| 00:10 | <zcorpan> | http://simon.html5.org/sandbox/html/google-chrome-comic |
| 00:16 | <zcorpan> | hmm doesn't work so well with history |
| 00:16 | <Lachy> | zcorpan, it's missing the CC-by-nc-nd notice |
| 00:17 | <Lachy> | I also put up a copy of the comic yesterday, though I didn't get creative enough to make easily browsable yet. http://lachy.id.au/dev/2008/google-chrome/ |
| 00:19 | <zcorpan> | Lachy: copied your notice |
| 00:19 | <Lachy> | I updated http://en.wikipedia.org/wiki/About_URI_scheme#Google_Chrome-specific_about:_addresses with the about:crash thing |
| 00:19 | <Lachy> | are there any others? |
| 00:19 | <Lachy> | about: uri schemes? |
| 00:20 | <zcorpan> | about:memory |
| 00:20 | <zcorpan> | about: |
| 00:20 | <Hixie> | there's about:hang iirc |
| 00:20 | <Hixie> | something like that |
| 00:20 | <Hixie> | to test the hanging protection |
| 00:21 | <Lachy> | about:hang didn't work |
| 00:21 | <Hixie> | there's one something like that |
| 00:21 | <Hixie> | look in the source :-) |
| 00:21 | <Lachy> | about: isn't google chrome specific |
| 00:22 | <Lachy> | I don't know where exactly to look, but may I could search for it |
| 00:22 | <Hixie> | about:memory, about:crash and the about:hang one whatever it's called are |
| 00:23 | <zcorpan> | i'd like an about:config |
| 00:25 | <Hixie> | i doubt there's much to configure in chrome |
| 00:25 | <Hixie> | they keep talking about how they have few prefs |
| 00:26 | <Lachy> | http://ejohn.org/blog/google-chrome-process-manager/#comment-320135 |
| 00:29 | <Hixie> | aha, about:shorthang |
| 00:31 | <Lachy> | didn't work for me |
| 00:31 | <Lachy> | what is supposed to do exactly? |
| 00:32 | <Hixie> | hang |
| 00:32 | <Hixie> | the tab |
| 00:32 | <Lachy> | oh, it does work |
| 00:42 | <Lachy> | about:internets LOL! |
| 00:43 | <takkaria> | chrome does badly on meebo.com, keeps on getting laggy as hell |
| 00:45 | <Hixie> | about:internets is pretty funny |
| 00:45 | <Hixie> | didn't know about that one |
| 00:45 | <Hixie> | pretty good demo of the chrome technology though |
| 00:45 | <Hixie> | shows how it works much like the screen saver control panel |
| 00:46 | <Hixie> | doesn't show up in the task manager though |
| 00:53 | <Hixie> | uh |
| 00:53 | <Hixie> | don't believe the issues graph |
| 00:54 | <Hixie> | as sweet as it would be to have dropped to just 97 e-mails... |
| 00:55 | <takkaria> | whatwg.org seems to not be responding so well |
| 00:56 | <Hixie> | yeah, it's swamped by requests for acid3 |
| 00:56 | <takkaria> | ah, I see |
| 00:56 | <takkaria> | what's the new lines on the issues graph represent? |
| 00:56 | <Lachy> | JohnResig, yt? |
| 00:57 | <Hixie> | takkaria: the script that collects the data crashed due to the load (i assume) and didn't count all the e-mails |
| 00:58 | <Lachy> | comment spammers are getting sneaky. I just noticed this comment on JohnResig's blog, and I have one saying exactly the same thing with the same URI, but with a different name http://ejohn.org/blog/google-chrome-process-manager/#comment-320147 |
| 00:58 | <Lachy> | it's sitting in my blog moderation queue. I almost approved it till I saw that |
| 00:58 | <JohnResig> | Lachy: yeah, it's real weird |
| 00:58 | <Lachy> | JohnResig, I recommend you delete it from yours |
| 01:03 | <Lachy> | I wonder if the about:internets thing also shows random teapots, like the real screensaver |
| 01:03 | Hixie | gets as far as "kill your babies" and decides that he's not going to read the rest of the message |
| 01:04 | <Hixie> | i'm getting very tired of dealing with the rudeness in public-html |
| 01:04 | <takkaria> | stevef I find consistently quite rude thesedays |
| 01:06 | <Lachy> | I don't even get what "kill your babies" is supposed to mean?! |
| 01:07 | <Hixie> | i dunno but if somebody said that on the whatwg list they'd get banned pretty quick. |
| 01:07 | <Hixie> | (for a week) |
| 01:10 | <takkaria> | stevef said something about jgraham "squealing like a pig" recently, which is the most outright rude I've seen anyone be for a while |
| 01:11 | <Philip`> | "stuck pig", I think |
| 01:11 | <Hixie> | wtf |
| 01:12 | <takkaria> | http://lists.w3.org/Archives/Public/www-archive/2008Aug/0154.html |
| 01:14 | <Hixie> | i love how they just want the spec to say something and have absolutely no care in the world for actually checking whether it makes sense |
| 01:15 | <Dashiva> | You mean sense isn't defined by what they want? |
| 01:15 | <Philip`> | (The meaning of "stuck" in that phrase seems quite archaic - the only place I vaguely remember the word 'stick' being used in that way was probably written by Tolkien) |
| 01:16 | <Dashiva> | Philip`: Isn't it the same stick as in 'stick it to them'? |
| 01:16 | <Philip`> | (as in "I'll stick you" and so on) |
| 01:16 | <Philip`> | Dashiva: I don't think so, in either of the ways that phrase could be interpreted |
| 01:17 | <Philip`> | (I think that phrase could be like "stick it to the police" or like "stick it to the wall") |
| 01:22 | <Philip`> | Hmm, I think I remembered mostly rightly |
| 01:22 | <Philip`> | "'Who says there's bad news?' shouted the soldier. / 'Ar! Who says there isn't?' / 'That's cursed rebel-talk, and I'll stick you, if you don't shut it down, see?'" |
| 01:37 | <takkaria> | Chrome is indeed pretty usable. firefox should adopt showing history in the awesomebar results, it's more useful than I thought it would be |
| 01:37 | <Philip`> | Will it still be useful when you have more than a few hours of history, and it's zillions of entries long? |
| 01:38 | <takkaria> | depends how well it scales |
| 01:38 | <takkaria> | er, fierfox already does it, actually, doesn't it? |
| 01:38 | <gavin_> | takkaria: hmm? firefox does show history in the awesomebar |
| 01:39 | <takkaria> | I appear to be being clueless today, please ignore me :) |
| 01:45 | <Lachy> | takkaria, Firefox does show navigation history in its awesome bar |
| 01:46 | <Lachy> | I should have read gavin_'s message first. Please ignore me too ;-) |
| 01:46 | <Lachy> | woah, wikipedia is claiming Google Chrome fails acid2 because the page reflows when resizing the window http://en.wikipedia.org/wiki/Google_Chrome#User_interface |
| 01:47 | <Lachy> | I'll delete it cause reflowing isn't a bug |
| 01:47 | <Hixie> | lol |
| 01:47 | <Hixie> | "as a result it fails to fully adhere with existing webstandards" |
| 01:47 | <Hixie> | that's pretty awesome in so many ways |
| 01:48 | <Hixie> | whoever wrote that should check the other browsers too :-) |
| 01:52 | <Hixie> | cool, pumping up the CPU allocation on hixie.dreamhost.com made it more responsive |
| 01:53 | <Lachy> | I edited it |
| 02:01 | Hixie | begins the long and arduous task of merging the forms stuff into html5 |
| 02:03 | <Lachy> | Hixie, Google Chrome is randomly failing tests 26 and 27 in acid3. Is there any reason it shouldn't be consistent with those? |
| 02:03 | <Lachy> | Test 26 failed: e2 - parent element doesn't exist after looping |
| 02:03 | <Lachy> | Test 27 failed: e2 - parent element doesn't exist after waiting |
| 02:04 | Hixie | looks at the test |
| 02:05 | <Hixie> | oh that's the GC test |
| 02:05 | <Hixie> | GC is inherently unpredictable |
| 02:07 | <Lachy> | does that mean that it's getting incorrectly garbage collected? |
| 02:08 | <Hixie> | i believe so |
| 02:08 | <Hixie> | i'd have to make a test case to be sure |
| 02:08 | <Hixie> | it's probably either a v8 bug or something that'll be fixed when they sync to webkit tot |
| 02:08 | <Lachy> | ok |
| 02:09 | <Lachy> | so is it failing in test 26, and then that causes 27 to fail automatically? |
| 02:09 | <Hixie> | probably |
| 02:09 | <Hixie> | i haven't debugged it to be sure |
| 02:09 | <Lachy> | that's what it looks like, since it's the same error and appears to be the same variable it's checking |
| 02:16 | Hixie | wonders why his setup doesn't have libgtk-x11 and wonders how to get it so he can upgrade from this mess of firefox2 to firefox3 which at some point suddenly started requiring libgtk-x11 |
| 02:18 | <Lachy> | LOL, the Incognito mode warnings are funny: |
| 02:18 | <Lachy> | Test 26 failed: e2 - parent element doesn't exist after looping |
| 02:18 | <Lachy> | Test 27 failed: e2 - parent element doesn't exist after waiting |
| 02:18 | <Lachy> | oops, didn't copy it... |
| 02:19 | <Lachy> | Be wary of: ... Malicious software that tracks your keystrokes in exchange for free smileys |
| 02:19 | <Lachy> | Surveillance by secret agents |
| 02:19 | <Lachy> | People standing behind you |
| 02:19 | <Hixie> | i wonder why we added <form accept> |
| 02:20 | Hixie | drops <form accept> for now |
| 02:21 | <Hixie> | actually i think i'm going to first spec only the basics and omit most of the new stuff (other than the new <input type>s and validation) |
| 02:22 | <Hixie> | i guess <form name> is useful with document.forms |
