| 00:01 | <zewt> | can browsers implement rel=noreferrer already so "anonymous link" sites can die horribly |
| 00:02 | <The_8472> | uhm... you know... you could just block cross domain referrers |
| 00:03 | <zewt> | uhm, that wouldn't magically make sites using anonymous link proxies stop doing so |
| 00:03 | <zewt> | i have GM scripts to undo some of them, but i shouldn't have to, and it doesn't work for all of them (not all of them encode the real URL as a query parameter) |
| 00:03 | <The_8472> | if that's your issue... there are addons that clean up the urls as long as the actual target is embedded in the link |
| 00:04 | <The_8472> | anyway, noreferrer only solves half of the problem. just look at google. they're linkjacking their own links to track what you've clicked on. |
| 00:05 | <zewt> | that's a separate (and roughly equally disgusting) problem |
| 00:05 | <The_8472> | it's not just the target sites that harvest data. it's the source sites too |
| 01:17 | <rniwa> | annevk: yt? |
| 01:17 | <rniwa> | annevk5: ^ |
| 01:21 | <Hixie> | ironically one of the hardest parts of this <dialog> thing is likely to be working out how to handle the positioning |
| 01:21 | <Hixie> | position:fixed or position:absolute? |
| 01:21 | <Hixie> | has to be absolute so you can scroll it if it's too big for the page |
| 01:21 | <Hixie> | but |
| 01:21 | <Hixie> | do you want to scroll the stuff under it in that case? |
| 01:21 | <Hixie> | how do you make sure it is visible when it appears in that case? |
| 01:22 | <Hixie> | how do we make the "cover" div under it actually cover everything if the dialog is position:absolute? |
| 01:22 | <Hixie> | what an odd mess |
| 01:22 | <Hixie> | oh well, for later. |
| 01:46 | <jamesr> | Hixie, what's <dialog>? modal dialog? |
| 01:52 | <hober> | indeed |
| 02:00 | <rniwa> | Hixie: can we get your response for http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2011-June/032078.html ? |
| 02:01 | <rniwa> | Hixie: I have a WebKit patch that's sort of blocked by this thread |
| 03:58 | <Hixie> | AryehGregor: can you help out rniwa? (see above) |
| 05:10 | <Yuhong> | Hixie: I'd love to comment on your posts, but I don't have a Google+ account. |
| 06:25 | <Hixie> | Yuhong: what's your e-mail address? i can send you an invite |
| 06:25 | <Hixie> | Yuhong: though it's really only fun to use g+ if your close friends and family are on it |
| 06:26 | <Yuhong> | I just wish to comment on your posts. |
| 06:26 | <Yuhong> | Anyway: yuhongbao_386 at hotmail.com |
| 06:28 | <Hixie> | i shared something with you so you should be able to join now |
| 08:43 | <hsivonen> | how do these bogus spec bugs that contain the source code of an HTML document come about? |
| 08:43 | <hsivonen> | do spammers expect some system somewhere to serve pasted source as an HTML page? |
| 08:43 | <hsivonen> | do people accidentally manage to paste view source contents into the comment form? |
| 08:50 | <zcorpan> | maybe people go "wtf is this", then paste whatever they have in their clipboard, then click some buttons |
| 08:52 | <zcorpan> | maybe the comment form needs an anti-bogo checkbox saying "I understand that submitting this form will file a bug in the W3C Bugzilla" |
| 09:04 | <zcorpan> | i don't like the a11yTF bugs that just say current text and suggested new text without mentioning what problem they are trying to solve or highlight what the difference is between the new and old text |
| 09:05 | <zcorpan> | so readers have to first carefully read both texts to find out what the difference is, and then make a guess at what problem they're trying to solve |
| 09:05 | <zcorpan> | which means that i just don't read those bugs at all |
| 09:05 | <zcorpan> | hi john, btw |
| 09:24 | <annevk> | Hixie, around now, somewhat |
| 09:27 | <annevk> | oh miwa was asking |
| 09:27 | <annevk> | I defined superglobal id/class |
| 09:28 | <annevk> | not really satisfied with it yet, but it's there |
| 11:24 | <matjas> | what’s the (hi)story behind the element@attribute syntax? |
| 11:25 | <matjas> | I knew about element/@attribute (XPath) but e.g. https://bugzilla.mozilla.org/show_bug.cgi?id=573322 uses element@attribute |
| 11:25 | <matjas> | where does it originate? |
| 11:28 | <annevk> | @attribute originates from XPath and people just went with it from there |
| 11:32 | <matjas> | I figured as much, thanks for confirming :) |
| 14:03 | <zcorpan> | if <button form formaction> etc is a security problem because todays filters use blacklists instead of whitelists, then surely <video src=? onloadstart="..."></video> is more of a security problem |
| 14:04 | <zcorpan> | or do blacklists uniformly block attributes that start with "on"? |
| 14:05 | <Ms2ger> | I sure hope so |
| 14:07 | <annevk> | people raised new event handlers as security issue |
| 14:09 | <jgraham> | In other news, using blacklists is a security problem |
| 14:09 | <zcorpan> | i wonder if spammers are going to love the hidden attribute |
| 14:10 | <zcorpan> | (where site authors/admins use modern browsers and don't see the spam while a non-zero amount of users use legacy browsers that don't support hidden and will see the spam) |
| 14:10 | <danj> | and in other news, splitting atoms in your kitchen is a bad idea |
| 14:11 | <Ms2ger> | Is it really?! |
| 14:12 | <danj> | yup - http://hosted.ap.org/dynamic/stories/E/EU_SWEDEN_NUCLEAR?SITE=AP&SECTION=HOME&TEMPLATE=DEFAULT&CTIME=2011-08-03-11-33-45 |
| 14:15 | zcorpan | adds download to html-elements |
| 14:16 | jgraham | notes that atom-splitting in the kitchen happens all the time |
| 14:18 | <zcorpan> | hsivonen: http://www.w3.org/Bugs/Public/show_bug.cgi?id=13604 - the dom wouldn't be identical anyway, since in text/html the textContent would contain "<![CDATA[" while in XML it wouldn't |
| 14:20 | <hsivonen> | zcorpan: not in foreign content |
| 14:21 | <zcorpan> | hsivonen: i thought the bug was discussing html script elements |
| 14:23 | <hsivonen> | zcorpan: I clarifyied in case I misunderstood |
| 14:41 | <annevk> | In case it comes up or someone is asking, I'm away Mon-Sat |
| 14:42 | <zcorpan> | we're not publishing next week right? |
| 14:43 | <annevk> | HTML? |
| 14:43 | <annevk> | I don't think so |
| 14:44 | <zcorpan> | ok |
| 15:19 | <Philip`> | hsivonen: Re http://krijnhoetmer.nl/irc-logs/whatwg/20110802#l-459 : I noticed http://www.w3.org/Bugs/Public/show_bug.cgi?id=13508 uses both Nadia and Aidan, and http://www.w3.org/WAI/PF/HTML/wiki/Spec_Review/Interactive_Elements says "Aidan is the opposite of Nadia" |
| 15:19 | <Philip`> | which seems an unlikely coincidence so I guess they were intentionally chosen to be palindromic |
| 15:23 | <zcorpan> | what's with the "06" in e.g. http://krijnhoetmer.nl/irc-logs/whatwg/20110804#l-761 ? |
| 15:24 | <timeless> | --note, the following messages were composed while offline (see comments about using bugzilla while offline...) -- |
| 15:24 | <timeless> | AryehGregor: mikesmith could read the irc logs :) |
| 15:24 | <timeless> | AryehGregor: there have been proposals for dealing w/ mass changes |
| 15:24 | <timeless> | sadly, as someone who does mass changes and deals w/ them, it's actually valuable to see each individual message |
| 15:24 | <timeless> | since i often screw up say 3% of the bugs i touch in mass changes |
| 15:24 | <timeless> | and there's no way you'll spot that while doing the mass change nor in a single email report |
| 15:24 | <timeless> | jgraham: in theory one could try to do it directly, but bugzilla keeps track of whom has seen which comment/change |
| 15:24 | <timeless> | which means it requires touching each user reference for each bug |
| 15:24 | <timeless> | otherwise the next touch to the bug will trigger the unsent changes |
| 15:25 | <timeless> | AryehGregor: depending on what you want to do w/ mass changes, you can open the query in a decent version of gmail and use <n> to quickly scan through them |
| 15:25 | <timeless> | or you can use select all (in that decent version of gmail) and mark as read |
| 15:25 | <timeless> | (there are lame versions of gmail -- in case you're wondering-- where you can't do these things, e.g. the java, html [/h], and mobile [/x]) |
| 15:25 | <timeless> | zewt: importance also goes the other way, someone will set a very important thing to trivial so that it won't flag a manager :) |
| 15:25 | <timeless> | Philip`: i make mass changes to thousands of bugs. having to click thousands links in a bugmail to get back my thousands of bugmails would not be fun |
| 15:25 | <timeless> | AryehGregor: there are a couple of reasons to want functional archives of bugzilla: |
| 15:25 | <timeless> | 1. bugzilla search sucks -- gmail search (free text) is much better |
| 15:25 | <timeless> | 2. when you make a mistake, it's much easier to undo it if you have the bugmail |
| 15:25 | <timeless> | 3. bugzilla data reflects current info (e.g. see when AryehGregor renamed his bugzilla account), bugmail reflects data as it was |
| 15:25 | <timeless> | 4. some people work offline (even gmail supports this) |
| 15:25 | <timeless> | AryehGregor: re calendar to find day of week, that's awesome |
| 15:25 | timeless | just hovers over the w7 system notification area and it tells me the day of week |
| 15:25 | timeless | is no longer offline |
| 15:26 | <jgraham> | I feel like someone just opened a time capsule |
| 15:27 | <jgraham> | But from yesterday, not 1979 |
| 15:27 | <jgraham> | Or whenever time capsules were mostly buried |
| 15:27 | <Philip`> | zcorpan: Probably ANSI escape codes or similar |
| 15:27 | <timeless> | i spent yesterday unpacking my shipped apartment |
| 15:27 | <timeless> | some of which had boxed from a previous move (5 years ago) |
| 15:28 | <timeless> | and some of which have stuff from the move before that (closer to 8 years ago) |
| 15:28 | <timeless> | so i know what you mean re time capsules :) |
| 15:28 | <timeless> | but, that's also why i wasn't replying in real-time... |
| 15:29 | <Philip`> | Oh, actually mIRC colour escapes |
| 15:29 | <Philip`> | since it's an 0x03 at the start |
| 15:29 | <zewt> | heh, mirc codes are pretty sadly amusing |
| 15:29 | <timeless> | can someone ask the blogger of http://blog.chromium.org/2011/08/connecting-web-apps-with-web-intents.html to fix their quotes to use ['] instead of the fancy ones they use? |
| 15:29 | <zewt> | apparently the guy invented them because he didn't know that ansi color codes already existed and had been in use on irc for years? |
| 15:30 | timeless | chuckles |
| 15:30 | timeless | isn't surprised |
| 15:31 | <zewt> | essentially, reinventing the wheel after you've already been selling cars for years |
| 15:31 | <Philip`> | timeless: I think my idea was that you'd get a single bugmail with a single link to click, which would then cause the thousands to be sent to you |
| 15:31 | <Philip`> | so it wouldn't be not fun |
| 15:32 | <timeless> | Philip`: subscriber preferences make more sense i think |
