | annevk: dispatching a canplaythrough event from the address bar makes it work |
| 18:42 | <annevk> | oh argh, event loop bugs |
| 18:43 | <zcorpan> | i'm not sure it's an event loop bug, just a bad assumption about when the event will be fired |
| 18:44 | <annevk> | but it always fires in other browsers at the right time? |
| 18:44 | <annevk> | or does it sometimes not work in Firefox/Chrome? |
| 18:45 | <zcorpan> | i guess it could sometimes not work in other browsers if the video is cached or if the script file is slow to load |
| 18:45 | <annevk> | not saying it's an event loop bug per se btw, more like an event loop issue |
| 18:46 | <annevk> | maybe other browsers do the event loop thingie for this in a certain way that guarantees this scenario works |
| 18:48 | <zcorpan> | i think it has more to do with other browsers only firing canplay and canplaythrough when the video actually can play (through) |
| 18:49 | <zcorpan> | while opera lies and fires them early because we haven't implemented it properly yet |
| 18:51 | <zcorpan> | still, it's a network race problem and the script should be smarter and check the state of the video when the script runs instead of blindly waiting for an event that might already have been fired |
| 18:53 | MikeSmith | reads http://www.ietf.org/mail-archive/web/http-state/current/msg00718.html (rough notes from http-state WG face-to-face meeting at IETF 77) |
| 19:08 | zcorpan | wants to be able to do elm.dispatchEvent(new Event('canplaythrough')) instead of var e = document.createEvent('Event'); e.initEvent('canplaythrough', false, false); elm.dispatchEvent(e); |
| 19:16 | <annevk> | there was some discussion on better DOM Event APIs |
| 19:16 | <annevk> | not sure what happened to it :/ |
| 19:22 | <Philip`> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cdiv%20style%3Dlist-style-image%3Aurl%28image%29%3E%3Caudio%20controls%3E |
| 19:22 | <Philip`> | Interesting styling in Firefox |
| 19:49 | <Hixie> | anyone know what the url to the hybi meeting audio is going to be? |
| 19:49 | <Hixie> | or what irc channel they're on? |
| 19:49 | <Hixie> | i guess they'd use a jabber channel |
| 19:49 | <othermaciej> | it's jabber |
| 19:51 | <Hixie> | any idea what the url to the jabber thing is? |
| 19:53 | <othermaciej> | I'm afraid not, but I think it may have been sent to the hybi list |
| 19:53 | <othermaciej> | I have to escape the meeting for a while |
| 19:55 | <annevk> | http://www.ietf.org/mail-archive/web/hybi/current/msg00822.html |
| 19:55 | <annevk> | Hixie, ^^ |
| 19:56 | <Hixie> | cool thanks |
| 19:59 | <annevk> | what's the program that handles audio streams for Ubuntu? |
| 19:59 | <annevk> | I think I nuked it the other day when trying to fix something |
| 19:59 | <Hixie> | dunno |
| 19:59 | <annevk> | and now I forgot the name and don't have sound :) |
| 19:59 | <Philip`> | Do you mean like PulseAudio or something else? |
| 19:59 | <Hixie> | quicktime failed for me, but itunes worked |
| 20:00 | <Hixie> | http://videolab.uoregon.edu/events/ietf/ietf778.m3u is the audio url |
| 20:00 | <annevk> | Philip`, cheers! |
| 20:00 | <Philip`> | "mplayer -playlist http://videolab.uoregon.edu/events/ietf/ietf778.m3u"; works for that kind of stream |
| 20:01 | <JonathanNeal> | Hola |
| 20:01 | <zcorpan> | [hidden]:not(colgroup):not(col):not(thead):not(tbody):not(tfoot):not(tr):not(td):not(th) { display:none } |
| 20:01 | <zcorpan> | colgroup[hidden], col[hidden], thead[hidden], tbody[hidden], tfoot[hidden], tr[hidden], td[hidden], th[hidden] { visibility:collapse } |
| 20:01 | <annevk> | cool |
| 20:02 | <zcorpan> | wonder if browsers would be ok with that in the ua style sheet |
| 20:02 | <annevk> | installing pulseaudio doesn't require a restart |
| 20:02 | <annevk> | i have to say, curb your enthusiasm was more fun |
| 20:02 | <JonathanNeal> | zcorpan, did you want me to change my css to reflect that? |
| 20:02 | <TabAtkins> | zcorpan: That's what I'm hoping for. |
| 20:02 | <zcorpan> | JonathanNeal: no, unrelated |
