| 00:05 | <eseidel> | annevk: but does getSVGDocment() === contentDocument... |
| 00:05 | <eseidel> | I would expect it to |
| 00:06 | <Hixie> | what do you call an attack that is described in a security vulnerability report |
| 00:06 | <Hixie> | it's not an actual attack |
| 00:06 | <Hixie> | it's not a hypothetical attack |
| 00:06 | <Hixie> | it's a... |
| 00:07 | <annevk> | eseidel, yeah, it should |
| 00:08 | <annevk> | invalid report? |
| 00:08 | <Hixie> | ? |
| 00:08 | <annevk> | nm |
| 00:12 | <Hixie> | you know, i should set up some sort of little exam |
| 00:12 | <Hixie> | that asks questions like: |
| 00:13 | <dveditz> | Hixie: "proof of concept"? |
| 00:13 | <Hixie> | Q. which of the following requirements is incorrect: |
| 00:13 | <Hixie> | 1. a character encoding saying UTF-16 must be treated as UTF-8 |
| 00:15 | <Hixie> | 2. an invalid UTF-8 byte sequence must cause the user agent to informed the user of the presence of an error |
| 00:15 | <Hixie> | 3. the content-type header of the resource must be ignored |
| 00:15 | <dveditz> | can we be typo nazis? |
| 00:15 | <annevk> | dveditz, doesn't matter in this case :) |
| 00:15 | <Hixie> | and people have to get 80% of the questions right before their input is taken into account |
| 00:15 | Hixie | comes off his high horse |
| 00:16 | <Hixie> | dveditz: proof of concept is perfect, thanks |
| 00:16 | <gsnedders> | Hixie: surely both 2 and 3 are incorrect? Unless substituting U+FFFD for that sequence constitutes informing the user; as the content-type header is sometimes respected (and so isn't totally ignored). |
| 00:16 | <dveditz> | it's a std term for non-attacks that demonstrate a vulnerability |
| 00:17 | <gsnedders> | </smartass> |
| 00:17 | <Hixie> | gsnedders: well, i really meant "which of the following requirements must not appear anywhere in the html5 spec" |
| 00:17 | <Hixie> | dveditz: yeah. couldn't recall the term :-) |
| 00:17 | <annevk> | Hixie, could be part of your book ;) |
| 00:17 | <gsnedders> | Hixie: MUST or must? :P |
| 00:18 | <gsnedders> | (yes, I had to do that) |
| 00:18 | <gsnedders> | Hixie: but the answer is therefore 2 |
| 00:18 | <annevk> | you alread closed </smartass>; now you're just being invalid |
| 00:18 | <annevk> | or maybe ill-formed |
| 00:18 | <eseidel> | annevk++ |
| 00:18 | <gsnedders> | annevk: it has an implicit opening element, as before |
| 00:19 | <gsnedders> | s/element/tag/ |
| 00:19 | <Hixie> | good god you guys are all so pedantic. awesome. i approve. |
| 00:20 | SadEagle | asks Hixie about the meaning of "saying" in #1 |
| 00:20 | <Hixie> | clearly you all pass this exam with flying colours |
| 00:21 | <gsnedders> | see, Hixie didn't even state the exam question well enough for me :P |
| 00:21 | <eseidel> | Hixie: are you training your army now? |
| 00:21 | <gsnedders> | now, back to a third of my computer exam (i.e., the coursework), which by now is due in today |
| 00:21 | <Hixie> | eseidel: my minions! mwuhahaha! |
| 00:21 | <Hixie> | er |
| 00:21 | <Hixie> | did i say that out loud |
| 00:22 | <gsnedders> | No. You typed it. |
| 00:22 | <gsnedders> | Actually, I don't know if you did. |
| 00:22 | <gsnedders> | But you certainly typed it. |
| 00:23 | <gsnedders> | You may be mistaking the two media. |
| 00:24 | <Philip`> | Is there any text-to-speech software that can be trained to do evil laughs? |
| 00:30 | <Dashiva> | I don't know, but I know there's a "rim shot" flash |
| 00:30 | <Dashiva> | Always a winner in good company |
| 00:30 | <SadEagle> | Philip`: just ask hixie to send you voice samples |
| 00:30 | SadEagle | wonders about aural CSS for that sort of thing |
| 00:30 | <Dashiva> | laugh-type: mad-scientist; |
| 00:31 | <annevk> | laugh-lang: en-GB--or-was-it-US-x-Hixie; |
