| 00:41 | <Yuhong> | http://d0cs4vage.blogspot.com/2011/06/insecticides-dont-kill-bugs-patch.html |
| 06:35 | <variable> | what is the appropriate way to mark up column headers in a table? |
| 06:36 | <ryanseddon> | thead with th and scope attribute set to col |
| 06:37 | <variable> | ryanseddon: what if both columns and rows are headers? |
| 06:38 | <ryanseddon> | put row headers in a th and give them scope="row" |
| 06:38 | <variable> | ah ok cool |
| 06:39 | <variable> | so it goes <thead> <th scope="row">row header</th>... <th scope="col">column head</th></thead> <tr><td>fooooo</td></tr> ? |
| 06:47 | <ryanseddon> | yeah that looks ok |
| 06:53 | <variable> | thanks! |
| 06:55 | <ryanseddon> | variable: you might want to look over this too http://www.usability.com.au/resources/tables.cfm |
| 08:16 | <MikeSmith> | http://scottbw.wordpress.com/2011/02/17/converting-chrome-installed-web-apps-into-w3c-widgets/ |
| 09:59 | <annevk> | hsivonen, did you see the changes to MetaExtensions on the wiki? |
| 10:00 | <hsivonen> | annevk: I saw something yesterday but was too focused on about:blank to act before I had to run away from my desk |
| 10:01 | <annevk> | Someone moved the dc stuff back to being okay afaict |
| 10:02 | <hsivonen> | annevk: and zapped a Bing keyword by accident apparently |
| 10:04 | <annevk> | I hit rollback and it automatically did things... |
| 10:04 | <annevk> | I thought you would get a confirmation page for that first |
| 10:04 | <annevk> | and you know, a button |
| 10:05 | <hsivonen> | annevk: why did you revert the dc stuff? |
| 10:06 | <annevk> | it was not incomplete? |
| 10:06 | <hsivonen> | annevk: a bit incomplete still, but at this point, it would make more sense to fix it rather than to revert, IMO |
| 10:07 | <annevk> | undid my changes |
| 10:09 | <annevk> | it looks like more was deleted by him btw |
| 10:10 | <annevk> | +1,338 and -2,553 ... |
| 10:10 | <annevk> | oh maybe he did not just move but also change the contents |
| 10:10 | <annevk> | or she |
| 10:11 | <hsivonen> | annevk: yes, (s)he made those entries suck less |
| 10:11 | <hsivonen> | annevk: I'll complete the spec trail now |
| 10:11 | <hsivonen> | annevk: I think it's ok to have a subset of dcterms registered |
| 10:11 | <hsivonen> | annevk: if the rest of the terms aren't interesting enough for anyone to try to register, let's leave them unregistered |
| 10:12 | <hsivonen> | I'm not going to spend my day registering the long tail of dcterms |
| 10:12 | <annevk> | so is WHATWG still doing <meta> and microformats.org only doing rel=""? |
| 10:13 | <zcorpan> | yes |
| 10:13 | <annevk> | k cool |
| 10:13 | <zcorpan> | that was the Decision |
| 10:16 | <hsivonen> | dcterms.issued and dcterms.modified suck because DC sucks and doesn't spec the date format |
| 10:30 | <hsivonen> | yay, of the registered dcterms keywords, only the underdefined date-related ones aren't redundant with HTML's built-in features |
| 10:39 | <annevk> | http://annevankesteren.nl/2011/06/initializing-events is the use of a hyphen in user-agent correct here? |
| 10:41 | <hsivonen> | annevk: no |
| 10:41 | <hsivonen> | annevk: user agent-generated |
| 10:41 | <hsivonen> | would be correct |
| 10:42 | zcorpan | finds "user-agent defined" in html5 |
| 10:43 | <annevk> | whoa |
| 10:43 | <hsivonen> | zcorpan: maybe that's allowed in en-US-x-Hixie |
| 10:43 | <annevk> | that looks weird |
| 10:45 | <jgraham> | hsivonen: user agent-generated looks odd to me |
| 10:45 | <jgraham> | whereas user-agent generated looks fine |
| 10:46 | <zcorpan> | annevk: events event |
| 10:46 | annevk | removes the hyphen |
| 10:46 | <hsivonen> | jgraham: what rule do you use to justify that kind of hyphenation? |
| 10:46 | <annevk> | zcorpan, ta |
| 10:47 | <hsivonen> | I left my copy of CMoS in Finland, so I can't cite chapter and verse, but the chapter and verse is probably in the IRC logs already |
| 10:47 | <hsivonen> | when annevk asked about the hyphenation of English compound adjectives the previous time |
| 10:48 | <annevk> | :) trips me up |
| 10:48 | <jgraham> | hsivonen: "it looks right". But I think in this case the hyphen has the effect of increasing the tightness of binding. So user agent-generated looks like user (agent generated) whereas user-agent generated looks like (user agent) generated. And indeed I believe that compound nouns often go user agent -> user-agent -> useragent |
| 10:48 | <hsivonen> | though CMoS should be taken with a grain of salt, since it advocates the use on an en dash is Civil War–era |
| 10:49 | jgraham | is very skeptical of style guides in general |
| 10:49 | <hsivonen> | jgraham: oh, I agree that the appearance of the tightness of binding sucks |
| 10:49 | <hsivonen> | s/is/in/ |
| 10:49 | <jgraham> | They often seem to be written by people with a very prescriptive view of linguistics |