| 15:32 | <timeless> | i haven't met someone who only sometimes wants all bugmail changes |
| 15:32 | <timeless> | although, i could be wrong |
| 15:32 | <timeless> | the problem is that bugzilla really can't create those emails later |
| 15:33 | <timeless> | e.g. if the mass changer is AryehGregor, and i get the email and then he changes his email address |
| 15:33 | <timeless> | and then i click the link asking for his mass changes to be split |
| 15:33 | <timeless> | the emails bugzilla would give me would have the *wrong* changer address in them |
| 15:33 | <timeless> | and if he changes his real name to claim he's on vacation |
| 15:33 | <timeless> | then i'd see that he sent them while on vacation |
| 15:33 | <timeless> | ... |
| 15:33 | <timeless> | the result is bugzilla would have to queue the possible email for all possible recipients |
| 15:33 | <timeless> | which would be really bad for storage for bugzilla |
| 15:34 | <Philip`> | Disk space is cheap, so queueing is easy |
| 15:34 | <Philip`> | They'd compress well too :-) |
| 15:34 | <timeless> | well, they would compress well yes |
| 15:34 | <zewt> | why? save the body of the outbound message, one per change |
| 15:34 | <timeless> | but bugzillas tend to use mysql |
| 15:34 | <timeless> | which probably sucks at compressing |
| 15:35 | <timeless> | they'd only compress well as a complete set, not as individual messages |
| 15:35 | <Philip`> | Concatenate all the messages, gzip them, then store them in a BLOB column, and hope no database administrator notices your terrible abuse of SQL |
| 15:35 | <timeless> | i suppose you could compose an MBox for each user |
| 15:35 | <zewt> | would also allow you to do things like request a full email audit trail of a bug later on |
| 15:35 | <gsnedders> | timeless: MySQL can compress parts |
| 15:36 | <hsivonen> | Philip`: mindblowing. Is there Krid, too? |
| 15:36 | <timeless> | gsnedders: databases are generally designed for pigeon holes, and compressing an individual hole doesn't gain much |
| 15:36 | <Philip`> | Alternatively, maybe emails could have some extra header which contains some globally unique mail-set identifier, and mail clients could have a "select all in same set as this message" feature |
| 15:36 | <zewt> | well, it would if you were storing a block of text repeated thousands of times, but again I don't see why you need to store them all |
| 15:36 | <timeless> | but i suppose if you arranged a portion for inidividual recipients... it might be doable |
| 15:37 | <timeless> | Philip`: so... |
| 15:37 | <Philip`> | then bulk change mails would get put into the same set |
| 15:37 | <timeless> | on that count, there is a mass change header in the mails |
| 15:37 | <timeless> | so if your client *isn't* gmail, you already have that! |
| 15:37 | <timeless> | it's in x-bugzilla-change-reason: iirc |
| 15:37 | <zewt> | gmail is horrifyingly stupid about headers :| |
| 15:37 | <Philip`> | Ah, excellent |
| 15:37 | <timeless> | zewt: ue[ |
| 15:37 | <timeless> | yep |
| 15:37 | <timeless> | so, we're back to gmail being the only disadvatanged client |
| 15:38 | <zewt> | (never mind that it literally has no concept of sorting) |
| 15:38 | <timeless> | yeah, which is another problem |
| 15:38 | <zewt> | gmail is good at a lot of things but it has these bizarre holes |
| 15:38 | <timeless> | i think gmail lists messages by delivery date |
| 15:38 | <timeless> | which means if i trigger the email flood after a reply to the flood |
| 15:38 | <timeless> | the message sequence will look bizarre |
| 15:38 | <timeless> | i already get that in Hg land |
| 15:38 | <zewt> | that's the user's problem, i'd say |
| 15:38 | <Philip`> | Someone who cares about mail should smuggle themselves into Google and fix Gmail |
| 15:39 | <timeless> | where smaller messages sent later are delivered before larger messages sent earlier! |
| 15:39 | <timeless> | (and are thus listed closer to 'beginning of time') |
| 15:39 | <zewt> | also no way in gmail to view messages *to me* :O |
| 15:40 | <timeless> | zewt: ? |
| 15:40 | <zewt> | eg. all mail minus mailing list mail |
| 15:40 | <timeless> | there's a >> or > indicator on things |
| 15:40 | <Ms2ger> | On the subject of email |
| 15:40 | <timeless> | although that might only apply to conversations as opposed to messages |
| 15:40 | <zewt> | i don't know of any way to search for "no labels" |
| 15:40 | <timeless> | perhaps you have to turn that on or off |
| 15:40 | <zewt> | (not exactly what I want, but close) |
| 15:40 | <timeless> | *that*'s indeed annoying |
| 15:40 | <Ms2ger> | Why does MS's Alex use plus-minus signs to quote? |
| 15:40 | <timeless> | i really would like <has-no-labels> |
| 15:41 | <timeless> | Ms2ger: hrm, good question |
| 15:42 | <timeless> | would it be bad if i asked the author of http://blogs.forbes.com/fredcavazza/2011/07/17/why-opposing-html5-and-flash-is-a-non-sense/ to fix his spelling of a Trademark? |
| 15:42 | <zewt> | also while randomly ranting about gmail, there's no way to search or sort my message size, so if you have a bunch of attachments using up gigs of space, there's pretty much no way to find them |
| 15:42 | <zewt> | (short of downloading the whole lot and loading it in another client) |
| 15:43 | <timeless> | zewt: yep |
| 15:43 | <timeless> | actually, i think you can try using an imap client |
| 15:43 | <timeless> | but perhaps we don't count that :) |
| 15:43 | <zewt> | timeless: "is a non sense" heh |
| 15:43 | <timeless> | ? |
| 15:43 | <zewt> | always funny when things like that are edited in the text but left unchanged in the url |
| 15:43 | <timeless> | oh |
| 15:43 | <timeless> | lol |
| 15:44 | <timeless> | cute |
| 15:44 | timeless | kicks firefox find in page for not finding in url |
| 15:44 | <zewt> | reminds me of the old ^H^H^H^H^H^H |
| 15:44 | <timeless> | yep |
| 15:44 | <timeless> | anyway, think he'd respond well to such a request? |
| 15:46 | <zewt> | dunno, i skimmed his "points", saw too much non sense(sic) in them and closed the tab |
| 15:46 | <zewt> | (we can't replace flash! it's been around for TWELVE YEARS!) |
| 15:46 | <timeless> | his statements aren't mostly wrong :) |
| 15:47 | <gsnedders> | HTML has been around for TWENTY ONE YEARS. Flash can't replace that! |
| 15:48 | <Philip`> | If HTML5 doesn't officially exist because it's not a published W3C standard yet, does Flash not exist either since it's also not one? |
| 15:49 | <Ms2ger> | Does that also mean MS Word doesn't exist? |
| 15:49 | <jgraham> | Does that also mean that Descartes doesn't exist? |
| 15:50 | <reggna> | Do I exist? |
| 15:51 | <timeless> | no |
| 15:52 | <jgraham> | I don't think |
| 15:52 | <zcorpan> | i am not a W3C recommendation, therefore I don't exist |
| 15:52 | <timeless> | zcorpan: we're working on rescinding those |
| 15:53 | <timeless> | therefore even if you were, you'd be subject to ceasing to exist within ... 4 months? |
| 15:53 | <timeless> | annevk: how's that going btw? :) |
| 15:53 | <zcorpan> | timeless: murder! |
| 15:53 | Ms2ger | asks ArtB |
| 15:54 | <timeless> | zcorpan: planned obsolesence |
| 15:55 | <annevk> | timeless, haven't looked too much into it yet |
| 15:55 | <timeless> | annevk: i was glad to see ian was responsive |
| 15:56 | <timeless> | hopefully we can get it done in closer to 2 months or even 6 weeks :) |
| 15:56 | <timeless> | afterwards, i think we can plan to go duck hunting and bag a larger set of lame ducks |
| 16:08 | <AryehGregor> | Hixie, I think the right answer here is obviously to make it logical order, so that "character" still moves forward one character and "line" still moves forward a bunch of characters. But I don't know the first thing about vertical writing, rniwa would probably have a much more informed opinion. |
| 16:08 | <AryehGregor> | I still don't really know what the use-cases are for modify(). |
| 16:10 | <gsnedders> | AryehGregor: one character or one grapheme? |
| 16:10 | <gsnedders> | Grapheme is surely what would be expected? How do you put the cursor in the middle of a grapheme. |
| 16:10 | <AryehGregor> | gsnedders, most likely a "grapheme cluster", in the terminology of UAX#29. |
| 16:11 | <AryehGregor> | That's not what I was talking about, though. |
| 16:11 | <gsnedders> | Yeah, I know. |
| 16:12 | AryehGregor | 's eyes glaze over slightly as he reads the definition of "extended grapheme cluster" |
| 16:12 | <gsnedders> | Line is probably still what you want with vertical writing, AIUI, just the positioning of the next line is different. |
| 16:13 | AryehGregor | has no idea if the rules make any sense whatsoever for his use-cases, because he has no idea how things like "the spacing (but dependent) vowel signs in Indic scripts" are actually used |
| 16:13 | <AryehGregor> | RTL is a piece of cake compared to some of the stuff you see in Unicode. |
| 16:16 | <Hixie> | AryehGregor: well, the stuff in question is in your spec now, not mine, so i wash my hands of it :-P |
| 16:16 | <AryehGregor> | :) |
| 16:16 | <AryehGregor> | I would have replied to rniwa, but he's apparently one of those annoying people who only stays on IRC when he's actually present. |
| 16:16 | <AryehGregor> | Rather than idling all the time. |
| 16:17 | <Ms2ger> | :( |
| 16:17 | AryehGregor | notes that it's Ms2ger's spec too, and in fact Ms2ger is the only listed editor, so it's only fair to push the problem off on him |
| 16:18 | <Hixie> | oh, didn't realise that |
| 16:18 | <Ms2ger> | Not anymore |
| 16:18 | <Hixie> | hah |
| 16:18 | <AryehGregor> | Drat. |
| 16:18 | Ms2ger | removes himself |
| 16:18 | <AryehGregor> | https://bitbucket.org/ms2ger/dom-range/changeset/eb238987d64e |
| 16:18 | <AryehGregor> | I didn't notice that. |
| 16:18 | <AryehGregor> | Oh well. |
| 16:18 | <AryehGregor> | We were planning to chop it up and redistribute it among other specs anyway. |
| 16:19 | <Ms2ger> | Yeah |
| 16:19 | <Ms2ger> | And that question is very much yours :) |
| 16:19 | <AryehGregor> | Of course, now Microsoft is complaining about Traversal being in Core, so they'd presumably complain about Range too. |
| 16:20 | <AryehGregor> | And I don't want to steal Selection for the editing spec yet because I'm getting lots of review right now and don't want to rock the boat. |
| 16:25 | <Hixie> | jgraham: i get 504s quite frequently these days |
| 16:25 | <Hixie> | jgraham: even without the annotations enabled |
| 16:31 | <jgraham> | Hixie: Hmm. Right now or a few days ago? |
| 16:31 | <Hixie> | had one a few minutes ago |
| 16:31 | <Hixie> | it's transient |
| 16:31 | <Hixie> | if i retry sufficient times it goes away |
| 16:31 | <jgraham> | Hmm, OK |
| 16:31 | <jgraham> | I will try to look at the problem |
| 16:32 | <Hixie> | btw i have the annotation stuff disabled right now because it was failing too often also |
| 16:47 | <zewt> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=13264 do bug trolls really need a serious response? heh |
| 16:47 | <Hixie> | that's not a troll, he's probably right |
