| 00:02 | <GPHemsley> | Nevermind, I don't feel like having a detailed discussion right now :P |
| 00:03 | <GPHemsley> | Though it's possible the ideal W3C that I have in my head is too far from reality. |
| 01:57 | <Yuhong> | GPHemsley: Yep, I have considered "HTML5" a misnomer even as a buzzword for a while now. |
| 02:00 | <Yuhong> | http://news.ycombinator.com/item?id=4422122 |
| 02:00 | <Yuhong> | On Mosaic IMG tag. |
| 02:37 | <Hixie> | there's a CLOCK on this wiki |
| 02:37 | <Hixie> | o_O |
| 02:53 | <MikeSmith> | Hixie: that's a timer so that I can time how quickly I can delete edits that other people make |
| 03:06 | <Hixie> | hah |
| 04:13 | <cgcardona> | http://jsfiddle.net/E9DSH/ shows the networkState as 1 (NETWORK_IDLE) when the Audio Object's progress event fires when the spec says it should be 2 (NETWORK_LOADING). Am I missing something? |
| 04:13 | <cgcardona> | Also sometimes the suspend event fires with a NETWORK_LOADING instead of a NETWORK_IDLE |
| 04:13 | <cgcardona> | in chrome ^^ |
| 04:13 | <cgcardona> | Also in Safari - so I'm guessing webkit. |
| 04:14 | <cgcardona> | I'm wondering if it's just some subtlety in the resource selection algorithm |
| 04:15 | <cgcardona> | I'm looking here: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#event-media-progress |
| 04:17 | <cgcardona> | Also it's interesting that it fires more than one time |
| 04:18 | <MikeSmith> | cgcardona: it seems to be showing 2 for me |
| 04:18 | <MikeSmith> | in that jsfiddle |
| 04:18 | <cgcardona> | MikeSmith: on each page refresh? |
| 04:18 | <cgcardona> | it does perhaps once out of every 15 or so for me |
| 04:18 | <cgcardona> | put all the other times it's a 1 |
| 04:19 | <GPHemsley> | Hixie: Yup, it's a UTC clock. And if you click it, you purge the page cache (if that's a problem). |
| 04:19 | GPHemsley | is not sure what Yuhong was referring to. |
| 04:19 | <GPHemsley> | annevk: I notice the URL spec doesn't cite WebIDL. |
| 04:20 | <MikeSmith> | cgcardona: it's same for me after a page refresh |
| 04:20 | <MikeSmith> | consistently 2 |
| 04:20 | <cgcardona> | how strange. in chrome? what version? |
| 04:21 | <MikeSmith> | latest canary and dev on OS X |
| 04:22 | <MikeSmith> | 24.0.1302 |
| 04:22 | <cgcardona> | ok cool MikeSmith thanks. I'm only getting 2 maybe 15% of the time. strange. |
| 04:23 | <MikeSmith> | also I guess you know that test doesn't seem to work in Firefox or Opera |
| 04:23 | <cgcardona> | right its got an mp3 extension |
| 04:23 | <cgcardona> | my actual audio library accounts for that and detects the codec |
| 04:23 | <cgcardona> | and adds the correct file type |
| 04:24 | <Hixie> | GPHemsley: funky |
| 04:28 | <GPHemsley> | Incidentally, if you don't like the clock, you can turn it off in My preferences > Gadgets |
| 04:28 | <GPHemsley> | (You'll find a list of a bunch of other gadgets that are installed, as well.) |
| 04:30 | GPHemsley | wonders if he should take away annevk's minor edit privileges. |
| 04:52 | <GPHemsley> | <3 http://ian.hixie.ch/bible/policies |
| 04:53 | <Hixie> | http://ian.hixie.ch/bible/handling-people is the more useful one |
| 04:53 | <Hixie> | knowing those makes it so much easier to tell when people are manipulating you |
| 04:55 | <Hixie> | (for standards work, spotting the ones in the last section of that file is especially useful) |
| 04:56 | <cgcardona> | epic |
| 04:59 | <cgcardona> | 'Not sure if Hixie really thinks my work is shit—or if he's trying to tell if I'm manipulating him' http://memegenerator.net/instance/28964300 |
| 05:03 | GPHemsley | notes Hixie's rules for pluralizing words that end in 'us' is (almost) completely non-interoperable: http://ian.hixie.ch/bible/english |
| 05:06 | <GPHemsley> | (The exception is 'radii'.) |
| 05:07 | <GPHemsley> | Also, "similar" is undefined in "Any word whose singular ends with a *sound* like "us" or similar," |
| 05:07 | <Hixie> | details |
| 05:07 | <a-ja> | no "All Y'all" in en-US-hixie ? |
| 05:08 | <GPHemsley> | as there are at least 3 different sounds represented in that list of words |
| 05:08 | <Hixie> | not if you pronounce them correctly! :-P |
| 05:08 | <GPHemsley> | bus/cactus/octopus |
| 05:08 | <GPHemsley> | all have different vowels in their last syllable |
| 05:09 | <Hixie> | no no no, all the same |
| 05:09 | <Hixie> | LALALA I CAN'T HEAR YOU |
| 05:09 | <GPHemsley> | ^_^ |
| 05:09 | <GPHemsley> | if you can't hear me, how do you know what the sounds are? |
| 05:10 | <GPHemsley> | a-ja: If he listed all the 2nd person plurals in en-US, he'd have to create a completely different spec |
| 05:11 | <a-ja> | point taken |
| 05:11 | GPHemsley | wonders if he could ever find evidence of Hixie using singular 'they'... |
| 05:12 | <Hixie> | not hard, especially when i'm trying to be feminist |
| 05:12 | <GPHemsley> | Another violation, then. |
| 05:12 | <Hixie> | indeed :-) |
| 05:13 | <GPHemsley> | tsk tsk |
| 05:16 | <GPHemsley> | There has been surprisingly little use of 'thru' in this channel. |
| 05:18 | <GPHemsley> | Alright, that's enough Hixie stalking for one night. |
| 05:18 | <GPHemsley> | Time for me to go to bed. |
| 05:44 | <hsivonen> | context for http://memegenerator.net/instance/28964300 ? |
| 05:55 | <cgcardona> | heh |
| 05:55 | <cgcardona> | hixie's comment like two lines above it |
| 05:55 | <cgcardona> | Hixie: knowing those makes it so much easier to tell when people are manipulating you |
| 05:56 | <cgcardona> | hsivonen: ^ |
| 05:56 | <cgcardona> | just a joke |
| 05:59 | <hsivonen> | cgcardona: oh I see |
| 06:00 | <hsivonen> | pattern matching the uri mailing list discussion to handling-people would be an amusing exercise |
| 06:07 | <hsivonen> | Is http://css-live.ru/ a legitimate site? |
| 06:07 | <hsivonen> | i.e. not a SEO scheme? |
| 06:09 | <MikeSmith> | hsivonen: never came across it before |
| 06:10 | <MikeSmith> | but link:css-live.ru shows only 32 links to it |
| 06:11 | <MikeSmith> | hmm mostly from the site itself and from "wordpressadmin.ru" |
| 06:13 | <hsivonen> | they ask for permission to translate articles |
| 06:13 | <MikeSmith> | oh |
| 06:13 | <hsivonen> | so far, my experience with offers to translate my stuff into Eastern European languages have been SEO schemes |
| 06:14 | <hsivonen> | hmm. the start and end of that sentence don’t match |
| 06:14 | <MikeSmith> | yeah same here |
| 06:14 | <hsivonen> | I guess I forgot what I was saying in mid-sentence |
| 06:14 | <MikeSmith> | heh |
| 06:15 | <MikeSmith> | from what I have seen, in many cases where W3C gets offers to translate something, it's from somebody doing it just for SEO reasons |
| 06:17 | <cgcardona> | MikeSmith: this shows what's happening with the audio object in greater detail. http://jsfiddle.net/CPrtb/ You'll need to have your console open so you can see the logs. |
| 06:18 | <cgcardona> | of course i don't have a button for the volume on there but those events fire as well. |
| 06:19 | <cgcardona> | admittedly verbose :p |
| 06:24 | <MikeSmith> | cgcardona: that music is nice until the dude starts single |
| 06:24 | <cgcardona> | Lp |
| 06:24 | <cgcardona> | :p |
