| 00:04 | <Philip`> | Is there a way to clear globalStorage in Firefox? |
| 00:04 | <Philip`> | (via the UI) |
| 00:05 | <annevk2> | btw, I can probably make *.philip.html5.org point back to the same server as philip.html5.org if you want |
| 00:05 | <annevk2> | why are you testing something that's removed from the specification? |
| 00:05 | <Hixie> | annevk2: that won't work |
| 00:05 | <Hixie> | annevk2: because philip.html5.org, if it's dreamhost, is a virtual server |
| 00:06 | <annevk2> | hmm yeah, I suppose it would be a bit unclear where the requests go |
| 00:06 | <Philip`> | It's apparently possible to set ServerAlias *.example.com in Apache, but I don't know if Dreamhost supports that |
| 00:07 | <Hixie> | you might be able to ask them for it |
| 00:07 | <annevk2> | Hixie, though actually, it seems all my domains have a unique value for the A record, so maybe it's possible |
| 00:07 | <Hixie> | annevk2: it's not a dns problem, it's an apache problem |
| 00:08 | <Philip`> | annevk2: Because it's kind of close to localStorage, which makes it easier to attempt to write tests for localStorage and to discover likely bugs |
| 00:11 | <gavin> | Philip`: at the bottom of Options->Advanced->Network, perhaps? |
| 00:12 | <gavin> | Philip`: there's also "Offline Website data" in Tools->Clear Private Data |
| 00:12 | <Philip`> | gavin: That Network -> Offline Storage thing seems different, since the list is empty and 'clear now' doesn't clear globalStorage |
| 00:13 | <Philip`> | gavin: Clear Private Data doesn't seem to clear it, either |
| 00:13 | <annevk2> | oops |
| 00:13 | <gavin> | uh, that sounds like a bug |
| 00:13 | <gavin> | I'm not really familiar with the various *storages |
| 00:13 | <gavin> | file a bug, cc dcamp? |
| 00:18 | <annevk2> | does anyone have a bot hitting html5.org a lot? |
| 00:19 | <Hixie> | not to my knowledge |
| 00:19 | <Philip`> | Oh, looks like clearing cookies works, as long as you restart the browser |
| 00:19 | <Hixie> | which pages? |
| 00:19 | <Philip`> | annevk2: Google might have one |
| 00:19 | Philip` | keeps getting a x\uFFFDy key which he seems unable to delete |
| 00:19 | <annevk2> | it's not really a problem, but there is this thing that hits its 170000 times a month and probably more and identifies itself as "Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)" |
| 00:20 | <annevk2> | and one that does it 20000+ and says "Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0) Fetch API Request" |
| 00:20 | <Philip`> | From the same IP? |
| 00:20 | <Hixie> | neither of those are google |
| 00:20 | <annevk2> | no |
| 00:20 | <Hixie> | unless we have a rogue server |
| 00:20 | <Hixie> | (which is possible) |
| 00:20 | <Hixie> | which IPs? |
| 00:20 | <annevk2> | Google respects robots.txt no? |
| 00:21 | <Hixie> | unless we have a bug |
| 00:21 | <Hixie> | (which is possible) |
| 00:21 | <annevk2> | it's hitting /tools/web-apps-tracker which is not allowed to be indexed |
| 00:21 | <Hixie> | which IPs? |
| 00:21 | <annevk2> | nr1 IP is 203.162.3.165 but that's only 6000 |
| 00:21 | <Hixie> | maybe that's why i was feeling the pain from html5.org :-) |
| 00:21 | <annevk2> | nr 2 is 203.160.1.70 with 4000 |
| 00:21 | <Hixie> | 203.162.3.165 is in vietnam, of all places |
| 00:22 | <Hixie> | 203.162.3.165 appears to be a spam harvester |
| 00:23 | <Hixie> | 203.160.1.70 looks like it's related |
| 00:24 | <annevk2> | maybe a simple solution for now would be to just ban IE 5.5 :) |
| 00:29 | <Hixie> | so does any browser support <form accept-charset> ? |
| 00:30 | <Philip`> | I thought they all did |
