/irc-logs / freenode / #whatwg / 2008-01-02 / end

Options:

  1. # Session Start: Wed Jan 02 00:00:00 2008
  2. # Session Ident: #whatwg
  3. # [00:00] <hsivonen> Validator.nu URI scrubbing makes the URI vanish
  4. # [00:00] <Hixie> macdome: informative note in 15.10.2.9 AtomEscape; the text specifically in that section, step 8.2, is the actual text for that
  5. # [00:01] <MacDome> thank you'
  6. # [00:04] <rubys> Phillip`: ok, I got that covered now
  7. # [00:05] * Quits: jgraham (n=james@81-86-215-9.dsl.pipex.com) ("This computer has gone to sleep")
  8. # [00:05] <hsivonen> Philip`: %ef%bf%bf goes unsanitized in Instiki as well. I sent a test URI to Jacques Distler.
  9. # [00:06] * Joins: jgraham (n=james@81-86-215-9.dsl.pipex.com)
  10. # [00:07] <hsivonen> interestingly, the Jena IRI lib blames the character on XML--not an IRI spec: http://validator.nu/?doc=http%3A%2F%2Fabout.validator.nu%2F&schema=%ef%bf%bf
  11. # [00:09] <MacDome> Hixie: thanks, I Updated the bug
  12. # [00:14] <Philip`> hsivonen: http://validator.nu/?doc=http%3A%2F%2Fphilip.html5.org%2Fmisc%2Fchars.html&showsource=yes&out=xml
  13. # [00:21] * Philip` wonders how many enterprisey XML-backended web sites would break horribly if you gave them data with these byte sequences
  14. # [00:27] * Quits: jgraham (n=james@81-86-215-9.dsl.pipex.com) ("This computer has gone to sleep")
  15. # [00:39] * Quits: rubys (n=rubys@cpe-075-182-087-110.nc.res.rr.com) (Read error: 113 (No route to host))
  16. # [00:44] * Quits: Lachy__ (n=Lachlan@cm-84.215.54.100.getinternet.no) ("Leaving")
  17. # [00:44] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
  18. # [00:50] * Quits: tndH (i=Rob@adsl-83-100-132-169.karoo.KCOM.COM) ("ChatZilla 0.9.79-rdmsoft [XULRunner 1.8.0.9/2006120508]")
  19. # [00:56] * Joins: roc (n=roc@121-72-34-26.dsl.telstraclear.net)
  20. # [01:01] * weinig is now known as weinig_
  21. # [01:02] * weinig_ is now known as weinig
  22. # [01:15] * Joins: roc_ (n=roc@121-72-42-231.dsl.telstraclear.net)
  23. # [01:25] * Quits: roc (n=roc@121-72-34-26.dsl.telstraclear.net) (Read error: 110 (Connection timed out))
  24. # [01:35] * Quits: roc_ (n=roc@121-72-42-231.dsl.telstraclear.net)
  25. # [01:39] * Parts: Hemebond (n=Hemebond@ip-118-90-36-41.xdsl.xnet.co.nz)
  26. # [01:48] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
  27. # [02:00] * Joins: heycam` (n=cam@clm-laptop.infotech.monash.edu.au)
  28. # [02:07] * Joins: doublec (n=Chris_Do@203-97-173-6.cable.telstraclear.net)
  29. # [02:45] * Quits: doublec (n=Chris_Do@203-97-173-6.cable.telstraclear.net) ("ChatZilla 0.9.79-rdmsoft [XULRunner 1.8.0.9/2006120508]")
  30. # [02:56] * Joins: G0k (n=hmason@c-67-164-171-32.hsd1.co.comcast.net)
  31. # [03:03] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
  32. # [03:13] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
  33. # [03:16] * Joins: roc (n=roc@121-72-42-231.dsl.telstraclear.net)
  34. # [03:20] * Quits: webben_ (n=benh@91.84.19.165)
  35. # [03:35] * Quits: roc (n=roc@121-72-42-231.dsl.telstraclear.net)
  36. # [04:35] * weinig is now known as weinig|foos
  37. # [04:35] * weinig|foos is now known as weinig|food
  38. # [05:40] * Quits: heycam` (n=cam@clm-laptop.infotech.monash.edu.au) ("bye")
  39. # [05:49] * weinig|food is now known as weinig
  40. # [06:02] * Quits: tantek (n=tantek@70-13-107-175.area2.spcsdns.net)
  41. # [06:41] * Joins: tantek (n=tantek@70-13-201-101.area2.spcsdns.net)
  42. # [06:44] * Parts: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
  43. # [06:44] * Quits: tantek (n=tantek@70-13-201-101.area2.spcsdns.net) (Client Quit)
  44. # [06:57] <G0k> Hixie: so what the status on server sent events?
  45. # [07:01] <Hixie> how do you mean?
  46. # [07:01] <G0k> about it being possibly removed?
  47. # [07:02] <Hixie> it hasn't seen much implementation from other browser vendors than opera, and it's redundant with other features
  48. # [07:02] <Hixie> so it might possibly be removed
  49. # [07:02] <G0k> which other feature makes it redundant?