| 02:22 | <Lachy> | Hixie, probably because when you specced it, WF2 was a delta spec and <form accept> is in HTML4, and you just defined how it was supposed to be handled |
| 02:22 | <Hixie> | oh. |
| 02:23 | <Hixie> | so it is |
| 02:23 | <Hixie> | well |
| 02:23 | <Hixie> | there's evidence that i'm not biased against html4 and like my own stuff! |
| 02:23 | <Hixie> | i thought it was retarded even when i thought i'd invented it! |
| 02:24 | <Lachy> | you were right, it is retarded :-) |
| 02:25 | <Lachy> | does any browser support <input accept="">? |
| 02:25 | <Lachy> | or are we still hoping that some browser might implement it? |
| 02:25 | <Hixie> | i've just dropped it |
| 02:25 | <Hixie> | or "not added" it, to be precise |
| 02:26 | <Lachy> | unless Opera added it as part of its WF2 stuff |
| 02:26 | <Lachy> | I don't have a build of Opera handy, so I'll have to check tomorrow if I remember |
| 02:27 | <Hixie> | does it even implement <input type=file accept>? |
| 02:30 | <Lachy> | I don't know, that's what I was wondering |
| 02:30 | <Hixie> | i'm not going to do the form and select seeding at the moment |
| 02:31 | <Hixie> | (they were only added for xforms parity, and don't seem to have really caught people's attention) |
| 02:36 | <Hixie> | ok, i've done the green summary box for <form> http://www.whatwg.org/specs/web-apps/current-work/#the-form |
| 02:38 | <Lachy> | Hixie, shouldn't it be in the Flow content category? |
| 02:39 | <Hixie> | uh yes |
| 02:39 | <Hixie> | that was copied from the <Div> element |
| 02:39 | <Hixie> | which is apparently also broken |
| 02:39 | <Hixie> | regenning |
| 02:40 | <Lachy> | why is onsubmit="" included as a global attribute, rather than a form specific attribute? |
| 02:41 | <Hixie> | all event attributes are global in html5 |
| 02:41 | <Lachy> | what about onreset? That's allowed on HTML4 <form>, but not listed in HTML5 |
| 02:41 | <Hixie> | simplifies implementations, makes it easier to catch bubbling events wherever convenient |
| 02:41 | <Hixie> | the list in html5 is woefully incomplete |
| 02:41 | <Lachy> | ok |
| 02:41 | <Hixie> | i haven't really tried making it complete yet |
| 02:42 | <Hixie> | e.g. all the media events are missing |
| 02:54 | <Lachy> | wow, I can't believe Karl's suggestion to use standards mode without a DOCTYPE?! |
| 04:20 | <kingryan> | hsivonen: you around? |
| 05:27 | <kingryan> | Hixie: are you around? |
| 05:27 | <Hixie> | yo |
| 05:28 | <kingryan> | i have parsing question |
| 05:28 | <Hixie> | shoot |
| 05:28 | <kingryan> | with content model PCDATA and input "</br/>" |
| 05:28 | <kingryan> | it seems that a parse error should be triggered |
| 05:28 | <kingryan> | and we have a tests that says so, but I can't figure out how |
| 05:29 | <kingryan> | instead i just get an end-tag |
| 05:31 | <Hixie> | what parse mode? |
| 05:31 | <Hixie> | or is "</br/> all the input? |
| 05:31 | <kingryan> | fragment |
| 05:31 | <kingryan> | with that as the entire input |
| 05:31 | <Hixie> | with what element as context? |
| 05:31 | <kingryan> | div |
| 05:32 | <Hixie> | ok let's see |
| 05:32 | <Hixie> | gah |
| 05:33 | <Hixie> | i can't handle chrome's lack of smooth scrolling to read the html5 spec |
| 05:33 | Hixie | goes back to safari |
| 05:33 | <Hixie> | ok so insertion mode becomes "in body" |
| 05:34 | <Hixie> | you get a parse error in response to the "br" end tag |
| 05:34 | <Hixie> | you then pretend you got a "br" start tag |
| 05:34 | <Hixie> | that creates a <br> |
| 05:35 | <Hixie> | you then get another parse error because the self-closing flag wasn't acknowledged |
| 05:35 | <kingryan> | hmm, i'm only running the tokenizer right now |
| 05:35 | <kingryan> | maybe that's the problem |
| 05:35 | <Hixie> | then you stop parsing |
| 05:35 | <Hixie> | oh |
| 05:35 | <Hixie> | well |
| 05:35 | <Hixie> | with just the tokeniser you get no parse errors |
| 05:35 | <kingryan> | that's what I thought |
| 05:36 | <kingryan> | i just wanted to be sure, before I blamed excors for checking in a bad test |
| 05:38 | <Hixie> | i wouldn't recommend implementing or testing the tokenizer as a separate object from the tree construction stage, to be honest; there are a number of optimisations that involve blurring the line |
| 05:38 | <Hixie> | the distinction is artificial |
| 05:38 | <kingryan> | yeah, that's probably a good idea |
| 05:38 | <kingryan> | but as it is, the tests in html5lib are separate |
| 05:39 | <Hixie> | the tests in html5lib that test just the tokenizer are specific to html5lib's tokenizer, then :-) |
| 05:39 | <kingryan> | i suppose |
| 05:39 | <kingryan> | its nice to be able to test the lower levels of the stack (tokenizer) independently |
| 05:39 | Hixie | wonders whether to spec the summary blocks for the other form elements first, or to go into detail for <form> first |
| 05:39 | <Hixie> | i guess |
| 05:39 | <Hixie> | seems dodgy to me though |
| 05:40 | <kingryan> | since a lot of errors at higher levels get disguised |
| 05:40 | <Hixie> | i mean the two bits are tightly related |
| 05:40 | <kingryan> | i agree |
| 05:40 | <kingryan> | full separation is a pipe-dream |
| 05:40 | <kingryan> | but at least basic tokenization can be tested independently |
| 07:01 | <Hixie> | http://news.cnet.com/8301-1023_3-10028842-93.html?part=rss&subj=news&tag=2547-1023_3-0-5 |
| 07:01 | <Hixie> | we get our logo on a rocket and that's the best we can do? |
| 07:01 | <Hixie> | looks like a smarties tube! |
| 07:24 | <hsivonen> | Hixie: Julian does have a good point about void elements |
| 07:25 | <nessy> | Hixie: are you involved with Chrome? |
| 07:26 | <Hixie> | nessy: no more so than any other browser |
| 07:26 | <nessy> | cool :) |
| 07:26 | <nessy> | hope you haven't been asked a million times already :) |
| 07:27 | <Hixie> | hsivonen: sure |
| 07:36 | <Hixie> | hsivonen: or at least, he has a point |
| 07:36 | <Hixie> | hsivonen: i'm not sure it's "good" |
| 07:46 | <virtuelv> | acidtests.org down? |
| 07:46 | <virtuelv> | err, no, just horribly slow |
| 07:47 | <Hixie> | a browser came out toady |
| 07:47 | <Hixie> | today |
| 07:47 | <virtuelv> | yeah |
| 07:47 | <virtuelv> | shame I couldn't get it to work in WINE |
| 07:49 | <virtuelv> | especially since I don't have compatible virtualbox modules |
| 08:07 | <Hixie> | so do we want to keep the form="" attribute? |
| 08:07 | <Hixie> | and if so, do we want to keep it on <fieldset>? |
| 08:07 | <Hixie> | seems weird for fieldset to be associated with a form |
| 08:07 | <Hixie> | what does that mean? |
| 08:15 | <zcorpan> | Hixie: yes we do :) |
| 08:15 | <zcorpan> | Hixie: but maybe not on fieldset, dunno |
| 08:26 | <Hixie> | i guess maybe it makes sense if you want to be able to hook into the onformchange or onforminput events |
| 08:34 | <hsivonen> | Hixie: his point is good in the sense that it does suck that something as trivial as void elements requires updates as HTML adds stuff |
| 08:34 | <Hixie> | "waah" |
| 08:34 | <Hixie> | it's a one line fix per new void element |
| 08:34 | <Hixie> | and they come at the rate of what, one a year on average? |
| 08:35 | <Hixie> | give me a break |
| 08:35 | <hsivonen> | yes and yes |
| 08:36 | <hsivonen> | we could try to get e.g. Genshi to update and see how easy it is to do upgrade evangelism with a library that isn't bound by the XSLT spec |
| 08:36 | <Hixie> | i'm tired of people whining about their minor inconveniences when we have people writing entirely new browser features (video, canvas, etc), validators, god knows what |
| 08:37 | <hsivonen> | I'm annoyed that the JDK keeps ballooning up with frozen code instead of the JDK being very lean and people updating their 3rd-party libs early and often |
| 08:40 | <Hixie> | i wonder why so few sites use fieldset |
| 08:41 | <Hixie> | how about multiple forms per control |
| 08:41 | <Hixie> | can we throw that overboard? |
| 08:42 | <hsivonen> | multiple forms per control? |
| 08:43 | <Hixie> | <input form="a b"> |
| 08:43 | <Hixie> | <form id=a> |
| 08:43 | <Hixie> | <form id=b> |
| 08:43 | <Hixie> | wf2 has that supposedly |
| 08:45 | annevk | hopes the html5lib tokenizer tests don't impose html5lib specifics to other implementations |
| 08:46 | <jmb> | annevk: I don't think they do |
| 08:47 | <annevk> | well, I'm specifically concerned about implementations trying to match our architecture, but that's prolly far fetched |
| 08:48 | <hsivonen> | annevk: I'll complain if you start testing something implmentation-specific |
| 08:48 | <jmb> | as will takkaria/I :) |
| 08:49 | <hsivonen> | jmb: are you working on Hubbub, too? |
| 08:49 | <jmb> | hsivonen: I started it :) |
| 08:49 | <hsivonen> | oh. nice |
| 08:51 | <annevk> | good :) |
| 08:54 | <wilhelm> | Never had any use for multiple forms per control. |
| 08:54 | <wilhelm> | But remember to put back <input size>. (c; |
| 08:58 | <Hixie> | doing <fieldset> at the moment |