| 20:02 | <JonathanNeal> | (minus the fact that :not() and those are unsupported. |
| 20:02 | <JonathanNeal> | :) |
| 20:03 | <TabAtkins> | :not() is supported! Just not across all browsers yet. |
| 20:05 | <zcorpan> | does ie9 support :not()? |
| 20:06 | zcorpan | guesses so since it supposedly passes all selectors tests |
| 20:06 | <TabAtkins> | Dunno, I don't have a Vista or W7 machine to test it on. |
| 20:07 | <annevk> | minutes for hybi are being kept here: http://etherpad.com/FWABRLTLpK |
| 20:07 | <annevk> | not sure if everyone can view it or not, there seem to be some limitations |
| 20:08 | annevk | wonders whether to suggest that Opera prefers -5 |
| 20:10 | <Dashiva> | annevk: Is that a typo? |
| 20:10 | <TabAtkins_> | Dashiva: No, just a non-sequitur; anne is talking about the hybi minutes. |
| 20:11 | <TabAtkins_> | Or, wait, maybe a typo? |
| 20:11 | <Dashiva> | -5 seems malformed, I'd expect -75 or -05 |
| 20:11 | <TabAtkins_> | (If he means -75). |
| 20:11 | <annevk> | Dashiva, they are zero padded? |
| 20:11 | <MikeSmith> | JonathanNeal: http://www.w3.org/html/check?doc=http%3A%2F%2Fdev.w3.org%2Fhtml5%2Ftests%2Fvalidation%2Ffull%2Finvalid%2Fmissing-attributes%2Fbase-missing-href-target.html |
| 20:11 | <annevk> | yup they are |
| 20:12 | <JonathanNeal> | :) |
| 20:19 | <annevk> | discussion so far is mostly procedural |
| 20:19 | <annevk> | guess that makes sense for a new WG |
| 20:19 | <annevk> | but it's boring |
| 20:33 | <annevk> | Hixie, WebSocket.send() should throw if DOMString cannot be converted to a Unicode stream, not just for surrogate characters |
| 20:34 | <Hixie> | what other inputs would fail that way? |
| 20:35 | <franksalim> | would a string with \0s in the middle fail like that? |
| 20:35 | <annevk> | you can get other invalid code points otherwise, no? |
| 20:35 | <gsnedders> | How do you do string concat in Perl? |
| 20:35 | <Hixie> | gsnedders: . operator |
| 20:36 | <Hixie> | franksalim: i don't think so |
| 20:36 | <Hixie> | annevk: i can't think of any others... which ones? |
| 20:37 | <franksalim> | Hixie, but you could get a bad websocket frame as currently speced, right? |
| 20:37 | <gsnedders> | Hixie: thx |
| 20:37 | <Hixie> | franksalim: why? |
| 20:37 | <Hixie> | or how? |
| 20:37 | <Hixie> | 0x00 0x00 0x00 0xFF is fine |
| 20:38 | <Hixie> | it's a 2-character 0x00 frame with data U+0000 U+0000 |
| 20:38 | <annevk> | Hixie, oh, maybe you're right, but maybe it should use http://dev.w3.org/2006/webapi/WebIDL/#idl-DOMString instead? |
| 20:38 | <annevk> | that is http://dev.w3.org/2006/webapi/WebIDL/#dfn-obtain-unicode |
| 20:39 | <annevk> | would make sense if we use either that consistently or consistently throw |
| 20:39 | <Hixie> | we could do that too |
| 20:39 | <Hixie> | i don't really mind either way |
| 20:39 | <franksalim> | Hixie, i see |
| 20:42 | annevk | can't really think of all the parts of the platform that use it |
| 20:42 | <MikeSmith> | JonathanNeal: fwiw, http://dev.w3.org/html5/markup/base.html updated as well |
| 20:42 | <annevk> | I guess this is one reason why it'd be nice to have an API for specs |
| 20:42 | <JonathanNeal> | That looks great, thanks MikeSmith. |
| 20:43 | <MikeSmith> | thanks again for catching it |
| 20:46 | <JonathanNeal> | Oh boy, now I can finally get that tee shit "All your <base /> are belong to us" |
| 20:46 | <JonathanNeal> | I mean ... shirt |
| 20:55 | <annevk> | update on hybi: at this point people are reiterating mailing list discussion |
| 20:56 | <franksalim> | annevk, probably because nothing was ever decided conclusively on list |
| 20:56 | <annevk> | i doubt we'll get there now :) |
| 20:56 | <franksalim> | i'm ever optimistic |
| 20:57 | <annevk> | nice |
| 20:57 | annevk | wants that back |
| 20:59 | gsnedders | lost a fair bit of unsaved data today from his computer crashing. gah. |
| 21:02 | <MikeSmith> | annevk, franksalim - who's scribing? |
| 21:03 | <franksalim> | MikeSmith, http://etherpad.com/FWABRLTLpK |