| 10:49 | <jgraham> | and no understanding of language evolution |
| 10:50 | <hsivonen> | the job of a style guide is to be prescriptive |
| 10:51 | <annevk> | maybe it should be user-agent-generated |
| 10:51 | <jgraham> | But languages and fashions change. And indeed aren't that consistent to begin with. So the style guide will necessarily be wrong |
| 10:51 | <hsivonen> | annevk: I think that would make sense but isn't what CMoS would say |
| 10:51 | <annevk> | from what I read here in the The Economist Style Guide user-agent generated seems fine |
| 10:52 | <hsivonen> | annevk: whoa! |
| 10:52 | <annevk> | they have "public-sector borrowing requirements" as one of the examples |
| 10:52 | <gsnedders> | Even descriptiveness is becoming harder nowadays… It's now far harder for new words to get words into the OED, for example, as you need three written, published, sources. The majority of publishers use spell-checkers in their editing process, so the words will be changed. |
| 10:52 | <zcorpan> | maybe -user agent generated- ? |
| 10:53 | <jgraham> | hsivonen: Given the above it is hardly surprising that two style guides disagree |
| 10:53 | annevk | forgot he had a style guide |
| 10:54 | <zcorpan> | it's easier in swedish. användaragentgenererad |
| 10:55 | <annevk> | no idea how this would go in Dutch |
| 10:55 | <jgraham> | zcorpan: Only because Swedes haven't heard of space characters :p |
| 10:55 | <hsivonen> | jgraham: maybe the APA Publication Manual is a phychology experiment |
| 10:56 | <zcorpan> | user agent gegenereerd? |
| 10:57 | <annevk> | zcorpan, heh, I was stuck at "user agent" |
| 10:57 | <annevk> | gebruikersagent :p |
| 10:58 | <zcorpan> | good thing with dutch is that you just use the english word if it doesn't translate well :) |
| 10:58 | <annevk> | yeah |
| 11:00 | <Aleksandr> | q |
| 11:27 | <annevk> | Is today a holiday or something? Almost nothing seems to be going on... |
| 11:34 | <karlcow> | annevk: http://en.wikipedia.org/wiki/June_23#Holidays_and_observances ;) |
| 11:36 | <karlcow> | oh it gave me the opportunity to discover this page http://en.wikipedia.org/wiki/Portal:Current_events |
| 11:57 | <annevk> | heycam is only on email... |
| 12:13 | <zcorpan> | http://www.bookofspeed.com/chapter4.html - "Any XML file including RSS and Atom feeds, SVG graphics, IE's .htc files, DTD (Document Type Definition) and basically anything else that ends in *ML" |
| 12:14 | <zcorpan> | now why would you have a DTD if you cared about performance |
| 12:32 | <jgraham> | zcorpan: Why would you have a DTD anyway? |
| 12:33 | <zcorpan> | yeah |
| 12:34 | <annevk> | published by W3C, must be good |
| 12:36 | <jgraham> | annevk: With that in mind I plan to add EmotionML to all my webpages so people know when I am being sarcastic. |
| 12:38 | <annevk> | seems you forgot some markup right there! |
| 12:38 | <annevk> | oh wait |
| 12:59 | <annevk> | Can someone help me with hg? |
| 12:59 | <annevk> | I mistakenly used hg update -c and now the work I did in my commits is lost |
| 13:00 | <annevk> | They are however still in my local log (have not pushed yet) so I should be able to turn things around I think if I know how |
| 13:03 | <Philip`> | annevk: Sounds like you just updated the working directory to a changeset that doesn't include your commits, so you should be able to update back to it by specifying your latest changeset |
| 13:17 | <karlcow> | rha hsivonen sent a link to the long version |
| 13:23 | <annevk> | Philip`, thanks; didn't really work out, but I resolved it somehow |
| 13:23 | <Philip`> | Oh |
| 13:29 | <annevk> | Philip`, my clumsyness, not your advice |
| 13:29 | <annevk> | ooh |
| 13:29 | <annevk> | partial is not the new [Supplemental] |
| 13:29 | <annevk> | s/not// |
| 13:32 | <annevk> | Ms2ger`, http://html5.org/specs/dom-parsing.html#dom-range-createcontextualfragment markup is borked |
| 13:45 | <MikeSmith> | it seems the state of things with Web Storage is sort of as was the state of things pre-transaction-support MySQL |
| 13:46 | <MikeSmith> | in that, lack of transaction support in MySQL made is quite possible to end up with a corrupted DB |
| 13:48 | <annevk> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=12230 was there ever a bug filed on Gecko? |
| 13:48 | <annevk> | and other browsers? |
| 13:53 | <gsnedders> | MikeSmith: MyISAM still supports no transactions, AFAIK. Just other table formats do. |
| 13:53 | <MikeSmith> | oh |
| 13:56 | <jgraham> | I see gsnedders is working towards a ghostbusters-like grammar |
| 14:05 | <The_8472> | myisam doesn't support transactions, but mutating actions take an exclusive table lock |
| 14:05 | <The_8472> | so you don't end up with concurrent writes |
