| 04:40 | <rillian> | cpearce: it looks like the spec currently lets you add alternate audio/video tracks either using fragments, or through the media.audioTracks[]/media.videoTracks[] interfaces |
| 04:40 | <rillian> | you can likewise add text tracks through media.textTracks |
| 04:40 | <rillian> | but you can't add alternate audio/video tracks through <track> |
| 04:40 | <rillian> | maybe that's just an oversight |
| 08:58 | <annevk> | w3.org is offline? |
| 08:58 | <annevk> | cannot connect to IRC or well, anything |
| 08:59 | <annevk> | just a hickup it seems |
| 09:02 | <rniwa> | annevk: I can access www.w3.org without any issues |
| 09:02 | <rniwa> | annevk: maybe try pinging 128.30.52.37? |
| 09:04 | <annevk> | it's working again |
| 09:04 | <annevk> | very temporary :) |
| 09:05 | <rniwa> | i see |
| 09:22 | <annevk> | alright |
| 09:22 | <annevk> | time to add Node.contains() |
| 09:25 | <Lachy> | Nice. HTML context menus now supported in Firefox nightly. https://twitter.com/#!/codepo8/status/108989867278614528 |
| 09:26 | <Lachy> | just looks like they're using <menuitem> instead of <command>. Not sure why though |
| 09:29 | <annevk> | see some debate in a bug report somewhere |
| 10:36 | <annevk> | so should we have <data-...> |
| 11:00 | <annevk> | Is there a name for an interface other interfaces inherit from that is never instantiated itself but only exists by virtue of the interfaces that inherit from it? |
| 11:01 | <Ms2ger> | Abstract? |
| 11:02 | <annevk> | So "CharacterData is an abstract interface and does not exist as node. It is used by Text, Comment, and ProcessingInstruction nodes. |
| 11:02 | <annevk> | " |
| 11:03 | hsivonen | looks forward to OO pedants sending feedback about "abstract interface" |
| 11:03 | <annevk> | If they suggest alternatives it will be all good :) |
| 11:04 | <annevk> | This is just a non-normative note to clarify things since every other heading in that section is about a node |
| 11:05 | <Ms2ger> | Except Node |
| 11:06 | <annevk> | I guess that should get the same note |
| 11:21 | <annevk> | http://software.hixie.ch/utilities/js/live-dom-viewer/?x%3Cscript%3Edocument.body.firstChild.removeChild%28document.body%29%3C%2Fscript%3E |
| 11:21 | <annevk> | Gecko: HIERARCHY_REQUEST_ERR, WebKit: NOT_FOUND_ERR |
| 11:21 | <annevk> | I don't think http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#concept-node-pre-remove actually needs step 1 given step 2 |
| 11:22 | <annevk> | Opera: NOT_FOUND_ERR |
| 11:23 | <annevk> | What do you think Ms2ger? |
| 11:23 | <Ms2ger> | wfm |
| 11:23 | <Ms2ger> | Please do update the test :) |
| 11:24 | <annevk> | I should check out the tests I guess |
| 11:25 | <annevk> | now they are separated from the specification |
| 11:26 | <annevk> | heh |
| 11:26 | <annevk> | added 563 changesets with 1341 changes to 478 files |
| 11:30 | <annevk> | Ms2ger, there is no test? |
| 11:30 | <Ms2ger> | Node-removeChild.html? |
| 11:30 | <annevk> | yeah it does not have a HIERARCHY_REQUEST_ERR in it |
| 11:31 | <annevk> | guess I can add something that makes Gecko fail |
| 11:31 | <annevk> | it also does a bunch of Attr (node) tests |
| 11:31 | <Ms2ger> | Not anymore :) |
| 11:35 | <annevk> | added a test for removeChild on a text node |
| 11:35 | <Ms2ger> | Thanks |
| 11:35 | <annevk> | someone else can file a bug on Gecko |
| 12:32 | <annevk> | I'm starting to think we should just go ahead and merge Range in |
| 12:32 | <annevk> | Is there much else to do that I am missing? |
| 12:33 | <annevk> | Also note it is prerequisite for adding modification listeners |
| 12:33 | <annevk> | Ms2ger, AryehGregor, ^^ |
| 12:34 | <Ms2ger> | sURE |
| 12:34 | Ms2ger | hits caps lock |
| 12:37 | <annevk> | does this count as written permission per the license? :) |
| 12:38 | <annevk> | whoa |
| 12:38 | <annevk> | https://bitbucket.org/ms2ger/dom-core/raw/tip/dom-core.html |
| 12:39 | <annevk> | isn't this some giant XSS hole in bitbucket? |
| 12:39 | <Ms2ger> | Yes |
| 12:42 | <annevk> | Ranges is actually pretty big |
| 12:47 | <Ms2ger> | 15 pages or so? |
| 12:49 | <annevk> | yeah something like that |
| 12:49 | <annevk> | 2000 lines :) |
| 13:11 | <hsivonen> | that bitbucket thing is scary. has it been reported to bitbucket? |
| 13:11 | <Ms2ger> | Not by me |
| 13:12 | <hsivonen> | and bitbucket has been created *after* people were supposed to know about origins |
| 13:18 | <annevk> | it looks like a regression |
| 13:19 | <annevk> | what is the "Assert:" stuff Ms2ger? |
| 13:19 | <annevk> | can that be removed? |
| 13:20 | <Ms2ger> | Sure |
| 13:20 | <gsnedders> | assert stuff? where?" |
| 13:21 | <Ms2ger> | DOM Range |
| 13:22 | <gsnedders> | Ah. The asserts there are pretty much all just for WebIDL things> |
| 13:22 | <gsnedders> | *? |
| 13:22 | <Ms2ger> | Yeah |
| 13:22 | <gsnedders> | I'd say they don't add anything, then, really. |
| 13:38 | <hsivonen> | the DOM Core spec exposed a bug in my new View Source impl... |
| 13:39 | <Ms2ger> | Oh? |
| 13:41 | <annevk> | someone with a good term for "ancestors and the node itself"? |
| 13:42 | <annevk> | XPath calls it ancestor or self |
| 13:42 | <hsivonen> | Ms2ger: it appears that I broke the way View Source communicates the charset of the page to the main thread when I fixed https://bugzilla.mozilla.org/show_bug.cgi?id=675499 |
| 13:45 | <annevk> | how about "reverse subtree" |
| 13:46 | <annevk> | (the current term is "ancestor containers") |
| 13:46 | <hasather> | annevk: wouldn't that be a supertree then? |
| 13:48 | <hasather> | (I guess tree is not a good word to describe it though) |
| 13:53 | <annevk> | the roof :) |
| 13:54 | <annevk> | hasather, it's the best short word so far |
| 13:54 | <annevk> | hasather, though very novel afaict |
| 13:55 | <annevk> | and it might confuse people who are into this: http://en.wikipedia.org/wiki/Supertree |
| 13:56 | <hasather> | annevk: yea, saw that too |
| 13:57 | <Philip`> | "subtree" sounds confusing because you'd think it's a tree with branches, when acually it's just a list |