| 16:47 | <Hixie> | he's certainly right for the html spec |
| 16:47 | <Hixie> | dunno what we can add for storage though |
| 16:48 | <Ms2ger> | http://rlv.zcache.com/chocolate_chip_cookie_jar_postcard-p239265891586692704trdg_400.jpg |
| 16:48 | <zewt> | suggesting pictures in web storage sure sounds 100% trollish to me, heh |
| 16:48 | <Philip`> | Trolls can be right |
| 16:48 | <AryehGregor> | Okay, I want to post something like this in the public-webapps thread, but I probably shouldn't: "Well, on the other hand, we now have a senior member of the W3C administration demanding that the W3C's editors only reference W3C standards, even if those editors feel parallel standards developed at other standards bodies are superior. It seems to me that's *definitely* for no reason other than politics, and also undermines and is disrespectfu |
| 16:48 | <AryehGregor> | l of the work of the WebApps WG. But hey, it's your call. If you alienate editors, it doesn't hurt anyone but the W3C." |
| 16:48 | <Ms2ger> | You shouldn't |
| 16:49 | <AryehGregor> | I guess I'll just remain silent. |
| 16:49 | <Hixie> | Ms2ger: hah, nice picture |
| 16:49 | <Ms2ger> | I would appreciate that in the spec ;) |
| 16:49 | <zewt> | the "too much reading" in particular is what makes it stand out as trolly to me, heh |
| 16:49 | <AryehGregor> | Although the hypocrisy of him saying we have to remove all references to the WHATWG standards and then accusing us of politics is galling. |
| 16:49 | <Ms2ger> | Maybe you could photoshop it onto the kitchen sink |
| 16:50 | <Hixie> | Ms2ger: hah |
| 16:50 | <Hixie> | Ms2ger: is it cc? |
| 16:50 | <Ms2ger> | Doubt it |
| 16:50 | <Hixie> | the sink is :-) |
| 16:50 | <Ms2ger> | I know |
| 16:50 | <Hixie> | got an attribution in there and everything |
| 16:50 | <Ms2ger> | Even microdata processors know that |
| 16:50 | <Hixie> | :-P |
| 16:51 | <AryehGregor> | I'm not sure I want to have a Bugzilla component in the W3C. That gives them leverage. |
| 16:51 | <AryehGregor> | Maybe I'll use Google Code instead. |
| 16:51 | <smaug____> | then I won't fire any bugs |
| 16:51 | <smaug____> | file |
| 16:51 | <AryehGregor> | If it's in Google Code, you mean? |
| 16:51 | <smaug____> | could whatwg have its own bugzilla? |
| 16:51 | <smaug____> | AryehGregor: right |
| 16:52 | <AryehGregor> | We could, but then everyone would have to make an account. |
| 16:52 | <AryehGregor> | Which is annoying. |
| 16:52 | <Ms2ger> | People sure would like that |
| 16:52 | <AryehGregor> | Lots of people already have W3C Bugzilla accounts, and practically everyone has a Google account. |
| 16:52 | <smaug____> | I'm about to give up my Google account ... |
| 16:52 | <Philip`> | http://www.flickr.com/photos/apreche/3442252639/ is CC BY |
| 16:53 | <AryehGregor> | I don't want people to have to remember Yet Another Password to file bugs. I guess we could set up something like the contributor⊙wo system, but that sounds like work. |
| 16:53 | <AryehGregor> | Anyway, I'll still accept feedback via e-mail regardless, and file it myself if I can't act on it quickly. |
| 16:53 | <Hixie> | i can give you the file-bug.cgi script if you want |
| 16:53 | <AryehGregor> | I suppose I'll keep the W3C Bugzilla component for now. |
| 16:54 | <AryehGregor> | I assume nobody's going to do anything as drastic as delete all the bugs without warning. |
| 16:54 | <AryehGregor> | So it should be safe. |
| 16:54 | <Hixie> | the worst that i could see happening is that the bugmail notifications would be turned off and the bugs would all be reassigned or something like that |
| 16:54 | <Hixie> | which would make recollecting them a huge pain |
| 16:54 | <Hixie> | but not impossible |
| 16:54 | <AryehGregor> | I really should avoid getting into political sniping anyway. |
| 16:55 | <Hixie> | (or all closed) |
| 16:55 | <AryehGregor> | I don't think anyone will do anything hostile without fair warning, at least not for now. |
| 16:56 | <shetech> | one hopes |
| 16:56 | <Hixie> | the w3c isn't evil |
| 16:56 | <Hixie> | they're just using an archaic process |
| 16:56 | <shetech> | Heh. I wasn't talking about w3c, but I'm occasionally a cynic. ;-) |
| 16:57 | <AryehGregor> | No, but every time something happens like plh demanding that W3C specs only reference W3C specs, that makes me want to have less and less to do with the W3C. |
| 16:57 | <gsnedders> | AryehGregor: only reference W3C specs when they publish the document, rather, which doesn't seem that unreasonable |
| 16:57 | <shetech> | Ah, politics. It's what makes the world go 'round. |
| 16:58 | <Hixie> | AryehGregor: oh the bureaucracy at w3c certainly makes me wish i could just ignore them entirely |
| 16:58 | <Hixie> | gsnedders: well except for the part where "publish" is itself unreasonable. :-P |
| 16:58 | <Hixie> | bbia |
| 16:58 | <Hixie> | b |
| 16:58 | <AryehGregor> | gsnedders, it seems unreasonable from my perspective as an editor that a W3C administrator would think he can waste my time telling me where my informative references have to point. |
| 16:59 | <dglazkov> | good morning, Whatwg! |
| 16:59 | <shetech> | Pardon me, but isn't that the w3c administrator's gig? To *pull* what they want? |
| 17:00 | <AryehGregor> | Well, yes. I kind of object to the existence of a W3C administration, basically. |
| 17:00 | shetech | snorts |
| 17:00 | <shetech> | (in a humorous way) |
| 17:00 | <AryehGregor> | We get along fine without any administration at the WHATWG. |
| 17:01 | <shetech> | yes. I heart the way whatwg is running itself. it sure makes my job easier! Oh, except for the fact that whatever I write is pretty much obsolete next week. :D |
| 17:01 | <shetech> | No matter, I like the whatwg approach better. |
| 17:01 | <shetech> | Makes more sense |
| 17:02 | <jgraham> | What is obsolete next week? |
| 17:02 | <shetech> | Not really anything *from* the whatwg... more that comments about browser behavior aren't "current" for very long |
| 17:03 | <jgraham> | Blame browsers for that not the WHATWG |
| 17:03 | <shetech> | I ran into one, for example, about differences between how Chrome and FF handle svg files using the <img> tag to place them. |
| 17:03 | <shetech> | Yes, browser fault, not whatwg |
| 17:03 | <shetech> | :) |
| 17:03 | <shetech> | I do think that the "living standard" approach just makes tons more sense in our agile world |
| 17:03 | <jgraham> | I mean who wants to live in a world where browsers fix bugs and get new, exciting features, really? |
| 17:03 | <jgraham> | ;) |
| 17:03 | <shetech> | HEH |
| 17:06 | <shetech> | besides, jgraham: IE doesn't. :P |
| 17:06 | <shetech> | that's what makes it so exciting! |
| 17:08 | <jgraham> | Well IE does now compared to 2004 |
| 17:08 | <jgraham> | Although of course lots of people still use 2004 era IE |
| 17:08 | <shetech> | True that. I worked on a team last year that still insisted on IE6. |
| 17:08 | <shetech> | Yoiks |
| 17:09 | <jgraham> | In related news, Microsoft are the only browser vendor who optimise for large enterprise customers (in pretty much any of their product line) |
| 17:10 | <jgraham> | (I mean Apple and Google don't really do "enterprise" except in very narrow ways, even on non-browser products) |
| 17:10 | <jgraham> | Possibly this explains the culture at Microsoft that likes numbered versions of things and formal processes and so on |
| 17:11 | <shetech> | Well, that's true enough. Apple especially, and Goog to a lesser extent are really intended for consumer level. |
| 17:11 | <shetech> | jgraham: would you be willing to provide a little more detail about how IE optimizes for large enterprise? Or am I making an incorrect connection? |
| 17:11 | <AryehGregor> | It would be nice if I could tell anolis not to do backlinks to some references. Specifically the ones in notes and things. |
| 17:13 | <jgraham> | shetech: Well I don't know since I neither use IE often nor know much about large enterprises. But I meant that they have very long support lifetimes and integrate with all the windows stuff for doing mass-deployment and locking down settings and so on |
| 17:13 | <Ms2ger> | Why? |
| 17:13 | <gsnedders> | shetech: Make it easy to deploy to thousands of system, deloy patches and other upgrades in such a way that fits in with standard workflows for testing new software, with long support times. |
| 17:14 | <jgraham> | I just said that! |
| 17:14 | <AryehGregor> | Why what? |
| 17:14 | <AryehGregor> | I don't want backlinks to non-normative stuff because I often use backlinks to figure out what's using an algorithm so I know if I can simplify or get rid of or replace it. |
| 17:15 | <AryehGregor> | But I want to use <span> and not manual <a> to link to the algorithm because otherwise the link will rot if I ever remove or rename it. |
| 17:15 | <shetech> | Hee. |
| 17:15 | <Ms2ger> | Mm |
| 17:15 | <shetech> | Thanks guys |
| 17:16 | <shetech> | jgraham and gsnedders, thanks. This is helpful (short and sweet). |
| 17:19 | <zewt> | AryehGregor: might be more useful to be able to visually separate normative and informative references in the backlinks popup |
| 17:19 | <Philip`> | AryehGregor: Would it be sufficient to use <a href=#whatever> if it detected and reported occurrences of #whatever with no corresponding id=whatever? |
| 17:19 | <AryehGregor> | zewt, also sounds like more work. |
| 17:19 | <zewt> | everything is work :) |
| 17:19 | <AryehGregor> | Philip`, that would work, yeah. It would be more verbose in the source code, but that's no big deal. |
| 17:20 | <Philip`> | (The spec splitter will warn about broken fragment links like that, I think) |
| 17:20 | <annevk> | AryehGregor, it's even better |
| 17:21 | <Philip`> | (http://www.whatwg.org/specs/web-apps/current-work/multipage/stdout.txt - like at the end of that) |
| 17:21 | <AryehGregor> | annevk, what is? |
| 17:21 | <annevk> | he's blocking publication of Progress Events because of a non-normative reference (used for an example) to the WHATWG copy of HTML |
| 17:21 | <AryehGregor> | Yeah, exactly. And anyone puts up with this nonsense why? |
| 17:22 | <annevk> | I might not much longer |
| 17:22 | <AryehGregor> | Nice. |
| 17:22 | <AryehGregor> | Or, not really nice, but hopefully it will cause the W3C to wake up a bit more. |
| 17:22 | <AryehGregor> | More like "necessary" than "nice". |
| 17:22 | <timeless> | doubtful |
| 17:23 | <jgraham> | Well it's not that surprising. plh's job depends on W3C being accepted as the biggest stanadsrds organisation in town, and companies like the patent policy |
| 17:23 | <jgraham> | *standards |
| 17:23 | <AryehGregor> | Is Opera willing to back you if you don't want to work anymore in the W3C? |
| 17:23 | <timeless> | fwiw, i'm pretty sure the w3 doc policy is pretty clear on referencing stable specs |
| 17:23 | <timeless> | and given that the whatwg specs are clearly marked as evolving |
| 17:23 | <AryehGregor> | jgraham, he seems not to realize that the way to do that is to make editors want to work there instead of bullying them into toeing the W3C party line. |