| 06:24 | <MikeSmith> | did Taylor Swift write this song? |
| 06:24 | <MikeSmith> | :) |
| 06:24 | <MikeSmith> | a |
| 06:24 | <cgcardona> | :) |
| 06:24 | <MikeSmith> | anyway, I see only NETWORK_LOADING in the console |
| 06:25 | <cgcardona> | never a NETWORK_IDLE? |
| 06:25 | <cgcardona> | wonder what the heck is going on over here? |
| 06:25 | <MikeSmith> | ah weird |
| 06:25 | <cgcardona> | i get NETWORK_IDLE for suspend, pause, timeupdate |
| 06:25 | <cgcardona> | canplay, canplaythrough |
| 06:26 | <cgcardona> | anyway - it's no biggie. I'm just trying to understand the Audio Object greater |
| 06:26 | <MikeSmith> | after a page reload yeah now I'm seeing _IDLE for play |
| 06:26 | <MikeSmith> | I see _LOADING for pause |
| 06:26 | <cgcardona> | i'm guessing it has to do with the media element load algorithm in a way I'm not understanding. |
| 06:27 | <cgcardona> | i've not read through the media element load algorithm too deeply so I'm gonna check it out now. |
| 06:41 | <MikeSmith> | hsivonen: fyi I finally just now pushed the datetime change to the repo, and re-deployed for all the W3C validator hosts |
| 06:43 | <hsivonen> | MikeSmith: ok. I wonder if I should redeploy now or after TPAC. |
| 06:43 | <MikeSmith> | no rush I guess |
| 06:44 | <hsivonen> | hmm. Robin is listed as unaffiliated as a TAG member |
| 06:44 | <MikeSmith> | I thought Robin had stepped down from the TAG |
| 06:44 | <hsivonen> | oh |
| 06:44 | <hsivonen> | I was just checking if getting hired by the W3C triggered some rule that prohibited continued participation on the TAG |
| 06:45 | <hsivonen> | so the answer is “yes”? and the TAG page is just out of date? |
| 06:46 | <MikeSmith> | I think there's no prohibition but it was probably more just about him making a choice about his time |
| 06:46 | <hsivonen> | I see |
| 06:46 | <MikeSmith> | but yeah I think it's probably out of date |
| 06:47 | <hsivonen> | so being on both the TAG and on the HTML WG was too much even for Robin. :-) |
| 06:48 | <hsivonen> | groups to try next: public-tracking |
| 06:49 | <MikeSmith> | that's a fun list |
| 06:49 | <annevk> | GPHemsley: I can only make big changes? |
| 06:49 | <MikeSmith> | hsivonen: Robin stepped down as co-chair of the DAP WG as well, btw |
| 06:50 | <annevk> | GPHemsley: and yeah, yeah, I'll add a reference; I'll do that Oprah thing at one point, references for everyone |
| 06:50 | <hsivonen> | MikeSmith: ok. I hadn’t noticed. |
| 06:50 | <hsivonen> | I was wondering why there was a different chair listed on the TPAC schedule |
| 07:09 | <annevk> | Hixie: btw, I found the bug with dfn.js; you cannot use ID names such as "constructor" |
| 07:09 | <annevk> | Hixie: I ended up renaming the section to "Constructors" |
| 07:38 | <hsivonen> | annevk: in your formulation for text/css encoding determination, you forgot the case where an ASCII-encoded @charset rule claims the encoding is UTF-16 (or BE) |
| 07:39 | <hsivonen> | annevk: also, the formulation checks the validity of the label only at the end |
| 07:39 | <hsivonen> | instead upon seeing each label candidate |
| 07:39 | <annevk> | hsivonen: yeah, Simon pointed out the second bug |
| 07:39 | <annevk> | my bad |
| 07:39 | <hsivonen> | TabAtkins: ^ |
| 07:40 | <annevk> | I can rewrite it I suppose, I believe TabAtkins likes to just copy & paste... |
| 07:40 | <hsivonen> | ok |
| 07:41 | <hsivonen> | I noticed the bug about checking the label only at the end after I had already written code to check the label immediately in each case |
| 07:42 | <hsivonen> | shows that I think I know what the spec is supposed to say instead of really following the spec. :-/ |
| 07:43 | <annevk> | I wonder if I should introduce "known label" or just use "get an encoding" in each step |
| 07:43 | <annevk> | I guess I'll try the latter first since Simon preferred that |
| 07:44 | <hsivonen> | annevk: it would be nice if the resulting text made it obvious that the BOM takes precedence without spec lawyering |
| 07:45 | <hsivonen> | can a non-ASCII byte sequence be valid in two of the three legacy Japanese encodings? |
| 07:45 | <hsivonen> | or in both of the legacy Korean encodings? |
| 07:52 | <annevk> | hsivonen: how about now: http://html5.org/temp/cssencoding.txt |
| 07:52 | <annevk> | oh shit |
| 07:52 | <annevk> | forgot about utf-16 again :/ |
| 07:53 | <annevk> | fixed that too |
| 07:53 | annevk | curses utf-16 |
| 07:55 | <annevk> | hsivonen: btw, I also added <?xml-stylesheet?>, Simon suggested it could have a charset attribute too |
| 07:56 | <annevk> | hsivonen: if that's bogus I'll remove it again; I really think step 4/5 ought to move elsewhere, they should be an argument to determining the encoding |
| 07:57 | <hsivonen> | annevk: I wonder if step 2 should change UTF-16 to UTF-8 like <meta> |
| 07:57 | <hsivonen> | no idea what’s more Web compatible |
| 07:57 | <hsivonen> | Gecko already supported xml-stylesheet charset |
| 07:59 | <hsivonen> | so looks OK except I’m not sure if step 2 should do the thing that <meta> does with UTF-16 |
| 07:59 | <annevk> | hsivonen: normally we just ignore utf-16 labels right? |
| 07:59 | <hsivonen> | do we? where? |
| 07:59 | <annevk> | hsivonen: cross-document |
| 07:59 | <hsivonen> | but this is not cross-document |
| 07:59 | <annevk> | hsivonen: should we ignore it for <link charset>? |
| 07:59 | <annevk> | hsivonen: yeah okay, I'm happy to align with <meta> |
| 08:00 | <annevk> | if you make the call I'll update my text file |
| 08:00 | <hsivonen> | annevk: I wouldn’t ignore it for <link charset> without data showing that ignoring is more Web-compatible |
| 08:00 | <annevk> | writing some minor text/css history here :p |
| 08:00 | <hsivonen> | annevk: the sheet operates in the origin of the including document, so the security scenario is not like iframe |
| 08:01 | <hsivonen> | annevk: let’s align with <meta> |
| 08:04 | <annevk> | hsivonen: reload |
| 08:04 | <hsivonen> | how does the process work at the CSS WG? Will this text go through an epic debate in a telecon next? |
| 08:04 | <hsivonen> | annevk: thanks. looks good |
| 08:04 | <hsivonen> | TabAtkins: http://html5.org/temp/cssencoding.txt |
| 08:07 | <annevk> | hsivonen: last time TabAtkins just put it in css3-syntax |
| 08:07 | <annevk> | hsivonen: I understand that Bert does not like css3-syntax, but that's a minor hurdle |
| 08:09 | <hsivonen> | meanwhile on www-tag: “Wrong on several counts. First, this isn’t what ‘meaning’ means.” |
| 08:09 | <annevk> | I was loving that tweet |
| 08:10 | <annevk> | subject too "Working without being ambushed by Ambiguity" |
| 08:12 | <annevk> | whoa, timbl started that thread |
| 08:12 | <annevk> | http://lists.w3.org/Archives/Public/www-tag/2012Oct/0086.html |
| 08:12 | <annevk> | pretty lengthy |
| 08:21 | <hsivonen> | From the W3C: “The Open Web Platform, which includes core W3C technologies such as HTML, CSS, SVG, XML, XSLT, XSL-FO, PNG, RDF, and many more, are already extensively in eBooks and eBook production.” |
| 08:26 | <annevk> | Open technologies for your walled gardens! We'll even promote them too! |
| 08:58 | <MikeSmith> | hsivonen_: left out TTML from that list |