  50. # [07:02] <othermaciej> we might implement it in WebKit
  51. # [07:02] <othermaciej> you could argue it is redundant with XMLHttpRequest
  52. # [07:03] <Hixie> it's redundant with networking and with multipart xmlhttprequest
  53. # [07:03] <othermaciej> (assuming it is spec'd clearly enough that you can rely on getting multiple responses for separate data chunks)
  54. # [07:03] <othermaciej> not even multipart
  55. # [07:03] <othermaciej> people use regular-type XHR for multiple server-sent messages over one connection
  56. # [07:04] <Hixie> true
  57. # [07:04] <othermaciej> but <event-source> is potentially a nice convenience
  58. # [07:04] <G0k> i have two concerns about it
  59. # [07:04] <Hixie> yeah, the question is should the platform have conveniences? or just the core platform features that can be used to implement the conveniences?
  60. # [07:05] <G0k> i think the DOM event metaphor is really strained
  61. # [07:05] <Hixie> if we do keep it, we'll certainly need to massively simplify it. it's way over-engineered right now.
  62. # [07:05] <othermaciej> it should have conveniences in cases where patterns are very commonly seen in library code
  63. # [07:05] <G0k> and it's impossible to implement backwardsly
  64. # [07:05] <G0k> yeah
  65. # [07:05] <othermaciej> thus the Selectors API
  66. # [07:05] <othermaciej> (and getElementsByClassName)
  67. # [07:05] <Hixie> yep
  68. # [07:06] <Hixie> maybe this is such a case
  69. # [07:06] <Hixie> maybe it isn't
  70. # [07:06] <G0k> hixie: what if we a) make it for Message events only and b) add a way to implement it based on slow download requests so that it can be emulated with XHR for older browsers?
  71. # [07:06] <Hixie> that's why it's marked as "may be removed". cos it may be removed. :-)
  72. # [07:07] <G0k> so like old browsers can add application/x-dom-event-not-a-stream to their Accept headers, and it will just end the connection after every event is sent
  73. # [07:07] <Hixie> G0k: (a) is easy; but i'm not sure i follow (b). If you have a solid proposal, feel free to mail it to the list for consideration.
  74. # [07:07] <Hixie> i wouldn't rely on Accept headers for anything
  75. # [07:07] <Hixie> but yeah
  76. # [07:08] <G0k> why not?
  77. # [07:08] <Hixie> that kind of stuff hasn't historically worked well
  78. # [07:09] <G0k> hm. i suppose we could do it the other way around, if it doesn't have the accept header for the event stream, the server would default to a non-streaming method
  79. # [07:11] <Hixie> i would recommend something that doesn't in any way involve accept or content-type headers working correctly
  80. # [07:12] <Hixie> if the web's history has taught us anything, it's that people don't understand or honour mime types.
  81. # [07:12] <othermaciej> G0k: you don't have to end the connection after every event to work with XHR
  82. # [07:12] <G0k> well even if the accept wasn't formed correctly, the worst thing that would happen is that it would degrade to slower behavior
  83. # [07:12] <othermaciej> G0k: I believe current-generation browsers could emulate server-sent events on top of XHR using the current HTML5 event stream format
  84. # [07:13] <othermaciej> (though I haven't actually tried it)
  85. # [07:13] <G0k> othermaciej: i've found the behavior of partial downloads kinda unpredictable though
  86. # [07:13] <othermaciej> the JS library folks are calling the persistent connection for messages from the server thing "COMET"
  87. # [07:14] <othermaciej> so probably a sensible thing to do would be to find what kind of API they offer as design input for a standards-track way of doing it
  88. # [07:14] <Hixie> COMET is just one implementataion, as i understand it
  89. # [07:14] <G0k> i think it's really critical to be able you can emulate this in older browsers
  90. # [07:14] <othermaciej> I thought COMET was the concept
  91. # [07:14] <G0k> i think comet is a big idea
  92. # [07:14] <G0k> yeah
  93. # [07:14] <othermaciej> http://en.wikipedia.org/wiki/Comet_(programming)#Browser_compatibility
  94. # [07:14] <othermaciej> here's the mechanisms people have used for it
  95. # [07:14] <othermaciej> but I dunno what kind of API is offered
  96. # [07:14] <Hixie> man people have got to stop making up new acronyms for stuff that already has names
  97. # [07:15] <G0k> i've seen many
  98. # [07:15] <othermaciej> wow, lots of info about WebKit in there
  99. # [07:15] <othermaciej> very educational
  100. # [07:16] <G0k> from my basic research, i think the most reliable implementations have used a "slow download with re-download" XHR or iframe method
  101. # [07:16] <G0k> the multipart thing would allow for constant connections, but only gecko really does that
  102. # [07:16] <othermaciej> streaming over a single connection works just as well
  103. # [07:17] <othermaciej> I'm not sure what is better about multipart
  104. # [07:17] <othermaciej> other than handling message boundaries inside the engine
  105. # [07:17] <G0k> i think that's the critical issue
  106. # [07:17] <othermaciej> it's not that hard to make up a protocol for message boundaries
  107. # [07:17] <G0k> i don't think any UAs really make guarantees about when they flush their buffers for XHR reads
  108. # [07:18] <othermaciej> neither WebKit nor Gecko will buffer indefinitely
  109. # [07:18] <othermaciej> I don't know about IE
  110. # [07:18] <G0k> no, but imagine a chat application
  111. # [07:18] <othermaciej> I assume the streaming XHR thing doesn't work in IE or people would use it
  112. # [07:18] <G0k> even if it's 15 seconds, that's bad
  113. # [07:19] <G0k> which is why people (meebo, for instance) use multiple requests
  114. # [07:19] <othermaciej> is it ever actually 15 seconds?