| 08:59 | <Hixie> | i'm just doing the element summaries first |
| 09:01 | <Hixie> | hsivonen: html5.validator.nu giving me io errors |
| 09:01 | <Hixie> | "http://www.whatwg.org/specs/web-apps/current-work/source-whatwg";: non-document-error io: The host did not accept the connection within timeout of 5000 ms |
| 09:01 | <Hixie> | oh |
| 09:01 | <Hixie> | wait |
| 09:01 | <Hixie> | that's probably my site being slow |
| 09:01 | <Hixie> | not yours |
| 09:01 | <Hixie> | nevermind |
| 09:04 | hsivonen | wonders what data Chrome sends to Google |
| 09:05 | <hsivonen> | apparently merely typing an URL leaks the URL to Google as a search term |
| 09:06 | <annevk> | quite a bit, from what I read yesterday on the privacy page |
| 09:07 | <annevk> | http://www.google.com/chrome/intl/en/privacy.html has the info on that |
| 09:07 | <Hixie> | get a wire sniffer and try it :-) |
| 09:10 | <Hixie> | the next fucking acid test is going to be 512 bytes long and a single file |
| 09:10 | <Hixie> | my server can't handle this load |
| 09:11 | <Hixie> | though on the plus side, with the limits i got dreamhost to set on the apache server, at least the machine itself hasn't gone down at all |
| 09:12 | <jruderman> | you had to help dreamhost make their server not crash due to your site? |
| 09:12 | <Hixie> | dreamhost's default configuration overtaxes the server under heavy load |
| 09:13 | <Hixie> | i told them how to fix it but i got the impression they didn't deploy the fix |
| 09:13 | <Hixie> | but i did get them to fix it in a hardcoded way just for my server |
| 09:13 | <Hixie> | basically my server is limited to 75 connections at a time |
| 09:13 | <Hixie> | so even under max load, the machine won't even blink and i can still check my e-mail and edit the site, etc |
| 09:13 | <Lachy> | I'm looking for sites that break when forced to render in standards mode instead of quirks mode, but it seems every site I visit frequently is already in standards mode. |
| 09:14 | <Lachy> | ... why can't we go back to the way it was where no-one but the early adopters used standards mode? It'd make this easier |
| 09:16 | <Hixie> | just go down the list of sites on reddit |
| 09:16 | <annevk> | Hixie, HTMLFormElement > HTMLFieldSetElement under fieldset |
| 09:16 | <Hixie> | thx |
| 09:17 | <hsivonen> | I should update my doctype page to cover Opera 9.5 and Chrome |
| 09:17 | <hsivonen> | did Konqueror go from 3.3 to 3.5 or was there a 3.4? |
| 09:18 | <hsivonen> | apparently there was |
| 09:18 | <hsivonen> | though it probably doesn't matter, since distro updates force Konq updates onto KDE users |
| 09:23 | jgraham | would just replace acidtests.org with a page that says "Yes, Gogle Chrome scores 75-77. Don't you have anything better to do?" ;) |
| 09:24 | <GregHouston> | Hixie: Dean Edwards ran into a similar issue with his kitchen server, and ended up serving a lot of his content from Google Code. http://dean.edwards.name/weblog/2007/03/google-it/ |
| 09:24 | <hsivonen> | ooh. Google Maps finally has walking directions |
| 09:24 | hsivonen | wants those for GMM |
| 09:26 | hsivonen | wants a way to move locations between Google Maps instances... |
| 09:27 | <hsivonen> | hmm. walking directions beta sucks for Mandelieu |
| 09:28 | <Hixie> | i believe walking directions suck for a lot of places |
| 09:28 | <Hixie> | though at least it's not as bad as it was before we released it |
| 09:28 | <Hixie> | where it would make you walk down highways |
| 09:29 | <annevk> | zcorpan, "I think rowspan and colspan can work the same in both (as specced in HTML5)." hmm, HTML5 doesn't do layout |
| 09:29 | <annevk> | cycling directions would've been more useful for me, but maybe they overlap enough |
| 09:29 | <Hixie> | GregHouston: yeah that wouldn't work for acid tests |
| 09:30 | <Hixie> | GregHouston: i need far more careful control over http headers and stuff |
| 09:30 | <Lachy> | http://farm4.static.flickr.com/3121/2821783957_b4633f072e_o.png |
| 09:33 | <annevk> | Hixie, FieldSet (uppercase S) |
| 09:34 | <Hixie> | thx |
| 09:36 | <KevinMarks> | I want cycling+public transport dirctions for bay area |
| 09:37 | <KevinMarks> | and tube+walk for london |
| 09:38 | <Hixie> | yeah just public transport with the minimum walking speed increased a bit and the maximum walking distance increased a bit would be nice |
| 09:40 | <hsivonen> | we already have a non-Ajaxy service for public transport + walking in Helsinki area, but it doesn't integrate with GPS |
| 09:41 | <hsivonen> | eww. sncf.fr is all Flashy |
| 09:41 | <hendry> | are there any html5 mirrors? :) |
| 09:42 | <annevk> | dev.w3.org/html5/spec/ |
| 09:43 | <annevk> | ouch, that one is ugly nowadays |
| 09:44 | <jruderman> | Hixie: google maps still has you park on highways |
| 09:44 | <jruderman> | http://maps.google.com/maps?f=q&hl=en&geocode=&q=san+jose+to+1981+landings+dr,+94043&sll=37.412974,-122.098392&sspn=0.009732,0.019698&ie=UTF8&t=h&z=12 |
| 09:44 | <Hixie> | well it doesn't do parking |
| 09:44 | <hsivonen> | wow. snfc.fr seems to have totally missed the point that the core use case is searching for schedules and booking tickets |
| 09:45 | <Hixie> | it gets you from an address to an address |
| 09:45 | <jruderman> | the directions to the mozilla office abruptly end at 'merge onto 101N' |
| 09:45 | <hsivonen> | sncf that is |
| 09:45 | <Hixie> | woah |
| 09:45 | <Hixie> | that's a bug |
| 09:45 | <Hixie> | wtf |
| 09:45 | Hixie | files a bug |
| 09:46 | <jruderman> | thanks :) |
| 09:46 | <jruderman> | of course, now i'm going to come to you whenever i find a non-security bug in google software, since everywhere else seems to be a black hole |
| 09:47 | <Hixie> | feel free |
| 09:49 | <Dashiva> | What about that chrome crasher, is that fixed yet? :) |
| 09:49 | <othermaciej> | jruderman: you haven't heard about the gracious new mozilla office hosting in the middle of 101? |
| 09:49 | <othermaciej> | very convenient and very generous of Google I say |
| 09:51 | <jruderman> | Hixie: the walking directions from my house to google are similarly busted -- it doesn't realize i have to cross the freeway to get there |
| 09:51 | <Hixie> | uri? |
| 09:52 | <Hixie> | walking directions in general aren't too hot right now |
| 09:52 | <Hixie> | they're still beta |
| 09:53 | <jruderman> | it also doesn't know the shortcut through the grass and trees just north of rengstorff&101, but i can forgive it for that ;) |
| 09:53 | <jruderman> | http://maps.google.com/maps?f=d&saddr=2434+Rock+St,+Mountain+View,+CA+94043&daddr=1981+Landings+Dr,+Mountain+View,+CA+94043&hl=en&geocode=&mra=ls&dirflg=r&date=09%2F03%2F08&time=1:47am&ttype=dep&noexp=0&noal=0&sort=&sll=37.421959,-122.090984&sspn=0.002433,0.004925&ie=UTF8&t=h&z=16&start=0 |
| 09:54 | <jruderman> | hmm, no 'beta' notice upon loading that url |
| 09:54 | <Dashiva> | "Take the hidden passage behind the bookshelf to the rear garden." |
| 09:54 | <Hixie> | jruderman: that's transit, not walking :-) |
| 09:55 | <Hixie> | but the bug is the same as the car bug |
| 09:55 | <Hixie> | we're geocoding the landings drive address wrongly |
| 09:56 | <Philip`> | Lachy: Try http://www.google.com |
| 09:56 | <Hixie> | oh i see what the problem is |
| 09:56 | <Philip`> | That gets broken layout if it's in standards mode |
| 09:56 | <Hixie> | people have "edited" the location of that address so that it is no longer on landings drive |
| 09:58 | <Hixie> | and the map data doesn't have the car park as a driveable area |
| 10:04 | <jruderman> | don't you have an AI that can look at an aerial photograph and tell you which sections look like parking lots? ;) |
| 10:05 | <Lachy> | Philip`, google.com doesn't noticably break in standards mode |
| 10:05 | <Hixie> | jruderman: google doesn't comment on future products and services |
| 10:05 | <jruderman> | bah :P |
| 10:07 | <Philip`> | Lachy: It does for me, in Opera - the "Advanced Search, Preferences, Language Tools" box gets totally misaligned |
| 10:07 | <jruderman> | Lachy: mozilla gets some bug reports that we resolve as INVALID because the page has a strict doctype, would work in quirks mode, but doesn't work in standards mode. a lot of them involve DOM differences rather than layout differences. |
| 10:07 | <Lachy> | oh, ok. that's just a minor break |
| 10:08 | <jruderman> | when javascript on a page breaks, it's usually a major break |
| 10:09 | <jruderman> | well, if we hear about the breakage, at least |
| 10:10 | <hsivonen> | Hixie: a public beta might be more effective if the UI had a link to a bug tracker |
| 10:13 | <annevk> | like on http://validator.nu/ ? :) |
| 10:13 | <hsivonen> | annevk: good point |
| 10:14 | <jruderman> | “Validation” and “validator” in the name and the user interface of the service refer to the ISO/IEC FDIS 19757-2 definition of “validator” (which performs validation), to the Schematron “validation” function (which is performed by a validator), and to the HTML 5 definition of “validator”. |