| 08:58 | MikeSmith | makes a note to add it later |
| 09:02 | <hsivonen> | sigh. what’s wrong with me. I’ve started hg pulling without hg qpopping first |
| 09:02 | <MikeSmith> | didn't know you were using queues |
| 09:02 | <MikeSmith> | I've doned that too |
| 09:09 | <jgraham> | hsivonen: What's wrong with you is that you're still using hg ;) |
| 09:09 | Ms2ger | kicks jgraham |
| 09:43 | <hsivonen> | I just took a look at our Character Encoding menu code |
| 09:43 | <hsivonen> | complex code is complex |
| 09:43 | <hsivonen> | RDF and all |
| 09:45 | <Ms2ger> | Yay RDF |
| 09:46 | <hsivonen> | uh oh. our plug-in host has code for translating Gecko charset names to Java charset names |
| 09:46 | <hsivonen> | I wonder what that is about |
| 09:53 | <annevk> | hsivonen: try to make your menu more like Chrome/Safari |
| 09:54 | <hsivonen> | annevk: what data is the non-inclusion of e.g. MacGreek in the Encoding Standard based on? |
| 09:55 | <hsivonen> | that is, Mac encodings other than roman and cyrillic |
| 09:55 | <hsivonen> | what about EUC-TW? based on not having a common name among browsers? |
| 09:55 | <annevk> | for single-byte encodings the data I used for my decisions is here: http://lists.w3.org/Archives/Public/www-archive/2012Apr/att-0058/spectable.html |
| 09:55 | <annevk> | (there's several variants of that table floating around, I think that's the latest) |
| 09:56 | <annevk> | for multi-byte encodings I looked on those common among browsers, indeed |
| 09:56 | <annevk> | basically the same way of decision making, just less data on the latter because there's not that many |
| 09:57 | <hsivonen> | annevk: that page doesn’t explain the non-inclusion of non-Roman, non-Cyrillic Mac |
| 10:00 | <annevk> | hsivonen: sorry |
| 10:00 | <annevk> | http://lists.w3.org/Archives/Public/www-archive/2011Dec/att-0021/encoding-labels.html is somewhat better at addressing that question I suppose |
| 10:00 | <annevk> | non-support by Chrome/Opera was the reason for removal of most of them it seems, sometimes non-support by Firefox/IE/Safari |
| 10:01 | <annevk> | (that table is hard to read though :/) |
| 10:02 | <hsivonen> | I see. I hope it’s the right call for MacGreek, etc. |
| 10:02 | <hsivonen> | most likely is |
| 10:02 | <annevk> | hsivonen: I think the reason is that Chrome has no support for x-mac-ce, x-mac-greek, x-mac-turkish |
| 10:03 | <hsivonen> | I just want to know the basis in case someone blames me if we remove those and some site breaks |
| 10:03 | <annevk> | hsivonen: and Opera on non-Mac platforms has no support for those either |
| 10:03 | <hsivonen> | annevk: interesting |
| 10:03 | <annevk> | (weird decision by Opera's encoding guy imo) |
| 10:03 | <hsivonen> | I’d be surprised if there were Mac*Roman* Web sites |
| 10:03 | <annevk> | (there's a bug that may or may not be fixed already) |
| 10:04 | <annevk> | hsivonen: they also did not turn up in the content studies, but macintosh and I think ukrainian did |
| 10:04 | <annevk> | but it's now almost a year since I did the research so memory is fussy |
| 10:05 | <hsivonen> | in Mac Opera pulling additional encodings from a platform API or something? |
| 10:06 | <annevk> | I guess I can say it's an ifdef, I hope they won't sue :) |
| 10:06 | <odinho> | annevk: oh my oh my oh my ;-) |
| 10:35 | <annevk> | hsivonen: should I subscribe to apps-discuss? |
| 10:36 | <hsivonen> | annevk: probably better not to |
| 10:37 | <hsivonen> | annevk: do you want to argue about how reality-aligned text/* registrations will be? |
| 10:37 | <hsivonen> | 386 opportunity |
| 10:37 | <annevk> | heh |
| 10:38 | <annevk> | I'm hoping GPHemsley will squash that argument in MIME Sniffing at some point |
| 10:38 | <hsivonen> | is GPHemsley editing MIME Sniffing? |
| 10:43 | <annevk> | he's considering it |
| 10:45 | <MikeSmith> | hmm isn't foo&;bar a document-conformance error? |
| 10:45 | <MikeSmith> | and parse error |
| 10:46 | <MikeSmith> | spec says an ambiguous ampersand is and ampersand followed by one or more alphanumeric chars followed by semicolon |
| 10:46 | <MikeSmith> | I think I assumed it was zero or more alphanumeric |
| 11:01 | <MikeSmith> | hmm yeah foo&;bar is not a parse error and not a document-conformance error either |
| 11:10 | <zcorpan> | annevk: looks good. now i'll check how well browsers match it |
| 11:15 | <annevk> | hsivonen: so I guess at some point I should write a xmlencoding.txt too? |
| 11:15 | <hsivonen> | possibly. that code in Gecko is semi-bogus |
| 11:16 | <hsivonen> | and hard to fix nicely |
| 11:36 | <annevk> | fair enough and actually, if we're gonna touch XML at all, we might as well improve it a whole lot |
| 11:49 | <MikeSmith> | argh is "&hyphen" still meant to be a parse error? |
| 11:51 | <MikeSmith> | &hypen; (with semi-colon) is listed in the named-character references tables but &hyphen is not |
| 11:52 | <MikeSmith> | for cases like ´ which is listed both with and without semicolon in the table, my workspace validator code is still correctly reporting a parse error |
| 11:53 | <MikeSmith> | but for ones like &hyphen which aren't listed without a semicolon I'm not reporting an error |
| 11:53 | <MikeSmith> | any more |
| 11:53 | <MikeSmith> | I wonder if I should be |
| 11:54 | <annevk> | you want me to try read the spec for that case? |
| 11:56 | <MikeSmith> | nah |
| 11:56 | <MikeSmith> | I just need to re-read it |
| 11:56 | <MikeSmith> | but I'm hungry |
| 11:56 | <MikeSmith> | so I'll eat and then re-read |
| 11:57 | <zcorpan> | don't forget to consume the hyphen at the end |
| 11:57 | <zcorpan> | i mean semicolon |
| 11:57 | <zcorpan> | if your meal does not end with a semicolon, it may or may not be a parse error |
| 11:58 | <MikeSmith> | hah |
| 11:58 | <MikeSmith> | clever |
| 11:59 | <zcorpan> | (possibly you need to unconsume all of it and then eat it again, but let's not worry about that) |
| 12:07 | <jgraham> | Please don't eat your meal again if you have to unconsume it because it didn't reach a semicolon |
| 12:07 | <jgraham> | That's not healthy |
| 12:58 | <zcorpan> | hsivonen: why wouldn't <p> precede <main> in conforming content? |
| 13:54 | <GPHemsley> | annevk: You tend to be a little overzealous in what you consider "minor" ;) |
| 14:02 | <GPHemsley> | annevk: In order to edit mimesniff, I'd have to be a member of the whatwg organization on github, right? |
| 14:02 | <annevk> | GPHemsley: yeah |
| 14:03 | <annevk> | GPHemsley: I can add you if that's what you're asking for |
| 14:03 | <annevk> | GPHemsley: I'm non-intrusive man, just look at uri⊙wo |
| 14:04 | <GPHemsley> | annevk: Sure, then, I suppose I'll give it a shot. |
| 14:04 | <annevk> | GPHemsley: username is GPHemsley? |
| 14:04 | <GPHemsley> | yup |
| 14:05 | <GPHemsley> | I still do wish we had more than one mailing list, though |
| 14:05 | <GPHemsley> | I really don't want to subscribe to a list where I won't read 95% of the content :/ |
| 14:06 | <annevk> | GPHemsley: just ask people to file bugs rather than email |
| 14:06 | <annevk> | GPHemsley: and archive stuff that does not have you in cc or to or something like that |