  115. # [07:19] <othermaciej> I don't actually know
  116. # [07:19] <G0k> i'm haven't benchmarked it
  117. # [07:20] <G0k> this wikipedia page seems to suggest that it never fires on webkit until at least 256 bytes are recieved
  118. # [07:20] <othermaciej> anyway it would be good for WebKit to support multipart/x-mixed-replace in XHR (I assume that is the kind people mean), but I am not entirely sure what correct behavior for that is supposed to be
  119. # [07:20] <othermaciej> I guess we'd have to reverse-engineer Gecko
  120. # [07:20] <othermaciej> G0k: I don't think that is true
  121. # [07:20] <othermaciej> older versions used to require some amount of data to be sent to prime the buffer initially
  122. # [07:21] <othermaciej> I think after that (and now right away) it will fire readystate 3 without a floor on bytes received
  123. # [07:21] <othermaciej> (could be wrong, but that's my recollection)
  124. # [07:21] <G0k> ok, but it's highly possible that opera and IE don't even do that
  125. # [07:22] <G0k> and i don't think the XHR specs even specify this
  126. # [07:22] <G0k> also imagine a proxy server
  127. # [07:22] <G0k> actually, don't, that makes no sense
  128. # [07:22] <Hixie> 58 tests done, 42 tests to go.
  129. # [07:22] <Hixie> more than half-way there!
  130. # [07:22] <G0k> woot
  131. # [07:23] <othermaciej> XHR specs don't specify this, but they could
  132. # [07:23] <othermaciej> used to be they didn't specify anything
  133. # [07:23] <G0k> ok but that won't make IE support it
  134. # [07:24] <G0k> i really really feel being able to emulate this kind of thing on old browsers is a splendid idea
  135. # [07:24] <othermaciej> x-mixed-replace is a pretty crappy wire protocol as far as message boundaries go
  136. # [07:24] <G0k> yeah
  137. # [07:25] <G0k> plus it makes server code potentially really complicated
  138. # [07:25] <othermaciej> really? seems no worse than any other server-sent message protocol
  139. # [07:26] <othermaciej> as far as impact on most of your server code
  140. # [07:26] <othermaciej> (presumably on some level the server has an API that says "post this message now" or something)
  141. # [07:27] <G0k> well yeah once you abstract that away
  142. # [07:27] <G0k> but now you have to set up boundaires
  143. # [07:28] <G0k> and escape your data appropriately
  144. # [07:30] <othermaciej> is there escaping?
  145. # [07:30] <othermaciej> I think you just need to make sure to pick a boundary string that won't appear in the message
  146. # [07:30] <othermaciej> which is what sucks
  147. # [07:30] <othermaciej> if there was escaping, you could let some abstraction layer pick the boundary and do escaping
  148. # [07:30] <G0k> yeah so how do you know that before you start sending
  149. # [07:31] <G0k> i mean in this particular case, i think the fact that boundaries start with --- and the dom event stuff doesn't makes it easier
  150. # [07:31] <othermaciej> I guess you can do "escaping" by making sure no line ever starts with --
  151. # [07:34] <G0k> still, this doesn't help us for older UAs that don't support mixed
  152. # [07:35] <othermaciej> no new feature helps for older UAs
  153. # [07:36] <G0k> but i'm saying if we had a mode which used mulitple connections
  154. # [07:36] <G0k> that could emulate server sent events even on UAs which only support single part XHR
  155. # [07:37] <othermaciej> I think you'd be better off using the Htmlfile control in IE and streaming XHR in other browsers
  156. # [07:37] <othermaciej> if what you want to do is put a nicer API on top of existing UA functionality
  157. # [07:38] <G0k> what is the Htmlfile control?
  158. # [07:39] * Joins: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de)
  159. # [08:05] * Quits: G0k (n=hmason@c-67-164-171-32.hsd1.co.comcast.net)
  160. # [08:08] * MacDome is now known as MacDomeSleep
  161. # [08:29] * MacDomeSleep is now known as MacDome
  162. # [08:32] * MacDome is now known as MacDomeSleep
  163. # [08:45] <Hixie> i want to test HTTP somehow in acid3
  164. # [08:45] <Hixie> but i have no idea how
  165. # [08:45] <Hixie> i don't want to do anything fancy on the server side...
  166. # [09:07] <jruderman> Hixie: what aspect of HTTP?
