| 02:43 | <Hixie> | MikeSmith: hey, neat, can we get the cross-referencing dfn.js set up for the multipage full copy too? |
| 02:43 | <Hixie> | what would i have to add to my pipeline? |
| 02:44 | <MikeSmith> | hey Hixie |
| 02:45 | <MikeSmith> | if you add it to your build that should be all that's needed |
| 02:45 | <MikeSmith> | what I'm doing for the author view is just hacking it in |
| 02:47 | <MikeSmith> | the splitter just copies the dfn.js reference into the split-out files |
| 02:47 | <MikeSmith> | afaict |
| 03:03 | <MikeSmith> | Hixie: I remembered something else I've been wanting to bug you about |
| 03:03 | <MikeSmith> | when you have a chance, please see: |
| 03:03 | <MikeSmith> | http://qa-dev.w3.org:8888/?doc=http%3A%2F%2Fdev.w3.org%2Fhtml5%2Ftests%2Fvalidation%2Ffull%2Finvalid%2Fmissing-attributes%2Flink-missing- |
| 03:03 | <MikeSmith> | href.html |
| 03:04 | <MikeSmith> | oops |
| 03:04 | <MikeSmith> | http://qa-dev.w3.org:8888/?doc=http%3A%2F%2Fdev.w3.org%2Fhtml5%2Ftests%2Fvalidation%2Ffull%2Finvalid%2Fmissing-attributes%2Flink-missing-href.html |
| 03:04 | <Hixie> | oh so dfn.js doesn't actually work cross-file? |
| 03:04 | <MikeSmith> | Hixie: well, no |
| 03:04 | <Hixie> | oh ok |
| 03:04 | <Hixie> | nevermind then |
| 03:04 | <MikeSmith> | no magic -- only works within the |
| 03:04 | <MikeSmith> | same file |
| 03:04 | <Hixie> | i disabled it on the whatwg copy because it wasn't particularly useful without that |
| 03:05 | Hixie | looks at the file |
| 03:05 | <MikeSmith> | yeah, it may be that I should disable it in the multipage author view too |
| 03:05 | <MikeSmith> | see the attributes list in the error message |
| 03:05 | <MikeSmith> | the "(required)" stuff |
| 03:06 | <MikeSmith> | I added that by patching the copy of the HTML5 spec that v.nu uses |
| 03:06 | <MikeSmith> | 10-line change |
| 03:06 | <MikeSmith> | http://pastebin.ca/1734693 |
| 03:07 | <MikeSmith> | so.. I wanted to as if you could consider making that change to your source |
| 03:07 | <MikeSmith> | or something similar |
| 03:07 | <Hixie> | html5 used to have those, i took them out because very few attributes really are required vs optional, most have complicated rules about whether to have them in or not |
| 03:08 | <Hixie> | so i figured it would be misleading to say "required" or not |
| 03:08 | <MikeSmith> | well, for the v.nu case, I think it's useful |
| 03:08 | <Hixie> | e.g. title="" is required on <link>... if rel="stylesheet alternate" |
| 03:08 | <MikeSmith> | right |
| 03:08 | <MikeSmith> | but v.nu catches those cases and reports them |
| 03:09 | <Hixie> | yeah but the spec would be misleading if it said "required" for href="" but not title="" |
| 03:09 | <Hixie> | because title="" isn't optional in all cases |
| 03:09 | <MikeSmith> | ah, I see your point |
| 03:10 | <MikeSmith> | so, OK, it may be that this is really only useful for the v.nu messages |
| 03:10 | <Hixie> | i agree that it's useful in validators, though maybe it should explicitly say what is missing, since i was a bit confused when first looking at the error message |
| 03:10 | <Hixie> | gotta go, bbiab |
| 03:10 | <MikeSmith> | OK |
| 03:11 | <MikeSmith> | but yeah, that's the core problem -- the relaxng/jing backend doesn't give useful error messages about missing attributes |
| 03:11 | <MikeSmith> | the backend doesn't report which attributes are missing |
| 05:57 | <zcorpan> | happy new year |
| 06:03 | <Hixie> | indeed |
| 06:03 | <Hixie> | happy 2010 :-) |
| 06:06 | jtbandes | wants nested CSS, e.g. #main { foo:bar;... p { baz:quux; ...} } == #main { foo:bar; ..} #main p {baz:quux;} |
| 06:09 | <zcorpan> | jtbandes: write a python script to translate it to standard css |
| 06:09 | <jtbandes> | Ew.. |
| 06:10 | <jtbandes> | I could, but that would make it almost as hard to deal with as non-nested CSS |
| 06:11 | <zcorpan> | why? |
| 06:11 | <zcorpan> | debugging gives a different view? |
| 06:11 | <jtbandes> | eh? |
| 06:12 | <zcorpan> | oh nm, i was thinking of dragonfly and similar |
| 07:54 | <Hixie> | othermaciej: any news on the microdata thing? |
| 08:35 | <MikeSmithX> | http://blog.jclark.com/2010/01/xml-namespaces.html |