| 14:06 | <annevk> | GPHemsley: but whatwg is not super high traffic |
| 14:06 | <GPHemsley> | It used to be... |
| 14:06 | <GPHemsley> | (I was subscribed a few years ago) |
| 14:07 | <GPHemsley> | but OK |
| 14:07 | <annevk> | GPHemsley: you're part of the GitHub team now |
| 14:07 | <GPHemsley> | I'll give it another shot |
| 14:08 | <GPHemsley> | annevk: Thanks! |
| 14:11 | <GPHemsley> | If anyone was worried, I've restored the favicon to the wiki. |
| 14:13 | GPHemsley | wonders what lxml is, since it has a lot of install warnings. |
| 14:14 | <annevk> | I kinda want to define scheme as including the ":" |
| 14:16 | GPHemsley | finds it curious that the new wiki search is case-sensitive |
| 14:17 | <annevk> | I guess nobody can stop me doing that, but I if anyone thinks of any issues, let me know! |
| 14:17 | <MikeSmith> | oh boy |
| 14:18 | <annevk> | MikeSmith: was uri⊙wo the mailing list we discussed the other day btw? |
| 14:18 | <MikeSmith> | annevk: I can't imagine that anybody's going to object to you re-defining what a scheme is |
| 14:18 | <MikeSmith> | annevk: yeah |
| 14:18 | <MikeSmith> | you posted to it didn't you? |
| 14:18 | <annevk> | MikeSmith: so I guess I sort of covered posting on there |
| 14:18 | <annevk> | MikeSmith: well someone else started, my name is in the thread's subject... |
| 14:19 | <GPHemsley> | anolis, html5lib, and mimesniff all have to have the same parent directory, right? |
| 14:19 | <MikeSmith> | annevk: hey I was thinking we should post to whatever mailing list there is for curl developers |
| 14:19 | <MikeSmith> | or curl developers and users |
| 14:19 | <MikeSmith> | and wget, etc. |
| 14:19 | <annevk> | GPHemsley: no, just xref and mimesniff |
| 14:19 | <annevk> | GPHemsley: anolis/html5lib you need to install using system Python |
| 14:20 | <MikeSmith> | as in post a heads-up saying Hey here's the new URL spec that it would be great to have some review and feedback on |
| 14:21 | <annevk> | MikeSmith: maybe I should write the thing first where you can terminate the parser if input does not match the URL syntax |
| 14:21 | <annevk> | MikeSmith: as far as I can tell that should be relatively straightforward |
| 14:23 | <GPHemsley> | annevk: I'm getting a "/bin/sh: anolis: command not found" error |
| 14:24 | <annevk> | GPHemsley: did you install Anolis? |
| 14:24 | <GPHemsley> | oh, hang on, I think it's just a PATH issue |
| 14:26 | <GPHemsley> | hmm, it says I need cssselect installed |
| 14:28 | <MikeSmith> | annevk: that sounds good to me |
| 14:28 | <MikeSmith> | (about the terminating the parser stuff) |
| 14:32 | <GPHemsley> | hmm... apparently if you run make without having changed anything, the compiled file disappears |
| 14:32 | <GPHemsley> | (that seems weird) |
| 14:34 | <annevk> | run |
| 14:34 | <annevk> | touch Overview.src.html |
| 14:34 | <annevk> | make |
| 14:34 | <annevk> | if it's still gone, something else is amiss |
| 14:36 | <GPHemsley> | annevk: Well, I made actual changes and ran make again, and it came back |
| 14:36 | <GPHemsley> | (recall that the file was already present before I ran make on an unchanged source) |
| 14:36 | <annevk> | sure, doesn't seem like a huge deal then |
| 14:43 | <dsadinoff> | To whom can we turn to solve the "I'd like to see frozen headers while scrolling a table" problem? There seem to be many half-baked solutions, but none that work reliably. Seems like a job for CSS, no? |
| 14:43 | <annevk> | dsadinoff: yeah, www-style⊙wo |
| 14:44 | <annevk> | dsadinoff: WHATWG isn't doing any CSS for now |
| 14:44 | <dsadinoff> | okay, sorry to bother you. |
| 14:44 | <MikeSmith> | dsadinoff: it's still OK to talk about CSS here :) |
| 14:44 | <annevk> | right, this channel is pretty open-ended |
| 14:47 | <dsadinoff> | Mostly, I'm looking for evidence that it's on someone's to-do list. It's kind of an HTML issue, in that I see it as the unfulfilled promise of HTML4's THEAD |
| 14:48 | <GPHemsley> | abarth, Hixie: Do you relinquish editorship of mimesniff? |
| 14:48 | <annevk> | dsadinoff: it's not HTML |
| 14:49 | <annevk> | dsadinoff: HTML has no control over UI or look & feel, CSS does to some extent |
| 14:49 | <dsadinoff> | THEAD seems like HTML to me. |
| 14:50 | <SimonSapin> | dsadinoff: feel free to ask on http://lists.w3.org/Archives/Public/www-style/ (the CSS mailing list) |
| 14:50 | <annevk> | MikeSmith: so I did redefined query/fragment to include the leading ?/# |
| 14:50 | <annevk> | -d |
| 14:50 | GPHemsley | mumbles something about anolis not letting him use @ |
| 14:51 | <annevk> | GPHemsley: not gonna protect you |
| 14:51 | <GPHemsley> | from the stupid ones... >_> |
| 14:51 | <dsadinoff> | Thanks everybody. there seems to be an interesting and germane discussion at www-style on "sticky positioning" . |
| 14:51 | <dsadinoff> | cheers! |
| 14:52 | <MikeSmith> | annevk: redefine all the things! |
| 14:52 | <annevk> | MikeSmith: it's not exactly on purpose, it was just the easiest |
| 14:53 | <annevk> | MikeSmith: and "redefining" scheme makes it easier too, and we do often talk about data: URLs and javascript: URLs |
| 14:53 | <annevk> | it fits usage too |
| 14:53 | <MikeSmith> | true |
| 14:53 | <annevk> | but I can see arguments either way, e.g. that syntax constructs should not be part of the value |
| 14:53 | <annevk> | but that's mostly a purity argument, not a pragmatic one |
| 14:54 | <annevk> | man, I think about this too much, better take a break |
| 14:55 | <GPHemsley> | :) |
| 14:55 | <GPHemsley> | annevk: So, are you doing away with the // or what? |
| 14:56 | <annevk> | GPHemsley: can't; http:example.org on http://foo/ results in http://foo/example.org |
| 14:57 | <GPHemsley> | it does? o_0 |
| 14:58 | <annevk> | but now this |
| 14:58 | <annevk> | http:example.org on https://foo/ results in http://example.org/ |
| 14:58 | <annevk> | you know why? because topic |
| 14:58 | <GPHemsley> | -_- |
| 14:59 | <Hixie> | GPHemsley: i haven't edited it since it came out of the html spec |
| 14:59 | GPHemsley | takes that as a yes |
| 14:59 | <annevk> | GPHemsley: generally you should feel free to submit patches to any of the specs in the whatwg repo |
| 15:00 | <annevk> | GPHemsley: Ms2ger, I, and others review the changes and will let you know if something goes amiss |
| 15:00 | <GPHemsley> | annevk: For mimesniff, though, I can commit directly, right? Or no? |
| 15:01 | <annevk> | commit-then-review is what we usually practice |
| 15:01 | GPHemsley | needs to clarify what you mean by "submit patches", as he hasn't seen anyone do that. |
| 15:01 | <GPHemsley> | ah, ok |
| 15:01 | <GPHemsley> | I come from Mozillaland, where patches are submitted to bug reports to be reviewed. |
| 15:02 | <Hixie> | the whatwg attitude is "trust and verify", we just assume that if someone goes crazy we'll revoke their rights and restore from backup |
| 15:02 | <GPHemsley> | k |
| 15:02 | <Hixie> | so far you seem pretty sensible, to me :-P |
| 15:02 | <annevk> | maybe we should do that at some point, but thus far for every 1000 commits there's only a couple that are wrong |
| 15:02 | <GPHemsley> | :) |