| 14:06 | <Philip`> | I think it's consistent with the view that each SQL statement is implicitly executed in an ACID transaction, you just can't do multi-statement transactions |
| 14:09 | <annevk> | Inbox 750 o_O |
| 14:10 | <annevk> | I think this is actually lower than before I went away |
| 14:17 | <koskoz> | !topic |
| 14:17 | <volkmar> | someone knows if there is an implementation of the HTML command api by any vendor? |
| 14:17 | <annevk> | haven't heard of any |
| 14:45 | <Ms2ger`> | annevk, thanks, fixed |
| 14:45 | <Ms2ger`> | (Though I kind of liked the orange :)) |
| 14:48 | <annevk> | Netherlands ftw |
| 14:48 | <Ms2ger`> | No comment ): |
| 14:48 | <Ms2ger`> | ) |
| 14:48 | Ms2ger` | tries that again |
| 14:48 | <Ms2ger`> | :) |
| 14:52 | <annevk> | Ms2ger`, BAD_BOUNDARY_POINTS_ERR or some such became INVALID_STATE_ERR? |
| 14:53 | <annevk> | I guess we should ask File API to also use DOMException and request codes they'd like to see added... |
| 14:54 | <Ms2ger> | Yeah, there was one one user |
| 14:54 | <Ms2ger> | Seemed a bit wasteful |
| 16:26 | <hsivonen> | hah. even at Google, marketing seems to trust PDF over HTML: http://www.google.com/help/hc/images/Google_Places_OneSheeter.pdf |
| 16:26 | <hsivonen> | does anyone know if Google Maps picks up business locations from coordinates in meta tags on Web sites? |
| 16:31 | <annevk> | hmm, heycam?! |
| 16:32 | <annevk> | http://dev.w3.org/2006/webapi/WebIDL/#dfn-partial-interface should be updated to say it can also be more than one spec |
| 16:56 | <annevk> | Ms2ger, you mentioned that matches bug no? |
| 16:56 | <annevk> | Ms2ger, IDL bug in cssom-view |
| 17:08 | <karlcow> | hsivonen: Note that we do not use locational meta tags (like "geo.position" or "distribution") or HTML attributes for geotargeting. While these may be useful in other regards, we've found that they are generally not reliable enough to use for geotargeting. |
| 17:08 | <karlcow> | — http://googlewebmastercentral.blogspot.com/2010/03/working-with-multi-regional-websites.html |
| 17:08 | <annevk> | Ms2ger, concept-document-character-encoding and concept-document-media-type? |
| 17:08 | <hsivonen> | karlcow: thanks |
| 17:09 | <karlcow> | I do not know what yahoo, or bing do |
| 17:13 | <annevk> | Ms2ger, also, we should probably introduce the term "attribute" to mean "Attr object" |
| 17:40 | <Ms2ger> | annevk, attribute makes sense, concept-s look good, not sure what CSSOM bug you mean |
| 17:41 | <annevk> | hmm |
| 17:41 | <annevk> | removing CaretPosition.range is non-trivial |
| 17:48 | <AryehGregor> | Hixie, what's the Bugzilla query you use for deciding the order of the bugs you look at? It might be helpful to put at <http://wiki.whatwg.org/wiki/Bugzilla_conventions>;. |
| 18:49 | <annevk> | http://www.w3.org/mid/BANLkTi=cfgCzsYkTX8a-68CPW6bdHNpdvA⊙mgc anyone a good name for this? |
| 18:50 | <annevk> | Gecko has "attachment" atm and WebKit "affinity" |
| 18:50 | <annevk> | neither seem particularly clear for a boolean |
| 18:51 | <Philip`> | Sticky-direction |
| 18:51 | <AryehGregor> | So it seems the status quo with copy/paste is that browsers refuse to support it and therefore everyone uses Flash? |
| 18:52 | <annevk> | Philip`, is that a serious suggestion? |
| 18:52 | annevk | cannot tell |
| 18:54 | <Philip`> | Doesn't matter whether it's serious, just whether it's any clearer than the alternatives |
| 18:59 | <rafaelw> | morning all. question about assigning null to HTMLInputElement.value |
| 19:00 | <rafaelw> | gecko reacts by setting the value property to the empty string. |
| 19:00 | <rafaelw> | webkit resets the value property the value of the attribute. |
| 19:00 | <rafaelw> | i'm having trouble locating the relevant place in the spec. |
| 19:01 | <annevk> | what does resetting the value property mean? |
| 19:01 | <annevk> | (also depends on an unresolved Web IDL issue) |
| 19:01 | <rafaelw> | if I had <input value="foo">, assigning myInput.value = null |
| 19:02 | <Philip`> | http://dev.w3.org/2006/webapi/WebIDL/#es-DOMString should be where it's defined, I think |
| 19:02 | <rafaelw> | gecko: myInput.value === ''; |
| 19:02 | <rafaelw> | webkit: myInput.value === 'foo' |
| 19:02 | <Ms2ger> | Should be "null", probably |
| 19:02 | <AryehGregor> | value is not nullable, so per spec, null will be cast to a string. |
| 19:02 | <AryehGregor> | Either "" or "null". |
| 19:03 | <AryehGregor> | It looks like current WebIDL says "null". |
| 19:03 | <Ms2ger> | It does |
| 19:07 | <smaug____> | webkit's behavior is very surprising |
| 19:07 | <smaug____> | would be good to test what Opera and IE do |
| 19:07 | <smaug____> | if they assign "" to the value |
| 19:07 | <smaug____> | would make sense to spec that |
| 19:09 | <rafaelw> | trying to get to my windows box right now =-) |
| 19:11 | <smaug____> | Opera behaves like Gecko |