| 10:14 | <jruderman> | what the heck is that trying to say? |
| 10:15 | <hsivonen> | jruderman: it's saying "SGML old timers, please don't email me telling what a validator is" |
| 10:15 | <hsivonen> | I guess I could move that to the FAQ section |
| 10:15 | <Philip`> | *F* AQ? |
| 10:16 | <Hixie> | hsivonen: maybe, though in practice we get most of what we need from beta from our logs. it's actually surprisingly easy to find out what the bugs are by looking for trends in how people are using the service |
| 10:16 | <hsivonen> | Philip`: more F than the others in the FAQ, sadly |
| 10:16 | <Philip`> | I thought frequently asked questions would be more along the lines of "my page doesn't validate! what do I do?" |
| 10:17 | <Dashiva> | "Your site doesn't display google properly when I enter google.com" |
| 10:23 | <hsivonen> | the about page getting old in general |
| 10:23 | <hsivonen> | it was written for XML-oriented early adopters with special needs |
| 10:23 | <hsivonen> | not for casual HTML5 authors |
| 10:24 | Philip` | wonders how many people use the fancy XML-oriented features, compared to the HTML5 features |
| 10:25 | <annevk> | at some point html5.validator.nu and validator.nu should be flipped too, imo |
| 10:25 | <jruderman> | hsivonen: when i enter google.com (and get google.fi) it says "Using windows-1252 instead of the declared encoding iso-8859-1." what does that mean? |
| 10:26 | <annevk> | maybe validator.nu and validator.nu/advanced |
| 10:26 | <hsivonen> | jruderman: It means that the bytes were decoded according to Windows-1252 (for reality and HTML5 compliance), but the page declared itself as ISO-8859-1 |
| 10:26 | <hsivonen> | jruderman: you get .fi, because validator.nu is hosted in Finland and Google tries to be smart |
| 10:26 | <jruderman> | hsivonen: what about the page made the validator decide to decode the bytes according to windows-1252? |
| 10:27 | <jruderman> | i guess i should grep the spec for 'iso-8859-1' |
| 10:27 | <hsivonen> | jruderman: the HTML5 spec says that when the page says ISO-8859-1, it must be decoded as Windows-1252 |
| 10:27 | <annevk> | iso-8859-1 is just an alias for windows-1252 |
| 10:27 | <annevk> | in the real world |
| 10:28 | Philip` | wonders if the validator should ever emit messages without telling the author how they can make the message go away |
| 10:28 | <hsivonen> | I'm surprised that Google still doesn't default to UTF-8 for unknown UAs |
| 10:29 | <hsivonen> | Philip`: should I add "Please upgrade to UTF-8" to all encoding errors? |
| 10:29 | <jruderman> | ok, that makes sense. it would be nice if the validator made it clear that was the reason (as opposed to there being something truly wrong with the page, like a content-type header not matching a meta http-equiv content-type or the bytes of the page) |
| 10:30 | <jruderman> | hsivonen: yes, especially the 'this is invalid UTF-8' encoding errors |
| 10:31 | <Philip`> | hsivonen: I guess most authors wouldn't have a clue what that meant or how to do that |
| 10:32 | <hsivonen> | I don't want to make the messages themselves longer, but I could add elaborations similar to the elaborations on RELAX NG errors |
| 10:32 | <hsivonen> | (which are out of date) |
| 10:32 | Philip` | isn't sure what the validator could actually say that would be sufficiently helpful |
| 10:33 | <jruderman> | "Support for UTF-32 is not recommended. This encoding is rarely used, and frequently misimplemented." is it more frequently misimplemented than UTF-16? |
| 10:34 | <hsivonen> | jruderman: I would expect so, but I have no data. |
| 10:34 | <hsivonen> | jruderman: interesting apps use UTF-16 internally, so developers are likely to notice bugs |
| 10:35 | <hsivonen> | jruderman: but UTF-32 is a checklist item that no one actually bothers to test seriously |
| 10:36 | <jruderman> | ahh |
| 10:36 | <jruderman> | most of the problems i've seen with UTF-16 involved surrogate pairs (or surrogates that should have been in pairs but weren't) |
| 10:37 | <jruderman> | although i also saw one awesome bug report where someone managed to set their *default* encoding to UTF-16 and then wondered why many pages appeared as random chinese characters |
| 10:37 | <hsivonen> | I'd expect most UTF-32 problems to involve conversion to surrogate pairs when the in-memory representation is UTF-16 |
| 10:37 | <hsivonen> | jruderman: browsers probably shouldn't have UI for selecting UTF-16 |
| 10:37 | <jruderman> | https://bugzilla.mozilla.org/show_bug.cgi?id=452934 |
| 10:37 | <jruderman> | yeah |
| 10:38 | <jruderman> | i noticed that google chrome had UTF-16 as an option for the default encoding, and i was mildly surprised |
| 10:38 | <hsivonen> | UTF-16 as an interchange format should die |
| 10:39 | Philip` | saw that Chrome seems to be using ICU, rather than relying on native platform APIs |
| 10:40 | <hsivonen> | Philip`: not a bad idea |
| 10:41 | <jruderman> | kk filed https://bugzilla.mozilla.org/show_bug.cgi?id=453423 |
| 10:42 | <jruderman> | i'm sad that people in some countries are still expected to fiddle with character encoding overrides |
| 10:42 | <jruderman> | it's kind of a security hole |
| 10:42 | <Philip`> | http://blogs.msdn.com/ie/archive/2008/09/02/ie8-security-part-vi-beta-2-update.aspx |
| 10:42 | <Philip`> | Implementing various security things per HTML5 |
| 10:43 | <Philip`> | Renamed authoritative=true to X-Content-Type-Options: nosniff |
| 10:43 | <hsivonen> | jruderman: I can understand that authors in Western Europe and the Americas don't care when Windows-1252 is the default, but I'm baffled that many CJK and Cyrillic authors don't care to make stuff right |
| 10:44 | <Hixie> | X-Content-Type-Options: nosniff |
| 10:44 | <Hixie> | lol |
| 10:44 | <Hixie> | what a joke |
| 10:44 | <Hixie> | i can't wait for X-Content-Type-Options: nosniff-seriously |
| 10:45 | <Hixie> | and X-Content-Type-Options: nosniff-pleeeease-please-i-mean-it-this-time-really |
| 10:45 | <jruderman> | hsivonen: exactly |
| 10:45 | <annevk> | I wonder if it only works for text/plain files as they seem to suggest in which case it might actually work |
| 10:45 | <jruderman> | Hixie: it will have to be a near header each time so that browsers that only support 'nosniff' won't revert |
| 10:46 | <Hixie> | oh right |
| 10:46 | Philip` | tests how well nosniff actually works |
| 10:46 | <Hixie> | X-Content-Type-Options-2: nosniff |
| 10:46 | <Hixie> | X-Content-Type-Options-3: nosniff |
| 10:46 | <Hixie> | Philip`: does it work with <img>, <script>, etc? |
| 10:46 | <Hixie> | Philip`: how about navigating to a GIF labelled as PNG? |
| 10:46 | <Hixie> | Philip`: or a text/plain video? |
| 10:47 | <jruderman> | X-Content-Type-Options-no-seriously: nosniff |
| 10:51 | <Philip`> | Hixie: It doesn't work on <img> - the Content-Type is still totally ignored |
| 10:52 | <Philip`> | Hixie: It doesn't work on <script> - the Content-Type is still totally ignored, and only the <script type> has an effect |
| 10:53 | <Philip`> | Hixie: It doesn't work on <style> - the Content-Type is still totally ignored, and only the <style type> has an effect |
| 10:54 | <annevk> | <style>? you mean <link rel=stylesheet>? |
| 10:55 | <Philip`> | Uh |
| 10:55 | <Philip`> | Yes |
| 10:56 | <annevk> | so far X-Content-Type-Options fails |
| 10:56 | <annevk> | just like whatwg.org btw |
| 10:56 | <Philip`> | It works correctly on <object type="text/html" data="some HTML with text/plain"> (i.e. nosniff makes it render as text instead of HTML) |
| 10:57 | <Hixie> | does it work for images in object or directly navigated? |
| 10:57 | <Philip`> | It doesn't work correctly on <object type="image/jpeg" data="some PNG with text/plain and nosniff"> (it still renders the PNG image) |
| 10:57 | <Hixie> | the most important one is really videos sent as text/plain |
| 10:57 | <Hixie> | how about without the type=? |
| 10:57 | <Philip`> | (although with type="image/png" it renders a broken X icon instead) |
| 10:57 | <Hixie> | -_- |
| 10:57 | <Hixie> | sounds like a roaring success |
| 10:57 | <Hixie> | someone tell public-html |
| 10:57 | <Hixie> | it'll get them off our backs for a few days |
| 10:58 | <Hixie> | :-) |
| 10:58 | <Philip`> | (It renders a broken X regardless of the nosniff) |
| 10:58 | <Philip`> | (unless I'm doing something silly?) |
| 10:58 | <Hixie> | png is handled differently to jpeg and gif |
| 10:58 | <Hixie> | so i wouldn't be surprised |
| 10:59 | <Philip`> | I would test it in the Live DOM Viewer, but it's not working :-( |
| 10:59 | <Hixie> | :-( |
| 11:03 | <Philip`> | Images (PNG and JPEG) work in <object> if type="image/jpeg" or type="image/gif", regardless of the content-type or nosniff |
| 11:03 | <Philip`> | and don't work if type="image/png" |
| 11:03 | <Hixie> | what if there's no type? |
| 11:03 | <annevk> | their code is buggy... |
| 11:03 | <Philip`> | Then they don't work at all |