| 13:59 | Philip` | has a book where "ancestor" includes the node itself, and "proper ancestor" doesn't |
| 14:02 | <annevk> | subtree is standard terminology |
| 14:02 | <annevk> | interesting |
| 14:03 | <Philip`> | subtree is standard when referring to the tree formed by the descendants of a node, but not when referring to the chain of ancestors, I think |
| 14:05 | <annevk> | we'd use supertree for the the ancestor part |
| 14:06 | <Philip`> | "supertree" sounds confusing because you'd think it's a tree with branches, when actually it's just a list |
| 14:07 | <annevk> | alternatives? |
| 14:07 | Philip` | would probably assume the supertree is the largest tree that contains the given node (i.e. the one rooted at the most distant ancestor) |
| 14:10 | <Philip`> | If "ancestor" is already defined to exclude the node itself, "ancestor or self" doesn't sound too bad (it's a bit ugly but short and clear), or maybe "inclusive ancestor" |
| 14:13 | <annevk> | inclusive ancestor sounds rather clear and works a bit better in this context |
| 14:16 | <woef> | As someone who has no idea what you guys are talking about, "ancestor and self" is much easier to understand than "inclusivve ancestor" |
| 14:19 | <annevk> | true enough, but understanding this term is the least of your worries when you implement range |
| 14:19 | <Philip`> | The term (and "ancestor" by itself) should be linked to a precise definition, so I guess it's not too bad if people have to click the link to find what you mean |
| 14:20 | <Philip`> | (It'd be worse if they thought they knew what it meant, so they didn't check the definition, but actually had the wrong idea) |
| 14:21 | <woef> | "Let's name it something nobody will be certain to understand and force them to look for the proper definition" |
| 14:21 | <Philip`> | Call it "concept #382" |
| 14:21 | <woef> | That needs some usability testing :p |
| 14:21 | <woef> | Philip`: hehe |
| 14:24 | <Philip`> | Then you just define "Concept #382: A concept #53 that is the target concept #53 or a concept #197 of it", and "Concept #197: A concept #53 that is a concept #382 of the target concept #53 but is not the target concept #53 itself", etc |
| 14:25 | <woef> | And throw in some false links to make sure they're paying attention and not just clicking along. |
| 14:41 | <annevk> | I guess I should land Range before I make too many changes so people can follow on what is being changed |
| 14:50 | <annevk> | went from 37 to 44 pages |
| 14:51 | <annevk> | Should AryehGregor become co-editor of DOM Core now? I guess that would be a bit too unwieldy |
| 14:51 | <zcorpan> | annevk: shouldn't it be s/and/or/ in https://bitbucket.org/ms2ger/dom-core/changeset/e126979296ce ? |
| 14:57 | <annevk> | why? |
| 15:00 | <hsivonen> | hmm. so IGs "endorse" bugs |
| 15:06 | <zcorpan> | annevk: because it now says that objects that implements all of those interfaces are nodes |
| 15:06 | <zcorpan> | annevk: whereas if an object implements only one of them is not a node |
| 15:07 | <zcorpan> | hsivonen: which bug? |
| 15:11 | <annevk> | thanks |
| 15:20 | <Ms2ger> | annevk, if you feel like updating Selection for your changes... ;) |
| 15:20 | <hsivonen> | zcorpan: http://lists.w3.org/Archives/Public/public-html/2011Sep/0004.html |
| 15:21 | <AryehGregor> | annevk, go ahead and merge Range in, sounds good to me. |
| 15:22 | <annevk> | Ms2ger, I was about to say, and then I got that email, damnit |
| 15:22 | <AryehGregor> | Don't take the Selection stuff, obviously. |
| 15:22 | <annevk> | right |
| 15:22 | <annevk> | I haven't |
| 15:22 | <AryehGregor> | I'll take that at some point soonish. |
| 15:22 | <AryehGregor> | Oh, I see, you took it already. |
| 15:22 | <AryehGregor> | Good. |
| 15:25 | <zcorpan> | hsivonen: thanks |
| 16:10 | <zcorpan> | heycam: hey. http://dev.w3.org/2006/webapi/WebIDL/#es-attributes - why does it say "Otherwise, it exists on the interface’s interface prototype object or on every object that implements the interface." ? why leave a choice between having it on the prototype and having it on the object? |
| 16:38 | <dglazkov> | good morning, Whatwg! |
| 16:42 | <annevk> | hey hey dglazkov |
| 17:28 | <espadrine> | good evening, dglazkov! |
| 17:30 | <dglazkov> | annevk: do you like to travel? |
| 17:30 | <annevk> | I do |
| 17:31 | <annevk> | weird habbit :) |
| 17:31 | <annevk> | habit(sp?) |
| 17:31 | <Ms2ger> | hobbit |
| 17:32 | <annevk> | too tall for a hobbit |
| 17:32 | timeless | makes that typo too |
| 17:32 | <annevk> | would love a home like Bilbo though |
| 17:32 | <timeless> | rabbit hobbit habit |
| 17:33 | <timeless> | annevk: your irc client doesn't have a spell checker? |
| 17:33 | timeless | is using Nightly's spell checker (w/ freenode-webchat) |
| 17:33 | <annevk> | yeah (ctrl+t, enter word, hit enter, and see what Google says), but I don't always use it |
| 17:34 | <timeless> | eww |
| 17:34 | <timeless> | i do ctrl+t, e, down, enter, type word |
| 17:34 | <timeless> | e = https://encrypted.google.com/ |
| 17:35 | <timeless> | but that's only when i don't have a text field available for Gecko's spellchecker |
| 17:35 | <annevk> | http://evolutionofweb.appspot.com/ is pretty cool |
| 17:36 | <timeless> | shiny |
| 17:36 | <MikeSmith> | nice |
| 17:36 | <MikeSmith> | annevk: who made that? |
| 17:36 | <annevk> | found it at http://chrome.blogspot.com/2011/09/happy-third-birthday-chrome.html |
| 17:37 | <MikeSmith> | ah great |
| 17:37 | <timeless> | it doesn't speak spanish :( |
| 17:38 | <hsivonen> | MikeSmith: looks like something Google contracted out |
| 17:38 | <timeless> | hrm |
| 17:38 | <timeless> | the screenshots of opera are odd |
| 17:38 | <MikeSmith> | hsivonen: OK |
| 17:38 | <hsivonen> | What's the deal with Netscape not having releases before 4 |
| 17:38 | <timeless> | opera v1 screenshot is w3.x |
| 17:39 | <timeless> | but 2.1 is a screenshot from wXP or newer |
| 17:39 | <timeless> | and 3 is from w7 |
| 17:39 | <timeless> | but 4 is from 9x |
| 17:39 | <hsivonen> | also, Firefox 3.6 is missing |
| 17:40 | <hsivonen> | as well as some significant Safari point releases |