| 17:23 | <timeless> | i think their complaint is technically valid |
| 17:24 | <timeless> | albeit unfortunate and stupid |
| 17:24 | <AryehGregor> | timeless, no one was objecting to it on process grounds. |
| 17:24 | <AryehGregor> | HTML5 isn't stable either. |
| 17:24 | <AryehGregor> | It's still only a Last Call. |
| 17:24 | <AryehGregor> | Plus, this is an informative dependency, so stability is not required (right?). |
| 17:24 | <timeless> | that i'll have to recheck |
| 17:24 | <AryehGregor> | Only normative dependencies have to be as stable as the thing that's referencing them. |
| 17:24 | <timeless> | i haven't memorized the requirements maze |
| 17:24 | <AryehGregor> | Realistically, no version of HTML is going to be "stable" for a long time. |
| 17:25 | <jcranmer> | html 4 |
| 17:25 | <timeless> | i'm a 3.2 fan |
| 17:25 | <Ms2ger> | I'm a 3.0 fan |
| 17:26 | <zewt> | i'm a ceiling fan :( |
| 17:26 | <jgraham> | HTML+ has been very stable |
| 17:26 | <jgraham> | Not implemented, but very stable… |
| 17:26 | <timeless> | http://www.w3.org/MarkUp/html3/ |
| 17:26 | <timeless> | HTML 3.0 Draft (Exprired!) Materials |
| 17:26 | shetech | laughed out loud at zewt's riposte. Randomly, in the middle of a shared office. |
| 17:26 | AryehGregor | also laughed out loud |
| 17:26 | Philip` | is reminded of the quote "There is this special biologist word we use for 'stable'. It is 'dead'." |
| 17:27 | <timeless> | Philip`: yeah, that's come up in some contexts here at times |
| 17:27 | <timeless> | i think possibly relating to people wanting a stable version of Firefox |
| 17:27 | <annevk> | AryehGregor, no idea |
| 17:27 | <timeless> | one whose UI and feature set doesn't change |
| 17:27 | <jcranmer> | There is a special CS term for "dead" |
| 17:27 | <jcranmer> | "Debian stable" |
| 17:27 | <timeless> | lol |
| 17:27 | <Ms2ger> | jcranmer++ |
| 17:27 | <annevk> | timeless, non-normative references can in theory reference anything |
| 17:28 | <timeless> | jcranmer++ |
| 17:28 | <AryehGregor> | http://aryeh.name/quotes.html |
| 17:28 | <annevk> | timeless, I was willing to change the reference though before he cried foul in Team-only space |
| 17:28 | <timeless> | wow |
| 17:29 | <annevk> | timeless, now I'm probably still willing, but also less enthusiastic about working in that environment |
| 17:29 | <timeless> | annevk: prioritize: 1. kill as many w3 specs as fast as possible. 2. do the minimal work to get your stuff done. |
| 17:29 | <Ms2ger> | Hmm? |
| 17:30 | <timeless> | that way when you aren't in w3, there's less of w3 left to haunt you, and you don't have to worry about trying to get rid of those pieces when you aren't allow to affect it |
| 17:30 | <timeless> | <zcorpan_> going to file:/// gives me an xml parse error in firefox :-( |
| 17:30 | <timeless> | was zcorpan using a localized firefox? |
| 17:30 | <Ms2ger> | I don't think that was it |
| 17:30 | <Ms2ger> | But we fixed it |
| 17:30 | <timeless> | or did he have localized filenames which violated the encoding of the xml system |
| 17:30 | <timeless> | it's usually one or the other |
| 17:31 | <AryehGregor> | I think it was a non-printable ASCII character. |
| 17:31 | <timeless> | ah |
| 17:32 | timeless | sighs |
| 17:33 | <timeless> | every time i connect my phone, there's an OS update available |
| 17:33 | timeless | ponders |
| 17:33 | <timeless> | at nokia i wasn't willing to do updates to my phone more than monthly |
| 17:33 | <timeless> | probably because it broke way too many things each time |
| 17:41 | <dglazkov> | non-printable character: an unfortunate outcome of one of the cocaine-laced brainstorms at Marvel. |
| 17:49 | <karlcow> | "[12:32] <annevk> timeless, I was willing to change the reference though before he cried foul in Team-only space" |
| 17:49 | <karlcow> | cough cough not very clever from you annevk |
| 17:53 | <AryehGregor> | smaug____, why are you getting rid of your Google account, BTW? |
| 17:53 | <Philip`> | dglazkov: Maybe the character's costume is covered in those special symbols that are hidden on banknotes, so that printers will detect and refuse to reproduce them |
| 17:55 | karlcow | suddenly imagine the Super Hero costume with a Unicode error message on the shirt. |
| 17:55 | <smaug____> | AryehGregor: "big companies have too much influence in Internet" |
| 17:56 | <AryehGregor> | smaug____, well, okay. I mean, there's a reason for that: they usually have the resources to do stuff better than small organizations. |
| 17:56 | <smaug____> | and they just happen to do evil things, even if they try not to |
| 17:57 | <AryehGregor> | Ideally stuff would be tied together a bit less and federated a bit more, but Google's better at that than most (although it could probably do a lot better than it is). |
| 17:57 | <mhausenblas> | annevk available? |
| 17:58 | <mhausenblas> | can you help me re a CORS question pls? |
| 17:58 | <mhausenblas> | now, a month ago someone pointed out that I have a syntax mistake in http://enable-cors.org/#how-apache |
| 17:59 | <mhausenblas> | people pointed out that Access-Control-Allow-Origin: "*" is wrong and it should be Access-Control-Allow-Origin: * |
| 18:00 | <mhausenblas> | now, I get a request from someone saying 'I spent a while trying to debug this with our apache servers.' and actually it is the other way round |
| 18:00 | <mhausenblas> | so, my inital version was right (?) |
| 18:00 | <mhausenblas> | any idea, annevk? |
| 18:01 | <mhausenblas> | hmmm ... I guess you're not around annevk - I leave you a message on G+ then ... |
| 18:01 | <Philip`> | http://dvcs.w3.org/hg/cors/raw-file/tip/Overview.html#access-control-allow-origin-response-header looks like it says there shouldn't be any quotes |
| 18:02 | <gsnedders> | mhausenblas: there shouldn't be any quotes |
| 18:02 | <mhausenblas> | hmmm |
| 18:02 | <Philip`> | and later it says "If the Access-Control-Allow-Origin header value is the literal "*" character and the credentials flag is false return pass and terminate this algorithm." |
| 18:02 | <mhausenblas> | right |
| 18:02 | <Philip`> | which sounds like it's a single character |
| 18:03 | <mhausenblas> | so, the first request was right |
| 18:04 | <mhausenblas> | I just wonder why the new request (from a guy who claims that he's debugged it with Apache) thinks it's not the case |
| 18:04 | <mhausenblas> | I must confess that the spec is rather clear, the BNF says it all |
| 18:04 | <mhausenblas> | an Apache bug, maybe? |
| 18:04 | <mhausenblas> | nah |
| 18:05 | <Philip`> | Maybe 'Header set Access-Control-Allow-Origin *' should be 'Header set Access-Control-Allow-Origin "*"' |
| 18:05 | <gsnedders> | mhausenblas: What' she saying it's not the case? |
| 18:05 | <gsnedders> | s/What's/Why's/ |
| 18:05 | <Philip`> | since http://httpd.apache.org/docs/2.0/mod/mod_headers.html indicates it sometimes wants the value to be double-quoted |
| 18:05 | <Philip`> | (and it'll strip the quotes off before sending the header, presumably) |
| 18:05 | <mhausenblas> | hmmm |
| 18:06 | <mhausenblas> | the person who says that the current status at http://enable-cors.org/#how-apache is wrong literally says: |
| 18:06 | <mhausenblas> | In the 'For Apache' section, please change : |
| 18:06 | <mhausenblas> | Header set Access-Control-Allow-Origin * |
| 18:06 | <mhausenblas> | to : |
| 18:06 | <mhausenblas> | Header set Access-Control-Allow-Origin "*" |
| 18:06 | <mhausenblas> | I spent a while trying to debug this with our apache servers. |
| 18:06 | <Philip`> | The Apache documentation sounds pretty vague about syntax |
| 18:06 | <mhausenblas> | now I'm a bit in trouble, ain't I? |
| 18:06 | <Philip`> | like what strings have to be double-quoted |
| 18:06 | <Philip`> | and how you could include a double quote inside a string if you wanted to |
| 18:07 | <mhausenblas> | I mean I changed it a month ago cause I was convinced - so blame it on a bad Apache doc? ;) |
| 18:07 | <AryehGregor> | There, I think this is productive: https://plus.google.com/100662365103380396132/posts/dzK5yHL1nA3 |
| 18:07 | <mhausenblas> | :P |
| 18:07 | <mhausenblas> | anyways, I guess I have to toy around with it myself to figure what exactly is happening |
| 18:08 | <mhausenblas> | thanks for all your help, Philip` and gsnedders! |
| 18:08 | <mhausenblas> | now: FAWM |
| 18:08 | <Philip`> | Shifting blame to the documentation is always a good idea :-) |
| 18:09 | <mhausenblas> | he he |
| 18:09 | <mhausenblas> | ok, catch you laters online |
| 18:09 | <mhausenblas> | cya |
| 18:09 | <zewt> | heh, for htaccess directives i usually end up looking for an example, since the apache docs are ... not so great |
| 18:09 | <Philip`> | AryehGregor: Is G+ automatically expanding the link and putting the "Reference to the HTML specification. This message : [ Message body ] [ Respond ] [ More options ] ..." garbage in there? |
| 18:10 | <AryehGregor> | Philip`, yes. |
| 18:10 | <AryehGregor> | It doesn't notice that's boilerplate. |
| 18:10 | <AryehGregor> | Clearly, this would be a great use for everyone to use and support <header> and <article>. |
| 18:12 | <TabAtkins> | zewt: Re sorting mail by size, just use the :has(attachment) or whatever criteria. |
| 18:17 | <AryehGregor> | smaug____, FWIW, the only reason I thought of using Google Code in this case is because in practice there's no signup system like OpenID or BrowserID or whatever that actually works in practice. It'd be awesome if BrowserID gets traction. |
| 18:17 | <zewt> | google can't even manage single-sign-in for their own products :| |
| 18:20 | <Philip`> | Bug trackers shouldn't really need to know people's identities anyway |
| 18:22 | <Philip`> | Let anyone file bugs without signing in, and just store some user ID in a permanent cookie so they can come back later and register with an email address and have all their bugs become associated with that, in case they decide they want to follow them by email |
| 18:22 | <zewt> | bugzilla is the worst about that |
| 18:23 | <Philip`> | (and have efficient spam-removal tools to deal with all the junk) |
| 18:23 | <TabAtkins> | And in the meantime you can still see what other bugs are associated with a particular user. |
| 18:23 | <zewt> | i signed in once and it was all "you have to change your password!!! and use a capital and a number this time!!!" it's a BUG TRACKER |
| 18:23 | <Hixie> | what Philip` describes is more or less what the whatwg file-bug.cgi script does |
| 18:23 | <Hixie> | using the ip address as the "unique" identifier |
| 18:23 | <zewt> | i am not going to use a super secure password on a bug tracker that i have no special permissions on. heh |
| 18:24 | Philip` | has several IP addresses per day, so that's not a very effective identifier |
| 18:25 | <Philip`> | (For a while I appended my initials to the spec bug messages, to make them easy to search for afterwards) |
| 18:26 | <Hixie> | hmmm |