| 11:03 | <Hixie> | even with the right type? |
| 11:03 | <Philip`> | Even with the right content-type, yes |
| 11:04 | <Hixie> | man |
| 11:04 | <Hixie> | they try so hard |
| 11:04 | <Hixie> | jesus <input> has a lot of attributes |
| 11:15 | <hsivonen> | does anyone have good ideas for parser selection in case I enabled file upload and textarea input on html5.validator.nu? |
| 11:15 | <Hixie> | parser selection? |
| 11:15 | <Hixie> | you mean how to decide on text/html vs xml? |
| 11:15 | <hsivonen> | Hixie: yes |
| 11:15 | <Philip`> | Decide on text/html |
| 11:15 | <Hixie> | well theoretically you could trigger on the UA-provided mime type, which will be based on the extension |
| 11:15 | <Hixie> | but i'd just use text/html |
| 11:16 | <hsivonen> | Hixie: UA-provided anything won't work for textarea |
| 11:16 | <Hixie> | ah for textarea just assume text/html |
| 11:16 | <Hixie> | or, provide ui |
| 11:17 | <hsivonen> | Hixie: providing UI sucks when the whole point of html5.validator.nu is that it has less UI |
| 11:17 | <Hixie> | yes |
| 11:17 | <Hixie> | so assume text/html |
| 11:17 | <Hixie> | autodetection will just cause confusion |
| 11:17 | <hsivonen> | ok |
| 11:18 | <Hixie> | <input>'s summary is ridiculous |
| 11:19 | <Hixie> | btw i keep hitting your five second timeout |
| 11:19 | <Hixie> | whatwg.org is responding, just very very slowly |
| 11:20 | <annevk> | Hixie, you checked in an empty document into dev.w3.org creating a massive diff (the generated HTML file is >10MiB) |
| 11:20 | <hsivonen> | Hixie: I'm reluctant to change timeouts, because normal non-broken non-tarpit sites respond in under 5 seconds |
| 11:21 | <hsivonen> | Hixie: I already have to tweak other DoS prevention measures every now and then to accommodate the spec's growth |
| 11:22 | <Hixie> | heh fair enough :-) |
| 11:22 | <Hixie> | annevk: yeah doesn't surprise me |
| 11:22 | <Hixie> | annevk: it'll fix itself |
| 11:22 | <annevk> | hsivonen, maybe you should whitelist whatwg.org :) |
| 11:23 | <Hixie> | nah |
| 11:23 | <hsivonen> | let's see if there are APIs for that... |
| 11:25 | <Hixie> | i'd rather you didn't whitelist whatwg.org, that'll just lead to confusing errors later when one site gets different behaviour than another |
| 11:25 | <hsivonen> | OK. |
| 11:25 | <hsivonen> | (and I didn't yet find a suitable API, either) |
| 11:29 | <annevk> | http://tapthehive.com/discuss/This_Post_Not_Made_In_Chrome_Google_s_EULA_Sucks |
| 11:32 | <hsivonen> | more to the point, the EULA for the *Services* sucks |
| 11:33 | <annevk> | "Your use of Google’s products, software, services and web sites (referred to collectively as the “Services” in this document" |
| 11:33 | <annevk> | -- http://www.google.com/chrome/intl/en/eula_text.html |
| 11:33 | <Hixie> | well that's dumb |
| 11:33 | <hsivonen> | oh. That's bad. |
| 11:33 | Hixie | will raise it with the lawyers |
| 11:34 | <Hixie> | though i expect they already know and are fixing it as we speak |
| 11:34 | <hsivonen> | as far as I can tell, the EULA nastiness is already known for Google Docs |
| 11:35 | <Hixie> | well for google docs it makes sense |
| 11:35 | <Hixie> | without it we could theoretically not show the document back to you |
| 11:36 | <hsivonen> | Hixie: other application providers scope the language more reasonable to the right required for performing the functionality of the service and only for the duration that the content resides in the service |
| 11:36 | <hsivonen> | someone made a comparison of Docs and competitors on the EULA point a while ago |
| 11:36 | <Hixie> | that would seem reasonable |
| 11:36 | <Hixie> | seems our lawyers are asleep |
| 11:37 | <Hixie> | which i guess makes sense |
| 11:37 | <Hixie> | i'll ping them tomorow |
| 11:37 | <Hixie> | tomorrow |
| 11:37 | <hsivonen> | Hixie: thanks |
| 11:38 | <hsivonen> | fortunately, Flickr has a lot of paranoid pro photographer users, which makes Yahoo! more careful on this point nowadays |
| 11:46 | <hsivonen> | zcorpan: do you have OCR that worked on the Google comic? |
| 11:47 | <hsivonen> | http://lists.w3.org/Archives/Public/wai-xtech/2008Sep/0010.html |
| 11:49 | <Hixie> | i love people who send e-mails with the subject line starting "OFF TOPIC" |
| 11:49 | <Hixie> | that just about says it all right there |
| 11:53 | <hsivonen> | hmm. Google Gears does show up in about:plugins, so I wasn't totally crazy to think it was an NPAPI plugin |
| 11:53 | <Hixie> | gears is a lot of things |
| 11:53 | <Lachy> | hah! That's the best response to John Foliot I've seen in a while :-) |
| 11:56 | <Lachy> | that article about Google Chrome's TOS doesn't seem right. The section he quoted from the TOS is talking about Services, and Google Chrome isn't a service. |
| 11:56 | <Hixie> | the eula does define "services" to include google software |
| 11:57 | <annevk> | zcorpan, for your viewer you want something like http://w3future.com/weblog/stories/2002/05/04/urisForDynamicPages.xml plus a timeout listener to take care of moving back and forth |
| 11:57 | <Lachy> | oh, you're right |
| 11:57 | <annevk> | zcorpan, see eg the source of http://anne.is.weggeweest.nl/image-viewer for example of such code |
| 11:58 | <Hixie> | the timeout error message i got from h.v.n changed |
| 12:01 | <Hixie> | nn |
| 12:02 | <annevk> | ooh, the Thomas Broyer tactic is also neat, we could just define a new XSLT 1.0 output method. That requires minimal effort and addresses the issue in theory |
| 12:02 | <hsivonen> | Hixie: perhaps your server is now responding in 5 seconds but takes more than 5 seconds between packets |
| 12:02 | <hsivonen> | or something |
| 12:07 | <hendry> | http://flickr.com/photos/hendry/2824592876/ # can chrome be really failing a cookies test.. hmm :) |
| 12:08 | <hsivonen> | oh. is WebKit not supporting dynamic SVG at all at this point? |
| 12:09 | <Philip`> | hendry: Clearly Google has recognised the privacy risk of tracking cookies, and so has decided to entirely disable them to protect their users |
| 12:10 | <hsivonen> | hmm. looks like the Finnish localizer of Chrome had no clue about the context of the strings or what common browser idioms are |
| 12:10 | <hendry> | Philip`: so JS access is disabled or something |
| 12:10 | <hsivonen> | Show Source is translated to something like Show Origin |
| 12:11 | Philip` | wonders whether the localisations are open source |
| 12:12 | <hsivonen> | also, for some reason, they served me a Finnish localization based on IP even though I used English Firefox and Windows XP |
| 12:13 | <hendry> | hsivonen: geolocation rocks :) |
| 12:13 | <hsivonen> | hendry: Swedish speakers must be thrilled |
| 12:14 | <Lachy> | hsivonen, google's geolocation localisation crap is a constant annoyance for me. Just make sure you get a cookie from google.com/ncr so they don't do it |
| 12:14 | <virtuelv> | I hate how this generally works on the web, where I get Norwegian pages forced upon me |
| 12:14 | <virtuelv> | no-bok, or no-nn is not even in my accept-language |
| 12:14 | <virtuelv> | (Google, I'm looking at you) |
| 12:15 | <Dashiva> | Is no-bok even a valid value? |
| 12:15 | <Lachy> | virtuelv, at least you understand norwegian. It's a lot worse for people travelling to foreign countries |
| 12:15 | <Dashiva> | oh, n/m |
| 12:15 | <Lachy> | though, even when I was in Australia, I still hated that it redirected me from google.com to google.com.au |
| 12:15 | <virtuelv> | Lachy: I realise that |
| 12:16 | <virtuelv> | the other thing bothering me is geotargetting of search results |
| 12:16 | <Dashiva> | At one point I had japanese first in my accept-language string without knowing it |
| 12:16 | <Dashiva> | It never kicked in until I tried to read apache docs |
| 12:16 | <annevk> | haha |
| 12:16 | <hsivonen> | with Maps, I want it to search Finland instead of the U.S. by default, but that doesn't magically happen |
| 12:17 | <virtuelv> | I really hate the fact that if I send a search result page to someone, and they won't be seeing the same results as me |
| 12:17 | <Philip`> | Dashiva: Maybe it's a typo for no-bork |
| 12:17 | <Lachy> | hsivonen, doesn't it do that if you search from maps.google.fi ? |
| 12:18 | <Philip`> | hsivonen: I've somehow set it to default to the exact area where I live, which seems to works well, though I've got no idea how I did that |
| 12:18 | <hsivonen> | Lachy: thanks. I hadn't tried going there. :-) |
| 12:18 | <hsivonen> | Google's DWIM isn't right |
| 12:19 | <Lachy> | I would do the same with maps.google.no, I have to use the .com so I get enlish, and then append "Norway" to the end of my queries |
| 12:19 | hsivonen | notes that Google optimizes Lapland out of the default view |
| 12:20 | <Philip`> | Google should have some heuristics to calculate the cluefulness of a user, and if they're sufficiently clueful then it would start to respect their browser language settings instead of just guessing from IP |