| 21:04 | <JonathanNeal> | Anahiem, why that's where I am, more or less. |
| 21:04 | <MikeSmith> | franksalim: I meant, who's the main person writing there? |
| 21:05 | <MikeSmith> | I realize it's world-writable |
| 21:05 | <franksalim> | i do not know |
| 21:05 | <MikeSmith> | k |
| 21:05 | <franksalim> | i can see him, but i do not know his name |
| 21:05 | <annevk> | this meeting just got a bit weirder |
| 21:09 | <JonathanNeal> | I wondered that, when I saw you had an etherpad link, I thought "somebody's gonna be like UNICORNS" |
| 21:11 | <annevk> | oh versioning |
| 21:16 | <annevk> | seems we went right past that discussion |
| 21:17 | <annevk> | though it was mentioned it's best practice to have versioning o_O |
| 21:17 | <jgraham> | (just as a personal opinion re: websockets, I don't think counting bytes in a string is at all hard compared to implementing the sockets stuff. So I don't think using simplicity as an argument for having terminators rather than explicit lengths is really a good one) |
| 21:19 | <jgraham> | Also, etherpad for live minutes seems good; better than irc |
| 21:19 | <annevk> | but it's one of the bits you'd likely implement yourself |
| 21:19 | <annevk> | whereas the sockets bit is likely copied from a tutorial on sockets |
| 21:19 | <franksalim> | as opposed to a tutorial on websockets? using a websocket library? |
| 21:20 | <jgraham> | annevk: To do anything non-trivial with sockets seems harder than just copying a tutorial |
| 21:20 | <jgraham> | at least all the tutorials I have read are like "this is a simple echo server" |
| 21:20 | <franksalim> | my opinion is that cut and paste code on the server is not nearly as much a problem as it is with markup |
| 21:21 | <annevk> | you should talk to one of the security guys from Yahoo! |
| 21:21 | <jgraham> | Which is quite different to having to actually do non-trivial stuff with multiple simultaneous clients |
| 21:22 | <franksalim> | i am not denying the existence of bugs |
| 21:23 | <jgraham> | The main issue is that it is reasonably likely that typical western users would only test with ascii. However having a testsuite avaliable would help with that a great deal |
| 21:25 | <TabAtkins> | Anyone have suggestions for a good irc program on linux? Preferably easily skinnable? |
| 21:26 | <Hixie> | irssi |
| 21:27 | <TabAtkins> | Argh, these repos are useless. >_< |
| 21:27 | <jgraham> | There is also supposed to be some thing that you can run as a daemon on one machine and connect to via text or graphical clients |
| 21:27 | <jgraham> | But I forget the name |
| 21:28 | <jgraham> | (if the curses nature of irssi puts you off) |
| 21:28 | <annevk> | Opera! |
| 21:28 | <tabatkin1> | Hrm, probably puts me off. |
| 21:28 | <tabatkin1> | Can I just run chatzilla on linux? |
| 21:28 | tabatkin1 | doens't know. |
| 21:29 | <Dashiva> | Run mIRC in wine? :P |
| 21:29 | <daedb> | Quassel? |
| 21:29 | <franksalim> | tabatkin1, don't like xchat? |
| 21:29 | <MikeSmith> | TabAtkins: XChat |
| 21:29 | <TabAtkins> | franksalim: Haven't tried it yet. At the moment, connecting with pidgin, which sucks. |
| 21:29 | <jgraham> | Seriously, being able to keep the client running all the time is teh awesome |
| 21:32 | <TabAtkins> | Argh, now what's the server name of the w3c? |
| 21:32 | <annevk> | irc:w3.org:80 |
| 21:32 | <annevk> | irc.w3.org:80 |
| 21:32 | <annevk> | i mean |
| 21:32 | <TabAtkins> | Hrm, xchat won't recognize it. |
| 21:33 | <jgraham> | Maybe I mean smuxi |
| 21:33 | <TabAtkins> | Looks like it's not successfully parsing out the port? |
| 21:34 | <jgraham> | iirc the way you enter ports in xchat is odd |
| 21:34 | <franksalim> | TabAtkins, works for me irc.w3.org/80 |
| 21:34 | <franksalim> | slash instead of : |
| 21:34 | <TabAtkins> | "/server irc.w3.org/80";? |
| 21:34 | <annevk> | o_O |
| 21:35 | <TabAtkins> | ?_? |
| 21:36 | gavin | would not expect irc.w3.org/80 to work anywhere |
| 21:36 | <TabAtkins> | Man, xchat sucks too. I DECLARE IT SO. |