| 18:26 | <Hixie> | should <dialog> have a close box by default or should it be up to authors to do that |
| 18:27 | <Hixie> | (s/by default/added by teh UA if the author asks for one/) |
| 18:27 | <timeless> | TabAtkins: that doesn't work |
| 18:27 | <Hixie> | i can't work out how we'd style it |
| 18:27 | <timeless> | we only want :has(attachment > 100mb) |
| 18:27 | <Hixie> | so i guess no close box by default! |
| 18:28 | <hober> | agreed |
| 18:29 | <AryehGregor> | <dialog>? |
| 18:29 | <TabAtkins> | timeless: Unless you get a ton of attachments, you can manually look through them for sizes. |
| 18:29 | <hober> | AryehGregor: see the Dialogs page on the WHATWG wiki |
| 18:29 | <timeless> | TabAtkins: the users i was looking at have attachments on virtually all messages |
| 18:30 | hober | has until some time tomorrow to write this CP |
| 18:30 | <timeless> | (perhaps many have a business card icon) |
| 18:30 | <TabAtkins> | timeless: That would be a problem, then. |
| 18:30 | <zewt> | searching for has:attachment gives me tons of noise |
| 18:30 | <timeless> | so yeah, :has(attachment) is useless |
| 18:30 | <zewt> | "smime.p7s" crap |
| 18:30 | <timeless> | yeah, either greeting cards or signatures |
| 18:30 | <zewt> | "vcf" files |
| 18:30 | <Hixie> | hober: pretty much entirely coincidentally, i plan to be working on this today (the mass move of bugs mikesmith did made this the least-recently changed bug!) |
| 18:30 | <TabAtkins> | timeless: No, it works great for me. It may not work for everyone. |
| 18:30 | <timeless> | s/greeting/business/ |
| 18:30 | <timeless> | TabAtkins: oh, it would work fine for timeless@gmail |
| 18:31 | <Hixie> | hober: so if you trust me, you can write a CP that just says "do what the whatwg spec does" and i'll try to write it up today :-) |
| 18:31 | <timeless> | but that's because 99.999% of mail there is bugmail :) |
| 18:31 | <timeless> | a simpler rule is -bugzilla-daemon |
| 18:31 | <hober> | Hixie: heh |
| 18:31 | <timeless> | that rule gets me anything that might be big :) |
| 18:31 | <Hixie> | hober: you can see what i'm thinking of doing in the Ideas section of that page |
| 18:31 | <hober> | Hixie: I can pop up there this afternoon for some whiteboarding if you think that would be useful |
| 18:31 | <Hixie> | hober: obviously, input is more than welcome |
| 18:31 | <hober> | Hixie: I think I have less ambitious goals than you in this case |
| 18:32 | <hober> | Hixie: i'm not sure if we should even try addressing the 'fat' tooltip case |
| 18:32 | <hober> | Hixie: at least not yet |
| 18:32 | <Hixie> | hober: i'm free 2 to 4 if you want, definitely happy to meet up if you want |
| 18:32 | <Hixie> | yeah i dunno if it should all go in at once |
| 18:32 | <Hixie> | i like to overdesign then cut out, makes me more confident the design can support extension later |
| 18:32 | timeless | grumbles |
| 18:32 | <timeless> | http://www.visionmobile.com/research.php#OGI |
| 18:33 | <timeless> | they demand my email address |
| 18:33 | <Hixie> | s/if you want//g |
| 18:33 | <hober> | Hixie: how about I swing by around 2? |
| 18:34 | <Hixie> | hober: sounds good to me. easiest is to meet in the building 43 lobby. |
| 18:34 | <Hixie> | hober: (i'm meeting with a guy from cisco about webrtc at 1) |
| 18:35 | <hober> | ok, see you there/then |
| 18:35 | <Hixie> | cool |
| 18:37 | Hixie | ends up interacting with several web dialogs just in trying to book the meeting, heh |
| 18:38 | <timeless> | Hixie: this is why you shouldn't allow web dialogs on the web :) |
| 18:39 | <Hixie> | wow, i actually managed to book the meeting room closest to the lobby |
| 18:39 | <Hixie> | that's not normal |
| 18:40 | <Hixie> | probably means the building is being refurbished or something and they just forgot to tell me |
| 18:40 | <timeless> | heh |
| 18:40 | <timeless> | did you book for friday? |
| 18:40 | <timeless> | maybe everyone is going to a party |
| 18:41 | <Hixie> | pretty sure i booked for today, yeah |
| 18:42 | <Hixie> | the party isn't til 4, why do you think that's when i have to stop the meeting with hober? :-P |
| 18:42 | <timeless> | heh |
| 18:44 | <Hixie> | hmm |
| 18:44 | <Hixie> | some of these dialogs... i guess all of them are the fat tooltips, as hober calls them |
| 18:44 | <Hixie> | some of these tooltips open anchored to the cursor position. |
| 18:44 | <timeless> | ? |
| 18:45 | <timeless> | grr |
| 18:45 | timeless | sighs |
| 18:45 | <Hixie> | maybe there should be a variant of show() that takes a MouseEvent object |
| 18:45 | <timeless> | every time i update my phone's os, Gmail loses its account data |
| 18:45 | <Hixie> | hmmmmm |
| 18:45 | <timeless> | and i get to reregister it |
| 18:46 | <Hixie> | i wonder what i meant by "have to handle showing an element that's descendant of display:none content; does that just not count?" |
| 18:47 | <Hixie> | should we use callbacks for dialog.showModal(), or require you to set an event handler? |
| 18:47 | <Hixie> | hmm |
| 18:48 | <hober> | sorry, meeting |
| 19:15 | <Hixie> | hmm, interesting |
| 19:15 | <Hixie> | there are some modal dialogs that aren't centered but are instead anchored |
| 19:15 | <Hixie> | hmm |
| 19:15 | <TabAtkins> | Hixie: I'm rewriting the element() function's text, and I'm not sure how the display of out-of-document elements should be talked about. |
| 19:15 | <TabAtkins> | I think it's actually HTML's responsibility to define that images, videos, and canvas can be displayed while out-of-document. |
| 19:17 | <hober> | Hixie: yes, you see login dialogs that are anchored to the 'log in' link/button |
| 19:18 | <TabAtkins> | hober, Hixie: That will be eventually covered by CSS Positioning. |
| 19:18 | <hober> | TabAtkins: indeed, and I will happily defer to that module when it's ready :) |
| 19:18 | <TabAtkins> | Yeah, just do something magical for now. I'll review for reasonableness when necessary. |
| 19:19 | <Ms2ger> | hober, I was wondering whether you edited any specs already :) |
| 19:20 | <hober> | Ms2ger: I'm technically listed as an editor of the 2d transforms spec, only because i go in and fix spec bugs when working on the transforms test suite |
| 19:20 | <Ms2ger> | To me, it sounds like you want to do more :) |
| 19:20 | <Hixie> | TabAtkins: cool, thanks (Re positioning) |
| 19:20 | <hober> | Ms2ger: heh |
| 19:21 | <Hixie> | TabAtkins: about element(), why is <img> special but not, say, <iframe>? |
| 19:21 | <Hixie> | TabAtkins: or <svg>? |
| 19:21 | <Hixie> | TabAtkins: i don't really understand html's role here. html defines what those elements represent, but beyond that... |
| 19:21 | <TabAtkins> | SVG defines its own stuff. |
| 19:21 | <Hixie> | TabAtkins: i can have a hook similar to "represent" for this case if you like |
| 19:22 | <Hixie> | TabAtkins: and you can use that |
| 19:22 | <TabAtkins> | But the host language is the only one that knows whether an element has a "natural" notion of what it measn to be rendered, outside the confines of a document. |
| 19:22 | <Hixie> | TabAtkins: "any element that foobars a bitmap..." |
| 19:22 | <Hixie> | TabAtkins: and the HTML spec can say "An img element foobars its image" or whatever |
| 19:23 | <TabAtkins> | Hmm, yeah, that would work. |
| 19:23 | <TabAtkins> | I just need a notion of width, height, and appearance. |
| 19:25 | <TabAtkins> | Hixie: Or if it just represents a raster image directly, I can pull what else I need from that. |
| 19:26 | <Hixie> | appearance? |
| 19:26 | <TabAtkins> | Like, an <img> looks like it's image. |
| 19:28 | <gsnedders> | People can live with thereshouldbenored.com being down for a bit? Good. :) |
| 19:28 | <timeless> | grrr |
| 19:28 | <Hixie> | as opposed to it being what? |
| 19:28 | <timeless> | i think google accounts is crashing my browser |
| 19:28 | <Hixie> | oh you mean "its image" |
| 19:29 | <Hixie> | wow it's not usually confusing to get that one wrong :-P |
| 19:29 | <Hixie> | i think all the ones we care about here are bitmaps right? |
| 19:29 | <Hixie> | canvas, video, and img |
| 19:29 | <TabAtkins> | Shit, sorry. Damn you, apostrophe-omiting rules! |
| 19:29 | <TabAtkins> | Yeah, think so. |
| 19:29 | <TabAtkins> | "expose a bitmap" might work. |
| 19:30 | <Hixie> | file a bug about me adding that term |
| 19:30 | <TabAtkins> | kk |
| 19:30 | <Hixie> | link to this log :-) |
| 19:30 | <Hixie> | in other news i added a bunch of screenshots from google products to http://wiki.whatwg.org/wiki/Dialogs |
| 19:31 | <TabAtkins> | What component do I put this in? |
| 19:31 | <timeless> | grr |
| 19:31 | timeless | kicks something |
| 19:32 | <timeless> | Hixie: you misspelled Gmail |
| 19:32 | <timeless> | s/GMail/Gmail/ |
| 19:32 | <timeless> | > GMail - Click "more" on the left, then Create New Label. |
| 19:32 | <TabAtkins> | s/.*/gmail/ |
| 19:33 | <Hixie> | TabAtkins: other hixie drafts |
| 19:39 | Ms2ger | omites an apostrophe at TabAtkins |
| 19:39 | <Ms2ger> | Rhymes with "smites" |
| 19:45 | <Hixie> | i think http://wiki.whatwg.org/wiki/Dialogs#Proposal covers the bulk of the use cases |
| 19:49 | <rillian> | where's the repository for the webvtt spec? |
| 19:50 | <Ms2ger> | rillian, svn.whatwg.org/webapps |
| 19:50 | <Ms2ger> | It's somewhere in source |
| 19:50 | <TabAtkins> | Ms2ger: The double-consonant rules are inconsistent in English. |
| 19:50 | <TabAtkins> | Otherwise we'd say "editting". I sometimes autocorrect away from doubles, precisely because of the word "editting". |
| 19:51 | <Ms2ger> | Everything is inconssistent in English |
| 19:52 | <AryehGregor> | TabAtkins, a reasonably reliable rule is that you double the last consonant if the word is accented on the last syllable, and don't double it otherwise. |
| 19:52 | <AryehGregor> | For when you're adding verb suffixes like -ed or -ing. |
| 19:52 | <TabAtkins> | Hmm, never thought of that rule. |
| 19:52 | <AryehGregor> | Not entirely reliable, but works okay. |
| 19:53 | <AryehGregor> | I think I know it because in a MUD in like 1999, I reported a typo in one of the room descriptions where it said "worshiped", which I thought should be "worshipped", and one of the GMs educated me. |
| 19:53 | <AryehGregor> | (I then promptly pointed out another room description where it said "worshipped" instead of "worshiped", which he corrected.) |
| 19:54 | <AryehGregor> | (MUDs are cool. The GMs can just do any random thing by typing commands, since they don't need to provide graphics.) |
| 19:54 | <TabAtkins> | Yeah, used to play one. |
| 19:55 | <rillian> | Ms2ger, around line 34836. thanks! |
| 19:56 | <Ms2ger> | Np |
| 19:59 | <Ms2ger> | http://xkcd.com/934/ |
| 19:59 | <Ms2ger> | Go us |
| 20:00 | <AryehGregor> | \o/ |
| 20:00 | <TabAtkins> | Wait, "someday" have xmonad? We already have linux in the browser in linux. |
| 20:00 | <rillian> | *is* there an xmonad firefox extension? |
| 20:00 | <gsnedders> | TabAtkins: Only on little endian hardware, because typed arrays leak harware endianness. |