| 12:22 | <Lachy> | maybe we need a new I-Really-Do-Accept: header, or something to indicate that the language prefs have been customised and that they should be believed |
| 12:22 | <Dashiva> | X-Authorative-Accept? |
| 12:23 | <Lachy> | though, if people download localised versions of browsers which have reasonable defaults, it's strange that they still think they can't be trusted |
| 12:24 | <annevk> | content negotiation failed, time to try something else |
| 12:25 | <Dashiva> | Wasn't there a lot of talk about meta language for content negotiation? |
| 12:27 | <annevk> | could be |
| 12:28 | <zcorpan> | hsivonen: no i typed it by hand |
| 12:29 | <zcorpan> | annevk: i got history navigation work ok i think |
| 12:30 | <Philip`> | gocr does alright at OCR |
| 12:30 | <Philip`> | "Gaagle Chrome hndho OpnSourBrawsr)" |
| 12:31 | <Philip`> | etc |
| 12:31 | <zcorpan> | Philip`: if that's what you get then i guess it's faster to type by hand |
| 12:31 | <Philip`> | zcorpan: I suppose so :-( |
| 12:33 | <annevk> | we could just have http://n.whatwg.org/xslt or something and map an output algorithm to that |
| 12:33 | <Dashiva> | You'd think google's ocr would also do spell checking :) |
| 12:33 | <hsivonen> | zcorpan: that's a lot of typing |
| 12:33 | <annevk> | maybe #html at the end |
| 12:33 | <Philip`> | Dashiva: I didn't know Google did OCR |
| 12:34 | <zcorpan> | hsivonen: naw |
| 12:36 | <Dashiva> | Philip`: I just assumed the g in gocr was google :) |
| 12:36 | <Lachy> | I don't think we should take on the XSLT WG's job to define a new XSLT output method. If they're not willing to do the work to keep up to date, that's their problem. Don't shift the responsibility on to use to bend over backwards to cater for them |
| 12:37 | <Lachy> | s/use/us/ |
| 12:37 | <Philip`> | Dashiva: Ah - not all software in the world that begins with 'g' was written by Google :-) |
| 12:37 | <Dashiva> | Lies |
| 12:38 | <jcranmer> | gedit? |
| 12:39 | <jcranmer> | gnibbles? |
| 12:39 | <Lachy> | gcc, gmake, etc. |
| 12:39 | <jcranmer> | gmplayer |
| 12:39 | <jcranmer> | gfloppy |
| 12:39 | <Lachy> | even gBrowser |
| 12:39 | <jcranmer> | 95% of GNOME, in fact |
| 12:40 | <Dashiva> | jcranmer: Yeah, but GNOME is a Google project, it starts with a G. |
| 12:40 | <Lachy> | LOL |
| 12:40 | <hsivonen> | I cleaned up http://about.validator.nu/ a bit. |
| 12:40 | <Lachy> | yeah, and GNU was started by google 25 years ago too |
| 12:41 | <hsivonen> | zcorpan: also, bookmarklets included. Thanks. (I hope I got the right versions.) |
| 12:41 | <hsivonen> | oops. the schema descriptions are broken still |
| 12:44 | <virtuelv> | and gimp is short for google image manipulation program |
| 12:46 | <jcranmer> | ... this is where I'm supposed to make a bash.org reference, but it's down ... |
| 12:47 | <zcorpan> | hsivonen: cool. looks like the right version |
| 12:48 | <zcorpan> | hsivonen: i noticed that it breaks the back button |
| 12:50 | <hsivonen> | zcorpan: is that something I can fix? |
| 12:51 | <hsivonen> | does fixing it need HTML5 history management in the browser? |
| 12:51 | <zcorpan> | hsivonen: maybe, by moving back the root element after the form has submitted or something. but maybe that would cause scripts to execute again? |
| 12:51 | <Philip`> | hsivonen: With the Text Field input, it's unhelpful that the validation results are scrolled off the bottom of the screen and invisible after clicking 'validate' |
| 12:52 | <Philip`> | (at least on a 1280x800 screen with various menus taking up some vertical space) |
| 12:52 | <Philip`> | (*menus and toolbars) |
| 12:52 | <zcorpan> | hsivonen: perhaps you could clone the root element instead of moving it |
| 12:52 | <hsivonen> | Philip`: WFM on full HD Cinema Display :-) |
| 12:52 | <Philip`> | hsivonen: :-P |
| 12:53 | <hsivonen> | zcorpan: OK. I'll file a bug. |
| 12:53 | <hsivonen> | (today is a documentation day, so I'll write more docs now) |
| 12:58 | <hsivonen> | Philip`: do you have a suggestion how to fix? should I use media queries to make the textarea smaller? |
| 13:00 | <zcorpan> | hsivonen: perhaps submit to a fragment id |
| 13:01 | <zcorpan> | but maybe that would be annoying? |
| 13:08 | <hsivonen> | that's a possibility, but if I don't tie it to MQ, it might annoy other users |
| 13:11 | <zcorpan> | yeah |
| 13:19 | hsivonen | notes that various Java XML tools roll their own character encoding support |
| 13:20 | <Dashiva> | I thought java was all about massive libraries |
| 13:20 | <hsivonen> | so much work when they could simply not support anything but UTF-8 for output |
| 13:20 | <takkaria> | there's some tests lying around in Hubbub's repository that aren't in html5lib's, I should probably fix that at some point |
| 13:21 | <hsivonen> | Dashiva: pre-1.4.2 Java sucked badly for charset stuff |
| 13:21 | <hsivonen> | Dashiva: and apparently, querying an encoding for the characters in can and can't encode doesn't work |
| 13:21 | <Dashiva> | That seems a bit lacking, yeah |
| 13:22 | <Dashiva> | (and even if you want to support non-utf8 you can still output utf8 and use a postprocessor) |
| 13:28 | <annevk> | hsivonen, what does the "3" stand for? |
| 13:29 | <annevk> | hsivonen, http://validator.nu/?doc=http%3A%2F%2Fanne.is.weggeweest.nl%2Fimage-viewer does not tell me what attributes are missing that I need to specify |
| 13:30 | <hsivonen> | annevk: yeah that sucks. |
| 13:30 | <hsivonen> | annevk: I have two excuses |
| 13:30 | <hsivonen> | 1) Hixie removed the "(required)" annotation from the spec |
| 13:30 | <hsivonen> | and |
| 13:30 | <hsivonen> | 2) oNVDL upstream was supposed to fix this |
| 13:31 | <hsivonen> | annevk: I changed the version number to 3 after some major feature |
| 13:31 | <hsivonen> | but since this is perpetual beta, the version number is pointless |
| 13:32 | <annevk> | maybe h1 span { display:none} then? :) |
| 13:51 | <Philip`> | hsivonen: Maybe put the green/red pass/fail box near the top of the page, above the URL/file/text input and all the various configuration settings, and have the individual messages down below like they are already? |
| 13:52 | <Philip`> | Then it'd be easy to see immediately whether the page is okay, and if it's not then you can scroll down to see the problems |
| 13:52 | <Philip`> | (That seems to be what validator.w3.org does) |
| 14:25 | <zcorpan> | Hixie: shouldn't it be "that cannot output HTML markup with the short doctype" or some such since they generally can output HTML without a doctype but a doctype is required |
| 14:28 | <Philip`> | Judging by the referrer list in http://krijnhoetmer.nl/irc-logs/, a lot of Chrome users are getting error 110 |
| 14:40 | <zcorpan> | hsivonen: when you see an xslt-compat doctype, you can point to XSLT4HTML5 :) |
| 14:48 | <Philip`> | Hmm, IE says <select>.type == "select-one", and <select multiple>.type == "select-multiple" |
| 14:55 | <Dashiva> | Philip`: Isn't that as expected? |
| 14:58 | <Philip`> | Dashiva: Oh, apparently it is - it just isn't something I've ever seen or heard of before |
| 14:58 | <Dashiva> | Yeah, it's pretty obscure |
| 14:59 | <zcorpan> | but nice |
| 14:59 | <zcorpan> | also see textarea |
| 14:59 | <Dashiva> | I expect it was inteded for traversing the elements collection just looking at type (textarea also has the property) |
| 15:29 | <zcorpan> | Hixie: attr-form-form should be attr-fieldset-form |
| 15:53 | <Philip`> | Oh, is about:internets meant to do something in Chrome other than show an empty white page with "The tubes are clogged" in the title? |
| 15:53 | <Philip`> | If so, I guess it's buggy since it doesn't work for me :-( |
| 15:53 | <takkaria> | what OS are you on? |
| 15:53 | <Philip`> | Vista |
| 15:53 | <annevk> | geez, whatwg.org is slow |
| 15:53 | <takkaria> | that's why |
| 15:54 | <Philip`> | Oh, okay |
| 15:54 | <takkaria> | on XP and before, it shows the pipes screensaver |
| 15:54 | <annevk> | http://img352.imageshack.us/my.php?image=gogolegearseasteregguz9.jpg |
| 15:56 | <Philip`> | If Google can't even write code that works on the latest version of Windows, I suppose I shouldn't expect they'll ever do well on Linux :-p |
| 15:57 | <Lachy> | Philip`, no, they probably just didn't care about Vista too much since no-one uses it anyway, since most people who get it use the downgrade option in order to upgrade to XP anyway |
| 15:58 | <Lachy> | Philip`, you could try building it yourself http://dev.chromium.org/ |
| 15:58 | <Lachy> | I'm going to try building it for OS X |
| 15:58 | Philip` | wonders at Lachy's definition of "most" |
| 15:59 | <Philip`> | Lachy: Apparently it sort of builds and runs some tests on Linux / OS X, but isn't at a stage where it actually has a browser window |
| 15:59 | <annevk> | I thought Mac and Linux were several modules sort, last I read |
| 15:59 | <annevk> | short, even |