| 19:11 | <smaug____> | http://mozilla.pettay.fi/moztests/inputvalue.html |
| 19:12 | <rafaelw> | IE assigns 'null' |
| 19:12 | <rafaelw> | it also set the attribute value to 'null' |
| 19:13 | <rafaelw> | which gecko didn't do. |
| 19:13 | <AryehGregor> | I should update my reflection tests to test this, now that we're clear on what we want for null. |
| 19:13 | <AryehGregor> | Gecko converted the null to an empty string, IE converted to "null". |
| 19:13 | <smaug____> | AryehGregor: is it already clear in which cases we want "null" and in which cases ""? |
| 19:14 | <AryehGregor> | smaug____, the current WebIDL spec says "null" is the default now. Exceptions need to be specced on a case-by-case basis by adding [TreatNullAs=EmptyString] or such. |
| 19:14 | <AryehGregor> | So it would be good to test so we can find the exceptions. |
| 19:14 | <smaug____> | WebIDL says what the default is, but do we know what the exceptions are |
| 19:15 | <AryehGregor> | No, which is why tests are good, to find out the exceptions. |
| 19:15 | <AryehGregor> | I'll do that shortly. |
| 19:17 | <rafaelw> | gecko's behavior seems the most useful. |
| 19:18 | <AryehGregor> | Making null become an empty string makes the most sense, yeah. |
| 19:18 | <AryehGregor> | But I'm not sure if it's compatible in general. |
| 19:20 | <annevk> | smaug____, I think that might be our recent change to align with Gecko/WebKit null-value handling in DOM bindings |
| 19:20 | <MikeSmith> | friends, I made an attempt at organizing current web-platform specs into categories |
| 19:20 | <MikeSmith> | http://people.w3.org/mike/web-platform.html |
| 19:20 | <MikeSmith> | comments welcome |
| 19:21 | <Philip`> | No WebIDL? |
| 19:21 | <MikeSmith> | oh |
| 19:21 | <MikeSmith> | will add it |
| 19:21 | <smaug____> | MikeSmith: why you have Web Audio API, but not Audio API? |
| 19:21 | <Philip`> | There's some obsolete links to /TR/ |
| 19:21 | <MikeSmith> | thanks |
| 19:22 | <Philip`> | (At least canvas) |
| 19:22 | <smaug____> | s/Audio API/Audio Data API/ |
| 19:22 | <MikeSmith> | smaug____: is there an actual spec for it? |
| 19:23 | <smaug____> | MikeSmith: https://wiki.mozilla.org/Audio_Data_API |
| 19:23 | <Ms2ger> | Bonus features? |
| 19:23 | Ms2ger | likes that title |
| 19:23 | <MikeSmith> | smaug____: thanks |
| 19:23 | <smaug____> | MikeSmith: Web Audio API isn't in anyway more "spec" than Audio Data API |
| 19:23 | <Ms2ger> | MikeSmith, ElementTraversal is now specified in DOM Core |
| 19:24 | <Ms2ger> | And you're missing XHR2 |
| 19:24 | <smaug____> | I assume neither of those APIs will become a recommendation |
| 19:24 | <MikeSmith> | Ms2ger: thanks |
| 19:24 | <Ms2ger> | And CSS |
| 19:24 | <Ms2ger> | Oh, you don't |
| 19:26 | <MikeSmith> | smaug____: will add audio data api when i get back to my vim |
| 19:26 | <smaug____> | ok, thanks |
| 19:26 | <MikeSmith> | smaug____: thanks for the heads-up about it? |
| 19:27 | <MikeSmith> | Ms2ger: has anybody implemented xhr2? |
| 19:27 | <smaug____> | XHR2? |
| 19:27 | <Ms2ger> | I think we're doing parts of it |
| 19:27 | <smaug____> | progress events for XHR are from XHR2 |
| 19:27 | <MikeSmith> | smaug____: s/about it?/about it./ |
| 19:28 | <MikeSmith> | Ms2ger: ok, will add it |
| 19:28 | <Ms2ger> | responseType, too? |
| 19:28 | <MikeSmith> | ? |
| 19:28 | <MikeSmith> | what spec? |
| 19:28 | <annevk> | Cross-origin XHR is XHR2 too |
| 19:28 | <annevk> | XHR is just to please some people, XHR2 is the real deal |
| 19:29 | <annevk> | (XHR is generated with some switches from the XHR2 draft) |
| 19:30 | <MikeSmith> | Philip`: i will change all the non-REC TR links to EDs instead |
| 19:30 | <Ms2ger> | MikeSmith, were you going to list HTTP and friends as well? |
| 19:30 | <MikeSmith> | no |
| 19:30 | <Ms2ger> | XML and xml:base? |
| 19:31 | <Ms2ger> | DOM Traversal |
| 19:31 | <MikeSmith> | i think i wont put anything un-sexy |
| 19:31 | <MikeSmith> | dom traversal? |
| 19:31 | <smaug____> | XSLT |
| 19:31 | <smaug____> | XPath |
| 19:31 | <MikeSmith> | hmm |
| 19:31 | <Ms2ger> | And http://dev.chromium.org/developers/web-platform-status, fyi |
| 19:31 | <MikeSmith> | i guess i got to add those |
| 19:32 | <smaug____> | Range is part of the same spec as traversal http://www.w3.org/TR/DOM-Level-2-Traversal-Range/Overview.html#contents |
| 19:32 | <MikeSmith> | as un-sexy as they may be |
| 19:32 | <Ms2ger> | smaug____, not anymore ;) |
| 19:34 | <smaug____> | Ms2ger: is there already somewhat stable and reviewed spec for Range? |
| 19:34 | <smaug____> | new spec |
| 19:35 | <Ms2ger> | For some value of somewhat :) |
| 19:36 | <smaug____> | surprising or not, the old DOM Range is pretty consistent spec |
| 19:36 | <smaug____> | internally |
| 19:36 | <smaug____> | terrible to read though |