| 15:02 | <annevk> | so it's not really worth the hassle |
| 15:03 | <GPHemsley> | ok, just wanted to clarify |
| 15:05 | <annevk> | GPHemsley: you already forgot you have to ask for forgiveness? :p |
| 15:05 | GPHemsley | is not used to having so much power. |
| 15:07 | <GPHemsley> | In that case, I hope abarth forgives me if his answer is no. :) |
| 15:11 | <GPHemsley> | it sure is easy to break anolis... |
| 15:11 | <GPHemsley> | have one undefined reference, and the whole thing explodes ;) |
| 15:12 | <GPHemsley> | How do I add a new reference? Do I have to edit xref? |
| 15:12 | <annevk> | Hixie: I think I asked this before, why is it "space characters" and not "ASCII whitespace"? |
| 15:13 | <annevk> | GPHemsley: yes |
| 15:13 | <MikeSmith> | annevk: so you think it would be useful to send a heads-up about the URL spec to the curl mailing list? "review is welcome" etc.? |
| 15:13 | <MikeSmith> | if so I can send one |
| 15:14 | <annevk> | MikeSmith: ah yeah, why not, maybe indicate the plan is to give the parser a strict option |
| 15:14 | <MikeSmith> | OK |
| 15:14 | <annevk> | MikeSmith: most of the groundwork for such an option is now in place, might add it tomorrow |
| 15:14 | <MikeSmith> | hai |
| 15:14 | <GPHemsley> | annevk: references.json? |
| 15:14 | <annevk> | MikeSmith: oh and maybe cc me and say they have to cc me if they want replies |
| 15:14 | <MikeSmith> | k |
| 15:14 | <annevk> | GPHemsley: yes, for a data-anolis-ref |
| 15:14 | <Hixie> | annevk: same reason it's not "characters that are spaces" or "spaces" or "whitespace characters" or "the bimbams" or any other equivalently defined opaque phrase :-) |
| 15:15 | <GPHemsley> | annevk: Are there other kinds? |
| 15:15 | <annevk> | Hixie: I might start using ASCII whitespace in my specs |
| 15:15 | <annevk> | Hixie: aligns better with a bunch of other terms I need for URL |
| 15:15 | <GPHemsley> | annevk: This one is unusual in that it's an academic reference, rather than a spec or other webpage |
| 15:16 | <annevk> | Hixie: and I have this wish of at some point having unified terminology... |
| 15:16 | <annevk> | GPHemsley: yeah, cross-spec references |
| 15:16 | <annevk> | GPHemsley: no that's fine, it can go in references.json |
| 15:16 | <GPHemsley> | it may have the record for longest ref abbreviation |
| 15:17 | <Hixie> | annevk: if you really care, file a bug, but it's a lot of work to change and unless you have a really compelling reason (cross-spec consistency isn't strong enough imho) i'll likely wontfix or later it... |
| 15:17 | <GPHemsley> | ...or maybe not |
| 15:17 | <GPHemsley> | annevk: Are their best practices for naming a ref? |
| 15:17 | <annevk> | Hixie: I was mostly wondering if you had a compelling reason for me not to pursue "ASCII whitespace" in my specs and maybe at some point once we're all bored fix specs to use that everywhere :) |
| 15:18 | <GPHemsley> | s/their/there/ |
| 15:18 | <annevk> | Hixie: so thanks :) |
| 15:18 | <annevk> | GPHemsley: uppercase, not too long, not really? |
| 15:18 | <GPHemsley> | annevk: Should there be? |
| 15:18 | <GPHemsley> | we have a lovely wiki... |
| 15:18 | <GPHemsley> | ;) |
| 15:19 | <annevk> | GPHemsley: heh, if you feel like finding the pattern and encouraging people to follow it, be my guest |
| 15:20 | <Hixie> | annevk: sounds fine to me. i could come up with an argument like "phrases with acronyms don't read as well" but it'd be a post-hoc rationalisation. |
| 15:21 | <annevk> | GPHemsley: as a warning of sorts, in general additional rules make it more complicated for people to contribute |
| 15:21 | <GPHemsley> | they're *guidelines* ;) |
| 15:22 | <annevk> | GPHemsley: I'm not opposed :) |
| 15:22 | <GPHemsley> | is there a way to specify a year/date in the references? |
| 15:22 | <annevk> | GPHemsley: don't think so |
| 15:22 | <annevk> | GPHemsley: put it in the title |
| 15:22 | GPHemsley | wonders if that's a problem. |
| 15:24 | <GPHemsley> | I'll leave it out for now. |
| 15:24 | <annevk> | MikeSmith: GPHemsley: good luck, gotta go train/cycle around |
| 15:24 | <GPHemsley> | k |
| 15:25 | <GPHemsley> | thanks |
| 15:25 | <Hixie> | as editor, you don't want the date in the refs, trust me |
| 15:25 | <MikeSmith> | enjoy annevk |
| 15:25 | <Hixie> | you'll just spend your whole time updating refs |
| 15:27 | <GPHemsley> | well, article dates tend not to change, but OK |
| 15:27 | <GPHemsley> | Hixie: Incidentally, what is RFCs.json for? Because apparently it's not used to cite as RFCs as references... |
| 15:31 | <MikeSmith> | it is used I think |
| 15:31 | <MikeSmith> | somehow |
| 15:31 | <MikeSmith> | or maybe not |
| 15:31 | <MikeSmith> | ms2ger would know |
| 15:33 | <Hixie> | hm, today will probably end up being a feature day |
| 15:34 | <Hixie> | the next few bugs are all feature requests |
| 15:34 | <Hixie> | allow-pointerlock for sandbox, and a bunch of appcache things |
| 15:35 | <GPHemsley> | Is it standard practice to list acknowledgements alphabetically by first name, or am I free to do whatever I want? :) |
| 15:36 | <GPHemsley> | Oh, and where does en-x-hixie stand on the spelling of acknowledgements? |
| 15:36 | <Hixie> | free to do whatever you want |
| 15:37 | <Hixie> | on pretty much anything |
| 15:37 | <Hixie> | and if your judgement is poor, you'll hear about it :-P |
| 15:37 | <GPHemsley> | well, the fact that you are not worried about my judgment makes me happy |
| 16:28 | <MikeSmith> | I pushed ampersand changes from my workspace to http://qa-dev.w3.org:8888/ |
| 16:28 | <MikeSmith> | behavior now should match the spec |
| 16:28 | <MikeSmith> | if anybody has time to test |
| 16:29 | MikeSmith | Memoserv help |
| 16:29 | <MikeSmith> | oops |
| 17:03 | <GPHemsley> | Oh, right, we're using W3C for bug tracking, aren't we? |
| 17:04 | <GPHemsley> | How do I get myself set as the default assignee/QA contact? |
| 17:28 | <Ms2ger> | GPHemsley, ask MikeSmith |
| 17:28 | GPHemsley | has asked so many questions today. |
| 17:28 | <GPHemsley> | Ms2ger: Which one are you referring to? |
| 17:29 | <Ms2ger> | The most recent |
| 17:29 | <Ms2ger> | <GPHemsley> How do I get myself set as the default assignee/QA contact? |
| 17:29 | <GPHemsley> | ah, thanks |
| 17:33 | <pablof> | i hear this Anne chick is hot, http://lists.webkit.org/pipermail/webkit-dev/2012-October/022583.html |
| 18:01 | GPHemsley | recalls making the same mistake. |
| 18:12 | <MikeSmith> | GPHemsley: for what component? MIME? |
| 18:12 | <GPHemsley> | yeah |
| 18:14 | <MikeSmith> | ok done |
| 18:18 | <GPHemsley> | thanks |
| 18:18 | GPHemsley | finds it interesting that whitespace between <p> and text is stripped but whitespace between <dt> and text is not. |
| 18:19 | <Hixie> | stripped by what? |
| 18:22 | <GPHemsley> | the browser |
| 18:24 | <tantek> | which browser? |
| 18:24 | <Ms2ger> | Not in the DOM, I presume |
| 18:27 | <Hixie> | not sure what you mean |
| 18:27 | <Hixie> | neither should be stripped |
| 18:31 | <GPHemsley> | if I have <p>\n\tFoo, the browser shows |Foo (where | is the edge of the box |
| 18:32 | <GPHemsley> | if I have <dt>\n\tFoo, the browser shows | Foo (where | is the edge of the box) |