| 17:40 | <hsivonen> | and significant Opera releases |
| 17:40 | <timeless> | there's a "credits and sources" link in the left bar at the bottom |
| 17:40 | <hsivonen> | so, unfortunately, this doesn't work as a more complete "Modern browsers ship" visualization :-( |
| 17:41 | <timeless> | they're also missing ie1/2 |
| 17:41 | <zewt> | these wavy lines are a bit too mysterious |
| 17:41 | <timeless> | but yeah, theyreally should have netscap 1, 2 and 3 |
| 17:41 | <hsivonen> | and the mozilla suite |
| 17:41 | <timeless> | they quasi cover that w/ netscape versions 6..8 |
| 17:41 | <timeless> | but yeah |
| 17:42 | <timeless> | have people followed the diginotar coverage? |
| 17:42 | <Ms2ger> | annevk, did you file a bug on HTML for the "root" thing? |
| 17:42 | <hsivonen> | also, the colored band for Java gets wider over time |
| 17:42 | <annevk> | Ms2ger, no |
| 17:42 | <hsivonen> | timeless: somewhat followed |
| 17:43 | <timeless> | hsivonen: it's interesting that there was an addons cert |
| 17:43 | <timeless> | i wonder if it'd be reasonable for browser vendors to add a requirement for all cas that want to list of the form: |
| 17:43 | <annevk> | Ms2ger, HTML uses concept-tree-root in a different manner? |
| 17:43 | <hsivonen> | timeless: addons cert? |
| 17:43 | <timeless> | if you ever try to issue a cert for any of our properties, you must contact us |
| 17:43 | <timeless> | http://www.computerworld.com/s/article/9219663/Hackers_may_have_stolen_over_200_SSL_certificates |
| 17:44 | <annevk> | Ms2ger, I do think we should maybe add subtree and root subtree for HTML |
| 17:44 | <timeless> | > Mozilla confirmed that a certificate for its add-on site had been obtained by the DigiNotar attackers. "DigiNotar informed us that they issued fraudulent certs for addons.mozilla.org in July |
| 17:44 | <Ms2ger> | annevk, I copied the definition from HTML |
| 17:44 | <hsivonen> | timeless: It seems like a no-brainer that CAs should check if the hostnames they are minting certs for already have certs from someone else |
| 17:45 | <timeless> | hsivonen: sadly |
| 17:45 | <timeless> | as for profit companies |
| 17:45 | <timeless> | it's in your interest to steal your competitor's customers |
| 17:45 | <zewt> | how exactly can you tell if someone already has a certificate? heh |
| 17:45 | <Philip`> | They informed Mozilla in July, or they informed Mozilla now that they knew they issued fraudulent certs in July? |
| 17:45 | <timeless> | zewt: pretty easy |
| 17:46 | <timeless> | if i ask for a cert for super.example.com |
| 17:46 | <timeless> | then you try to visit https://super.example.com |
| 17:46 | <timeless> | if you can connect securely, then someone has a cert! :) |
| 17:46 | <timeless> | Philip`: my guess is informed nowish |
| 17:46 | <zewt> | i have certs that i don't use for https |
| 17:46 | <timeless> | and revoked in july |
| 17:46 | <zewt> | (ftp, etc) |
| 17:47 | <Philip`> | Shouldn't there be some requirement that if they revoke certs (presumably because they know they were invalidly issued), they at least inform the rightful owners of those domains (immediately, not a month later)? |
| 17:47 | <timeless> | Philip`: yeah well, um |
| 17:47 | <zewt> | can someone have an epiphany and come up with a new signing system that fixes the increasing breakage we have now, heh |
| 17:47 | <timeless> | i don't think i've ever seen that in requirements |
| 17:47 | <timeless> | mozilla has a crypto policy newsgroup |
| 17:48 | <timeless> | it can be proposed |
| 17:48 | <timeless> | and i'd imagine it'd get pretty decent support |
| 17:48 | <Philip`> | Seems like it's important to make it more expensive for a company to cover up its mistakes than for it to admit to them immediately |
| 17:48 | <timeless> | zewt: we don't have increasing breakage |
| 17:48 | <timeless> | so much as more people willing to spend the minimal effort to attack a system |
| 17:49 | <timeless> | Philip`: oh, that we're doing |
| 17:49 | <zewt> | sure we do; more and more entities with the ability to sign certificates; certificate exception dialogs becoming more and more pointlessly annoying :) |
| 17:49 | <timeless> | mozilla & co have killed diginotar |
| 17:49 | <timeless> | zewt: you played that card earlier |
| 17:49 | <timeless> | please don't replay a previously played card :) |
| 17:50 | <zewt> | perfectly valid when the card is correct :P |
| 17:50 | <Philip`> | Would the same killing have happened if they had admitted it immediately? (in which case they'd have no incentive to not try hiding it) |
| 17:50 | <timeless> | (there's a Hixie quote to be made here) |
| 17:50 | <timeless> | Philip`: past experience w/ CAs who have screwed up and fessed up hasn't resulted in death |
| 17:51 | <timeless> | which indicates, more or less, if you are honest and forthcoming you have historically gotten leeway |
| 17:51 | <Philip`> | I suppose that's good, then |
| 17:51 | <Philip`> | (although less good than if they didn't screw up) |
| 17:51 | <timeless> | past experience is of course not a commitment to future behavior |
| 17:51 | <timeless> | but, the goal is to provide carrot and stick |
| 17:52 | <timeless> | we've played carrot a few times, and stick once (now) |
| 17:52 | <zewt> | that's another breakage of the system: if a CA screws up and is dropped from browsers, there may be a *lot* of collateral damage |
| 17:52 | <timeless> | zewt: in this case, Vasco (recent parent of diginotar) indicated it has minimal business in this area |
| 17:52 | <zewt> | not frequently, but it's unpleasant that the possibility exists (of course, it's inherent to the system we have) |
| 17:52 | <timeless> | i haven't checked on their pricing model, but you should be able to calculate customer count |
| 17:53 | <zewt> | (not to suggest i know of any alternative model that doesn't have all of these problems, of course) |
| 17:53 | <timeless> | and yes, ideally those customers will ask for money back from their vendor (diginotar) |
| 17:53 | <timeless> | and take their business elsewhere (quickly, or lose customers while they dawdle) |
