| 00:29 | <Hixie> | ok the problem i dug myself into is now definitely tractable |
| 00:29 | <Hixie> | i might be able to check this in soon even |
| 00:30 | <Hixie> | only 8 known remaining issues with the patch |
| 00:30 | <Dashiva> | Where did you find all the dirt needed to fill in the hole? |
| 00:31 | <Hixie> | i found that my aforementioned mine support structure could be used to build an above-group structure as well |
| 00:36 | <Philip`> | Hixie: About magic sequences: What if I have e.g. a perfectly sensible plain text file (which should be opened in a text editor or displayed directly in the browser) and a perfectly sensible CSV file (which should be opened in a spreadsheet program), and they contain identical bytes? |
| 00:39 | <Philip`> | or what if I have a zip file containing some arbitrary data that should be downloaded and saved to disk, and an ODF file (which is structurally just a zip file) that should be opened in a word processor? |
| 00:46 | <Hixie> | shepazu: yt? |
| 00:47 | <Hixie> | shepazu: will DOM3 Events have Event.returnValue ? |
| 00:47 | <Hixie> | Philip`: ODF came after the web, so it could have been designed with this structure in mind |
| 00:48 | <Hixie> | Philip`: i agree that csv files are a minor problem |
| 00:48 | <Hixie> | but it's not like what we have now is without problems either |
| 00:49 | <Hixie> | imho the problems of magic sequences are far less serious that today's problems |
| 00:49 | <Hixie> | and would mostly have gone away through common practice (e.g. a CSV2 format that has a blank line at the top that says "CSV") by now |
| 00:50 | <Hixie> | (and text/csv wasn't registered til 2005) |
| 00:52 | Hixie | tries to work out how to deal with the differences between DOM3 Events and what is needed for 'beforeunload' |
| 01:22 | Hixie | gets a headache trying to work out how to spec all the legacy cruft in http://mxr.mozilla.org/mozilla-central/source/dom/src/events/nsJSEventListener.cpp (specifically ::HandleEvent) |
| 01:39 | <Hixie> | hmmmmmm |
| 01:39 | <Hixie> | shepazu: nevermind, looks like I can get away with just having returnValue on a custom Event object |
| 01:44 | <Hixie> | hmmm... yes.... yes... |
| 01:44 | <Hixie> | have a custom Event object with a mustable returnValue |
| 01:44 | <Hixie> | and onbeforeunload would set it if it hadn't been set already |
| 08:42 | Hixie | checks in a multi-1000 line patch to the spec |
| 09:33 | <Hixie> | i'm totally addicted to http://tetrisapp.appspot.com/ |
| 09:33 | <Hixie> | this is getting ridiculous |
| 09:42 | <macnetic> | hi |
| 09:42 | <Hixie> | hi |
| 09:42 | <macnetic> | I have a question about cache-control meta tags... |
| 09:42 | <macnetic> | the way I read http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.9.1 , only one of these tags should be used |
| 09:43 | <macnetic> | and if private is used, you're basically telling the browser it's okay to cache, right? |
| 09:44 | <macnetic> | so if you first use no-cache, and then private, it wouldn't be surprising if the browser caches the page? |
| 09:45 | <Hixie> | well |
| 09:46 | <Hixie> | no-cache doesn't mean it shouldn't be cached |
| 09:46 | <Hixie> | it means it shouldn't be used without checking with the server first |
| 09:47 | <ap> | note that the browser can still put the page into b/f cache, which is not affected by HTTP headers |
| 09:47 | <Hixie> | that too |
| 09:50 | <macnetic> | b/f cache? |
| 09:51 | <Hixie> | back-forward |
| 09:51 | <macnetic> | ic |
| 09:51 | <macnetic> | d'oh |
| 09:52 | <macnetic> | but am I right that private will override other tags that's supposed to avoid caching? |
| 09:53 | <Hixie> | i can't see anything in http that specifies that |
| 09:53 | <macnetic> | hmm |
| 09:53 | <macnetic> | ok |
| 09:53 | <macnetic> | thanks |
| 09:53 | <Hixie> | http is notoriously vague about edge cases |
| 09:53 | <Hixie> | not as bad as html4, but still |
| 09:53 | <macnetic> | heh |
| 09:54 | <macnetic> | so implementations may vary, then? |
| 09:54 | <macnetic> | it just seems odd to tell the browser it can cache the page when you're trying to avoid caching... |
| 09:55 | <macnetic> | the scheduling software we use generates time table web pages |
| 09:55 | <macnetic> | when it was upgraded the browser we use starting caching the pages |
| 09:56 | <macnetic> | I thought maybe the addition of a private tag to the already existing no-cache tag could be the explanation... |
| 09:56 | <Hixie> | implementations do vary, yeah |
| 09:56 | <Hixie> | if you want to avoid caching just set Cache-Control: no-cache |
| 09:56 | <macnetic> | I would |
| 09:57 | <macnetic> | but this is shrinkwrap software :( |
| 09:57 | <Hixie> | ah well |
| 09:57 | <Hixie> | file a bug with your upstream provider :-) |