  167. # [09:08] <Hixie> dunno
  168. # [09:08] <jruderman> Hixie: acid 2 tested browsers' recognition of 404 errors quite nicely, i hear ;)
  169. # [09:08] <Hixie> ooh, yeah, could test <object> and 404 stuff
  170. # [09:09] <jruderman> how about <object type> vs server-provided content-type?
  171. # [09:09] <Hixie> that one's controversial
  172. # [09:10] <jruderman> oh, you're trying to stay *away* from controversy? :P
  173. # [09:10] <Hixie> in acid tests? yeah, ideally
  174. # [09:11] <Hixie> acid tests are controversial enough as it is without them risking their credibility
  175. # [09:11] <jruderman> hehe, ok
  176. # [09:11] <Dashiva> They're supposed to be borderline useful, not purely spec nitpicking :)
  177. # [09:11] <jruderman> according to https://bugzilla.mozilla.org/show_bug.cgi?id=395110#c17 safari get the <object type> thing "wrong"
  178. # [09:11] <Hixie> yeah but html5 might make that "right"
  179. # [09:11] <jruderman> so i'm not surprised to hear that it's controversial
  180. # [09:11] <Hixie> it's unclear what's going to happen with that
  181. # [09:11] <jruderman> interesting
  182. # [09:12] <Hixie> html5 already says to ignore content-type for <script>
  183. # [09:12] <Hixie> and <img> has some thing where all image/* types are treated equivalently iirc
  184. # [09:13] <Hixie> content-type, in retrospect, is such a bad idea
  185. # [09:13] <jruderman> why is it a bad idea?
  186. # [09:13] <Hixie> so few people get it right
  187. # [09:13] <Hixie> it gets in the way of people more often than it helps
  188. # [09:14] <Hixie> doesn't really help the human race
  189. # [09:14] <Hixie> we shoudla done everything with unambiguous magic byte sequences
  190. # [09:14] * weinig is now known as weinig|zZz
  191. # [09:41] <othermaciej> Hixie: it's hard to test http in any serious way without using XMLHttpRequest
  192. # [09:42] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
  193. # [09:42] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Client Quit)
  194. # [09:43] * Quits: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de) (Connection timed out)
  195. # [09:44] * Joins: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de)
  196. # [09:44] * Joins: hdh (n=hdh@58.187.94.168)
  197. # [10:00] * Joins: mpt (n=mpt@nat/canonical/x-9a61a2c6958bba26)
  198. # [10:06] * Quits: haavard- (n=hkmoen@pat-tdc.opera.com)
  199. # [10:17] <Hixie> othermaciej: yeah
  200. # [10:22] <Hixie> oh jesus
  201. # [10:22] <Hixie> i try to catch a webkit bug and a gecko bug with one carefully crafted test, and IE7 responds by crashing.
  202. # [10:23] <Hixie> *sigh*
  203. # [10:24] <Lachy> are you trying to avoid including crash bugs in the acid test?
  204. # [10:24] <Hixie> yes
  205. # [10:24] <Hixie> pissing off the media is tactically bad
  206. # [10:24] <Hixie> and crashing a reporter's browser would piss him off
  207. # [10:24] <Hixie> :-)
  208. # [10:25] * Joins: ROBOd (n=robod@89.122.216.38)
  209. # [10:27] <othermaciej> Hixie: obviously you need to make pH 0 acid test of all crash bugs
  210. # [10:27] <Hixie> man that would be a bitch to write :-)
  211. # [10:27] <Hixie> so apparently it's only my copy that crashes
  212. # [10:28] <Hixie> i have two people who've tested IE7 on XP and it didn't crash
  213. # [10:29] <Lachy> oh nice, MS are disabling support for older MS Office formats. http://it.slashdot.org/article.pl?sid=08/01/01/137257 - Hooray for proprietary formats! :-) (too bad if you have some in your archive you need to read one day)
  214. # [10:30] <othermaciej> good thing we're not reading that through an article.xaml or article.swf url then
  215. # [10:30] <Hixie> the last test i added to acid3 shows a really weird bug in firefox
  216. # [10:30] <Hixie> check out the dot in the "i" of "Acid3"
  217. # [10:40] * Joins: webben (n=benh@nat/yahoo/x-6d4b47c4a1290b8d)
  218. # [10:42] <Lachy> Hixie, is the support-a.png image supposed to be a cat?
  219. # [10:44] * Lachy wonders if Hixie would ever use a picture of any other type of animal?
  220. # [10:45] <Lachy> gotta go, back soon
  221. # [10:45] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) ("This computer has gone to sleep")
  222. # [10:49] <Hixie> heh
  223. # [10:49] <Hixie> support-a.png returns a 404 of a cat, yes
  224. # [10:58] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
  225. # [11:01] <Hixie> right
  226. # [11:02] <Hixie> i worked around the IE crash
  227. # [11:02] <Hixie> it still shows the IE error messages
  228. # [11:02] <Hixie> oh wait i might be able to work around that too
  229. # [11:03] <Hixie> yes!
  230. # [11:03] <Hixie> haha!