| 00:32 | annevk | should get some sleep |
| 00:32 | <annevk> | nn |
| 00:32 | <gsnedders> | nn |
| 00:32 | gsnedders | needs less urgent deadlines |
| 00:32 | <Dashiva> | Is hx taken as language code? Otherwise we could use that |
| 00:33 | <Dashiva> | The whole en-GB-x-hixie thing just leads to copy-paste markup |
| 00:33 | <Dashiva> | It's just a talisman anyhow |
| 00:33 | <Philip`> | gsnedders: I find deadlines are only urgent because I never do any work before they are :-) |
| 00:34 | <gsnedders> | Philip`: heh. I've been too ill to. I normally do spend the entire day doing nitpicking editing, not writing half the damned thing. |
| 00:34 | <kingryan> | Lachy: votelinks have very low adoption |
| 00:34 | <kingryan> | probably not worthy a mention in a reference on HTML |
| 00:46 | <Hixie> | hsivonen? |
| 00:46 | <Hixie> | is http://hsivonen.iki.fi/test/moz/analysis.txt the best and most up to date analysis? |
| 05:50 | <Hixie> | ok |
| 05:50 | <Hixie> | next complaint |
| 05:50 | <Hixie> | could people please tell me what section they're talking about in the body of the message? |
| 05:51 | <Hixie> | i don't have the subject line when i do these mass replies :-) |
| 06:02 | <jruderman_> | maybe you should tweak your software so you do have subject lines |
| 06:02 | <Hixie> | far easier just to complain |
| 06:37 | <Hixie> | i love how i have a dozen algorithms in the spec, each of which uses a different set of typographic conventions to do the same thing |
| 06:37 | <Hixie> | some use subscripts |
| 06:37 | <Hixie> | some use foo[n] syntax |
| 06:37 | <Hixie> | some use elaborate english phrases |
| 07:40 | <hsivonen> | I'm sick and I'm not really here, but http://hsivonen.iki.fi/test/moz/analysis.txt is the latest analysis of that kind (a month old, though) |
| 07:51 | <Hixie> | thanks |
| 08:49 | <Hixie> | i don't understand why my site is doing gzip compression |
| 08:49 | <Hixie> | i have: |
| 08:49 | <Hixie> | <IfModule mod_gzip.c> mod_gzip_on No |
| 08:49 | <Hixie> | </IfModule> |
| 08:49 | <Hixie> | (er, but on three lines) |
| 09:11 | <Hixie> | hm |
| 09:11 | <Hixie> | anyone else tried posting to the whatwg list recently? |
| 09:32 | <Hixie> | what should <embed ===> parse as? |
| 09:39 | <jgraham_> | Hixie: I just tried posting to the whatwg list and got an error response suggesting some disk is full |
| 09:39 | jgraham_ | wonders if putting full python tracebacks in the errors is such as good idea |
| 11:34 | <annevk> | so it seems Garrett's problem comes down to the CSSOM not following IE in standards mode |
| 11:35 | <annevk> | and endorsing what other browsers do with respect to special casing the HTML body element which he considers wrong |
| 12:01 | <Philip`> | Oh, it's a leap year? |
| 12:02 | Philip` | fixes his watch, since it's clearly not designed to cope with the complexities of time |
| 12:03 | Philip` | wonders when Time5 will be done |
| 12:03 | <annevk> | mine actually says FR 29 |
| 12:07 | <Philip`> | Hixie: It seems quite possible that you're using mod_deflate instead of mod_gzip, and so you might want something like RemoveOutputFilter |
| 14:16 | Philip` | sees that http://www.mozilla.org/projects/intl/UniversalCharsetDetection.html contains lots of U+FFFD symbols unless you use a manual override for iso-8859-1 |
| 17:01 | Philip` | doesn't have any interesting content-type conflicts in his data :-( |
| 17:01 | <Philip`> | (I see about 25 sites sending multiple content-types, out of 125K pages, but they're all either sending two the same or are sending one "text/html" and one "text/html; charset=...") |
| 17:08 | <Philip`> | http://www.greenvironment.fi - a dozen sites send exactly those Limerick HTTP headers |
| 17:36 | <gsnedders> | Philip`: horrific. it should be X-Limerick. |