| 15:59 | <annevk> | yeah, what Philip` said |
| 15:59 | <Lachy> | my definition of "most" is based on nothing more than anecdotal evidence about people really disliking vista and choosing to downgrade |
| 16:00 | <Lachy> | oh, that sucks. |
| 16:00 | <Philip`> | http://googlemac.blogspot.com/2008/09/platforms-and-priorities.html |
| 16:12 | <Philip`> | http://blogs.msdn.com/ie/archive/2008/09/02/ie8-security-part-vi-beta-2-update.aspx#8922060 - "Will the authoritative asttribute still be supported or is it dropped in favor of the new header? I already implemented it here and there, so that's why I'm asking." - implementing features before they've even been released in a beta is not necessarily a great idea |
| 16:17 | <hallvors> | annevk: wasn't content-negotiation a nice idea, just badly spec'ed with even worse implementations? |
| 16:24 | <Philip`> | kingryan: If you happen to read the logs and see this before the email I sent, I think your change to remove the parse error from </br/> is wrong, since the spec says "When an end tag token is emitted with its self-closing flag set, that is a parse error." |
| 16:31 | <annevk> | hallvors, I agree it's a nice idea, though on the other hand I'm not sure if multiple content types really makes sense, languages sort of does though |
| 16:31 | <annevk> | hallvors, though even there it's dodgy as the various translated resources might be crappy compared to the original one |
| 16:32 | <annevk> | eg, http://fr.wikipedia.org/wiki/Terra_Amata and http://en.wikipedia.org/wiki/Terra_Amata |
| 16:37 | <annevk> | Philip`, though ideally a validator only reports "</br> not allowed" |
| 16:44 | <Philip`> | "As of Beta 2, CSS expressions are not supported in IE8 Standards Mode." - ooh, only just noticed that |
| 17:26 | <Philip`> | http://marketshare.hitslink.com/report.aspx?sample=21&qprid=43&qpcustom=Chrome+0.2 - 1%? It's beating Opera already :-( |
| 17:27 | <gsnedders> | I like how it rises as people start trying it out, then drops back down :P |
| 18:16 | <jcranmer> | Hixie: it seems multi-page HTML 5 is broken ATM? |
| 18:16 | <jcranmer> | http://www.whatwg.org/specs/web-apps/current-work/multipage/stderr.txt |
| 18:20 | <jcranmer> | also, slight typo in that you're missing a comma, but typographical errors are unimportant right now |
| 18:40 | <Philip`> | jcranmer: Looks like my spec-splitter script failed (presumably timed out) while trying to download the spec to split |
| 18:41 | <jcranmer> | Philip`: 404 error, actually |
| 18:41 | <jcranmer> | I think |
| 18:41 | <Philip`> | and so then it ran the splitter script, which failed because there was no input, and then it zipped up the output files, of which there were none because it had failed, and then uploaded those and the old multipage files got replaced by the zero new ones |
| 18:43 | <Philip`> | jcranmer: It failed with "curl: (52) Empty reply from server" |
| 18:43 | Philip` | tells the script to try again |
| 18:44 | <Philip`> | which'll take a couple of minutes |
| 18:45 | <Philip`> | kingryan: If you didn't happen to read the logs, see http://krijnhoetmer.nl/irc-logs/whatwg/20080903#l-909 :-) |
| 18:46 | <Philip`> | jcranmer: multipage is back now |
| 18:46 | <jcranmer> | ^_^ |
| 18:47 | <Philip`> | Anybody who knows html5lib: Is http://www.whatwg.org/specs/web-apps/current-work/multipage/stderr.txt important and can I make it shut up? |
| 19:07 | <gsnedders> | Um, is javascript: defined anywhere (like, to not be a URI) |
| 19:08 | <gsnedders> | (so javascript:alert("/?#") works) |
| 19:16 | <zcorpan> | http://tools.ietf.org/html/draft-hoehrmann-javascript-scheme-00 ? |
| 19:17 | <zcorpan> | "Designers of protocol elements that accept resource identifiers as defined in this document should consider this case and, where compatibility is a concern, define a pre-processing step that percent-encodes all '#' characters before the content of the protocol element is processed as 'javascript' resource identifier." |
| 19:50 | <kingryan> | Philip`: reading now |
| 19:50 | <kingryan> | about that: afaict, you're right, it should be a parse error, its just difficult (impossible?) to tell in the tokenizer |
| 19:50 | <kingryan> | http://krijnhoetmer.nl/irc-logs/whatwg/20080903#l-300 :) |
| 19:52 | <Philip`> | kingryan: No it's not - you see that it's an end tag, and that it's got the self-closing flag, so you emit a parse error |
| 19:52 | <Philip`> | (I think) |
| 19:52 | <kingryan> | Philip`: you may be right, but i had a difficult time implementing that |
| 19:53 | <kingryan> | it could be that it was 2am and I was tired though |
| 19:53 | <Philip`> | if (end-tag && self-closing-flag) emitParseError(); :-P |
| 19:54 | <Philip`> | When you're emitting tag tokens, you must have some way of knowing whether it's start/end, and that seems to be about all it needs |
| 19:54 | <kingryan> | right |
| 19:55 | <Philip`> | (This is separate from the bit that checks whether the self-closing flag was acknowledged, or whatever the terminology is, which does seem impossible in the tokeniser) |
| 19:56 | <kingryan> | yeah, its probably not impossible, but it seems to be more than what the spec is asking for in the tokenization section |
| 20:45 | <gsnedders> | http://stuff.gsnedders.com/referrers/referrers.html |
| 20:45 | <gsnedders> | All kinds of bizarre things find my site |
| 21:00 | <Hixie> | Philip`: just search for the word "fetching" in the spec, that should give you every case that can initiate HTTP |
| 21:00 | <Hixie> | and if i missed any, that's a bug |
| 21:00 | <Hixie> | uh, search for "fetch", even |
| 21:01 | <Hixie> | Philip`: did you test to see if IE8 will treat video binary data as text/plain if sent with text/plain, nosniff? |
| 21:09 | <Hixie> | hsivonen: i'm glad to inform you of http://www.google.com/google-d-s/addlterms.html |
| 21:11 | <Hixie> | and http://news.cnet.com/8301-13860_3-10031703-56.html?part=rss&subj=news&tag=2547-1_3-0-5 |
| 21:11 | <Philip`> | Hixie: I didn't test that |
| 21:12 | <webben> | Hixie: good stuff :) |
| 21:12 | Philip` | isn't sure where to find a video that IE would render to test with |
| 21:13 | <Philip`> | Didn't anyone from Google bother to check the EULA before releasing Chrome? |
| 21:16 | <Hixie> | Philip`: well the real question is does it offer it for download or show it |
| 21:20 | <Hixie> | i wish opera was consistent in its messaging |
| 21:21 | <Hixie> | i understand people in one company having different opinions |
| 21:21 | <Hixie> | but people having different facts is very confusing |
| 21:21 | <wilhelm> | Hm? |
| 21:21 | <Hixie> | ed and chaals are disagreeing on whether opera supports namespace-less svg and on whether it's important enough to need to support it in svg |
| 21:21 | <Hixie> | er |
| 21:21 | <Hixie> | in svg-in-html |
| 21:28 | <Hixie> | <label> elements are associated with a form? really? |
| 21:28 | <Hixie> | that seems silly |
| 21:29 | <annevk> | Charles talks about image/svg+xml implying the SVG namespace if no default namespace declaration is set and the root element is not in a namespace and Erik is talking about the DOCTYPE augmenting the markup |
| 21:29 | <annevk> | Jeff then confuses the two |
| 21:30 | <Hixie> | chaals also said that we didn't need to handle <svg> without namespace, and ed said we did. |
| 21:30 | <Hixie> | given text/html has no doctypes inline, we either have to handle it or we don't |
| 21:30 | <annevk> | I'm not sure Charles knows about the DOCTYPE augmenting the markup |
| 21:31 | <annevk> | (the quote from Charles is also from July...) |
| 21:32 | <Hixie> | one would assume he does know everything there is to know here, given how authoritatively he speaks on the subject |
| 21:33 | <gsnedders> | Hixie: It's possible to sound authoritative on anything |
| 21:34 | <annevk> | I'm also not sure whether Charles considers a DOCTYPE to augment the markup to be ok, but frankly, it all doesn't really matter |
| 21:34 | <Hixie> | well, it matters in that if ed is right, the svgwg proposal is fundamentally flawed |
| 21:36 | <annevk> | no comment |
| 21:39 | <annevk> | though fwiw, Erik is correct |
| 21:39 | <annevk> | I'm not sure why Jeff didn't simply test instead of digging up some old quote |
| 21:40 | <Hixie> | i look forward to seeing how the svgwg updates their proposal then |
| 21:44 | <hober> | Is their proposal work public? That is, how can someone follow along with edits to their proposal? |
| 21:44 | <Hixie> | i think so |
| 21:45 | <Hixie> | check their wg mailing list and minutes |
| 21:46 | <annevk> | seems Julian disagrees with my testing |
| 21:47 | <smedero> | I believe the draft work on the SVG in HTML proposal lives here: http://www.w3.org/Graphics/SVG/WG/wiki/SVG_in_text-html |
| 21:48 | <smedero> | Or so I gleaned from looking at their tracker instance... I might be wrong. |
| 21:48 | <Hixie> | they haven't done anything since i sent the comments then |