| 18:32 | <Ms2ger> | Well, yes |
| 18:32 | <Ms2ger> | That's the margin on the dt |
| 18:32 | <Ms2ger> | No |
| 18:32 | <Ms2ger> | I'm stupid |
| 18:32 | <Ms2ger> | And I can't tell the difference between d ant t |
| 18:33 | <GPHemsley> | and this is in Gecko |
| 18:33 | <Ms2ger> | Not here |
| 18:35 | <GPHemsley> | really? |
| 18:35 | <GPHemsley> | hmm |
| 18:36 | <GPHemsley> | I wonder if it has anything to do with dt:before { content: 'asdf'; } |
| 18:37 | <GPHemsley> | oh, or maybe it's the inline-block |
| 18:42 | <Ms2ger> | Oh, and RFCs.json is just a list you can copy from if you need a reference to an RFC and you're too lazy to write it up yourself |
| 18:42 | <Ms2ger> | (I made it while working on the HTML References section) |
| 18:44 | <Hixie> | GPHemsley: in the absence of other styles, that's a bug |
| 18:45 | <GPHemsley> | Hixie: Alright, I'll investigate it later. |
| 18:45 | <GPHemsley> | Ms2ger: Ah, OK. That's what I used it for. :) |
| 18:45 | <Hixie> | GPHemsley: :before acts like a text node that was inserted at the start of the element, so if that doesn't end with a space, and it's in-flow, then that will do it |
| 19:00 | <Hixie> | anyone know what referer browsers use when fetching things listed in appcache manifests? |
| 19:06 | <ap> | Hixie: I think that Safari doesn't send any |
| 19:09 | <annevk> | pablof: pics or it didn't happen |
| 19:09 | <Hixie> | ap: interesting. |
| 19:09 | <annevk> | GPHemsley: Acknowledgments without e is per W3C style guide iirc |
| 19:09 | <GPHemsley> | does that matter? :) |
| 19:10 | <annevk> | GPHemsley: lexicographical by first name is because of nick names and complex sort order for last names in e.g. the Netherlands |
| 19:10 | <Hixie> | ap: any opinion on whether we should change that? turns out the spec is bogus on this (it essentially has a null deref), so i need to define something. The three options I can see are nothing, the manifest, and the Document that triggered the update. The manifest probably makes the most sense. |
| 19:10 | <GPHemsley> | annevk: Yeah, that one I decided to leave. |
| 19:10 | <annevk> | GPHemsley: I don't think an "e" matters a whole lot, just nice to be consistent |
| 19:10 | <Hixie> | i sort by first name because that's what emacs's "sort-lines" does |
| 19:11 | <GPHemsley> | Hixie: Ah, so that's why you leave the 'and' tacked on to the end of a line |
| 19:11 | <Hixie> | yeah :-) |
| 19:12 | <GPHemsley> | annevk: I'm of the opinion that 'dg' doesn't make the right sound without the 'e' after it, so... |
| 19:12 | <ap> | Hixie: agreed. that should be a relatively easy (albeit low priority) fix to use manifest for Referer for us, and I don't think that it could break anything |
| 19:13 | <Hixie> | ap: cool, i'll do that then, assuming my testing finds nothing substantial in other browsers. thanks. |
| 19:13 | <annevk> | GPHemsley: you're not an American? |
| 19:13 | <GPHemsley> | oh, I am |
| 19:13 | <GPHemsley> | but I'm also a linguist |
| 19:13 | <GPHemsley> | and a standards enthusiast |
| 19:14 | <Hixie> | woah, woah |
| 19:14 | <GPHemsley> | when you put the last two together, the third goes out the window ;) |
| 19:14 | <annevk> | from Wikipedia's manual of style "acknowledgement vs acknowledgment: acknowledgement is preferred in British English,[9] acknowledgment in American English." |
| 19:14 | <Hixie> | standards enthusiast? |
| 19:14 | <Hixie> | i'm not sure that's allowed |
| 19:14 | <GPHemsley> | uh oh |
| 19:14 | GPHemsley | runs |
| 19:14 | <annevk> | Hixie: what are you gonna do? :p |
| 19:14 | <Hixie> | damnit, there's nothing i CAN do! |
| 19:15 | <Hixie> | i will look at him or her sternly! |
| 19:15 | <Hixie> | (not falling into this trap of assuming people are a particular gender again!) |
| 19:15 | <GPHemsley> | now now |
| 19:15 | <GPHemsley> | everyone settle down |
| 19:16 | <GPHemsley> | would it ease your mind if s/standards/specification/ ? |
| 19:17 | <Hixie> | not sure |
| 19:17 | <Hixie> | what kind of specification? |
| 19:17 | <GPHemsley> | what are my options? |
| 19:18 | <Hixie> | orange specifications and yellow specifications. |
| 19:18 | <GPHemsley> | orange |
| 19:18 | <Hixie> | hmmmmmm ok then |
| 19:19 | <Hixie> | in completely unrelated news, can i just say that debugging appcache is a huge pain in the neck |
| 19:19 | <Hixie> | because you have to reload things like 5 times to make sure you have hte latest files |
| 19:20 | <GPHemsley> | oh, and FTR, I prefer male pronouns |
| 19:20 | <Hixie> | for use when refering to you, or to others? |
| 19:20 | <GPHemsley> | when referring to me |
| 19:20 | <Hixie> | just checking |
| 19:20 | <GPHemsley> | for everyone else, I just use "them" |
| 19:21 | <Hixie> | seems wise |
| 19:21 | Hixie | finds out why his file isn't getting cached |
| 19:21 | <Hixie> | forgot to update the manifest to point to it. |
| 19:21 | <annevk> | Hixie: http://url.spec.whatwg.org/#writing should now match IRI except that I allow "iprivate" anywhere rather than just in "iquery" |
| 19:21 | <Hixie> | good times |
| 19:21 | GPHemsley | will write an English orthography spec one day. |
| 19:22 | <Hixie> | annevk: is a URI just a URL that only contains US-ASCII? (ignoring what escapes expand to?) |
| 19:22 | <annevk> | Hixie: and that definition of URL units should at some point be aligned with the code point ranges of the HTML syntax (for code points above ASCII) |
| 19:22 | <annevk> | Hixie: yeah |
| 19:22 | <Hixie> | cool |
| 19:22 | <Hixie> | hm. i wonder if it would make IETF folk happier or less happy if you actually put that definition of URI in the spec, come to the think of it. |
| 19:23 | <Hixie> | probably not worth it, on the balance. |
| 19:23 | <GPHemsley> | Do we care about line length? |
| 19:24 | <annevk> | GPHemsley: readable diffs are nice |
| 19:24 | <GPHemsley> | well, I've put each sentence on a separate line |
| 19:24 | <GPHemsley> | I'm just wondering if I should also wrap lines |
| 19:24 | <Hixie> | i wrap to 100 chars these days |
| 19:24 | <annevk> | my line length is pretty short at the moment, 77 |
| 19:24 | <Hixie> | used to wrap to 70 but my computers got bigger |
| 19:25 | <annevk> | heh |
| 19:25 | <GPHemsley> | if I wrapped, it'd be at 78, I think |
| 19:25 | <GPHemsley> | so, did I hear two votes for wrapping? |
| 19:25 | <annevk> | oh it's 76 |
| 19:26 | <Ms2ger> | Hixie, which does make your diffs a lot less useful :) |
| 19:26 | <GPHemsley> | I'm hearing 3 votes for wrapping |
| 19:27 | <annevk> | Hixie: Progress Events is part of XMLHttpRequest now |
| 19:28 | <Ms2ger> | GPHemsley, wrap, but only if you do at 80 :) |
| 19:28 | <GPHemsley> | Ms2ger: What's wrong with 78? |
| 19:28 | <Ms2ger> | It's two characters short of 80 |
| 19:28 | <GPHemsley> | that leaves room for line endings, though |
| 19:29 | <Ms2ger> | Why not 79, then? :) |
| 19:29 | <GPHemsley> | I suppose the real question is, what are the usecases for wrapped lines? |
| 19:29 | <GPHemsley> | Just in case Windows gets involved |
| 19:34 | <Hixie> | Ms2ger: for now, yeah. |
| 19:34 | <Hixie> | Ms2ger: transition pain. |