| 17:54 | <timeless> | anyway, that's the best economy i can offer today |
| 17:54 | hsivonen | wonders if phone vendors are taking any action to zap diginotar from the stock browsers on their phones |
| 17:54 | <timeless> | hsivonen: generally no |
| 17:54 | <timeless> | speaking from experience at nokia |
| 17:54 | <timeless> | we asked about issuing updates |
| 17:54 | <timeless> | and were turned down |
| 17:54 | <hsivonen> | timeless: yeah, it looks like only iPhone and the Nexus series get any reasonable updates |
| 17:54 | <timeless> | (This was from the previous CA disaster, and we sent things up the flagpole as hard as we could) |
| 17:55 | <timeless> | i can't speak for my current employer |
| 17:55 | <hsivonen> | though I don't know if either iOS or Android on Nexus got updates following the Comodo case |
| 17:55 | <timeless> | hrm, although i could test a current build of our platform :) |
| 17:56 | <timeless> | (for the record, previous disaster = comodo) |
| 17:57 | <hsivonen> | I wish companies that ship mobile OSs acted more like real software vendors |
| 17:57 | <timeless> | oh, fwiw MS hasn't gotten around to sending out kill bits for wXP |
| 17:57 | <timeless> | (ms would kinda like people to stop using XP...) |
| 17:58 | <zewt> | i've never even seen a browser update in android of any kind except during a full OS update |
| 17:58 | <timeless> | the problem w/ classic hardware vendors is that they're hardware vendors |
| 17:58 | <zewt> | it's bizarre but google doesn't seem to care much about the android browser |
| 17:58 | <timeless> | they, like CAs don't have recurring revenues from past customers |
| 17:59 | <timeless> | and thus the cost of doing support/maintenance is something which isn't factored in and doesn't make sense to them |
| 17:59 | <zewt> | CAs do, since certs expire |
| 17:59 | <timeless> | zewt: yeah, technically CAs do |
| 17:59 | <timeless> | but only kinda |
| 17:59 | <zewt> | heh reminds me of some fraud godaddy has |
| 17:59 | <zewt> | they automatically set your cert to auto-renew (at the full rate, like $50/year), without asking or telling you |
| 17:59 | <zewt> | (presumably buried in some 100-page "agreement") |
| 18:00 | <zewt> | you have to go through a zillion menus to even find that it's on and get rid of it |
| 18:00 | <hsivonen> | timeless: I wonder how many iPod Touch users bought OS updates. Has it been proven that mobile hardware companies cannot sell software updates? |
| 18:00 | <timeless> | hsivonen: i didn't buy my update |
| 18:00 | <hsivonen> | timeless: Apple has managed to sell software updates for desktop/laptop hardware |
| 18:01 | <timeless> | the license agreement for my Finnish iPod as presented by iTunes was in Finnish |
| 18:01 | <timeless> | and thus, I couldn't accept it |
| 18:01 | <timeless> | so, i never bought the update |
| 18:01 | <timeless> | i'm now in .CA |
| 18:01 | <hsivonen> | timeless: you are weird |
| 18:01 | <zewt> | heh |
| 18:01 | <Philip`> | When the TLS stuff was designed, was it expected that the response to fraudulent certificates would require multiple independent software vendors to ship updated versions, or was there meant to be a more elegant/robust way of handling it? |
| 18:01 | <zewt> | software loves showing me things in japanese, even though my system language is english (because they incorrectly use the system codepage) |
| 18:01 | <timeless> | and once I get around to plugging in one of my computers (probably my G5), I'll see about buying the update |
| 18:01 | <hsivonen> | timeless: (so am I. I, too, today rejected a piece of software due to not having the time to wade through their legal stuff) |
| 18:01 | <hsivonen> | (software from cisco) |
| 18:02 | <timeless> | Philip`: do you mean TLS or SSL? |
| 18:02 | <Philip`> | timeless: I don't know |
| 18:02 | <timeless> | TLS is a rather recent thing (since it's SSL3.1/3.2) |
| 18:02 | <Philip`> | timeless: Whatever makes the question make most sense |
| 18:02 | <timeless> | the assumption in PKI / SSL |
| 18:02 | <timeless> | was that the vendors would be few |
| 18:02 | <timeless> | and would have decent and proper CRLs |
| 18:03 | <timeless> | there was a limited amount of handwaving involving how CRLs would be deployed to devices |
| 18:03 | <timeless> | but it mostly assumed devices would be connected enough to be able to retrieve them |
| 18:03 | <timeless> | it also mostly assumed that you wouldn't have Rogue Countries |
| 18:03 | <timeless> | or rather |
| 18:03 | <timeless> | it accepted that you couldn't defeat a Rogue Country if you wanted to |
| 18:03 | <timeless> | and thus it was out of scope |
| 18:04 | <timeless> | Basically, if the US wanted to do something evil in VeriSign or RSA or whichever, it was assumed it could, but that was a risk one was willing to take |
| 18:04 | <timeless> | plus in the US at least, sunshine and leaks were probably assumed to catch such things |
| 18:05 | <timeless> | which actually is still applying to the current Rogue Countries |
| 18:05 | <timeless> | as zewt notes, the system we have is the best anyone can really think of to date |
| 18:05 | <timeless> | it isn't perfect |
| 18:06 | <hsivonen> | when VeriSign bought Network Solutions, I sent my bank a question asking them what they are going to do now that the entity that can tamper with DNS and the entity that guards against DNS tampering are the same |
| 18:06 | <hsivonen> | they actually forwarded it to someone technical |
| 18:07 | <hsivonen> | who called me and said the situation wasn't good but they'd just acknowledge the situation and their powerlessness about it |
| 18:10 | <AryehGregor> | timeless, there's an easy solution: pin particular CAs for sites, using STS. AFAIK, Chrome already does this for Google sites, so anyone who tried to use the forged cert against Chrome users would trigger unrecoverable failure. |
| 18:10 | <timeless> | hsivonen: you're looking forward to DNSSEC, eh? :) |
| 18:11 | <AryehGregor> | Alternatively, get certs-via-DNSSEC working properly and supported in all browsers, then only allow that. |