  231. # [11:03] <Hixie> take that, silly crashing browser that i wasn't even trying to show a bug in!
  232. # [11:03] <Hixie> i managed to make it not crash or show errors without even using anything outside of DOM1 Core!
  233. # [11:03] <Hixie> and without doing anything unjustifiable!
  234. # [11:04] <Hixie> and it still shows the wacky rendering in firefox!
  235. # [11:04] <Hixie> 41 tests to go
  236. # [11:06] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) ("Leaving")
  237. # [11:10] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
  238. # [11:12] * Quits: mpt (n=mpt@nat/canonical/x-9a61a2c6958bba26) (Read error: 113 (No route to host))
  239. # [11:13] * Joins: mpt (n=mpt@nat/canonical/x-e7b7c62b1b9fe400)
  240. # [11:23] <hsivonen> Philip`: hmm. looks like the Xalan serializer is broken. :-(
  241. # [11:24] <hsivonen> Philip`: also, the Sun UTF-8 decoder is bogus for not catching CESU-8
  242. # [11:30] <hsivonen> I wonder if there's somekind of project management theory on when it's worthwhile to write your own XML serializer instead of work around the bugs in someone else's
  243. # [11:40] * Quits: dolphinling (n=chatzill@rbpool1-46.shoreham.net) ("ChatZilla 0.9.79-rdmsoft [XULRunner 1.9b3pre/2007121408]")
  244. # [11:48] <hsivonen> aargh. People should write UTF-8-only serializers. The UTF-8 output of the Xalan serializer is broken because it tries to be too smart about output encodings
  245. # [12:11] * Joins: OmegaJunior (n=ZJr@a82-95-48-162.adsl.xs4all.nl)
  246. # [12:18] <hsivonen> http://java.sys-con.com/read/478303_2.htm
  247. # [12:19] <hsivonen> "HTML5 will eclipse XHTML" according to Microsoft Senior Evangelist
  248. # [12:19] <Hixie> didn't that happen already?
  249. # [12:25] * Joins: jgraham (n=james@81-86-215-9.dsl.pipex.com)
  250. # [12:26] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
  251. # [12:31] <hsivonen> Philip`: https://issues.apache.org/jira/browse/XALANJ-2419
  252. # [12:38] * Quits: OmegaJunior (n=ZJr@a82-95-48-162.adsl.xs4all.nl) (Read error: 110 (Connection timed out))
  253. # [12:39] * Quits: jgraham (n=james@81-86-215-9.dsl.pipex.com) ("This computer has gone to sleep")
  254. # [12:47] <Philip`> hsivonen: Fun
  255. # [12:58] * Joins: tndH_ (i=Rob@adsl-87-102-34-81.karoo.KCOM.COM)
  256. # [12:58] * tndH_ is now known as tndH
  257. # [12:59] * Quits: zcorpan (n=zcorpan@pat.se.opera.com) (Read error: 104 (Connection reset by peer))
  258. # [12:59] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
  259. # [13:01] <hsivonen> Philip`: I could turn my HTML5 serializer into a doctypeless namespaceless XML serializer with relative ease
  260. # [13:01] * Quits: webben (n=benh@nat/yahoo/x-6d4b47c4a1290b8d)
  261. # [13:01] <hsivonen> Philip`: the problem is that to really replace the Xalan serializer, I'd need to write the code that reconstructs the namespace declarations
  262. # [13:02] <hsivonen> and it seems to me that the time it would take could be used for feature work
  263. # [13:06] <Hixie> 40 tests to go
  264. # [13:08] * Quits: zcorpan (n=zcorpan@pat.se.opera.com) (Read error: 104 (Connection reset by peer))
  265. # [13:09] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
  266. # [13:12] <hsivonen> so, should a filter that takes out extension namespaces before validating zap the tags or zap the subtree?
  267. # [13:12] * Joins: jgraham (n=james@81-86-215-9.dsl.pipex.com)
  268. # [13:13] <hsivonen> I'm inclined to say subtree, but I could argue why you'd want to zap just tags
  269. # [13:14] * Joins: roc (n=roc@121-72-34-250.dsl.telstraclear.net)
  270. # [13:14] <Philip`> http://validator.w3.org/check?uri=http%3A%2F%2Fphilip.html5.org%2Fmisc%2Fchars.html&charset=iso-8859-1&output=soap12
  271. # [13:15] <Philip`> Why is well-formedness so totally impossible to get right?
  272. # [13:16] <hsivonen> Philip`: fun. Are you planning on reporting the bug?