| 20:01 | <kbrosnan> | rillian: https://addons.mozilla.org/en-US/firefox/addon/pentadactyl/ if you are cool with vim |
| 20:04 | <rillian> | that one's cute |
| 20:07 | <Hixie> | if anyone has a few minutes to review http://wiki.whatwg.org/wiki/Dialogs#Proposal before hober and i discuss it this afternoon that'd be great |
| 20:07 | <Hixie> | bbiab |
| 20:08 | <smaug____> | I hope that isn't going to the spec before it gets some reviewing |
| 20:08 | <AryehGregor> | It can be reviewed once it's in the spec, no? The WHATWG spec has stability annotations for this reason. |
| 20:11 | <smaug____> | what are the stability levels |
| 20:12 | <Ms2ger> | The boxes in the left margin that all say "Last Call" |
| 20:12 | <smaug____> | I mean, what all can read there |
| 20:12 | <smaug____> | "Ready for first implementations" in one |
| 20:13 | <AryehGregor> | smaug____, e.g.: http://www.whatwg.org/specs/web-apps/current-work/multipage/video-conferencing-and-peer-to-peer-communication.html#the-data-stream |
| 20:13 | <AryehGregor> | "Experimental draft" |
| 20:13 | <smaug____> | are those "levels" explained somewhere? |
| 20:13 | <AryehGregor> | No, they're pretty informal. |
| 20:13 | <AryehGregor> | I think anyone's allowed to change them. |
| 20:14 | <Ms2ger> | Yeah |
| 20:14 | <smaug____> | there seems to be for example "Ready for first implementations" stuff which is quite stable, and things which will quite likely to change |
| 20:14 | <AryehGregor> | They're meant to be just warnings like "don't assume this is stable and you should start implementing it right away". |
| 20:16 | <smaug____> | I mean some "Ready for first implementations" is stable, some "Ready for first implementations" is unstable |
| 20:16 | timeless | misses muds |
| 20:16 | timeless | played one in college |
| 20:16 | <smaug____> | so it is hard to interpret stability annotations |
| 20:17 | <AryehGregor> | You can ask if you're unsure. |
| 20:17 | <Philip`> | AryehGregor: It is worshipped, at least in proper English |
| 20:17 | <Philip`> | I think http://www.whatwg.org/specs/web-apps/current-work/status-documentation.html defines the stability markers |
| 20:19 | <smaug____> | AryehGregor: ask who? If no one has implemented the feature, it is hard to say much about its stability ;) |
| 20:19 | <AryehGregor> | smaug____, ask the editor. |
| 20:19 | <Ms2ger> | If no one has implemented it, it's unstable :) |
| 20:19 | <AryehGregor> | He'll know if he's gotten feedback about it, etc. |
| 20:20 | <AryehGregor> | Also, what Ms2ger said. |
| 20:20 | <Ms2ger> | At least, I think that's what smaug____ would say ;) |
| 20:21 | timeless | sighs |
| 20:21 | <timeless> | hsivonen: it's only 10:30pm where you are... maybe you're awake? :) |
| 20:21 | timeless | needs someone to beat the obvious drum |
| 20:21 | <smaug____> | yeah, I'd say all the stuff which hasn't been implemented should be marked "unstable - likely to change" ;) |
| 20:22 | <smaug____> | timeless: I think hsivonen is about to start his vacation today or tomorrow |
| 20:22 | <timeless> | aww :( |
| 20:22 | <timeless> | i hope he enjoys it |
| 20:22 | <timeless> | but the world will suffer while he does |
| 20:22 | <timeless> | :) |
| 20:23 | <Ms2ger> | It won't be reminded of the evilness of doctypes |
| 20:23 | <AryehGregor> | smaug____, the annotations also say which browsers have implemented the features. |
| 20:24 | <timeless> | i have a bunch of people arguing for APIs to enable individual web pages to do things which should be a device/platform detail |
| 20:24 | <timeless> | and i'm having trouble making the argument because i'm just one person and people have started ignoring me :( |
| 20:24 | <smaug____> | AryehGregor: yeah, those annotations are very useful |
| 20:24 | <Ms2ger> | If someone made them say that :) |
| 20:25 | <smaug____> | AryehGregor: but with my bad English, I can't say if "Work in progress" is more stable than "Experimental draft" for example |
| 20:26 | <AryehGregor> | smaug____, . . . to be honest, I'm not sure offhand either. I think "experimental" is less stable. |
| 20:26 | <smaug____> | :) |
| 20:26 | <Ms2ger> | The entire spec is a WIP, duh |
| 20:26 | <timeless> | i think i'd agree w/ AryehGregor fwiw |
| 20:27 | <timeless> | experimental should probably mean "someone wrote something and has no idea if it works |
| 20:27 | <timeless> | and they're trying to get people to experiment with it to see if it does" |
| 20:27 | <timeless> | whereas work in progress hopefully means "someone is writing something and iterating with an implementation" |
| 20:27 | <The_8472> | today i noticed that multiplebgs + gradients is way more flexible than borders (including border-images). |
| 20:28 | <smaug____> | timeless: in this case "Work in progress" certainly doesn't mean that something is implementing the feature |
| 20:29 | <timeless> | smaug____: which case is this? |
| 20:29 | <smaug____> | but in fact that there are new proposals coming which will replace that feature |
| 20:29 | <timeless> | the one where the editor is also making a js based impl |
| 20:29 | <smaug____> | Undo handling |
| 20:29 | <timeless> | is still an impl |
| 20:29 | <timeless> | i'm not sure i'd call undo handling a WIP |
| 20:29 | <timeless> | afaict those are just a bunch of proposals |
| 20:30 | <moo-_-> | is there a way to save references to local files in local storage and retrieve them later on? |
| 20:30 | <timeless> | (am i wrong?) |
| 20:30 | <moo-_-> | not the file data itself, I am talking about photos here |
| 20:31 | <timeless> | moo-_-: you want to save a file path? |
| 20:31 | <timeless> | so that you can later retrieve the then current underlying data? |
| 20:31 | <moo-_-> | timeless: yes, for opening the file again if the page is refreshed |
| 20:31 | <timeless> | if the file moves, i presume you're happy to get an error? |
| 20:31 | <moo-_-> | timeless: yes |
| 20:31 | <moo-_-> | np |
| 20:31 | <timeless> | (i'm pretty sure the answer is no, fwiw) |
| 20:31 | <moo-_-> | timeless: just not lose the user initiated file interaction |
| 20:31 | <smaug____> | I sure hope you can get access only to the filename, not to the full path |
| 20:32 | <timeless> | so, offhand, i claim it isn't a useful feature |
| 20:32 | <timeless> | just because i grant you access to a file on Jan 1 |
| 20:32 | <timeless> | doesn't mean I'll remember i granted that access on Dec 12 |
| 20:32 | <timeless> | if you happen to remember the filename and an abstract |
| 20:32 | <timeless> | then when you suggest I might want to provide that file again on Dec 12 |
| 20:32 | <timeless> | i'll either be happy to do so, or decide i've changed my mind |
| 20:32 | <timeless> | and that option will make me happy |
| 20:35 | <timeless> | moo-_-: in short, if browsers can't make it easy for user's to safely and conveniently initate file interactions, |
| 20:35 | <timeless> | ... that's a bug in browsers |
| 20:35 | <moo-_-> | timeless: I am more worried about that references survive browser close/refresh |
| 20:35 | <moo-_-> | window.URL.createObjectURL() |
| 20:35 | <timeless> | and will need to be fixed |
| 20:35 | <moo-_-> | I need to know if I can reuse result of this call across page loads |
| 20:36 | <timeless> | > Browsers will release these automatically when the document is unloaded; |
| 20:36 | <timeless> | from MDN [https://developer.mozilla.org/en/DOM/window.URL.createObjectURL] |
| 20:36 | <timeless> | in short, i think it's pretty clear that you can't reuse the result across page loads |
| 20:37 | <moo-_-> | hmmm |
| 20:37 | <timeless> | does my logic seem flawed? |
| 20:37 | <timeless> | (i need to pack up and go home, so please lemme know soon) |
| 20:37 | <timeless> | -- otherwise, have a good weekend |
| 20:39 | <moo-_-> | http://dev.w3.org/2006/webapi/FileAPI/#lifeTime |
| 20:39 | <moo-_-> | yep |
| 20:39 | <moo-_-> | :< |
| 20:39 | <moo-_-> | need to then upload images to the server first |
| 20:39 | <moo-_-> | suuuuuuucks |
| 20:45 | <moo-_-> | I was secretly hoping |
| 20:45 | <moo-_-> | that if the user is uploading multiple photos |
| 20:45 | <moo-_-> | he/she can close browser and continue later |
| 20:45 | <moo-_-> | I wonder if Flash has something to work around the problem |
| 20:46 | <zewt> | need to improve file api + structured clone + history api to make that possible |
| 20:46 | <zewt> | it's already possible by spec i think, but nobody supports all that together... |
| 20:46 | <timeless> | moo-_-: do you understand my argument against that? |
| 20:46 | <timeless> | yes, it means that it'd be hard to write a BitTorrent client |
| 20:47 | <jgraham> | We should replace the status markers with a simple 1-5 numeric system where 1 means "I pity the fool who implements this" and 5 means "this is stable enough to put on a space mission" |
| 20:47 | <zewt> | apis should definitely allow continuing to use files later |
| 20:47 | <timeless> | jgraham: how often would we get to 3? |
| 20:47 | <moo-_-> | timeless: would there be problem to have origin or page URL and having object URLs valid within this origin? |
| 20:47 | <jgraham> | Well none of the spec would be a 5 yet, obviously |
| 20:47 | <timeless> | moo-_-: ? |
| 20:47 | <timeless> | moo-_-: your original question was about local resources, right? |
| 20:48 | <zewt> | it's tricky to implement correctly; I think the rule Chrome *tries* to follow is that you can continue to use the file so long as the underlying file hasn't been changed (but I don't see how it can efficiently enforce that), and I'm sure it still doesn't support structured cloning to persistent APIs |
| 20:48 | <timeless> | (file:) |
| 20:48 | <moo-_-> | timeless: so that object URLs generated from files would be bound to page itself, not document session |
| 20:48 | <jgraham> | I presume space missions have now come around to the idea of silicon-based-RAM and such things, but the internet is still about a decade too new |
| 20:48 | <moo-_-> | so if you visit the page again you can access the files previously opened |
| 20:48 | <zewt> | for example, it's very annoying that you can't stash open files in History, so that if the session is restored a web app can reopen open files and continue the state it was in |
| 20:48 | <timeless> | zewt: i claim that even a file which hasn't changed might turn out to be a problem if you wait long enough |
| 20:49 | <timeless> | it could turn out that it has stuff which was reclassified as confidential |
| 20:49 | <timeless> | i'd rather the page upon a resume cause the browser to show the user the list of previously authorized objects |
| 20:49 | <timeless> | and allow the user to control whether to allow access |
| 20:49 | <zewt> | timeless: while I understand the direction you're looking in, I'm not (from the cases I've heard so far) inclined to say that it's enough of a worry to outweigh the uses |
| 20:50 | <zewt> | clients are always free to say "restoring this page will continue granting access to these resources, is that OK?", though I'd be surprised if any actually did |