| 00:31 | <annevk2> | making some UTF-8 page, declare windows-1252 in that attribute, and see what happens? |
| 00:31 | <annevk2> | s/making/make/ |
| 00:31 | <Philip`> | since I'm fairly sure I've had to use it to make forms always submit UTF-8 |
| 00:31 | <Hixie> | i can't get forms to submit as anything but utf-8 |
| 00:32 | <annevk2> | http://twitter.com/waka/statuses/953993561 |
| 00:33 | annevk2 | has not tried charsets+forms, but thinks that what Hixie says is weird |
| 00:33 | <Hixie> | fixed |
| 00:34 | <Philip`> | http://philip.html5.org/misc/form.html |
| 00:35 | <Philip`> | First goes to http://philip.html5.org/misc/form.html?foo=t%E9st, second goes to http://philip.html5.org/misc/form.html?foo=t%C3%A9st |
| 00:35 | <Philip`> | in Opera 9.5 and Firefox 3 |
| 00:35 | <annevk2> | confirmed |
| 00:35 | <Philip`> | ...but not IE8 |
| 00:35 | <Hixie> | add <input type=hidden name=_charset_> to those forms? |
| 00:36 | <Philip`> | Done |
| 00:37 | <Hixie> | they all send back C3 A9 when i paste it into the live dom viewer |
| 00:38 | <annevk2> | Opera is consistent, IE6 does iso always for Philip`'s example |
| 00:38 | <Philip`> | What if you do accept-charset=iso-8859-1 explicitly? |
| 00:38 | <Hixie> | same |
| 00:38 | <Hixie> | always utf-8 |
| 00:38 | <Hixie> | i wonder why |
| 00:39 | <annevk2> | so for a UTF-8 page accept-charset is ignored? |
| 00:39 | <Hixie> | could be |
| 00:39 | <annevk2> | it makes some sense |
| 00:39 | <Philip`> | Hmm, IE8 is a bit buggy - now that I have three identical forms on that page, it puts inconsistent spacing between them :-/ |
| 00:39 | <Hixie> | Philip`: can you make a copy of that page for utf-8? |
| 00:40 | <annevk2> | and utf-16 pretty please? |
| 00:40 | <Hixie> | also if you could cook me dinner that'd be awesome |
| 00:41 | <Philip`> | http://philip.html5.org/demos/charset/form/ |
| 00:42 | <annevk2> | the utf-16 page is utf-8 |
| 00:43 | <Philip`> | That's because I haven't finished writing it :-) |
| 00:43 | <Philip`> | Ahh, help! |
| 00:43 | <Philip`> | I pressed alt+right in IE8, and now the menu bar and toolbar and everything has slid off the top of the screen |
| 00:43 | <Philip`> | and I can't work out how to get it back |
| 00:44 | <annevk2> | the UTF-8 one works identical to the ISO encoded one in Opera/Firefox |
| 00:44 | <annevk2> | that is, accept-charset is acknowledged |
| 00:44 | <gsnedders> | BenMillard: If you don't have it, get my phone number from http://stuff.gsnedders.com/http-parsing.html |
| 00:44 | <gsnedders> | BenMillard: probably the only safe way to contact me until TPAC now |
| 00:46 | <Philip`> | annevk2: (UTF-16 is there now) |
| 00:47 | <Hixie> | i don't understand what IE is doing |
| 00:47 | <Hixie> | IE8b2 |
| 00:47 | <Hixie> | it doesn't work at all half the time |
| 00:47 | <Philip`> | Hixie: I've got your dinner ready here, you'd better come now or it'll get cold |
| 00:48 | <Hixie> | :-) |
| 00:48 | <annevk2> | hmm, in Opera default will say _charset_ UTF-8 |
| 00:48 | <annevk2> | in Firefox it says UTF-16LE but seems to submit UTF-8 |
| 00:53 | Philip` | adds POST to those pages |
| 00:53 | <Hixie> | oh i know what's confusing me |
| 00:53 | <Hixie> | <meta charset> is ignored if you're document.write()ing |
| 00:53 | <Hixie> | so all these docs are utf-8 when i put them in the dom viewer |
| 01:10 | <Hixie> | hmm |
| 01:10 | Hixie | learns of a security vulnerability caused by flash allowing Content-Type to be set cross-domain |
| 01:11 | <olliej> | yay flash! |
| 01:50 | <weinig> | Hixie: really?! |
| 01:52 | <sicking> | Hixie, "set cross-domain"? |