| 19:39 | <annevk> | smaug____, replacement is here: http://html5.org/specs/dom-range.html |
| 19:40 | <annevk> | smaug____, needs review still I believe |
| 19:40 | <Ms2ger> | And, well, speccing |
| 19:43 | <MikeSmith> | I had in mind to limit this list to specs that actually expose methods to web applications/developers |
| 19:44 | <MikeSmith> | in the more general sense of methods |
| 19:44 | <MikeSmith> | stuff they can use in scripts or in markup |
| 19:49 | <AryehGregor> | So first of all, it looks like Gecko converts null to "" instead of "null" across the board for setAttribute(), which contradicts all other browsers. |
| 19:50 | <Ms2ger> | Gecko converts null to "" pretty much everywhere |
| 19:51 | <Ms2ger> | The main exceptions are a few things I changed and code that works on the JS values directly |
| 19:58 | <AryehGregor> | Seems WebKit is also a fan of converting null to "", except where it unsets the attribute. |
| 19:58 | <Ms2ger> | For webkit, the exception would be indexeddb |
| 19:58 | <AryehGregor> | Opera follows spec on this point everywhere. IE does in most cases too, but I think there are some exceptions. |
| 20:04 | <annevk> | pushed a few https://bitbucket.org/ms2ger/dom-core edits |
| 20:04 | <annevk> | we now have concept-document/element/attribute; hurray |
| 20:05 | <AryehGregor> | Okay, so I have a list of all the reflected attributes where IE9 casts null to "". Do we want to spec these? In all those cases, the only browser that casts to "null" is Opera. |
| 20:05 | AryehGregor | files a bug regardless |
| 20:05 | <annevk> | interesting |
| 20:05 | <annevk> | I thought we changed to use "" pretty much exclusively |
| 20:09 | <Ms2ger> | annevk, doesn't look like I have IANACHARSET defined |
| 20:10 | <annevk> | one day it will be obsolete I guess |
| 20:14 | <AryehGregor> | Okay, so why does WebIDL follow IE/Opera here instead of Gecko/WebKit? Don't we usually prefer to follow Gecko/WebKit? |
| 20:14 | <Ms2ger> | IE complained |
| 20:15 | <Ms2ger> | And it matches JS |
| 20:15 | <annevk> | Gecko complained too |
| 20:17 | <annevk> | "[TreatNullAs=EmptyString] DOMString" shouldn't that be "[TreatNullAs=EmptyString] DOMString?" ? |
| 20:17 | <annevk> | note the trailing ? |
| 20:17 | <Ms2ger> | No |
| 20:18 | <Ms2ger> | If you have DOMString?, you handle null in the algorithm |
| 20:20 | <annevk> | k |
| 20:21 | <AryehGregor> | Gecko complained that we convert null to "" instead of "null", when Gecko itself consistently converts null to ""? |
| 20:22 | <annevk> | Gecko did not want the IDL default to be "" |
| 20:22 | <annevk> | consequently, I suspect they want to change their API |
| 20:22 | <annevk> | s |
| 20:22 | <annevk> | at least, I hope they do |
| 20:28 | <mpilgrim> | i just finished fixing webkit's indexeddb to make null -> "null", don't you dare change it again |
| 20:29 | <mpilgrim> | even though i hate it and think it's a stupid default |
| 20:32 | <Dashiva> | This is one place where exceptions have the upper hand :) |
| 20:32 | <Ms2ger> | The web? |
| 20:39 | <Dashiva> | Ms2ger: No, places where you don't want to handle edge cases |
| 20:50 | <annevk> | http://dev.w3.org/2006/webapi/WebIDL/#dfn-partial-interface should be updated to say it can also be more than one spec |
| 20:51 | <annevk> | heycam, ^^ |
| 20:52 | <heycam> | I think the idea of things being defined before other things probably needs to be revised in general |
| 20:52 | <heycam> | not everyone is putting all their IDL in one big fragment |
| 20:52 | <heycam> | and there's no real difference between multiple fragments within the one spec and multiple specs |
| 20:53 | <heycam> | I am interested to see if people other than Hixie want to use partial |
| 20:53 | <heycam> | and whether then I think it's an appropriate use of it or not :) |
| 20:53 | <annevk> | I told you once before http://dev.w3.org/csswg/cssom-view/ uses it |
| 20:53 | <annevk> | http://html5.org/specs/dom-range.html uses it too |
| 20:53 | <annevk> | http://html5.org/specs/dom-parsing.html too |
| 20:54 | <Ms2ger> | And http://html5.org/specs/dom-parsing.html |
| 20:54 | <annevk> | hah |
| 20:54 | <heycam> | I see |
| 20:55 | <heycam> | to me it seems like that should be `[NoInterfaceObject] interface WindowCSS { ... }; Window implements WindowCSS;` |
| 20:55 | <heycam> | but I will think about it some more |
| 20:55 | heycam | has a telcon now |
| 20:55 | <heycam> | oh congrats on bug 666666 btw Ms2ger :) |
| 20:56 | <Ms2ger> | heycam, no, we're not in the nineties anymore |
| 20:56 | <Ms2ger> | heycam, and thanks ) |
| 20:56 | <Ms2ger> | :) |
| 20:58 | <annevk> | heycam, whoa please no |
| 20:59 | <annevk> | soooo ugly |
| 20:59 | <annevk> | heycam, you could name it "additional" instead of "partial" though |
| 21:33 | <twisted> | yo |
| 21:33 | <twisted> | #ja-mainnav div.main div.ja-megamenu ul.megamenu li.active a:hover { |