| 19:35 | <Ms2ger> | Hixie, man, do you know how long your spec is? This is going to take ages ;) |
| 19:35 | <Hixie> | not as long as one might imagine |
| 19:35 | <Ms2ger> | $ wc -l source |
| 19:35 | <Ms2ger> | 125073 source |
| 19:37 | <annevk> | GPHemsley: readable diffs |
| 19:38 | <GPHemsley> | annevk: I can't tell if you're objecting or agreeing. |
| 19:38 | <Hixie> | Ms2ger: how many of those are lines less than 80 characters, that start with one or more spaces, and that are not followed by a blank line or a line with a different indent? |
| 19:38 | <annevk> | GPHemsley: that's the use case for wrapped lines and why I want them |
| 19:38 | <GPHemsley> | ohh |
| 19:38 | <GPHemsley> | what makes a diff unreadable? |
| 19:38 | <Ms2ger> | Hixie, I bet you can write an elisp one-liner to tell me that |
| 19:39 | <annevk> | GPHemsley: long lines, because it's line-based diffing typically |
| 19:39 | <Hixie> | Ms2ger: perl one-liner maybe. i don't do elisp. :-) |
| 19:40 | Ms2ger | gasps |
| 19:40 | <annevk> | GPHemsley: given 80 or so characters it's relatively easy to spot a change, if lines go over 9000 not so much |
| 19:40 | <GPHemsley> | annevk: so the problem is that the lines go off the screen? because one could also make the case that long single-lines make it easier to compare the lines |
| 19:40 | <Hixie> | i wrote a 1000-line multithreaded scheme app once (a command-line lambda calculus interpreter) and then i realised if i was going to do any more scheme or lisp i'd run out of parentheses, so i went to perl where they're optional. |
| 19:40 | <GPHemsley> | (assuming the same wrapped-line diff spans multiple lines) |
| 19:41 | GPHemsley | wonders how you run out of parentheses |
| 19:41 | <Ms2ger> | GPHemsley, well, sure, but you have to search for the difference within the line |
| 19:41 | <Hixie> | GPHemsley: you only get assigned a fixed number when you're born |
| 19:41 | <Ms2ger> | And that's easier for shorter lines |
| 19:41 | <GPHemsley> | Hixie: Well, shoot. I might run out, then. I use them all the time! |
| 19:42 | <Hixie> | gotta be careful with them, man! |
| 19:42 | <GPHemsley> | Ms2ger: Well, if you also keep one sentence per line, they're already relatively short. |
| 19:42 | <Ms2ger> | True |
| 19:42 | <Hixie> | http://www.hixie.ch/tests/adhoc/html/offline/005.html tells me opera and webkit return "" and gecko returns the Document. Wonder what Gecko returns if two docs are refreshing the same appcache at the same time... |
| 19:42 | <GPHemsley> | But this is all hypothetical now, because I've already begun the wrapping process on this spec |
| 19:42 | <Ms2ger> | The issue is mainly if you put entire paragraphs on one line |
| 19:43 | <GPHemsley> | right |
| 19:43 | <Hixie> | (fwiw, my main use case for wrapping is just that it makes editing easier. same reason why i don't do one-sentence-per-line, which is better for diffs: my editor setup makes it easier to wrap paragraphs than sentences, and it's easier to read regularly wrapped paragraphs than irregular-length sentences.) |
| 19:43 | <GPHemsley> | although these sentences do tend to go very long |
| 19:46 | <Ms2ger> | If your sentences get too long, use a nice <ol> :) |
| 19:47 | <Hixie> | i guess IE9 doesn't do appcache |
| 19:47 | <Hixie> | does IE10/ |
| 19:47 | <Hixie> | ? |
| 19:47 | <Hixie> | anyone got IE10 who can test http://www.hixie.ch/tests/adhoc/html/offline/005.html for me? |
| 19:48 | Ms2ger | has some version in a vm |
| 19:48 | <Hixie> | i wonder what the referrer for the manifest itself should be |
| 19:48 | <Hixie> | one of the documents at random, i guess, the first time, and then the manifest the second time, maybe |
| 19:48 | <Hixie> | or just the document both times |
| 19:50 | <Ms2ger> | FAIL |
| 19:50 | <Ms2ger> | And then ...005.html |
| 19:50 | <Ms2ger> | Hixie, ^ |
| 19:55 | <Hixie> | cool, thanks |
| 19:56 | <Hixie> | ap: so looks like two browsers do "" and two browsers use the Document's URL (presumably a random one, or the "first" one, if there are multiple at the same time) |
| 19:57 | <Hixie> | ap: which suggests maybe changing it to the manifest isn't a good move |
| 19:57 | <Hixie> | ap: do you mind if we use the (a) document url? |
| 20:01 | <Hixie> | ap: i've cc'ed you on https://www.w3.org/Bugs/Public/show_bug.cgi?id=17687 |
| 20:06 | <GPHemsley> | hmm... |
| 20:07 | <GPHemsley> | Ms2ger: Anolis seems to randomly insert a blank line with a bunch of whitespace in the middle of my table |
| 20:07 | <GPHemsley> | oh |
| 20:07 | <GPHemsley> | it's just poorly correcting a tag mismatch |
| 20:07 | <Ms2ger> | Blame html5lib :) |
| 20:07 | <GPHemsley> | ah, heh |
| 20:11 | <GPHemsley> | Well, that diff is messy and mostly useless. |
| 20:11 | <GPHemsley> | But I swear it's mostly whitespace. |
| 20:11 | <GPHemsley> | A few bits of punctuation and tag clean-up, too. |
| 20:11 | <GPHemsley> | But mostly whitespace. |
| 20:14 | GPHemsley | wonders if Alfred HÎnes should really be Alfred Hönes. |
| 20:15 | <GPHemsley> | (It wouldn't be Alfred HŒnes, would it?) |
| 20:16 | <GPHemsley> | yeah, I think I was right the first time |
| 20:18 | <annevk> | the name in the doc results in a bunch of results on Google |
| 20:19 | <annevk> | so I think it's correct |
| 20:19 | <GPHemsley> | yeah, but his ASCII signature and search for the umlaut version suggest otherwise |
| 20:21 | <annevk> | I guess you could email him what he prefers, but I reckon abarth paid attention to that |
| 20:23 | <GPHemsley> | IDK, the whole list came at once |
| 20:24 | <GPHemsley> | I get the feeling it was copy&pasted from an e-mail with encoding problems. |
| 20:24 | <GPHemsley> | https://github.com/whatwg/mimesniff/commit/886dd7e31f32f37a4d9a6933b867a72b95555db4 |
| 20:24 | <GPHemsley> | more specifically https://github.com/whatwg/mimesniff/commit/886dd7e31f32f37a4d9a6933b867a72b95555db4#L0R1136 |
| 20:25 | <annevk> | I doubt that's an encoding problem though, it seems very intentional |
| 20:25 | <GPHemsley> | I mean an encoding problem in the source material. |
| 20:27 | <annevk> | me too |
| 20:27 | <GPHemsley> | encoding problems are rarely intentional ;) |
| 20:27 | <annevk> | I cannot think of anything that turns ö into Î |
| 20:28 | <GPHemsley> | well, see e.g. http://www.ietf.org/mail-archive/web/dnsop/current/msg09665.html |
| 20:28 | <GPHemsley> | "From: Alfred Hönes" |
| 20:28 | <GPHemsley> | I'm thinking something turned oe -> Πfor some reason |
| 20:28 | <GPHemsley> | then MacRoman got interpreted as ISO 8859-1 or equivalent |
| 20:41 | <Velmont> | gsnedders: You said mozilla had a html5 parser lib for embedding when html5lib got too slow? |
| 20:43 | GPHemsley | wonders if the print template shouldn't up the font size. |
| 20:44 | <GPHemsley> | And maybe reduce the margins... |
| 20:49 | GPHemsley | wanders off to proofread |
| 20:53 | <ap> | Hixie: I doubt that there is any compatibility impact from any of the choices in this case. If we are trying to remain compatible however, I'd prefer going with WebKit behavior - we have more engine specific content, and would be much more likely to suffer. People use appcache on iOS quite a bit |