| 18:11 | <AryehGregor> | For sites that opt in. |
| 18:11 | <timeless> | AryehGregor: you didn't read https://twitter.com/#!/moxie__/statuses/108567203829387264 ? |
| 18:11 | <hsivonen> | timeless: actually I am considering that the current system already fails if the curator of the root goes rogue |
| 18:11 | <zewt> | of course, it seems more likely that dnssec will be implemented, and both dnssec and tls certs will be accepted for most sites, giving two independent trees which are both points of failure... |
| 18:12 | <AryehGregor> | timeless, you can remove registrars with DNSSEC. Why not? Just have the registrar one step up revoke the cert, or not renew it. |
| 18:12 | <AryehGregor> | zewt, you could have a DNSSEC record that says "only use DNSSEC certs to access this site". |
| 18:12 | <hsivonen> | AryehGregor: did you read the whole tweet thread with Moxie and Dan Kaminsky? |
| 18:12 | <AryehGregor> | Or, just use STS and bake a list of sites and approved CAs for each site into the browsers. |
| 18:12 | <AryehGregor> | hsivonen, no. |
| 18:13 | <hsivonen> | AryehGregor: looks like we are going to have to replay it here |
| 18:13 | <timeless> | heh |
| 18:13 | <zewt> | but is there a TLS equivalent to say "never use dnssec"? |
| 18:13 | <hsivonen> | AryehGregor: https://twitter.com/#!/moxie__/status/108331615004000256 |
| 18:13 | <AryehGregor> | zewt, you can use STS and have all browsers ship with hardcoded lists of all major sites (i.e., all likely attack targets). |
| 18:14 | <timeless> | hsivonen: appreciated, since i'm not a fan of twitter |
| 18:14 | <AryehGregor> | hsivonen, nothing is realistically going to protect us against ICANN or the IANA going bad, unless we abandon centralized DNS. |
| 18:14 | <zewt> | special casing to make the rest of the world second-class citizens? that's horrible |
| 18:15 | <AryehGregor> | hsivonen, basically they're not likely to do bad stuff unless the US forces them to, and if the US wants to learn about Google sites it has easier ways to do that. |
| 18:15 | <AryehGregor> | Like subpoenas. |
| 18:15 | <AryehGregor> | I mean, it's a possibility, but you can't defend against everything. |
| 18:16 | <AryehGregor> | Stopping Iran from intercepting Google-bound traffic is feasible, stopping the US or VeriSign from doing it is not so feasible. |
| 18:16 | <hsivonen> | zewt: umm. the rest of the world is already second-class even on the legislative level |
| 18:17 | <AryehGregor> | Heck, the US government could nationalize Google if it really felt like it. Eminent domain, right? |
| 18:17 | <timeless> | technically yes |
| 18:17 | <zewt> | ... |
| 18:17 | <Philip`> | Are there ways to independently verify that browser vendors aren't doing bad stuff (like shipping binaries which whitelist some fake certificates), so that they're unlikely to be a point of failure? |
| 18:18 | <zewt> | my website should not be a second-class citizen compared to any other website |
| 18:18 | <AryehGregor> | If the scenario you're worried about is the US government doing evil things to US corporations, I'm pretty sure you lose either way. |
| 18:18 | <AryehGregor> | zewt, the list could be one that anyone can add themselves to. |
| 18:18 | <timeless> | Philip`: well, mostly |
| 18:18 | <timeless> | certainly with firefox you can build it yourself |
| 18:18 | <timeless> | and compare what you have against what is shipped |
| 18:18 | <AryehGregor> | E.g., it could just be a matter of Googlebot finding STS headers and building a list of all the ones with suitably long expiration dates. |
| 18:18 | <AryehGregor> | And publishing it. |
| 18:18 | <zewt> | ... but then you're just expanding the problem, since now you have to verify that whoever's modifying that record is authorized to do so |
| 18:18 | <zewt> | back to square one |
| 18:19 | <AryehGregor> | That's the browser vendor, who you have to trust anyway. |
| 18:19 | <AryehGregor> | They could be installing a backdoor for all you know. |
| 18:19 | <timeless> | it's moderately painful since firefox has whole-program-optimization |
| 18:19 | <AryehGregor> | Entities you realistically have to trust to some degree here: the US government, your browser vendor, your OS vendor, possibly your computer's manufacturer. |
| 18:20 | <timeless> | woohoo |
| 18:20 | <AryehGregor> | Entities you should not have to trust: every single two-bit company in the world that gets certified as a CA by someone. |
| 18:20 | <timeless> | after 1 email and ~1minute, i can now see bug reports i file myself! |
| 18:20 | timeless | likes this company |
| 18:20 | <timeless> | at nokia, it'd probably have taken a month :) |
| 18:20 | <timeless> | or 3 |
| 18:20 | <timeless> | or 9 |
| 18:20 | <timeless> | (possibly just getting a reply saying "do you still need this?") |
| 18:21 | <Ms2ger> | And why? |
| 18:21 | <AryehGregor> | Why what? |
| 18:21 | <Ms2ger> | Why one would want to see such bug reports |
| 18:23 | <zewt> | (i don't understand the question--of course you want to be able to see your own bug reports) |
| 18:24 | <hsivonen> | zewt: Opera seems to disagree :-) |
| 18:24 | <zewt> | causing me to not (often) submit opera bug reports :) |
| 18:24 | <zewt> | (but we've been over that :) |
| 18:24 | <timeless> | Ms2ger: why would one want to see one's own bug reports? |
| 18:24 | timeless | shrugs |
| 18:24 | <AryehGregor> | If I have long-running scripts, how can I stop the long-running script alerts? Is there some simple way I can spin the event loop in the middle or similar? |
| 18:24 | <timeless> | zewt: i switched to sending opera bug reports by email to opera employees |
| 18:25 | <timeless> | they reply eventually with updates :) |
| 18:25 | AryehGregor | sends them by IRC in this channel |
| 18:25 | <timeless> | (this also more or less works for google bug reports) |
| 18:25 | <Ms2ger> | setTimeout(..,0)? |
| 18:25 | <zewt> | AryehGregor: run a synchronous xhr to a php script that pauses? :P |
| 18:25 | <AryehGregor> | zewt, :( |
| 18:25 | <Ms2ger> | zewt-- |
| 18:25 | <Philip`> | Maybe you could use alert() to spin the event loop |