| 20:50 | <moo-_-> | zewt: the spec says they MUST not http://dev.w3.org/2006/webapi/FileAPI/#lifeTime :) |
| 20:51 | <zewt> | moo: not talking about URIs, talking about structured clones of blobs and (more specifically) File |
| 20:51 | <moo-_-> | zewt: aha |
| 20:51 | <moo-_-> | File object itself? |
| 20:51 | <zewt> | you can store structured clones of objects with apis like indexeddb and History |
| 20:51 | <zewt> | and structured clone supports Blob and File |
| 20:52 | <timeless> | zewt: so... |
| 20:52 | <timeless> | my general feeling is that while I don't support the EUs stupid privacy review stuff (see recent spam) |
| 20:52 | <timeless> | is that specs need to be written to favor user's privacy and protection |
| 20:52 | <zewt> | (not sure what you're referring to in particular) |
| 20:54 | <timeless> | http://www.enisa.europa.eu/act/application-security/web-security/a-security-analysis-of-next-generation-web-standards/ |
| 20:54 | <timeless> | for lack of a better url |
| 20:54 | <timeless> | they triggered a bunch of useless spam to various WGs |
| 20:54 | <zewt> | i don't disagree that privacy is important, but I don't have a feeling that this will cause significant real-world issues; and environments which have stronger, stricter requirements (NDAs, corporate policy, government), browsers are free and should be encouraged to expose stricter modes |
| 20:54 | <timeless> | often about specs which were nearly dead |
| 20:54 | <zewt> | such as like I said, warning before re-granting permissions after a session restore |
| 20:56 | <timeless> | zewt: note that i'm not advocating 1990s java style security dialogs |
| 20:56 | <zewt> | right |
| 20:56 | <zewt> | i'm also not meaning to brush off what you're saying--squinting at how things can go wrong is security 101 |
| 20:57 | <timeless> | i think chrome's download approach is possible the right way |
| 20:57 | <timeless> | basically for any object that's exposed, stick it in the tray with a red color instead of green and an arrow going the other way |
| 20:57 | <timeless> | if the use doesn't care about the notification, they can click a hide button at the edge of the tray |
| 20:58 | <timeless> | but the problem is that you almost have to let the user interact with the tray before the page is allowed to touch the objects |
| 20:58 | <timeless> | otherwise a page is likely to steal all resources before the user makes any decisions |
| 20:58 | <timeless> | (which doesn't help the user) |
| 20:58 | <zewt> | well, it would discourage leaving permissions active in the first place |
| 20:58 | <zewt> | you'd be unlikely to leave a site having access to a file for a year and forgetting about it |
| 20:58 | <timeless> | you can do it by having the tray focus when the first resource is accessed by the page |
| 20:59 | <timeless> | zewt: depends on whether i visit the site often |
| 20:59 | <timeless> | fwiw, i visited gmail recently and got a bar at the top... |
| 21:00 | <zewt> | bear in mind, though, that the site can always stash the file somewhere when you first grant access |
| 21:00 | <zewt> | if i open a file today, close the site and come back in a month, the site may already (the first time) read it from disk and saved the raw data somewhere else |
| 21:00 | <timeless> | http://www.google.com/support/forum/p/gmail/thread?tid=141cc0ce0197cfba&hl=en |
| 21:01 | <timeless> | zewt: sure |
| 21:01 | <timeless> | but various things to consider |
| 21:01 | <timeless> | - the site http://site.example.com/ might have been run by a trusted vendor on Jan 1 |
| 21:01 | <timeless> | the domain might have expired and might not be run by an evil guy on Dec 12 |
| 21:01 | <zewt> | well, that's a more broad question |
| 21:02 | <timeless> | as an example: http://398west.ca/ |
| 21:02 | <zewt> | it applies to every sensitive API and so far nobody has come up with an answer, really |
| 21:02 | <timeless> | Google still thinks (in Google Maps), that 398 West exists |
| 21:02 | <zewt> | (the only possibly-workable approach so far is packaged web apps, and that sucks, badly) |
| 21:02 | <timeless> | actually, those are more vulnerable, not less :) |
| 21:03 | <timeless> | If the site doesn't properly use HTTPS certs, it's easy for it to communicate to the replacement server |
| 21:03 | <zewt> | those you can at least grant permissions to as a unit and know when they change ... but they go fundamentally against the web |
| 21:04 | <timeless> | anyway... |
| 21:04 | <timeless> | i'm not opposed to a way for a browser to provide a token which is `potentially valid` |
| 21:04 | <timeless> | which a site could try to use in the future |
| 21:05 | <timeless> | which the user agent would then show the user a request for a resource with the option to select the previously selected resource and a note indicating when it was selected |
| 21:05 | <timeless> | in the case where the user wants to continue to share the resource, the cost isn't particularly high |
| 21:05 | <timeless> | in the case where the user wants to provide a different resource (e.g. a newer draft of a document) |
| 21:05 | <timeless> | the cost isn't particularly high |
| 21:05 | <zewt> | that applies to every API that's sensitive enough to require permission, even eg. geolocation |
| 21:05 | <timeless> | indeed :) |
| 21:06 | <zewt> | (geolocation being something that most people don't consider particularly sensitive, but some people very strongly do) |
| 21:06 | <timeless> | and if the user doesn't want to share anything anymore, the user can do something useful |
| 21:06 | <timeless> | yeah |
| 21:06 | <timeless> | enisa thing about geolocation |
| 21:06 | <timeless> | it was hilarious |
| 21:06 | <zewt> | anyway, my point there is that yeah, having the site change out from under you is an attack to worry about, but it's a broader problem that should probably be solved in the broader scope, not specifically for Files (or any other single API) |
| 21:06 | <timeless> | they provided a script to do a binary search to find out the user's last geolocation timestamp |
| 21:07 | <timeless> | the response was "or you could just get the property from the object directly" |
| 21:07 | timeless | should find that one |
| 21:07 | <zewt> | that's not to say I have any particularly good ideas for doing so (that scale to lots of APIs and lots of sites and different types of users) |
| 21:07 | <timeless> | http://lists.w3.org/Archives/Public/public-geolocation/2011Aug/0003.html |
| 21:07 | <timeless> | is the thread |
| 21:12 | <zewt> | heh gmail is still doing the "message may not have been sent by" nonsense |
| 21:12 | <zewt> | very poor that they're leaving such a horrible user-mistraining bug unfixed for so long... |
| 21:12 | <timeless> | ? |
| 21:12 | <timeless> | what's wrong w/ that? |
| 21:12 | <zewt> | This message may not have been sent by: timeless⊙gc Learn more Report phishing |
| 21:12 | <zewt> | showing false positive warnings trains users to ignore warnings |
| 21:12 | <timeless> | zewt: well, yeah, err |
| 21:12 | <timeless> | it isn't entirely wrong |
| 21:12 | <zewt> | I certainly never pay attention to those warnings since every one I've seen so far has been nonsense |
| 21:12 | <Philip`> | It used to say it for most mail sent by @google.com people to the WHATWG list |
| 21:13 | <Philip`> | Now it says it for many @gmail.com people |
| 21:13 | <timeless> | in the case of timeless⊙gc, my mail isn't sent by timeless⊙gc |
| 21:13 | <timeless> | it's sent by timeless....⊙gc |
| 21:13 | <timeless> | which list are you reading? |
| 21:13 | <zewt> | that one's on whatwg |
| 21:14 | <timeless> | most likely the list you're reading a list which is not properly dealing w/ magical header things |
| 21:14 | <timeless> | i have a domain @last.com which sends to flast⊙gc |
| 21:14 | <timeless> | it doesn't do the right thing, so any mail to first⊙lc which is forwarded to flast⊙gc properly gets a warning from gmail |
| 21:14 | <timeless> | because the forwarder is broken |
| 21:16 | <timeless> | for namecheap, it looks like the solution is to sign up for gapps, and they even wrote a howto: http://community.namecheap.com/forums/viewtopic.php?f=6&t=577 |
| 21:16 | timeless | should consider it |
| 21:16 | <zewt> | i've used gapps for a long time |
| 21:17 | <zewt> | works well for the google products that just don't work with it (like g+) |
| 21:17 | <timeless> | right, so SPF records / DKIM signatures. |
| 21:17 | <timeless> | yeah yeah |
| 21:17 | <timeless> | i'd actually like a list from you of what works some day |
| 21:17 | <timeless> | but i need to shave and go home |
| 21:17 | <timeless> | (i don't have any mirrors at home yet, so ...) |
| 21:17 | <zewt> | g+ is the only thing i've hit that doesn't these days |
| 21:17 | <zewt> | they merged auth for google and gapps which fixed a lot of stuff |
| 21:19 | <zewt> | used to be some bizarre design bug that caused you two end up with two accounts with the same address, that took them forever to clear up, heh |
| 21:20 | <timeless> | does the whatwg mailing list add a footer to the message? |
| 21:20 | <timeless> | if it does, that's probably the problem |
| 21:20 | <zewt> | though i forward other addresses to it and they work fine for me |
| 21:20 | <zewt> | (though I don't post to lists often with my other addresses so) |
| 21:21 | <zewt> | nope |
| 21:21 | timeless | ponders |
| 21:21 | <timeless> | could you possibly pastebin the source of one of my bad messages? |
| 21:21 | <zewt> | one of the dangers of google; when something goes wrong, you sort of have to figure it out yourself :) |
| 21:21 | timeless | is curious |
| 21:21 | <timeless> | oh, i have a general idea of what's going wrong |
| 21:22 | <timeless> | it's either SPF or DKIM sig being voided |
| 21:22 | <timeless> | my bet is the latter, but .. |
| 21:22 | <zewt> | http://pastebin.com/6N3y9wA3 |
| 21:23 | <zewt> | http://mail.google.com/support/bin/answer.py?hl=en&ctx=mail&answer=185812 sounds like it |
| 21:24 | <timeless> | h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:content-type :content-transfer-encoding; |
| 21:24 | <timeless> | is what was hashed |
| 21:24 | <timeless> | Sender: whatwg-bounces⊙lwo |
| 21:24 | <timeless> | is what was voided |
| 21:24 | <timeless> | so yeah. |
| 21:24 | <timeless> | sadly i'm not sure what the correct fix is |
| 21:24 | <timeless> | the namecheap best solution is to change the mx |
| 21:24 | <timeless> | that doesn't work for whatwg of course |
| 21:25 | <zewt> | first impression is that Sender shouldn't be authenticated at all, but it's been a long time since I've dug very deeply into smtp |
| 21:25 | <timeless> | zewt: well, the signature clearly lists it as one of the fields they're signing |
| 21:25 | <timeless> | so you can't do that |
| 21:25 | <zewt> | right, I'm saying the signature shouldn't include that, since it's (probably) normal for mailing lists to change it |