| 21:48 | <Hixie> | which isn't surprising, that was relatively recently |
| 21:48 | <smedero> | around here, shepazu would be the person to ask |
| 21:49 | <Hixie> | anyway. gotta go. |
| 21:49 | <Hixie> | bbiab. |
| 21:50 | <shepazu> | yes, it's public, and we will be updating it with hixie's feedback shortly.... we're a bit busy with other things for the next week or two, then we will address his comments |
| 21:51 | <annevk> | http://annevankesteren.nl/test/svg/doctype.xml has the DOCTYPE SVG test so people can check for themselves |
| 21:53 | <annevk> | http://ajaxian.com/archives/adding-custom-tags-to-internet-explorer-the-official-way "I'm going to do more testing on this functionality today to see how deep it goes, but if true it makes it easier to create browser shims for Internet Explorer for things like SVG, MathML, etc., including HTML 5 (if we namespace the HTML 5 elements, required to get this to work)." yes we know, no we don't want that! :) |
| 21:56 | <jgraham> | Philip`: Re: your question about 3 hours ago, the warning is only important if you care abou <!DOCTYPE hTmL> vs <!DOCTYPE html> and you can make it shut up by messing about with the warnings module |
| 21:59 | <annevk> | euh Hixie, where you mean "label" in the parser you wrote "input" |
| 21:59 | <annevk> | + <dt>A start tag whose tag name is "input"</dt> |
| 22:04 | <virtuelv> | ugh, this is hacky, http://ajaxian.com/archives/adding-custom-tags-to-internet-explorer-the-official-way |
| 22:06 | <Philip`> | "The reason I'm looking for an alternative to the createElement trick is I've found that it doesn't work with nested custom tags" - seems to work fine with nested tags when I try it |
| 22:07 | <Philip`> | (except for the old problem of IE8b2 not styling them, but it doesn't style custom prefix:name tags either) |
| 22:14 | <Philip`> | Is there some way to stop Opera trying to download all feeds simultaneously? It maxes out my (slow) connection for five or ten seconds, which is bad if I'm trying to play online games :-( |
| 22:16 | <wilhelm> | Philip`: The closest you'll get is that you can change the update interval, I think. |
| 22:19 | <Philip`> | Sadly it doesn't let me set each interval to a distinct prime number |
| 22:19 | <Philip`> | so it's still going to end up trying to download lots at once |
| 22:22 | <zcorpan> | is it time for publication soon? |
| 22:22 | <annevk> | few weeks |
| 22:22 | annevk | tries to find the date |
| 22:23 | <zcorpan> | oh i thought it was more like one week |
| 22:23 | <annevk> | I hilited it in http://krijnhoetmer.nl/irc-logs/html-wg/20080901 |
| 22:23 | <zcorpan> | no rationale? |
| 22:23 | <annevk> | nope |
| 22:24 | <annevk> | works for me, WF2 might be integrated by that date |
| 22:25 | <zcorpan> | yeah |
| 22:25 | <zcorpan> | finally |
| 22:26 | <annevk> | lots of new elements to debug and report bugs on :) |
| 22:26 | <Philip`> | Has the Forms Task Force officially been declared useless yet? |
| 22:27 | <smedero> | yes |
| 22:27 | <smedero> | i'd have to dig up the "non-binding" telecon decision. |
| 22:27 | annevk | saw Philip` already started by checking out <select>.type and all |
| 22:27 | <Philip`> | annevk: That wasn't related to WF2 at all |
| 22:27 | <Philip`> | I was just looking at strings in IE, and saw that |
| 22:27 | <annevk> | meh |
| 22:28 | <annevk> | Opera does it too |
| 22:28 | <Philip`> | Forms are boring, so someone else can look at them :-p |
| 22:28 | <Philip`> | annevk: I'd hope so, since the DOM spec says to do so |
| 22:29 | <Philip`> | (but I never knew that before today) |
| 22:29 | <annevk> | well, same here |
| 22:30 | <smedero> | this somewhat tracks the decision process: http://www.w3.org/html/wg/tracker/actions/56 |
| 22:30 | <annevk> | cheers smedero |
| 22:31 | <smedero> | i'm having trouble finding the transcript for the 7/10 telecon.... oh well |
| 22:51 | <gsnedders> | Hixie: ping |
| 23:05 | <Hixie> | gsnedders: pong |
| 23:06 | <Hixie> | shepazu: cool. there's no rush, btw, i'm gonna be swapmed with wf2 stuff for a while. |
| 23:06 | <shepazu> | Hixie: ok, cool.. still, we will try to be as prompt as possible |
| 23:07 | <Hixie> | cool |
| 23:08 | jcranmer | wants <input type=""> widely supported NOW |
| 23:09 | <jcranmer> | especially date, datetime, and time... |
| 23:09 | <Hixie> | i want <datagrid> |
| 23:09 | <Hixie> | and websocket |
| 23:10 | <jcranmer> | <input type="gurh, unknown value"> should fall back to a single-line textbox? |
| 23:10 | <Hixie> | yes |
| 23:11 | <Philip`> | But you shouldn't rely on that, since WF3 might define the "gurh, unknown value" input type |
| 23:11 | <BenMillard> | jgraham, I just tried using http://james.html5.org/tables/table_inspector.html on http://www.wisc.edu/about/facts/budget.php and got a huge error at http://james.html5.org/cgi-bin/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fwww.wisc.edu%2Fabout%2Ffacts%2Fbudget.php&source=&algorithm=smartheaders |
| 23:12 | <BenMillard> | jgraham, the page does have a few parse errors: http://validator.w3.org/check?verbose=1&uri=http%3A%2F%2Fwww.wisc.edu%2Fabout%2Ffacts%2Fbudget.php |
| 23:12 | <jcranmer> | Philip`: it'd be about fallback for <input type="date"> in my case |
| 23:26 | <jgraham> | BenMillard: known bug. Turns out that people using "XHTML" trip up people like me who try to repurpose XML libraries for HTML because the XML libraries have special restrictions needed to stop people unwittingly getting namespaces in XML wrong |
| 23:27 | <jgraham> | It turns out to be harder than I would like to fix because lxml (the library in question) gives a choice of unsutiable behaviours |
| 23:27 | <jgraham> | If you have enough time you can work out how to aportion blame between me, the lxml authors, the page authors and the W3C for the problem :) |
| 23:28 | <BenMillard> | jgraham, I'll leave that in your more capable hands. ;) |
| 23:29 | <Hixie> | jgraham: html5 gives some ideas on how to deal with these issues |
| 23:29 | <BenMillard> | jgraham, can you remember the thread were people demanded you make the Table Inspector accessible, with headers+id? I remember it but can't find it... |
| 23:30 | <jgraham> | Hixie: What does HTML5 have to say about dealing with libs that won't allow : in element names? |
| 23:31 | <jgraham> | BenMillard: I remmber it existed and that I did what they asked. I think it was Leif asking |
| 23:31 | <Hixie> | jgraham: http://www.whatwg.org/specs/web-apps/current-work/#coercing |
| 23:32 | <BenMillard> | jgraham, closest I've reached is: http://lists.w3.org/Archives/Public/public-html/2008Mar/0082.html |
| 23:33 | <jgraham> | BenMillard: If you copy the first table from that page into the table inspector source view it works nicely |
| 23:33 | <jgraham> | s/source view/textarea/ |
| 23:33 | <jgraham> | Hixie: Oh nice. Is that new or have I just been oblivious for a long time? |
| 23:34 | <Hixie> | jgraham: relatively new |
| 23:34 | <Hixie> | jgraham: hsivonen asked for it |
| 23:34 | <Hixie> | iirc |
| 23:34 | <jgraham> | Hixie: That's clearly because hsivonen rocks :) |
| 23:34 | <Hixie> | that's the rumour |
| 23:36 | <Philip`> | When I view the spec, I get "Fatal network error: Error: Internal Server Error (500)" |
| 23:36 | <Philip`> | in an alert box |
| 23:37 | <Hixie> | that's unfortunate |
| 23:37 | <Hixie> | wfm |
| 23:38 | <Hixie> | oh |
| 23:38 | <Hixie> | it ran out of memory trying to log you in |
| 23:39 | Hixie | gives the server a little more ram |
| 23:40 | <BenMillard> | jgraham, I've found Leif's message using Outlook's search within seconds. Note to self: try that first next time. |
| 23:40 | <jgraham> | It somehow never occurs to me to search local email. I guess because that doesn't give me a URL |
| 23:41 | <Hixie> | Philip`: should work now |
| 23:41 | <BenMillard> | jgraham, it does give you enough information to then have a happier time with the web search interface, though. |
| 23:41 | <jgraham> | BenMillard: Indeed |
| 23:42 | <BenMillard> | jgraham, Leif's message is here: http://lists.w3.org/Archives/Public/public-html/2007Sep/0153.html |
| 23:43 | <BenMillard> | jgraham, I'm still searching for a reply where you say that feature is definitely in the inspector, though. |
| 23:44 | <jgraham> | BenMillard: I replied off-list |
| 23:44 | <Philip`> | Hixie: Does work now - thanks :-) |
| 23:44 | <jgraham> | (I may also have replied on list) |
| 23:44 | <Philip`> | jgraham: The X-Archived-At header in W3C mail gives you a URL |
| 23:47 | <jgraham> | Philip`: Interesting. I guess using view source on the mail isn't so bad |
| 23:47 | <BenMillard> | jgraham, I guess this will do: http://lists.w3.org/Archives/Public/public-html/2008Mar/0082.html |
| 23:49 | <jgraham> | BenMillard: That sounds reasonable |
| 23:50 | <annevk> | jgraham, Archived-At gives you a URL, at least for w3.org lists |
| 23:50 | <annevk> | oops, didn't scroll down far enough |