| 01:53 | <Hixie> | as in, for a cross-domain request |
| 01:56 | <sicking> | Hixie, ah |
| 01:56 | <sicking> | Hixie, i thought they only allowed that when there was a crossdomain.xml file there? |
| 01:57 | <Hixie> | apparently not, though even if that was that was the case, given how easy it is to trick a server into serving a crossdomain.xml file... |
| 02:35 | <Hixie> | WF2 integration status: 100% [#############] Please Stand By... |
| 02:35 | <Hixie> | WF2 integration status: Complete. |
| 02:36 | <Hixie> | well, mostly. i still need to go through and copy examples over and do a thorough review of the material in wf2 to make sure it's all moved over. |
| 02:37 | <Hixie> | but next week i'll start focusing on the feedback that built up over the last 6 weeks |
| 03:17 | <wakaba> | Hixie: accept-charset DOM attribute -> acceptCharset? |
| 04:21 | <dglazkov> | ping annevk2 |
| 08:07 | <erlehmann_> | sometimes, old people are awesome - "man, the last time I had to throw together some HTML, I was constantly doing things like \begin{center}." |
| 08:26 | <annevk2> | Hixie, not WF2 feedback from the last few years? :) |
| 09:29 | <annevk2> | it's a bit sad that external data providers for <select> and <datalist> are gone |
| 09:36 | <zcorpan> | no one-liner google suggest? :( |
| 09:37 | <annevk2> | indeed |
| 09:38 | <annevk2> | though hopefully now that it's somewhat simpler other browsers will implement the new form controls |
| 09:38 | <annevk2> | can always be added back in in forms v3 |
| 09:45 | <annevk2> | Web Forms 2.0 had a five year anniversary since the first official draft September 5 |
| 13:35 | <hsivonen> | Philip`__: yes, it it typical for French books to have the spine text upside-down when the cover of the book faces up |
| 13:36 | Philip` | wonders what caused the differing conventions |
| 13:36 | <hsivonen> | French book spine text is one of the great mysteries of local customs |
| 13:37 | <Dashiva> | Are their bookshelves RTL? |
| 14:28 | <hsivonen> | http://help.yahoo.com/l/us/yahoo/search/webcrawler/slurp-14.html |
| 14:36 | <Philip`> | Seems nicer than Google's approach of using magic comments to indicate regions of the page that should be processed differently |
| 14:51 | <annevk2> | maybe we should host Anolis on Google App Engine? |
| 14:51 | <Philip`> | That won't work, since the CPU quota per request is tiny |
| 14:56 | <annevk2> | Philip`, where is that stated? |
| 14:57 | annevk2 | only sees a daily limit |
| 14:58 | <Philip`> | annevk2: Not sure, but I've seen people on blogs complaining that it doesn't like them spending more than fractions of a second on a request |
| 14:58 | <annevk2> | pointers? |
| 15:00 | <annevk2> | hmm |
| 15:00 | <annevk2> | "Google App Engine does not support any data structure that is larger than 1MB in size." |
| 15:00 | <annevk2> | and "Not only are your calls limited to returning in 10 seconds but if you actually try to do anything too stressful within those calls, you will quickly hit the short-term high CPU quota limit." |
| 15:00 | <annevk2> | from http://aralbalkan.com/1504 |
| 15:04 | <annevk2> | there's http://www.10gen.com/ which seems to be competing with GAE |
| 15:05 | <Philip`> | What's wrong with where it's currently hoste? |
| 15:05 | <Philip`> | s//d/ |
| 15:05 | <annevk2> | my version is too slow :) |
| 15:05 | <annevk2> | but yeah, prolly not much |
| 15:06 | <Philip`> | Parallelised scalable hosting things won't help when the problem is that a single serial Python script is too slow |
| 15:40 | <hsivonen> | how didn't this issue surface during the long ARIA naming debate: http://annevankesteren.nl/2008/10/ie8-bad-update#comment-6641 |