| 20:59 | <annevk> | dglazkov: getElementById should return Element, not HTMLElement, on ShadowRoot |
| 21:00 | <annevk> | dglazkov: same goes for the ShadowRoot constructor I guess |
| 21:02 | <annevk> | dglazkov: I guess I should give it some more careful review at some point, need more time :/ |
| 21:09 | <annevk> | oh boy |
| 21:09 | <annevk> | http://lists.w3.org/Archives/Public/www-archive/2012Oct/0084.html |
| 21:13 | <Velmont> | Anyone in Paris for TTWF? Hmm. It is kinda late, need to find food. Should potentially have an IRC-channel for testtwf |
| 21:13 | <annevk> | Velmont: going to TPAC? |
| 21:14 | <Velmont> | annevk: Indeed am, mr :] |
| 21:14 | <annevk> | Velmont: I'm prolly at some Hilton hotel in Lyon around 6PM Sunday |
| 21:14 | <annevk> | unless you know, things go wrong |
| 21:14 | <Velmont> | annevk: I have used some precious space in my rucksack for your t-shirt. |
| 21:15 | <annevk> | awesome! |
| 21:19 | <Velmont> | sicking: So... Can you think of any obvious missing IDB tests? |
| 21:21 | <Velmont> | I have already written all the obvious stuff I could think of, -- but I should really have a list ready for test the web forward authors. |
| 21:24 | <sicking> | Velmont: hmmm |
| 21:24 | <sicking> | Velmont: we do have a lot of tests in our testsuite. But our tests heavily depends on 'yield' so i don't think they would be easy to run in other browsers |
| 21:24 | <sicking> | Velmont: …until they support ES6 that is |
| 21:26 | <sicking> | Velmont: did you write tests for all different cursors? |
| 21:27 | <sicking> | Velmont: i.e. backward/forward/nondupe on both indexes and objectStores |
| 21:27 | <sicking> | Velmont: and that nondupe ones skips dupes as appropriate |
| 21:27 | <sicking> | Velmont: our tests are all here: http://mxr.mozilla.org/mozilla-central/source/dom/indexedDB/test/unit/ |
| 21:28 | <sicking> | Velmont: oh, and doing mutations as you are using all of these cursors |
| 21:28 | <Velmont> | sicking: I did look at them a bit when I did investigation. |
| 21:29 | <Velmont> | sicking: I **think** I have for back/forward/nodupe, -- quite a few tests on that. -- I have some for doing mutations as I'm doing most of those cursors... |
| 21:29 | <Velmont> | sicking: The biggest problem I have is information gathering/overload/etc probably. |
| 21:29 | <sicking> | Velmont: "overload"? |
| 21:29 | <Velmont> | sicking: I started marking up my tests with the assertions they test. But that was tedious and boring. |
| 21:31 | <Velmont> | overload as in I can't keep the overview over what tests I've written and which ones I've yet to do in my head any more. |
| 21:31 | <sicking> | Velmont: ah |
| 21:32 | <Velmont> | And all the lists I've started creating is half-baked :P -- Hmm. Should probably continue tagging the tests... |
| 21:32 | <Velmont> | sicking: I have some hacky javascript that colours the spec green for the assertions it finds in the tests. |
| 21:33 | <sicking> | Velmont: tests for what happens after a database or transaction is closed is likely going to find you bugs. I.e. checking that calling various functions works or throws as expected |
| 21:34 | <sicking> | Velmont: OOoh, error handling in general would be great |
| 21:34 | <Velmont> | sicking: Yeah, -- I only have an easy test for that I think. |
| 21:34 | <Velmont> | (not only one easy test for error handling in general, but that specific close case :P) |
| 21:34 | <Velmont> | sicking: Ahh. And ordering of exceptions ... |
| 21:34 | <Velmont> | sicking: If we only had a spec for that :P |
| 21:35 | <sicking> | Velmont: indeed! :) |
| 21:35 | <Velmont> | sicking: Then I could have someone write the test for it :D |
| 21:36 | <Velmont> | Hmm. 1) find all the cases where exception ordering matter 2) list them 3) make a helper that does that, and reports what exception fired 4) find commonalities / make a nice list over how you think stuff should behave. |
| 21:37 | <Velmont> | Could always actually have them help with researching it, helping to spec it. Guess that would be cool for someone that is interested. |
| 22:12 | <Velmont> | Soo... MikeSmith or whoever does magic, -- can webapps and webappsec hg repositories be mirrored on github? |
| 22:25 | GPHemsley | can't help but think that the mimesniff algorithm could be greatly simplified. |
| 22:26 | <Velmont> | Uhhh... That email on html-public now, "not safe". What a strange strange email. |
| 22:43 | GPHemsley | notes that it seems everyone violates the GIF parsing spec with regard to spec version numbers. |
| 22:49 | <GPHemsley> | does anyone know off-hand if HTTP specifies how to handle the case when a given header is sent more than once? |
| 22:50 | <GPHemsley> | oh, it seems that certain headers allow more than one per request |
| 22:51 | <Velmont> | Yep. :-) And some don't. |
| 23:14 | <karlcow> | GPHemsley: yes for example, you may have multiple "Via: " headers |
| 23:16 | <karlcow> | I do not remember from the top of my head in case of multiple same header when only one is expected what the spec says. |
| 23:18 | <GPHemsley> | what I'm concerned with is whether I have to define the behavior or if it's already defined in HTTP |
| 23:20 | <karlcow> | "HTTP does not define |
| 23:20 | <karlcow> | specific error handling mechanisms except when they have a direct |
| 23:20 | <karlcow> | impact on security, since different applications of the protocol |
| 23:20 | <karlcow> | require different error handling strategies." — http://tools.ietf.org/html/draft-ietf-httpbis-p1-messaging-21#section-2.5 |
| 23:21 | <karlcow> | GPHemsley: a simple question on HTTP mailing-list can help, just exposing the case and asking if it is defined. |
| 23:28 | <karlcow> | Though GPHemsley see section 3.2 http://tools.ietf.org/html/draft-ietf-httpbis-p1-messaging-21#section-3.2 |
| 23:29 | <zewt> | whoever decided that 1+null == 1, but 1+undefined == NaN, needs to be shot into the sun |
| 23:29 | <karlcow> | Paragraph starting with "Multiple header fields with the same field name MUST NOT be sent in a" |
| 23:30 | <karlcow> | "Multiple header |
| 23:30 | <karlcow> | fields with the same field name can be combined into one "field-name: |
| 23:30 | <karlcow> | field-value" pair, without changing the semantics of the message, by |
| 23:30 | <karlcow> | appending each subsequent field value to the combined field value in |
| 23:30 | <karlcow> | order, separated by a comma." |
| 23:30 | <GPHemsley> | "The order in which header fields with |
| 23:30 | <GPHemsley> | the same field name are received is therefore significant to the |
| 23:30 | <GPHemsley> | interpretation of the combined field value" |
| 23:31 | <karlcow> | yup |
| 23:31 | <GPHemsley> | but it doesn't say what that significance is |
| 23:31 | <karlcow> | but it doesn't say… |
| 23:31 | <karlcow> | voila! |
| 23:31 | <karlcow> | :) |
| 23:31 | <GPHemsley> | so that doesn't help me :P |
| 23:31 | <karlcow> | I guess it might be on case by case. |
| 23:31 | <karlcow> | If a value make sense or not too. |
| 23:32 | <karlcow> | makes |
| 23:33 | <karlcow> | GPHemsley: I would write down your specific use case and send it to ietf-http-wg⊙wo |