| 18:25 | <AryehGregor> | Ms2ger, will that slow stuff down if I do it a lot? |
| 18:25 | <zewt> | alert("Please click OK") |
| 18:25 | <timeless> | AryehGregor: on Gecko you can use a magic thing |
| 18:25 | <AryehGregor> | Like by interpreting 0 as 15 or something? |
| 18:26 | <timeless> | generators |
| 18:26 | <timeless> | if you can manage to get them to work |
| 18:26 | <Ms2ger> | 4, and only if nested, AIUI |
| 18:26 | AryehGregor | observes that if the tab is in the background, Chrome doesn't whine about long-running scripts, so doesn't see that it's worth the effort |
| 18:26 | <zewt> | well, the theory is if you have long-running scripts they should be in workers |
| 18:26 | <zewt> | of course, that's often hard in practice |
| 18:28 | <timeless> | anyone here have /. mod points? |
| 18:28 | timeless | needs something down-modded |
| 18:30 | Philip` | appears to have 13 |
| 18:30 | <timeless> | http://it.slashdot.org/comments.pl?sid=2407244&cid=37271796 |
| 18:32 | <Philip`> | Why do you object to that? |
| 18:32 | <timeless> | one sec |
| 18:33 | <zewt> | "Educate people" heh |
| 18:33 | <timeless> | eh? this would only fix something if the certificate client hard fails when it can't get CRL or OCSP working and only if mapping diginotar.nl happened to magically map www.diginotar.nl (which is where some of the CRLs live), plus service.diginotar.nl validation.diginotar.nl crl.pkioverheid.nl (which are home to OCSP/CRLs) |
| 18:33 | <zewt> | any security mechanism that requires educating users fails |
| 18:33 | <timeless> | Philip`: ^ is my response, but basically that hosts line doesn't do anything |
| 18:35 | <hsivonen> | 386 time clearly |
| 18:35 | <timeless> | ? |
| 18:36 | <Ms2ger> | xkcd.com/386 |
| 18:37 | <timeless> | oh yes |
| 18:37 | <timeless> | that's one of only 3 my browser knows |
| 18:41 | <hsivonen> | timeless: I'm quickly losing my DNSSEC enthusiasm as I read more about it |
| 18:42 | <hsivonen> | (my SIP enthusiasm also went down when I started reading about SIP) |
| 18:43 | AryehGregor | is still insufficiently disillusioned, perhaps |
| 18:43 | AryehGregor | is still enthusiastic about DNSSEC :) |
| 18:45 | AryehGregor | sighs very loudly about CSSOM incompatibilities |
| 18:45 | Ms2ger | makes AryehGregor a co-editor |
| 18:46 | <AryehGregor> | Is anyone even trying to follow this? http://dev.w3.org/csswg/cssom/#serializing-css-values |
| 18:46 | <AryehGregor> | It seems way too vague to follow anyway. "Where CSS component values of the value can be omitted without changing the meaning of the value (e.g. initial values in shorthand properties), omit them. If this would remove all the values, then include the first allowed value." |
| 18:46 | <AryehGregor> | This needs to be defined inline with each property, methinks. |
| 18:46 | AryehGregor | works around it, grumble grumble |
| 18:48 | <hsivonen> | AryehGregor: DNSSEC does seem attractive as a defense against small-time wifi hijacking, though |
| 18:48 | <AryehGregor> | It has a lot of uses. |
| 18:49 | <AryehGregor> | That's one, yeah. |
| 18:49 | <AryehGregor> | I'm also excited about being able to have TLS without extra CAs. |
| 18:50 | <zewt> | also if it allows delegating subdomains |
| 18:50 | <zewt> | the inability to do that with tls is ridiculous |
| 18:51 | <hsivonen> | btw, now that Moxie got mentioned: what's the business model of Whisper Communications? how do they make money? |
| 18:58 | <smaug____> | is Hixie the only one who has access to the server which has acid3 test? |
| 18:58 | <AryehGregor> | smaug____, I assume so. |
| 18:59 | <hsivonen> | smaug____: I believe there are others who can access the *server* but probably not the relevant directory |
| 18:59 | <smaug____> | that is unfortunate |
| 18:59 | <AryehGregor> | Why? |
| 19:00 | <smaug____> | is something happens to Hixie |
| 19:00 | AryehGregor | discovers that he massively messed up his clipboard somehow by trying to copy and paste something huge into a terminal on a remote server |
| 19:00 | <smaug____> | but anyway, I wait Hixie to change the test |
| 19:00 | <AryehGregor> | If something happens to Hixie and his Dreamhost account goes down, yeah, that will be fun. :) |
| 19:00 | <hsivonen> | smaug____: good luck. :-/ |
| 19:00 | <AryehGregor> | Tons of stuff is on that. Like whatwg.org. |
| 19:00 | <smaug____> | hsivonen: I think this change is agreed |
| 19:00 | <hsivonen> | smaug____: nice |
| 19:01 | <smaug____> | doctype.ownerDocument test should be removed |
| 19:03 | <timeless> | hsivonen: sorry to hear that re SIP |
| 19:03 | <timeless> | yeah, i'm not sure where i stand on SIP |
| 19:03 | <timeless> | i like it in theory, and some of my devices do a good job |
| 19:03 | <timeless> | but sadly the uptake of classic SIP just isn't there |
| 19:03 | <timeless> | for a while, Skype was the best hope |
| 19:04 | <timeless> | there actually was some progress on that front, someone is selling Skype ATAs - http://voip.about.com/b/2011/09/01/skype-gets-fully-residential.htm |
| 19:06 | <timeless> | so... |
| 19:06 | <timeless> | i'm not really sure how dnssec is much different from having decent Sub CAs offered by each ISP |
| 19:06 | <timeless> | or rather each DNS registrar |
| 19:07 | <timeless> | not actually a Root CA, just a sub CA |
| 19:07 | <timeless> | hsivonen: does whisper charge for speaking engagements? :) |
| 19:12 | <zewt> | guhh |
| 19:12 | <zewt> | gmail just logged me out *while i was writing an email* |
| 19:12 | <AryehGregor> | Is a draft saved, at least? |
| 19:12 | <zewt> | yeah |
| 19:12 | <zewt> | it's not even "we logged you out because", it's just surprise! you're at the login page |
| 19:14 | <timeless> | zewt: at least your message was saved |
| 19:14 | <timeless> | i've had other webmail things which helpfully eat my message |
| 19:15 | <timeless> | and yeah, i've hit that logged out case w/ gmail |
| 19:21 | <timeless> | hsivonen: hrm |
| 19:21 | <timeless> | one of my devices doesn't trust www.diginotar.com |
| 19:21 | <timeless> | or because of an update |
| 19:21 | <timeless> | of course, that could be because it never trusted it |