  273. # [13:17] <Philip`> hsivonen: I don't really care enough about getting it fixed to bother reporting it
  274. # [13:18] * Philip` thinks someone should make an Apache filter which detects XML responses and tests for well-formedness and if there's a problem it emails the server owner
  275. # [13:28] <hsivonen> Philip`: reported
  276. # [13:32] * Joins: jgraham_ (n=jgraham@81-86-215-9.dsl.pipex.com)
  277. # [13:35] <zcorpan> http://validator.nu/?doc=http%3A%2F%2Fphilip.html5.org%2Fmisc%2Fchars.html&out=xhtml
  278. # [13:36] * Joins: doublec (n=Chris_Do@203-97-173-6.cable.telstraclear.net)
  279. # [13:39] <hsivonen> zcorpan: yeah, known bug. https://issues.apache.org/jira/browse/XALANJ-2419
  280. # [13:41] * Quits: roc (n=roc@121-72-34-250.dsl.telstraclear.net) (Read error: 110 (Connection timed out))
  281. # [13:51] <Philip`> hsivonen: Thanks
  282. # [13:51] <Philip`> (I should probably stop being so lazy...)
  283. # [13:54] * Quits: jgraham (n=james@81-86-215-9.dsl.pipex.com) ("This computer has gone to sleep")
  284. # [13:55] * Joins: jgraham (n=james@81-86-215-9.dsl.pipex.com)
  285. # [14:10] * Joins: gsnedders (n=gsnedder@host86-138-198-209.range86-138.btcentralplus.com)
  286. # [14:13] <krijnh> Ping
  287. # [15:14] * Disconnected
  288. # [15:14] * Attempting to rejoin channel #whatwg
  289. # [15:14] * Rejoined channel #whatwg
  290. # [15:14] * Topic is 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!'
  291. # [15:14] * Set by gsnedders on Tue Dec 18 21:41:19
  292. # [16:15] * Disconnected
  293. # [16:15] * Attempting to rejoin channel #whatwg
  294. # [16:15] * Rejoined channel #whatwg
  295. # [16:15] * Topic is 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!'
  296. # [16:15] * Set by gsnedders on Tue Dec 18 21:41:19
  297. # [16:16] <Philip`> Evil plan:
  298. # [16:16] <Philip`> * Find a project which uses git
  299. # [16:16] <Philip`> * Commit something, and put U+FFFE in the commit message
  300. # [16:17] * MacDome is now known as MacDomeOut
  301. # [16:17] <Philip`> and then nobody can use gitweb any more, unless they use IE
  302. # [16:17] <Philip`> (git says "Warning: commit message does not conform to UTF-8." but it commits anyway)
  303. # [16:17] <hdh> probably put some dummy repo on repo.or.cz
  304. # [16:20] * Quits: hdh (n=hdh@58.187.94.168) (Read error: 104 (Connection reset by peer))
  305. # [16:32] <takkaria> Philip`: why's that a good idea?
  306. # [16:32] * Joins: billmason (n=billmaso@ip156.unival.com)
  307. # [16:34] <Philip`> takkaria: Evil plans aren't meant to be good ideas :-p
  308. # [16:42] * Joins: webben (n=benh@nat/yahoo/x-255a4b2afbd413ef)
  309. # [16:45] * Joins: dbaron (n=dbaron@pool-72-94-185-124.phlapa.fios.verizon.net)
  310. # [16:47] <kig> MacDomeOut: re: tablet, the cheapest intuos 3 is $200, but maybe easier to find someone in vicinity with one..
  311. # [16:52] <zcorpan> hsivonen: your xhtml 1.0 schema seems to allow <noscript> in inline contexts
  312. # [16:56] <hsivonen> zcorpan: did XHTML 1.0 ban that?
  313. # [16:56] <zcorpan> hsivonen: html4 and xhtml1 allows noscript in block contexts only
  314. # [16:56] <hsivonen> zcorpan: ok. thanks
  315. # [16:58] * Quits: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de) (Remote closed the connection)
  316. # [16:58] <zcorpan> <!ENTITY % Block "(%block; | form | %misc;)*"> ... <!ENTITY % misc "noscript | %misc.inline;">
  317. # [16:58] <hsivonen> <script> works as inline, doesn't it?
  318. # [16:58] <zcorpan> yes
  319. # [16:59] <zcorpan> <!ENTITY % Inline "(#PCDATA | %inline; | %misc.inline;)*"> ... <!ENTITY % misc.inline "ins | del | script">
  320. # [17:03] <hsivonen> zcorpan: fix deployed. thanks
  321. # [17:09] <zcorpan> hsivonen: <h1><noscript><p>... still validates for me
  322. # [17:10] <hsivonen> hmm
  323. # [17:16] * Joins: phsiao (n=shawn@nat/ibm/x-06beb976794d8c0d)
  324. # [17:23] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) ("This computer has gone to sleep")
  325. # [17:29] * Joins: aroben_ (n=adamrobe@76.111.161.65)
  326. # [17:29] * Quits: aroben (n=adamrobe@unaffiliated/aroben) (Read error: 104 (Connection reset by peer))
  327. # [17:45] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
  328. # [17:49] * Philip` wonders how best to implement X3D's <Text/>
  329. # [17:50] <Philip`> I guess I have to do it as a textured quad, probably using mozDrawText in Firefox and using drawImage(svg) in Opera, or something like that...
  330. # [17:52] <kig> unless you want to write a freetype font server and ....