| 21:25 | <timeless> | a dkim aware mailinglist will need to avoid munging any dkim signed fields |
| 21:26 | <timeless> | from google's perspective, the mailinglist is at fault |
| 21:26 | <timeless> | from my perspective, i'd like to know what a mailing list can do instead |
| 21:26 | <timeless> | sadly finding *that* info is painful |
| 21:26 | <timeless> | one approach is to take my message and stick it as a MIME attachment |
| 21:26 | <timeless> | (as the only part) |
| 21:26 | <zewt> | if mailing lists typically change Sender, then google's signatures are clearly at fault (i don't know, off-hand, whether changing that header is actually typical or if whatwg is doing something unusual) |
| 21:26 | <timeless> | and stick a <mailinglist> envelope around it |
| 21:28 | <zewt> | if every mailman installation does that, then google is clearly at fault in practice |
| 21:28 | <timeless> | i think mailman basically needs to change |
| 21:28 | <zewt> | (i'm assuming whatwg is just a standard-issue mailman install like the other billion of them) |
| 21:28 | <timeless> | it needs to strip out the DKIM signature if it's valid |
| 21:28 | <timeless> | and replace it with its own DKIM signature |
| 21:28 | <timeless> | which is correct |
| 21:28 | <timeless> | it's claiming sender=whatwg |
| 21:29 | <timeless> | so it should just resign it with sender=whatwg |
| 21:29 | <timeless> | http://wiki.list.org/display/DEV/DKIM |
| 21:30 | <zewt> | timeless: but practically speaking, if every mailman list in the world has this problem, then google is in error showing a warning like this before the problem is solved (or at least in the process of being solved, eg. a mailman update available) |
| 21:30 | <timeless> | well |
| 21:30 | <zewt> | (i'm assuming there isn't, since it looks like whatwg is already on the latest version) |
| 21:30 | <timeless> | could i interest you in working on it w/ me? |
| 21:30 | <timeless> | i'm seriously contemplating trying to fix mailman |
| 21:31 | <timeless> | i'm fairly certain that no one has fixed it |
| 21:31 | <timeless> | by `fixing`, i mean `writing a patch which implements one strategy` |
| 21:31 | <timeless> | my proposed strategy is this: |
| 21:31 | <timeless> | take the existing signature, if it's valid, store it with a new prefix, munge the message, and resign it for the mailer |
| 21:32 | <timeless> | if the signature isn't valid, add a note that it wasn't valid and <dunno, possibly just drop the message on the floor, or optionally send it to moderation, or munge the message and add a broken signature and send it off> |
| 21:33 | <timeless> | i'd do the code under the mailman license (whatever that is), and offer it to mailman and any lists i use |
| 21:34 | <timeless> | (yes, this has bothered me for a while, but the goal of google here is to get people to implement a fix) |
| 21:34 | <timeless> | i think ideally google would have a way to let individuals whitelist individual mailing lists in the interim |
| 21:34 | <timeless> | until people like you/me fix the mailing list software to stop breaking signatures |
| 21:34 | <zewt> | i'm not familiar with the nuts and bolts of how DKIM works |
| 21:34 | <timeless> | zewt: don't worry about it, we'd figure it out |
| 21:35 | <timeless> | the documentation seems pretty clear |
| 21:35 | <timeless> | and it seems fairly straightforward |
| 21:36 | <timeless> | fwiw, the last link i offered indicates that mailman 2.1.9 strips DKIM |
| 21:36 | <zewt> | but then that it was reverted by default |
| 21:37 | <zewt> | it gives me very low confidence in DKIM that it's being deployed without how it affects mailing lists being very clearly defined and understood |
| 21:37 | <timeless> | i think your average user probably doesn't use mailing lists :) |
| 21:37 | <zewt> | (which it clearly isn't, given this big list of possible options) |
| 21:37 | <timeless> | well, it means that in theory at the risk of using whatever security bugs are in 2.1.9, someone could use it, or flip the toggle |
| 21:39 | <timeless> | ok, so mailman is python :) |
| 21:40 | <timeless> | http://hewgill.com/pydkim/ |
| 21:40 | <timeless> | seems like there's a happy library for it |
| 21:40 | <timeless> | offhand, it seems like this shouldn't be too hard |
| 21:41 | <timeless> | ignoring politics |
| 21:41 | <zewt> | here's a mail from you on webapps, which doesn't trigger it: http://pastebin.com/ZNMwPvLK i forget which list software they use |
| 21:44 | <timeless> | the problem is the SPF record |
| 21:44 | <timeless> | w3 has one |
| 21:44 | <timeless> | whatwg does |
| 21:44 | <timeless> | n't |
| 21:44 | <timeless> | Received-SPF: neutral (google.com: 208.97.161.172 is neither permitted nor denied by best guess record for domain of whatwg-bounces⊙lwo) client-ip=208.97.161.172; |
| 21:44 | <timeless> | Received-SPF: pass (google.com: domain of public-webapps-request⊙lwo designates 128.30.52.56 as permitted sender) client-ip=128.30.52.56; |
| 21:44 | <timeless> | so whatwg could solve this by adding an SPF |
| 21:44 | <timeless> | that seems like a simple thing to fix |
| 21:44 | <timeless> | contact hixie or someone and ask them to add the necessary spice |
| 21:44 | <timeless> | http://old.openspf.org/dns.html |
| 21:44 | timeless | is amused to see `rim.net` in the examples |
| 21:46 | <timeless> | anyway, it'd be great if you poked Hixie about it |
| 21:46 | <timeless> | if not, i'll try to remember to do so |
| 21:47 | <zewt> | saying Hixie is probably enough to poke him about it :) |
| 21:48 | <timeless> | w3.org text = "v=spf1 a mx ptr mx:inrialpes.fr ip4:128.30.52.0/22 ip4:133.27.175.5 ip4:193.51.208.64/28 ip 4:212.89.8.176/29 ip4:212.89.8.80/28 mx:fundacionctic.org a:pec.etri.re.kr include:spf.keio.w3.org ~ all" |
| 21:48 | <timeless> | is the relevant part of the w3.org dns record :) |
| 21:48 | <timeless> | fwiw, dreamhost.com has a spf record |
| 21:49 | <timeless> | so if Hixie 's box is in dreamhost (Which |
| 21:49 | <timeless> | it probably is), and he's willing to let *anyone* in dreamhost send mail as whatwg |
| 21:49 | <zewt> | pretty sure the list is |
| 21:49 | <timeless> | he could use a lazy include instead of writing the right thing |
| 21:49 | <timeless> | (not a great idea, but...) |
| 21:50 | <timeless> | http://wiki.dreamhost.com/SPF |
| 21:50 | <timeless> | dreamhost even has a wiki page for it :) |
| 21:51 | <timeless> | right, in theory `v=spf1 include:dreamhost.com a mx ptr -all` would do the right thing |
| 21:51 | <timeless> | if you have a dreamhost account, you could test that pretty easily :) |
| 21:51 | timeless | really goes home (kinda) |
| 21:52 | <gsnedders> | (as in, the next room?) |
| 22:43 | <rniwa> | AryehGregor: yt? |
| 22:43 | <AryehGregor> | rniwa, yep. |
| 22:43 | <rniwa> | AryehGregor: what do you think of edit-action event? |
| 22:43 | <AryehGregor> | rniwa, not sure. Does such a thing exist yet, or is this just a proposal? What are the use-cases? |
| 22:43 | <rniwa> | AryehGregor: no |
| 22:44 | <rniwa> | AryehGregor: so, some editors implement their own editing commands |
| 22:44 | <AryehGregor> | Right. |
| 22:44 | <rniwa> | AryehGregor: or replace the UA's by their own |
| 22:44 | <rniwa> | AryehGregor: but there's currently no way of doing this |
| 22:44 | <rniwa> | AryehGregor: other than manually intercepting keydown |
| 22:44 | <rniwa> | AryehGregor: and other events |
| 22:45 | <AryehGregor> | So just an event that fires whenever execCommand() is run? |
| 22:45 | <AryehGregor> | Or, no. |
| 22:45 | <AryehGregor> | Are you talking about things like "user types text" or "user inserts a line break"? |
| 22:45 | <rniwa> | AryehGregor: yes |
| 22:45 | <AryehGregor> | Ah, okay. |
| 22:45 | <rniwa> | AryehGregor: we should fire an event for both |
| 22:45 | <rniwa> | AryehGregor: that's why I said user action |
| 22:45 | <AryehGregor> | Right, I see. |
| 22:45 | <rniwa> | AryehGregor: could be beforeEditingCommand |
| 22:45 | <rniwa> | AryehGregor: or beforeEditingAction |
| 22:45 | <rniwa> | AryehGregor: I mean the name could be anything |
| 22:46 | <rniwa> | AryehGregor: but the point is that we'll give some standard name for user editing actions |
| 22:46 | <AryehGregor> | And the default action would be to run insertText or insertParagraph or whatever. |
| 22:46 | <rniwa> | AryehGregor: e.g. inserting new line should be InsertLineBreak / InsertParagraph, etc... |
| 22:46 | <rniwa> | AryehGregor: right |
| 22:46 | <AryehGregor> | Why is this better than intercepting keydown etc.? |
| 22:46 | <rniwa> | AryehGregor: I mean I'm certain UA already does this one way or another |
| 22:46 | <rniwa> | AryehGregor: intercepting keydown is problematic because it depends on each platform |
| 22:47 | <rniwa> | AryehGregor: e.g. ctrl+y doesn't necessary mean undo |
| 22:47 | <AryehGregor> | Hmm, I see. |
| 22:47 | <rniwa> | AryehGregor: it's cmd+y |
| 22:47 | <AryehGregor> | Oh, right. |
| 22:47 | <AryehGregor> | Or Ctrl-Shift-Z. |
| 22:47 | <rniwa> | AryehGregor: yeah |
| 22:47 | <rniwa> | e.g. |
| 22:47 | <rniwa> | etc... |
| 22:47 | <AryehGregor> | Okay, that makes sense. |
| 22:47 | <AryehGregor> | Also for insertLineBreak. |
| 22:47 | <rniwa> | also, the user agent may provide some UI such as context menu that triggers editing commands |
| 22:48 | <rniwa> | e.g. Safari provides Format in its context enu |
| 22:48 | <rniwa> | menu* |
| 22:48 | <rniwa> | and lets user bold, italicize, etc.. .text |
| 22:48 | <AryehGregor> | Interesting. |
| 22:48 | <rniwa> | AryehGregor: I guess we can talk more when we meet later this month :) |
| 22:49 | <AryehGregor> | So it would make sense to just fire an event for every execCommand() invocation, whether internal to the browser or not. |
| 22:49 | <AryehGregor> | Sure. :) |
| 22:49 | <rniwa> | AryehGregor: right |
| 22:50 | <rniwa> | AryehGregor: but I'm not sure if all browsers go through execCommand internally when executing user editing actions |
| 22:50 | <rniwa> | AryehGregor: they might do something slightly different when actions are user triggerd |
| 22:50 | <AryehGregor> | Right, that just has to be specced. |
| 22:50 | <rniwa> | yeah |
| 22:51 | <rniwa> | AryehGregor: but I don't think I should be doing that in my spec |
| 22:51 | <AryehGregor> | Right. |
| 22:51 | <rniwa> | AryehGregor: your spec is probably a better place to do |
| 22:51 | <rniwa> | AryehGregor: if you can give standarized names to each editing action |
| 22:51 | <rniwa> | AryehGregor: then I can refer to that in my spec |
| 22:51 | <rniwa> | AryehGregor: I really want to tie all of this to new mutation events replacement though |
| 22:52 | <AryehGregor> | Sounds good. |
| 22:52 | <rniwa> | it seems like the discussion on that thread is bikeshedding to some extent now |
| 22:52 | <rniwa> | but ideally, script should be able to get a list of mutations that happened for a given edit action |
| 22:54 | <AryehGregor> | Okay, go to go. See you around. |
| 22:54 | <rniwa> | k |
| 22:54 | <rniwa> | ttyl |