| 21:33 | <twisted> | doesn't work |
| 21:33 | <twisted> | but |
| 21:33 | <twisted> | #ja-mainnav div.main div.ja-megamenu ul.megamenu li.active a { |
| 21:33 | <twisted> | does... |
| 21:33 | <twisted> | and I'm totally beat |
| 21:34 | <twisted> | css validator gives me a lexicon error |
| 21:34 | <twisted> | tried different encodings, (ascii, latin1, utf8 (default)) |
| 21:35 | <paul_irish> | twisted: come over to #html5 and ask that and we can help you out there |
| 21:40 | <paul_irish> | where can i find out more on why html5 video doesnt have an alpha channel? |
| 21:42 | <bga_> | i never hear that any video codeс supports rgba color |
| 21:42 | <bga_> | and i dont know why somebody need rgba video |
| 21:43 | <paul_irish> | in some webapps it'd be more effective to have small videos for effects (that i can make in After Effects) rather than do that sort of thing in canvas |
| 21:45 | <bga_> | i understand |
| 21:46 | <bga_> | paul_irish there is apng |
| 21:47 | <paul_irish> | yeah.. and you can Chroma Key out a solid color via 2d canvas pixel data |
| 21:48 | <paul_irish> | but.. i want that .. natively |
| 21:48 | <bga_> | or 2 videos 1st is rgb 2nd is greyscale as alpha channel and mix using canvas %) |
| 21:48 | <paul_irish> | yeah |
| 21:50 | <bga_> | paul_irish i guess patch to webm and theora is easiest way |
| 21:51 | <paul_irish> | i will email my webm dude. |
| 21:59 | <KevinMarks> | quicktime supports alpha channel video; has for over a decade |
| 22:02 | <KevinMarks> | it is often better to have a separate alpha channel matte rather than make it part of the video pixels |
| 22:33 | <Hixie> | dbaron: these media query listener things are implemented? |
| 22:34 | <dbaron> | Hixie, yes, in Gecko and I think in chrome (though my testcase didn't send the notifications in Chrome for some reason... though the addListener succeeded) |
| 22:34 | <Hixie> | aah |
| 22:34 | <Hixie> | when reading anne's e-mail i had no idea that (a) we were talking about running JS and (b) we were talking about matching existing behaviour |
| 22:34 | <Hixie> | makes more sense now |
| 22:34 | <dbaron> | Hixie, I'm not sure if the problem is the cssom-view spec's reference to the term or whether the term should be defined differently... |
| 22:35 | <Hixie> | so does gecko just fire these as soon as it can? i.e. it's a high priority task? |
| 22:35 | <Hixie> | or is there more subtlty? |
| 22:35 | <Hixie> | what happens if the callback causes it to change again, or can JS not trigger a media query change? |
| 22:36 | <Hixie> | oh, i see in your mail that you mention sometimes this happens synchronously |
| 22:36 | <dbaron> | what happens is that when things that can change media queries happen we post an event to the event queue saying that such things have happened (and then suppress other such events until that event fires) |
| 22:36 | <jamesr> | can you set media queries on iframes? |
| 22:36 | <Hixie> | not in the event loop at all |
| 22:36 | <dbaron> | then, *either* when that event fires *or* when we flush style |
| 22:36 | <dbaron> | we go through *all* of the media queries and see if they've actually changed |
| 22:36 | <dbaron> | and notify any of them that have |
| 22:37 | <Hixie> | sounds like cssom needs to define all the places that "flush style" as calling this and emptying the media query change task queue |
| 22:37 | <dbaron> | the test I sent on the thread is interesting because between two lines in the script the mq was false |
| 22:37 | <jamesr> | dbaron: is it possible for one handler to change the value of a later (or earlier) media query? |
| 22:37 | <dbaron> | Hixie, I'm not sure we really want to do that |
| 22:37 | <Hixie> | i thought you said it's what you _did_ do? |
| 22:37 | <Hixie> | as far as the event loop goes, if it's just that when tasks run they are high-priority, then the spec already supports that, just use a different task source |
| 22:37 | <dbaron> | yes, but we may change flushing in order to optimize |
| 22:37 | <Hixie> | ah |
| 22:37 | <Hixie> | well |
| 22:37 | <dbaron> | we want to flush as little as possible |
| 22:38 | <Hixie> | if it's running script, it needs to be interoperable, which means predictable, i.e. the spec has to say what all the UAs do |
| 22:38 | <jamesr> | dbaron: and what about when you have to resolve styles synchronously due to some script action? is that separate from the 'flush style' step in gecko? |
| 22:38 | <Hixie> | which i think is incompatible with changing it |
| 22:38 | <Hixie> | anyway, long story short, sounds like this is all not my problem :-P |
| 22:39 | <dbaron> | jamesr, we accumulate all the changes to notify before sending any of them, so a change made in one listener doesn't affect who gets notified that pass, although the script could cause another flush and nest a new pass inside the outer one |