  331. # [17:54] * aroben_ is now known as aroben
  332. # [17:56] <Philip`> I'd like to avoid relying on a server, though I think I'll end up requiring one anyway so I can do cross-domain XHR :-(
  333. # [18:04] * Joins: hober (n=ted@unaffiliated/hober)
  334. # [18:23] <Philip`> What is window.Text? (I can't find any documentation of it...)
  335. # [18:23] <Philip`> Oh, looks like http://www.w3.org/TR/DOM-Level-2-Core/core.html#ID-1312295772
  336. # [18:24] <Philip`> Doesn't sound very useful, so I'll overwrite it with my own global Text variable
  337. # [18:37] * Quits: weinig|zZz (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
  338. # [18:45] * Joins: rubys (n=rubys@cpe-075-182-087-110.nc.res.rr.com)
  339. # [18:48] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Remote closed the connection)
  340. # [18:49] * Joins: tantek (n=tantek@68-25-236-75.area2.spcsdns.net)
  341. # [18:52] * Quits: mpt (n=mpt@canonical/launchpad/mpt) ("This computer has gone to sleep")
  342. # [18:54] * Joins: maikmerten (n=maikmert@L926c.l.pppool.de)
  343. # [18:57] * Joins: csarven (n=nevrasc@modemcable130.251-202-24.mc.videotron.ca)
  344. # [19:10] * Quits: webben (n=benh@nat/yahoo/x-255a4b2afbd413ef) (Read error: 110 (Connection timed out))
  345. # [19:16] * Quits: gsnedders (n=gsnedder@host86-138-198-209.range86-138.btcentralplus.com) ("Partying in teh intarwebs")
  346. # [19:18] * Parts: rubys (n=rubys@cpe-075-182-087-110.nc.res.rr.com)
  347. # [19:25] * Parts: zcorpan (n=zcorpan@pat.se.opera.com)
  348. # [19:41] <kig> hohoo, bounding box scaling working. next is snap to original scale and shift-constrained rotate and scale
  349. # [19:41] <Philip`> Well, this would work better if texImage2DHTML didn't always crash when I use it
  350. # [19:42] <kig> ssss
  351. # [19:43] <kig> what is it supposed to do?
  352. # [19:43] <Philip`> It's supposed to not crash
  353. # [19:43] <Philip`> and, preferably, it should do some texture binding stuff
  354. # [19:43] <Philip`> (when given an image or canvas to convert into a GL texture)
  355. # [19:44] <kig> ah
  356. # [19:45] * Quits: aroben (n=adamrobe@unaffiliated/aroben)
  357. # [19:47] <kig> do you have the texture bound?
  358. # [19:48] * Joins: aroben (n=aroben@unaffiliated/aroben)
  359. # [19:48] <kig> since all texImage2DHTML seems to do is get the image surf data and call glTexImage2D with it
  360. # [19:49] <Philip`> I was stupid and forgot to call bindTexture; but now I am calling it and it still crashes :-(
  361. # [19:51] <kig> :I
  362. # [19:51] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
  363. # [19:52] * Philip` tries building a debug version, and hopes he has enough disk space
  364. # [19:53] <kig> paste a snippet?
  365. # [19:54] <kig> (not that crashing the browser as a failure mode is very nice)
  366. # [19:56] * Joins: weinig (n=weinig@17.203.15.140)
  367. # [20:00] * Joins: aroben_ (n=aroben@unaffiliated/aroben)
  368. # [20:03] <Philip`> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!doctype%20html%3E%0D%0A%3Cimg%20src%3Dhttp%3A%2F%2Fphilip.html5.org%2Ftests%2Fcanvas%2Fsuite%2Fimages%2Fred-16x16.png%20id%3Di%3E%3Ccanvas%20id%3Dc%3E%3C%2Fcanvas%3E%0D%0A%3Cscript%3E%0D%0Awindow.onload%20%3D%20function%20()%20%7B%0D%0A%20%20var%20gl%20%3D%20document.getElementById('c').getContext('moz-glweb20')%3B%0D%0A%20%20var%20ids%20%3D%20gl.genTextures(1)%3B%0D%0A%20%20gl.bindTexture(gl.T
  369. # [20:03] <Philip`> Oops, bit long
  370. # [20:03] <Philip`> http://tinyurl.com/28eceb
  371. # [20:03] * Joins: aroben__ (n=aroben@76.111.161.65)
  372. # [20:09] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
  373. # [20:12] <kig> that sure looks like it should work
  374. # [20:14] <Philip`> I was going to look at normal texturing after adding some primitive text, but I guess I'll have to give up on both of those ideas for now :-(
  375. # [20:19] * Quits: aroben_ (n=aroben@unaffiliated/aroben) (Connection timed out)
  376. # [20:20] * aroben__ is now known as aroben
  377. # [20:29] <Philip`> Oh, great, now I get errors when compiling Mozilla on Vista because Vista guesses that nsinstall.exe is a program installer because it has "install" in its name, and so it requires admin privileges which it can't get when running on the command line and so it dies
  378. # [20:34] * Joins: kingryan (n=kingryan@dsl092-219-050.sfo1.dsl.speakeasy.net)
  379. # [20:39] * Quits: hober (n=ted@unaffiliated/hober) (Remote closed the connection)
  380. # [20:41] * Quits: dbaron (n=dbaron@pool-72-94-185-124.phlapa.fios.verizon.net) ("8403864 bytes have been tenured, next gc will be global.")