| 19:24 | <timeless> | err www.diginotar.nl |
| 19:25 | <timeless> | iirc some phone vendors are slow to add CAs |
| 19:25 | <timeless> | which in some cases is a good thing :) |
| 19:35 | <AryehGregor> | Nice, I froze the Chrome UI for once. |
| 19:35 | <timeless> | congrats |
| 19:35 | <AryehGregor> | By accidentally trying to paste a super-giant URL into the URL bar. |
| 19:36 | <AryehGregor> | (several hundred KB at least) |
| 19:36 | <AryehGregor> | (not actually a URL, just some random text) |
| 19:36 | <timeless> | so... |
| 19:36 | <timeless> | clipboard is actually somewhat special anyway |
| 19:36 | AryehGregor | has to force-quit |
| 19:36 | <timeless> | iirc there historically hasn't been a good async api for it or something |
| 19:36 | <timeless> | (at least on windows) |
| 19:36 | <AryehGregor> | First time that's happened to me in Chrome for a long time. |
| 19:36 | <AryehGregor> | I'm on Linux. |
| 19:36 | <timeless> | which meant one could get very stuck |
| 19:36 | <timeless> | oh, the story on x11 is probably worse |
| 19:36 | <timeless> | i don't think there are any good x11 apis :) |
| 19:38 | <zewt> | AryehGregor: heh, "view image" on a canvas in firefox does that |
| 19:38 | <zewt> | opens it as a gigantic data: |
| 19:40 | timeless | remembers a time when the urlbar would stop painting text when it got too much content |
| 19:40 | <zewt> | (doesn't crash, just chugs) |
| 19:40 | timeless | also remembers a time when too many characters in the urlbar would kill x servers |
| 19:40 | <timeless> | (there's at least one bugzilla bug on that in case people don't believe me) |
| 20:09 | <zewt> | am I the only one that finds the firefox "save password" thing almost always disappears before I can do anything with it now? |
| 20:09 | <zewt> | "save password? psyche!" |
| 20:20 | <AryehGregor> | . . . seriously? WebKit computes "font-style: oblique" to "font-style: italic"? |
| 20:20 | <AryehGregor> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!doctype%20html%3E%0A%3Cspan%20style%3Dfont-style%3Aoblique%3E%3C%2Fspan%3E%0A%3Cscript%3Ew(getComputedStyle(document.querySelector(%22span%22)).fontStyle)%3C%2Fscript%3E |
| 20:20 | <AryehGregor> | . . . |
| 20:20 | <Ms2ger> | Unsurprising |
| 20:20 | <AryehGregor> | Really? |
| 20:30 | AryehGregor | stabs browser CSSOM implementations. STAB STAB STAB STAB. |
| 20:31 | <AryehGregor> | Guess what node.style.length is for <span style="text-decoration: line-through">? |
| 20:31 | <AryehGregor> | . . . 4. |
| 20:31 | <AryehGregor> | -moz-text-blink, -moz-text-decoration-color, -moz-text-decoration-line, -moz-text-decoration-style. |
| 20:31 | <AryehGregor> | This is a regression, too, I'm pretty sure. |
| 20:31 | AryehGregor | works around it, grumble grumble |
| 20:34 | <AryehGregor> | Could someone check what this outputs in Firefox 4 and/or 5 and/or 6 and/or 7 for me? http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!doctype%20html%3E%0A%3Cspan%20style%3Dtext-decoration%3Aline-through%3E%3C%2Fspan%3E%0A%3Cscript%3Ew(document.querySelector(%22span%22).style.length)%3C%2Fscript%3E |
| 20:34 | <AryehGregor> | In 8 it's 4. |
| 20:34 | <AryehGregor> | It should be 1. |
| 20:35 | <Philip`> | In 6.0 it's 4 |
| 20:35 | <AryehGregor> | Hmm, so not a recent regression. |
| 20:35 | <Philip`> | 6.0 is pretty recent |
| 20:35 | <AryehGregor> | Conceivably it's been since 4 or 5, yeah. |
| 20:36 | <annevk> | AryehGregor, want to take over CSSOM? |
| 20:36 | <AryehGregor> | annevk, no. Have no time. |
| 20:36 | <AryehGregor> | More than enough editing stuff to do for now. |
| 20:36 | <AryehGregor> | Maybe someday. |
| 20:36 | <annevk> | AryehGregor, the serializing stuff is going to be taken over by the individual modules btw |
| 20:36 | <AryehGregor> | Yeah, it definitely should be. |
| 20:36 | <AryehGregor> | There needs to be a central place where it defines some useful primitives, but for nontrivial properties the details need to be specced per-property. |
| 20:37 | <annevk> | if you have details for the primitives and the grammar you should be okay |
| 20:38 | <annevk> | which is what CSSOM has now, it's just not very detailed (and the way CSS is written makes it hard to hook into things) |
| 20:43 | <zewt> | when at first you don't succeed, mail listserv commands to the list again |
| 20:43 | <zewt> | (don't most lists try to detect that and stop it?) |
| 20:52 | <annevk> | karlcow, don't get your comment on twitter |
| 21:18 | <AryehGregor> | jgraham, did you ever figure out why testharness.js was producing no output for my reflection tests anymore? |
| 21:19 | AryehGregor | is observing something similar in another file too, it seems |
| 21:19 | <AryehGregor> | Oh, never mind for the other thing. |
| 21:20 | <AryehGregor> | I wasn't running the tests before the load event. |
| 21:20 | <AryehGregor> | Reflection tests are still a question, though. :) |
| 21:36 | <AryehGregor> | . . . How is a JS file's encoding determined? |
| 21:45 | <annevk> | AryehGregor, based on the referring file, BOM, HTTP |
| 21:45 | <annevk> | and maybe a charset attribute on <script> |
| 21:45 | <annevk> | (not in that order) |
| 21:45 | <AryehGregor> | In what order of precedence? |
| 21:47 | <annevk> | HTTP, charset="", BOM, referring file |
| 21:47 | <annevk> | defined in HTML |
| 21:48 | <AryehGregor> | k, thx. |
| 21:49 | <zewt> | bom in .js? D : |
| 21:57 | <annevk> | WRONG_DOCUMENT_ERR is actually useful for Range, who'd have thought |
| 21:58 | <AryehGregor> | Is it? |
| 21:58 | <AryehGregor> | When is it thrown? |
| 21:58 | <AryehGregor> | compareBoundaryPoints. |
| 21:58 | <annevk> | comparePoint throws it per spec |
| 21:58 | <AryehGregor> | Makes sense there, I guess. |
| 21:59 | <AryehGregor> | Would make more sense to return a special value, but okay. |
| 21:59 | <annevk> | yeah dunno, I haven't actually tested any of this |
| 21:59 | <AryehGregor> | You noticed I have like a zillion Range tests? |
| 21:59 | <AryehGregor> | You should steal those. |
| 21:59 | <AryehGregor> | (note: they might be slightly hard to understand in some cases) |