| 21:36 | <gavin> | unless you wanted to join channel #80 on irc.w3.org:6667 |
| 21:36 | TabAtkins | goes to try and get chatzilla working, since he knows how that works. |
| 21:37 | <hober> | what ever happened to port 194? |
| 21:40 | <tabatkins__> | There, chatzilla is up. Woo! |
| 21:40 | <tabatkins__> | Finally a sensible client. |
| 21:40 | <Dashiva> | I'm not going to take the obvious opening there... |
| 21:40 | <tabatkins__> | Now I just need to go grab my skin from my other laptop and I'll be set. |
| 21:49 | <Philip`> | "Bug 7034" |
| 21:49 | <Philip`> | Hooray for descriptive mailing list thread titles |
| 21:50 | <othermaciej> | if you post on that thread I'd suggest subject change |
| 21:53 | Dashiva | ponders making the list archives autolink bugs and issues |
| 22:15 | <annevk> | fixed the title |
| 22:16 | <annevk> | oh hey |
| 22:16 | <annevk> | shelley is leaving us again |
| 22:16 | <annevk> | http://twitter.com/shelleypowers/status/10993810486 |
| 22:16 | <annevk> | http://twitter.com/shelleypowers/status/10995234211 |
| 22:17 | <annevk> | "The spec is absolute crap" |
| 22:18 | <annevk> | if it was, surely authors would not like it and browser vendors would not implement it? |
| 22:18 | <TabAtkins_> | She uses a peculiar definition of "crap". |
| 22:18 | <Hixie> | dunno, never stopped them before :-P |
| 22:18 | <annevk> | hehe good point |
| 22:22 | <othermaciej> | when it comes to browser engineering, we spell quality with a *capital* K |
| 22:23 | <Dashiva> | Kwality? Or Kuality? |
| 22:24 | <Hixie> | Kaqlity! |
| 22:24 | <Hixie> | er |
| 22:24 | <Hixie> | i can't even spell my misspellings |
| 22:24 | <Hixie> | nevermind |
| 22:24 | <Hixie> | on another note: the hybi meeting was somewhat uneventful, fwiw, though i was a little concerned that the response to someone mentioning versioning was "well versioning is easy, we'll just copy some boilerplate from another draft to add version support" or some such |
| 22:25 | <annevk> | yeah... or maybe we don't |
| 22:25 | <othermaciej> | Dashiva: we feel that the exact spelling is an opportunity for differentiation and should be left up to the implementation |
| 22:34 | <TabAtkins_> | hsivonen: Is Leif making stuff up when he suggests that you are saying that @hidden should, generally, have a default CSS of something other than display:none? |
| 22:47 | <othermaciej> | TabAtkins_: the spec certainly says the default rendering of @hidden should be display: none |
| 22:47 | <TabAtkins_> | I know. Leif is suggesting that hsivonen said something different, which I suspect is him being crazy. |
| 22:47 | <Dashiva> | Hum, Leif repeated the <p /> thing... |
| 22:49 | Philip` | finds the effort/reward ratio means he chooses never to read anything Leif ever says |
| 22:49 | <Philip`> | (which is a shame since sometimes he says things that seem useful and correct) |
| 22:49 | <TabAtkins_> | Philip`: Yeah, generally true. This thread I'm in was just so batshit that I had to correct it. |
| 22:50 | <Hixie> | someone was wrong on the internet? |
| 22:51 | <Dashiva> | Is being wrong a violation of any conformance criteria? |
| 22:51 | <TabAtkins_> | Only in my dreams. |
| 22:52 | <Philip`> | You dream about conformance criteria? |
| 22:53 | <Hixie> | don't you? |
| 22:53 | <TabAtkins_> | ...yes? |
| 22:53 | <TabAtkins_> | That's why you're not a spec editor, Philip` |
| 22:57 | <jgraham> | Philip` never sleeps so he can't dream |
| 22:59 | <Dashiva> | There is an installed base of RDFa-in-HTML that we can't risk destabilizing, how nice |
| 22:59 | <dbgi> | anyone in here who knows freebsd well and enjoys setting up and hosting web sites? |
| 23:00 | <Hixie> | does anyone implemented SharedWorker? I want to test zcorpan's suggested new examples |
| 23:02 | <annevk> | we have but it's not out in public yet I think |
| 23:02 | <annevk> | doesn't Chrome have it too? |
| 23:02 | <annevk> | anyways, bed time; nn |
| 23:02 | <Hixie> | nn |
| 23:03 | <Dashiva> | Sounds like a plan... |