| 22:39 | <dbaron> | jamesr, and "flush styles" is what resolving styles synchronously due to a script action triggers |
| 22:40 | <jamesr> | ok. i think the webkit impl would (or possibly does, i haven't looked) be slightly different, then, although perhaps it's not observable |
| 22:45 | <AryehGregor> | Hixie, what's the Bugzilla query you use for deciding the order of the bugs you look at? It might be helpful to put at <http://wiki.whatwg.org/wiki/Bugzilla_conventions>; so we can look it over and close some of the bugs if we get bored. |
| 22:45 | <Hixie> | i don't use bugzilla for that, i use my imap folders |
| 22:46 | <Hixie> | but it can be approximated to the sort order you get with "last change" |
| 22:46 | <AryehGregor> | Ah, okay. |
| 22:46 | <Hixie> | looks like the field is actually called "Changed" not "last change" |
| 22:46 | <Hixie> | but you get the idea |
| 22:59 | <Hixie> | clair: thanks for the work on the wiki |
| 23:00 | <clair> | No worries :) |
| 23:00 | <clair> | Looks like you've made use of it, do you still want more code or is that enough? |
| 23:01 | <Hixie> | i haven't studied what we have their yet |
| 23:01 | <Hixie> | there even |
| 23:01 | <Hixie> | so no idea :-) |
| 23:01 | <clair> | Ah right, I just saw that you wrote a few notes |
| 23:01 | <Hixie> | yeah that was just some stuff that i was thinking about in the shower |
| 23:01 | <Hixie> | figured i should write it down somewhere |
| 23:01 | <Hixie> | no idea if it makes any sense or not |
| 23:02 | <clair> | What's the best way for me to say "But this would be nice/I'd like to be able to do this"? on your notes? |
| 23:02 | <clair> | heh, some bits do at a glance :) |
| 23:02 | <Hixie> | if you have an agenda, best thing to do is to find examples of dialogs that do what you want to do and include them :-) |
| 23:03 | <clair> | I just would prefer modals to show centre of page rather than at cursor position, that's all really ;) |
| 23:04 | <clair> | (Though of course having them at cursor position is also useful..) |
| 23:04 | <Hixie> | ah that's where screenshots would be useful |
| 23:04 | <Hixie> | (if you do attach screenshots, make the window as small as you can while still showing the effect, we don't want 3000x2000 screenshots with a 300x200 dialog in the middle!) |
| 23:05 | clair | grins |
| 23:05 | <clair> | I'm sure I'll find some dialogs that I like :) |
| 23:06 | <Hixie> | :-) |
| 23:06 | <Hixie> | from what i've seen, facebook is a good source of dialogs |
| 23:06 | <Hixie> | though i don't have an account myself |
| 23:06 | <Hixie> | also google docs seems to be pretty dialog-happy |
| 23:07 | <benschwarz> | Hey Hixie |
| 23:07 | <Hixie> | benschwarz! |
| 23:07 | <Hixie> | 'sup |
| 23:07 | <karlcow> | Big Lebowsky is dialogs happy too |
| 23:07 | <benschwarz> | Doing good man. Hows things? |
| 23:07 | <clair> | There's a javascript (jquery maybe) library that copies facebook dialogs - facebox or something. Although to today's standard's/facebook design it might be a bit rubbish now |
| 23:08 | <clair> | Should dig out my facebook and google docs accounts and take a fresh look |
| 23:09 | <benschwarz> | just updated the developer spec now Hixie |
| 23:09 | <benschwarz> | no updates from me. just spec content |
| 23:10 | <Hixie> | benschwarz: doing good, actually making positive progress against the feedback onslaught for once |
| 23:10 | <Hixie> | benschwarz: cool |
| 23:10 | <Hixie> | clair: i expect the big places (facebook, google, apple, microsoft) will have rolled their own dialogs and may be a good place to look at to get a variety of options |
| 23:11 | <Hixie> | clair: my assumption (may be false) is that most smaller sites will all use the same libraries so will have less variation |
| 23:11 | <Hixie> | clair: but i could be completely wrong |
| 23:11 | <benschwarz> | Hixie: I was thinking last night - |
| 23:11 | <benschwarz> | appCache / manifest being static… |
| 23:12 | <benschwarz> | I want my mobile devices to download less, where possible, and maybe use polyfills conditionally, without loading them also |
| 23:12 | <clair> | Hixie: Good point there... doesn't apple use some open source library that isn't jquery? |
| 23:13 | <benschwarz> | do you think using javascript to construct a manifest then base64 it to the browser (html tag) would work? |
| 23:14 | <Hixie> | clair: dunno |
| 23:14 | <Hixie> | benschwarz: i'm confused. what are you trying to cache? |
| 23:15 | <benschwarz> | Hixie: nothing too special, but it seems a shame to download (and store) files in cache that certain devices dont require |
| 23:15 | <benschwarz> | …which can mean detection / UA sniffing on the server side |
| 23:16 | <benschwarz> | just thinking aloud… wondering if you could construct a manifest in JS client-side |