| 08:45 | <zcorpan> | MikeSmith: any news on xml-stylesheet checking? |
| 08:50 | <zcorpan> | MikeSmith: "^text/xsl(;.*)?$" could be "^text/xsl(;|$)" instead, i think |
| 08:52 | <jtbandes> | Just a tip, with the disclaimer that I know nothing about the context here: isn't it better to use \A \Z rather than ^$ ? |
| 08:54 | <zcorpan> | dunno |
| 08:54 | <zcorpan> | MikeSmith: && should be invalid |
| 09:08 | <othermaciej> | Hixie: will be soon |
| 09:10 | <Hixie> | cool |
| 09:26 | <MikeSmith> | zcorpan: thanks |
| 09:26 | <MikeSmith> | jtbandes: what's \A \Z ? |
| 09:27 | <MikeSmith> | zcorpan: about news, you mean news about getting it deployed on validator.nu? |
| 09:35 | <zcorpan> | MikeSmith: naw i wondered if you've done anything more |
| 09:36 | <MikeSmith> | zcorpan: I think I've got it passing all test cases, and have implemented all the warnings cases |
| 09:36 | <MikeSmith> | I sent the diff to hsivonen for review |
| 09:36 | <zcorpan> | cool |
| 09:37 | <MikeSmith> | but it's a pretty big diff, so I reckon it will take him some time to review it |
| 09:37 | <zcorpan> | i made a few more comments in the bug |
| 09:37 | <MikeSmith> | ah, OK |
| 09:39 | <MikeSmith> | I'll make the chararacter-ref changes |
| 09:39 | <MikeSmith> | and push out to qa-dev when I have it done |
| 09:40 | <MikeSmith> | James Clark's latest blog entry is a good read |
| 09:40 | <MikeSmith> | on namespaces |
| 09:40 | <MikeSmith> | he mentions the whatwg wiki mechanism for rel values |
| 09:41 | <Hixie> | a mechanism which mark is planning on making orders of magnitude more complicated |
| 09:42 | <Lachy> | MikeSmith, where is James Clark's blog? |
| 09:43 | <Lachy> | found it. http://blog.jclark.com/ |
| 09:43 | <MikeSmith> | yep |
| 09:43 | <MikeSmith> | http://blog.jclark.com/2010/01/xml-namespaces.html |
| 09:43 | <MikeSmith> | is the direct link to the post |
| 09:44 | <MikeSmith> | he also favors dotted/reverse DNS syntax for names |
| 09:46 | <Lachy> | it's interesting how in step 2 he claims that "The one true way of naming things on the Web is with a URI." and then step 3 identifies the problem with that, but instead of fixing the problem in step 2, a layer of abstraction is added instead. |
| 09:48 | <MikeSmith> | Lachy: I don't think he's claiming that. I think instead he's just saying that statement seems reasonable to many people |
| 09:48 | <Lachy> | oh, good, he says "But this is where I believe XML Namespaces went wrong." |
| 09:49 | <MikeSmith> | yeah, a good deal of the rest of blog post is really about saying that it turns our #2 is not so reasonable |
| 09:50 | <MikeSmith> | that there needs to be a distinction between how to name things, and how to name types of things |
| 09:51 | <Hixie> | imho #3 and #5 are equally wrong, and #6 in true but shouldn't be on the list given that namespaces as written doesn't address it properly (and the list is supposed to be "how namespaces ended up being as it is today") |
| 10:23 | <zcorpan> | http://socialcomputing291.wordpress.com/2010/01/02/mini-project-on-html5with-css-and-javascript/ "it was fun to do it because HTML 5 is so easy to learn" is good; "I have tried to do the offline saving by using manifest functionality of HTML 5 but no luck." less so |
| 10:42 | <annevk> | weird that abarth would say "You're right, as you always are." |
| 10:42 | <annevk> | I always think he's always right |
| 12:20 | <foolip> | gsnedders: ping |
| 14:00 | <gsnedders> | foolip: pong |
| 14:00 | <gsnedders> | Lachy: you back in Oslo? |
| 14:15 | gsnedders | realizes he can't comment and moderate on one post on /. gah |
| 14:20 | <Lachy> | gsnedders, yes |
| 14:21 | <gsnedders> | Lachy: I CC'd some bugs to you, FYI |
| 14:21 | <gsnedders> | *nudge* |
| 14:21 | <Lachy> | gsnedders, ok. I'll look at those next week |
| 15:33 | <gsnedders> | hmm, I can't push to html5lib |
| 15:35 | <gsnedders> | Just get auth failed. |
| 15:38 | <gsnedders> | Oh, just copying and pasting password failed. |
| 15:39 | <gsnedders> | I've got something stuck on my clipboard. WTF? |