| 17:38 | <gsnedders> | Philip`: it seems to be as common as cneonction headers |
| 17:41 | <Philip`> | Most of the Cneonctions come from imdb.com |
| 17:42 | <Philip`> | but not all |
| 17:42 | <Philip`> | http://labs.google.com/papers.html - Cneonction: close |
| 17:42 | <Philip`> | http://labs.google.com/ - NnCoection: close |
| 17:44 | <Philip`> | Hixie: ^ bug report :-) |
| 17:45 | <Philip`> | gsnedders: I only see Cneonction on 11 different domains |
| 17:51 | <Philip`> | Oh, and 11 is about a dozen, so you were right that it's as common |
| 17:51 | <Philip`> | +/- some huge margin of error in this data |
| 18:09 | <gsnedders> | Philip`: I don't even want to guess how many million percent :P |
| 18:14 | <Philip`> | gsnedders: It's nowhere near that huge an error :-p |
| 18:14 | <Philip`> | at least when "error" means "difference between measured sample mean and true population mean, where population is the ~4.5M things in dmoz.org that the sample was taken from and isn't necessarily representative of "the web"" |
| 18:15 | <Philip`> | but I can't remember how to do the statistics and I can't be bothered to find out now :-) |
| 18:15 | <gsnedders> | if you define the population to be so small, I guess |
| 18:15 | <gsnedders> | Philip`: I'm meant to know for physics :P |
| 18:16 | <Philip`> | Why do you say it's small? |
| 18:16 | <gsnedders> | dmoz.org, and not the entire web |
| 18:17 | <Philip`> | The size of the population doesn't really matter - as long as it's enough times larger than the sample size, there's no benefit in making it much larger again |
| 18:18 | <Philip`> | (whereas there's always a benefit in making the sample size larger) |
| 18:18 | <Philip`> | so the (non)representativeness is the much more significant issue |
| 18:19 | <Philip`> | but that's easy to avoid by just saying "(n+/-e)% of pages from dmoz.org have some property" and if somebody tries to generalise that to the rest of the web then that's their fault and not mine |
| 18:34 | kingryan | is learning that he's "part of the problem" :) |
| 19:04 | gsnedders | laughs |
| 19:04 | <gsnedders> | post to whatwg needs to be moderated: "Too many recipients to the message" |
| 19:06 | <Lachy> | gsnedders, which post are you referring to? |
| 19:20 | <jgraham> | gsnedders: I got that on the content type sniffing thread. OTOH, I didn't mean to send it to whatwg at all |
| 19:33 | <gsnedders> | Lachy: to the whatwg mailing list |
| 20:34 | <annevk> | Hixie, Firefox is sniffing Content-Encoding stuff too now? |
| 20:35 | <Hixie> | no but apparently they are aware of pages they want to render correctly that they are not rendering correctly because of this, and so they plan to change to not be affected by the presence or absensce of that header |
| 20:36 | <annevk> | k |
| 20:50 | <Hixie> | aw, apparently I "haven't the faintest idea of how to write a specification that |
| 20:50 | <Hixie> | will last beyond the current fad-of-the-day in implementations" |
| 20:51 | <Hixie> | i wonder what it means for a specification to last... does a specification last if nobody implements it correctly? |
| 20:51 | <SadEagle> | Hixie: beats writing unimplementable specs |
| 20:51 | <Hixie> | can't disagree with that |
| 20:51 | <annevk> | you don't really want static specs anyway |
| 20:53 | <annevk> | Hixie, fwiw, https://bugzilla.mozilla.org/show_bug.cgi?id=408098 |
| 20:55 | <gsnedders> | Hixie: where did that come from? |
| 20:57 | <Hixie> | latest mail to public-html |
| 20:57 | <gsnedders> | Hixie: re: your mention of my parsing draft, I really need to scrap what I've written already :) |
| 20:57 | <Hixie> | right, time to shower and go to work. bbl. |
| 21:55 | <eseidel> | Hixie: is Acid3 final now? |
| 22:30 | <Hixie> | eseidelDesk: ye, but for some reason it still hasn't been announced |