| 23:18 | <Hixie> | benschwarz: ah |
| 23:18 | <Hixie> | benschwarz: the manifest url is used to decide what's going on, so no, not in the current model |
| 23:18 | <clair> | Couldn't they be constructed server side using UA sniffing? |
| 23:19 | <clair> | (although I realise that kind of detection is bad) |
| 23:26 | <jamesr> | you could have a non-cached loader page that issued a request (or a redirect or something) to the appriopriate cached URL for that ua/etc variation |
| 23:32 | <Yuhong> | http://groups.google.com/group/alt.html/msg/30ef82a1fbab905a?hl=en |
| 23:32 | <Yuhong> | They have clearly stated that they will only support 90% or the standards for fear of |
| 23:32 | <Yuhong> | "problems with backwards-accessibility". Bull. If they hadn't screwed up prior implementations of HTML and CSS that wouldn't be |
| 23:32 | <Yuhong> | an issue. " |
| 23:33 | <Yuhong> | Sounds familiar? |
| 23:35 | <Yuhong> | Here is the story: |
| 23:35 | <Yuhong> | Mozilla and IE5 for Mac faced the same problem too, and they decided to solve it by inventing DOCTYPE switching.[` |
| 23:36 | <Yuhong> | IE6 for Windows copied it. |
| 23:37 | <Yuhong> | The problem is then they stagnated IE for 5 years and it gained a monopoly, so guess what people did with ie6's standards mode? |
| 23:37 | <gsnedders> | Moz didn't implement it till after IE, though dbaron proposed it on mailing list before IE impl it. |
| 23:37 | <gsnedders> | AFIAK. |
| 23:37 | <dbaron> | doctype switching? |
| 23:37 | <dbaron> | not my idea |
| 23:38 | <gsnedders> | dbaron: Any idea who? My memory is obviously dodgy. :) |
| 23:39 | <Yuhong> | The problems caused by it of course led to the proposal for HTML versioning to the HTML WG, and later the invention of X-UA-Compatible, another mode switch. |
| 23:44 | <Yuhong> | Anyway, besides <li> being used for indent, <dd>, <dl> and several other tags was used for Indent. |
| 23:44 | <Yuhong> | <Hx> was used to change font size in the Mosaic age. |
| 23:46 | <dbaron> | gsnedders, well, Gecko has had some form of recognizable doctype sniffing since http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=1999-05-15+22%3A26&maxdate=1999-05-15+22%3A29&cvsroot=%2Fcvsroot (see the nsParser.cpp changes) |
| 23:46 | <dbaron> | gsnedders, though really there was a little code before that |
| 23:46 | <dbaron> | gsnedders, I think tantek or Todd Fahrner might know who came up with the idea... |
| 23:46 | <dbaron> | or rickg |
| 23:47 | <Yuhong> | Here is an example of this abuse: http://groups.google.com/group/comp.infosystems.www.authoring.html/msg/ee3198e46d567819?hl=en |
| 23:52 | <Hixie> | what's the opposite of an idempotent operation? |
| 23:52 | <Hixie> | specifically, what would one call an operation that can only be done once? |
| 23:52 | <AryehGregor> | What does "can only be done once" mean? |
| 23:52 | <AryehGregor> | What stops you from doing it twice? |
| 23:52 | <Hixie> | a word meaning "irreversible and not-idempotent" |
| 23:53 | <Hixie> | well e.g. you can only break an egg once |
| 23:53 | <Yuhong> | Of course when exactly DOCTYPE switching happened don't matter for this story, what matters that IE6 for Win adopted it and then stagnated for five years. |
| 23:53 | <Hixie> | once you've broken an egg, you can't break it again |
| 23:53 | <Hixie> | so breaking an egg is a .... operation |
| 23:53 | <gsnedders> | dbaron: I remember reading something linking to some 98 post on moz mailing list about it, though I was unaware Gecko's impl went back anywhere near that far |
| 23:53 | <AryehGregor> | Well, there's not going to be a mathematical term for it. The concept doesn't make sense for pure functions. |
| 23:53 | <AryehGregor> | It's all about side effects. |
| 23:53 | <Hixie> | doesn't have to be a mathematical term :-) |
| 23:54 | <gsnedders> | Today's interesting question: what's the most complex JS you can come up with that you can prove with static-analysis is pure? |
| 23:54 | <Hixie> | define "complex" |
| 23:54 | <Yuhong> | Which basically screwed it up. |
| 23:55 | <gsnedders> | Hixie: Does something interesting. |
| 23:56 | <Hixie> | return 'insert text of a very interesting book here'; |
| 23:56 | <gsnedders> | Hixie: Something more complex than function(a, b) { return a * b; } |
| 23:56 | <Hixie> | wait, what do you mean by "pure"? |
| 23:57 | <AryehGregor> | "No side effects". |
| 23:57 | <Hixie> | is throwing an exception a side-effect? |
| 23:57 | <gsnedders> | Hixie: No side effects, and state doesn't effect its result. |
| 23:57 | <gsnedders> | Hixie: No, that's a return value. |
| 23:58 | <Hixie> | is triggering GC or allocating memory a side-effect? |
| 23:59 | <Hixie> | are available memory or amount of time the script has been executing so far "state"s? |