| 15:44 | <annevk2> | in retrospect ariavaluenow etc. would've been better |
| 15:45 | <annevk2> | also more consistent with on* |
| 15:50 | <takkaria> | you wonder why IE can't special-case aria attributes, really, in pre-IE8 mode |
| 15:50 | <hsivonen> | I also wonder why people from MS were silent on this |
| 15:50 | <annevk2> | that should be doable too, yes |
| 15:51 | <annevk2> | hsivonen, yeah, see my comment further down, no idea :/ |
| 15:51 | <hsivonen> | and if non-MS people would have caught this had ARIA been specced in a more open way |
| 15:51 | <annevk2> | they're still silent on it |
| 15:55 | Philip` | notes that IE8 returns undefined, not null, when getting nonexistent keys |
| 16:04 | <takkaria> | I wonder if the IE team ever tries to run other people's testsuites |
| 16:05 | <Philip`> | I'm not aware that any other people have made testsuites for all this stuff |
| 16:25 | <takkaria> | I thought firefox had tests for localStorage, I may be wrong though |
| 17:12 | <annevk2> | Firefox does not support it... |
| 17:14 | <annevk2> | oh wait, maybe it did land like a week ago |
| 17:14 | <annevk2> | bug is not closed though |
| 17:52 | <Philip`> | https://bugzilla.mozilla.org/show_bug.cgi?id=422526 |
| 17:52 | <Philip`> | Doesn't seem to be checked in yet |
| 17:52 | <Philip`> | unless I'm just looking in the wrong place |
| 19:02 | Philip` | sees http://www.legendscrolls.co.uk/webstandards/html5 |
| 19:25 | <erlehmann> | http://www.secondose.com/wp-content/uploads/2008/10/lovestory.jpg |
| 21:06 | <gpy> | im hungry |
| 21:06 | <gpy> | how is that supposed to be solved.. |
| 22:46 | <tiglionabbit> | oh I just thought of something. Of the new form fields that are included for the sake of client-side validation, we have url, but we don't have phone number. Quite often I see the annoying pattern of 3 text fields with javascript jumps between them. It would be great to see a standard phone number field that eliminated this bad practice |
| 22:48 | <Philip`> | tiglionabbit: How would you validate a phone number? |
| 22:49 | <tiglionabbit> | start with the country code, or intuit it from the user's locale if you can't, then check if it follows the correct pattern |
| 22:49 | <Philip`> | Could the pattern be anything more complex than /\d+/? |
| 22:49 | <tiglionabbit> | what would be really nice would be to see a form field that behaves like a cell phone does when you start entering a number |
| 22:50 | <tiglionabbit> | formatting it with hyphons and parenthesis as you type |
| 22:50 | <tiglionabbit> | that are not actually part of the text |
| 22:51 | <Philip`> | I'd assume that's a huge amount of work to cover every locale, and would be out of date quite quickly |
| 22:51 | <tiglionabbit> | do cell phones have this problem? |
| 22:52 | <Philip`> | They're phones, so they're designed to dial numbers, so I guess they have the motivation and resources to do all that work :-) |
| 22:55 | <tiglionabbit> | hm. I know you have to type the hyphons in Japan, but.. what other countries don't follow the 3-3-4 pattern |
| 22:57 | <Philip`> | What if I want to enter a phone number like 999? |
| 22:59 | Philip` | (in the UK) usually formats phone numbers like 01234 567890 |
| 23:01 | <tiglionabbit> | true, there are local numbers like 411 |
| 23:02 | <tiglionabbit> | what I mean to say though is, regardless of how complicated it is, it has been done -- cel phones are a reference implementation |
| 23:03 | <tiglionabbit> | and it can just fall back to a text field otherwise |