| 10:13 | <macnetic> | that's what I'm trying to do, but when I quoted the standard and gave my interpretation of no-cache/private/public, they just replied that they'd checked our server and found the clock was 20 minutes off, and when they corrected that the web pages seemed to update properly... |
| 10:14 | <macnetic> | even though we'd previously said that it showed info for _yesterday_ |
| 10:14 | <macnetic> | how they made the connection between 20 minutes and yesterday I don't quite get, but... :-\ |
| 10:14 | <macnetic> | thanks anyway! :) |
| 10:22 | <Hixie> | sorry couldn't help more! |
| 10:25 | <Hixie> | nn |
| 11:06 | <MikeSmith> | had lunch with a guy today who told me (in the context of learning/mastering Japanese), "The only way to truly learn a new language is to write a dictionary for the language." |
| 11:13 | MikeSmith | wonders where zcorpan my be |
| 11:14 | annevk5 | can see him |
| 11:15 | <annevk5> | he's disabled today |
| 11:15 | <annevk5> | no laptop |
| 11:15 | <annevk5> | same for jgraham |
| 11:15 | annevk5 | doesn't know why |
| 11:19 | <MikeSmith> | annevk5: if/when you see him today, please tell him I finally got around to splitting up the input and button stuff |
| 11:19 | <MikeSmith> | http://www.w3.org/html/wg/markup-spec/#input.text-toc |
| 11:19 | <MikeSmith> | http://www.w3.org/html/wg/markup-spec/#button.submit-toc |
| 11:19 | <MikeSmith> | fwiw |
| 11:20 | <annevk5> | i'll tell him to read archives |
| 11:58 | <gpy> | annevk5: yo whats up? |
| 12:50 | <annevk5> | gpy, not much really :) |
| 12:50 | <annevk5> | gpy, standards meeting in Sweden |
| 12:56 | <gpy> | ah |
| 12:56 | <gpy> | where |
| 12:56 | <gpy> | im in dk |
| 12:58 | <annevk5> | Linköping |
| 13:02 | <MikeSmith> | gpy: do the restaurants in Denmark still serve tons of cigarette smoke along with every meal? |
| 13:02 | <MikeSmith> | that's one thing I really like about Denmakr |
| 13:02 | <MikeSmith> | you don't have to buy your own cigarettes |
| 13:03 | <MikeSmith> | you can just go to any restaurant and enjoy everybody else's cigarette smoke |
| 13:05 | <gpy> | annevk5: ah okay |
| 13:05 | <gpy> | MikeSmith: im not danish, im swedish, i just work here, wich is good, salarys are higher + a 40% profit in currency ;D |
| 13:07 | <MikeSmith> | gpy: excellent. I bet cigarettes are cheaper there too. |
| 13:07 | <MikeSmith> | in Norway you can buy half a cigarette |
| 13:07 | <MikeSmith> | and it will only cost you 20 kroner |
| 13:09 | <gpy> | MikeSmith: a little |
| 13:12 | <annevk5> | party tonight, then back home tomorrow for a few days :) |
| 13:12 | <annevk5> | well, "days" |
| 14:02 | <MikeSmith> | virtuelv sporting the jailhouse beard |
| 14:02 | <MikeSmith> | http://flickr.com/photos/hendry/3101837007/in/set-72157611171664224/ |
| 14:02 | <MikeSmith> | PHB waving for mercy ("Don't hurt me Arve!") |
| 14:05 | <MikeSmith> | hendry: thanks for posting photos |
| 16:35 | Philip` | doesn't like how Amazon advertises EC2/S3 prices without including tax |
| 16:35 | <Philip`> | It adds an unexpected six cents onto my bill :-( |
| 17:23 | <hsivonen> | r2529 looks like fun |
| 18:20 | <gsnedders> | does anything support the mid URI scheme? |
| 18:21 | <hsivonen> | gsnedders: Outlook? |
| 18:21 | <gsnedders> | ah |
| 18:21 | <hsivonen> | oh. mid. sorry. |
| 18:21 | <hsivonen> | I was thinking of something else |
| 18:22 | <gsnedders> | :) |
| 18:42 | <hsivonen> | Hixie: does Gecko already comply with your latest script execution spec update? |
| 18:44 | <gsnedders> | hsivonen: http://hsivonen.iki.fi/producing-xml/ needs the section on XML 1.0 updated now that 5th ed. is a REC |
| 18:45 | <hsivonen> | gsnedders: yeah, it should instruct people to always target 4th ed. |
| 20:11 | <Hixie> | you should all play tetrisapp.appspot.com |
| 20:11 | <Hixie> | hsivonen: no |
| 20:11 | <Hixie> | hsivonen: only IE does |
| 20:11 | <Hixie> | hsivonen: other browsers allow you to tall into scripts that are "frozen" in the spec |
| 20:12 | <hsivonen> | Hixie: any other differences? |
| 20:13 | <Hixie> | the way onbeforeunload fires around document.open() wasn't very interoperable |
| 20:14 | <Hixie> | and in general this whole area is not very interoperable in the edge cases |
| 20:14 | <hsivonen> | finding the differences should be fun... |
| 20:25 | <Hixie> | hahahahahahaha |
| 20:25 | <Hixie> | lastweekinhtml5 is funny |
| 20:25 | <Hixie> | given that the whatwg has more non-american involvement than american involvement |
| 20:39 | <jwalden> | hm? where's the "amerocentric" note in <http://blog.whatwg.org/this-week-in-html-5-episode-15>;? |
| 20:39 | <smedero> | http://lastweekinhtml5.blogspot.com/ |
| 20:39 | <smedero> | (you're looking in the wrong place. :-D) |
| 20:41 | <jwalden> | oh, "last" week |
| 20:43 | <jwalden> | man, this is just weird |