| 16:27 | <cardona507> | so localStorage is tied to my entire domain? meaning that if I have a page in directory A that has something stored in localStorage, that stuff will show up in directory B's localStorage? That is what appears to be happening to me. If so, how would we host multiple apps on a domain without them polluting each other? |
| 16:31 | <annevk> | use a prefix |
| 16:38 | <cardona507> | thanks annevk - what do you mean? |
| 16:40 | <cardona507> | like www2.foo.com? |
| 16:40 | <cardona507> | or foo.bar.com? |
| 16:45 | <Philip`> | cardona507: Use a prefix in the keys that you use when accessing localStorage |
| 16:45 | <cardona507> | thank Philip` - And happy new year everyone |
| 16:45 | <Philip`> | localStorage.a_whatever = 1; localStorage.b_whatever = 2; // no conflicts |
| 16:46 | <cardona507> | nice to know |
| 17:18 | <foolip> | gsnedders: I was going to ask if python html5lib is decent enough if I want to parse a file, modify the DOM a bit and then pretty-print the result |
| 17:19 | <foolip> | obviously yes to step one and two, but the last? |
| 17:42 | <Philip`> | foolip: It has an HTML5 serializer, which should do what you want |
| 17:42 | <Philip`> | and it has various configuration options depending on what you consider pretty |
| 17:43 | <Philip`> | (though it never does anything DOM-altering like inserting whitespace to make the output prettier) |
| 17:44 | <foolip> | hmm, guess I'll have to try then |
| 17:45 | <Philip`> | The multipage spec splitter uses html5lib to parser and modify and print pages, which sounds like what you want |
| 17:45 | <Philip`> | s/parser/parse/ |
| 17:45 | <Philip`> | (Oh, actually it uses lxml to parse because html5lib is painfully slow) |
| 17:46 | <Philip`> | (http://code.google.com/p/html5/source/browse/trunk/spec-splitter/spec-splitter.py) |
| 17:53 | <gsnedders> | (But that has occasionally caused Hixie grief, but seeming it makes stuff so much quicker he's perfectly willing to use lxml to parse) |
| 18:23 | <foolip> | gsnedders: do you have any pretty-priting experience, given that you generate specs in your free time? |
| 18:23 | <foolip> | i.e. does there exist a finished library that is HTML-nostic and knows where it is safe to insert whitespace and not, etc |
| 18:25 | <foolip> | in other news, transforming between different mapping coordinate systems requires a lot of math :/ |
| 18:28 | <Philip`> | gsnedders: Hixie doesn't have a choice of parser with the spec-splitter |
| 18:28 | <Philip`> | since (as far as I'm aware) it's running on my server and I've configured it to use one that doesn't steal all my CPUs |
| 18:39 | <gsnedders> | foolip: Basically the only spec I generate in my spare time is HTML5 in testing Anolis :) |
| 18:39 | <gsnedders> | foolip: and I really don't care about what the output looks like |
| 18:39 | <foolip> | ok |
| 18:39 | <gsnedders> | Philip`: He could nag you |
| 18:39 | <gsnedders> | Philip`: He seems to be quite good at that |
| 18:49 | <Philip`> | gsnedders: I could ignore him |
| 18:49 | <Philip`> | gsnedders: I think I'm quite good at that |
| 18:49 | <Philip`> | foolip: For transforming between different mapping coordinate systems, why not use a GIS library that implements it all already? :-) |
| 19:02 | <foolip> | Philip`: that's plan A, but the math is kind of fun to look into |
| 19:05 | Philip` | had fun using PostgreSQL/PostGIS, since he could simply store latitude/longitude in the database and create an index over the transformation into a square grid system |
| 19:07 | foolip | is trying to help dad convert the swedish RT90 system into something that is useable with google maps |
| 19:31 | <cardona507> | if I want to see the coords that are returned from getCurrentPosition() in my console, can I? I tried navigator.geolocation.getCurrentPosition().position.coords.latitude; - is it possible? |
| 22:21 | <payman`> | gsnedders: ping |
| 22:28 | <foolip> | hey payman`, are you back in lkpg? |
| 22:37 | <payman`> | foolip: no, still in Wien :D |
| 22:39 | <foolip> | having a good time? |
| 22:39 | <foolip> | when are you back? |
| 22:40 | <payman`> | yes, everything is cool! :) i will be back on jan 5th |
| 22:41 | <gsnedders> | payman`: pong |
| 22:42 | <payman`> | gsnedders: backpong! |