  381. # [20:43] * Joins: aroben_ (n=aroben@unaffiliated/aroben)
  382. # [20:44] <jruderman> Hixie: https://bugzilla.mozilla.org/show_bug.cgi?id=410460 ;)
  383. # [20:48] <gavin> I filed that yesterday when Hixie was asking about it in #devs
  384. # [20:48] <gavin> he knows about it :)
  385. # [20:58] * Joins: cgriego (n=cgriego@216.138.69.206)
  386. # [20:59] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
  387. # [21:02] * Quits: aroben_ (n=aroben@unaffiliated/aroben) (Connection timed out)
  388. # [21:03] * Joins: Philip`_ (n=philip@zaynar.demon.co.uk)
  389. # [21:04] * Quits: weinig (n=weinig@17.203.15.140)
  390. # [21:08] * Quits: Philip` (n=philip@zaynar.demon.co.uk) (Read error: 110 (Connection timed out))
  391. # [21:09] * Joins: aroben (n=aroben@unaffiliated/aroben)
  392. # [21:19] * Quits: jgraham (n=james@81-86-215-9.dsl.pipex.com) ("This computer has gone to sleep")
  393. # [21:37] * Quits: jruderman (n=jruderma@c-67-180-15-227.hsd1.ca.comcast.net)
  394. # [21:43] * Joins: roc_ (n=roc@202.0.36.64)
  395. # [21:45] * Philip`_ is now known as Philip`
  396. # [21:45] * Quits: maikmerten (n=maikmert@L926c.l.pppool.de) (Remote closed the connection)
  397. # [21:47] * Joins: gsnedders (n=gsnedder@host86-138-198-209.range86-138.btcentralplus.com)
  398. # [21:49] * Joins: jgraham (n=james@81-86-215-9.dsl.pipex.com)
  399. # [21:53] * Quits: ROBOd (n=robod@89.122.216.38) ("http://www.robodesign.ro")
  400. # [21:53] * Quits: cgriego (n=cgriego@216.138.69.206) (Read error: 113 (No route to host))
  401. # [22:08] * Joins: jruderman (n=jruderma@corp-241.mountainview.mozilla.com)
  402. # [22:09] * Joins: weinig (n=weinig@17.203.15.140)
  403. # [22:15] <gsnedders> hmmm. "ANSI" seems to be an alias for US-ASCII in the real world.
  404. # [22:22] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) ("This computer has gone to sleep")
  405. # [22:24] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Read error: 110 (Connection timed out))
  406. # [22:25] * Joins: didymos (i=jho@rapwap.razor.dk)
  407. # [22:34] <gsnedders> No, for windows-1252
  408. # [22:35] <gsnedders> That can't be right :\
  409. # [22:36] * Joins: cgriego (n=cgriego@216.138.69.206)
  410. # [22:37] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
  411. # [22:39] * Joins: doublec (n=Chris_Do@203-97-173-6.cable.telstraclear.net)
  412. # [22:41] <Hixie> hsivonen: http://groups.google.com/group/google-caja-discuss/browse_thread/thread/712c772b93777a64/88dcb5f7799358af?hl=en&q=whatwg#88dcb5f7799358af (search for validator.nu) in case you didn't know
  413. # [22:49] * Quits: doublec (n=Chris_Do@203-97-173-6.cable.telstraclear.net) ("ChatZilla 0.9.79-rdmsoft [XULRunner 1.8.0.9/2006120508]")
  414. # [22:56] * Quits: gsnedders (n=gsnedder@host86-138-198-209.range86-138.btcentralplus.com) ("Partying in teh intarwebs")
  415. # [22:57] * Joins: gsnedders (n=gsnedder@host86-137-236-187.range86-137.btcentralplus.com)
  416. # [22:59] <hsivonen> Hixie: interesting. I wasn't aware. thanks. (however, it isn't called libhtmlparser and it's under the MIT license, not MPL)
  417. # [23:01] <Philip`> (There's no version 1.5.1 either)
  418. # [23:04] * Joins: webben (n=benh@91.84.19.165)
  419. # [23:04] * Quits: webben (n=benh@91.84.19.165) (Client Quit)
  420. # [23:05] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
  421. # [23:05] * Joins: webben (n=benh@91.84.19.165)
  422. # [23:08] * Joins: aroben (n=aroben@unaffiliated/aroben)
  423. # [23:17] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
  424. # [23:40] * Joins: hober (n=ted@unaffiliated/hober)
  425. # [23:42] <Philip`> Hmm, texImage2DHTML doesn't crash now that I've compiled it myself
  426. # [23:42] <Philip`> but gl.uniformi doesn't work because it has unimplemented code
  427. # [23:44] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
  428. # [23:44] <Philip`> and the only way to use textures is by using uniformi
  429. # Session Close: Thu Jan 03 00:00:00 2008

The end :)