| 21:59 | <AryehGregor> | (note: that also might be an understatement) |
| 22:00 | <AryehGregor> | https://bitbucket.org/ms2ger/dom-range/src/tip/test/ |
| 22:00 | <AryehGregor> | Nothing tests compareBoundaryPoints yet, though, it looks like. |
| 22:01 | <annevk> | yeah cool |
| 22:01 | <annevk> | guess we should move those to the DOMCore repo at some point |
| 22:02 | <AryehGregor> | It would make sense. |
| 22:04 | <nlogax> | might as well ask here too. :) can i get rid of the default drag&drop cursor? (a green (+) thingie on os x, some equally ugly thing on windows) |
| 22:04 | <nlogax> | other than setting dropEffect |
| 22:58 | <roc> | grr ... why is offline GMail a Chrome app and not a Web app? https://chrome.google.com/webstore/detail/ejidjjhkpiempkbhmpbfngldlkglhimk |
| 23:03 | <AryehGregor> | More general: why doesn't the Chrome Web Store support other browsers? |
| 23:03 | <AryehGregor> | "Sorry, we don't support your browser just yet. You'll need Google Chrome to install apps, extensions and themes." |
| 23:03 | <AryehGregor> | That makes it sound like they intend to add support eventually, but . . . |
| 23:06 | <AryehGregor> | I imagine they will add support for other browsers eventually, since to do otherwise seems inconsistent with their general approach to the web. But it's pretty annoying that they seem not to view that as urgent. |
| 23:06 | AryehGregor | has no more idea than anyone what the Chrome Web Store or Gmail people are thinking, obviously |
| 23:07 | <roc> | who knows, but there must certainly be a huge temptation to follow Apple and Microsoft into proprietary app stores |
| 23:08 | <AryehGregor> | Why? Google makes virtually all its money off ads, and Firefox displays ads as well as Chrome does. |
| 23:09 | <AryehGregor> | What money it does make directly off the store will only be increased if it allows more browsers to use it. |
| 23:10 | <AryehGregor> | Unlike Microsoft and Apple, Google doesn't have a history of trying to obtain lock-in, and in fact has often tended to do the opposite -- make switching to other services as easy as possible. |
| 23:10 | <AryehGregor> | E.g., Gmail supports unlimited forwarding, IMAP, etc. out of the box, which (AFAIK) some other major webmail providers don't. |
| 23:11 | <roc> | yeah, they've been good |
| 23:11 | <roc> | but if they get lots of good stuff into the Chrome app store, then that increases Chrome usage |
| 23:12 | <roc> | and that translates into power over the Web, and app developers |
| 23:12 | <zewt> | google's mobile browser support has been terrible for web apps, in my experience |
| 23:12 | <zewt> | seems like they really don't care |
| 23:13 | <roc> | we'll see how things go down |
| 23:13 | <zewt> | just trying to implement a basic full-screen web app in android's browser (at least as of 2.3.x) is a nightmare |
| 23:14 | <roc> | but I really want offline GMail and I don't want to have to switch to Chrome to get it (nor do I want other people to have to switch to Chrome to get it, of course) |
| 23:14 | <zewt> | "installed web apps" should be considered an oxymoron |
| 23:15 | <zewt> | all of the "chrome apps" and all that nonsense just kills what's great about web apps |
| 23:22 | <roc> | offline GMail uses WebSQL, which is dead as a standard |
| 23:22 | <zewt> | (because using a language that every programmer in the world already intuitively understands is overrated; we Need More Wheels) |
| 23:25 | <jamesr_> | the new offline gmail uses websql? i thought it just used filesystem (although i'm not directly involved) |
| 23:25 | <roc> | in fact, the "SQL" every programmer in the world intuitively understands is different things to different people and often quite different to the SQLite 3.6.19 or whatever that WebSQL is |
| 23:25 | <roc> | jamesr_: khuey looked at it and that's what he says |
| 23:25 | <roc> | but regardless of the detabable merits of WebSQL, its deadness as a standard is a fact |
| 23:26 | <jamesr_> | good thing offline gmail isn't part of the web, then |
| 23:27 | <roc> | two wrongs don't make a right |
| 23:27 | <zewt> | roc: it doesn't matter if it's different in the details, the basic language (and most of the day-to-day as well) is well-understood, which makes learning variants easy |
| 23:27 | <AryehGregor> | roc, there's not any realistic alternative, though, given the low adoption so far of IDB. Not to mention the fact that it's staggeringly complicated to use for even the most trivial use-cases. |
| 23:27 | <jamesr_> | i have no idea what lead to that decision |
| 23:27 | <AryehGregor> | One line of localStorage that's instantly understandable to any web developer is so many lines of IDB that I've given up every time I tried it. |
| 23:28 | <roc> | AryehGregor: it is supported in Chrome though, so if you're going to write a Chrome app ... |
| 23:28 | <AryehGregor> | Well, true. |
| 23:28 | <zewt> | (similar to a major reason svn was so successful: it was very intuitive to cvs users, so you only had to learn the differences, where eg. git being so different gives it a much higher learning curve) |
| 23:28 | <AryehGregor> | But I suspect it's a lot harder to learn and use, especially given that anyone writing this sort of app already knows some sort of SQL. |
| 23:29 | <AryehGregor> | At least, I'm pretty sure I could pick up WebSQL a heck of a lot more easily. |
| 23:29 | <AryehGregor> | I wouldn't be surprised if WebSQL has way more features and better performance at this point, too. |
| 23:30 | <AryehGregor> | Not that I'm saying people should use WebSQL instead of IDB, but there are pretty clear reasons for doing so. |
| 23:34 | <AryehGregor> | Does anyone know if there's a Mozilla bug for supporting Node.contains()? |
| 23:34 | <AryehGregor> | If not, I'll file one. |
| 23:35 | <AryehGregor> | Found it: https://bugzilla.mozilla.org/show_bug.cgi?id=683852 |
| 23:37 | <AryehGregor> | Okay, does anyone know what the status of my reflection tests are in the HTMLWG? |
| 23:37 | <AryehGregor> | They're still not "approved", it seems? |