| 00:00 | <gsnedders> | Dashiva: Plenty with no attributes |
| 00:01 | <gsnedders> | Hmm… I wonder how much would be saved by cutting out the .items call |
| 00:01 | <Philip`> | gsnedders: Optimise it to zero cost by never executing it |
| 00:01 | <Dashiva> | Dunno if you'd save anything by using iteritems |
| 00:02 | <gsnedders> | Philip`: :P |
| 00:02 | <gsnedders> | Dashiva: I wasn't meaning that, but that may work a bit |
| 00:03 | <Philip`> | gsnedders: I'm being mostly serious - that's the best way I've ever found to make things faster |
| 00:03 | <gsnedders> | Philip`: How can I never execute it? |
| 00:03 | <Dashiva> | Is the instance_string_subs list long? |
| 00:04 | <gsnedders> | Dashiva: no |
| 00:04 | <gsnedders> | Dashiva: Normally only three |
| 00:04 | <gsnedders> | Dashiva: Six at most |
| 00:04 | <Dashiva> | Are all elements generic enough to affect all of text, tail and attribs? |
| 00:05 | <gsnedders> | Dashiva: All are to affect text and tail, IIRC |
| 00:06 | <Philip`> | gsnedders: By finding all the places you're executing it, and the redesigning so you don't have to |
| 00:06 | <Philip`> | s/the/then/ |
| 00:06 | <gsnedders> | Philip`: Any way to find whether I need to run it is more expensive (at least for the whole loop) |
| 00:07 | <Dashiva> | gsnedders: Well, I'm guessing the tree elements are your biggest factor, so everything you can cut on processing per node will give a relatively good payoff |
| 00:07 | <Dashiva> | If you can eliminate half the subs for attributes, hey |
| 00:07 | <gsnedders> | Dashiva: .iteritems() makes no diff |
| 00:09 | <Philip`> | gsnedders: You could do something like only run it over a small section of the document where you define that the substitutable text must occur, not the entire thing |
| 00:12 | <gsnedders> | Dashiva: Any condition that eliminates that for loop for attributes makes no difference |
| 00:13 | <Dashiva> | It could be that the actual valid substitutions are dominating |
| 00:13 | <Dashiva> | And then there's nothing you can do |
| 00:13 | <gsnedders> | I think "if identifier in value:" catches most of it |
| 00:13 | <gsnedders> | It's the regex that is the really expensive bit |
| 00:14 | <Dashiva> | Replace regexp with string substitution :P |
| 00:15 | <gsnedders> | Dashiva: That may be possible… |
| 00:24 | <Hixie> | ok all the types are done! |
| 00:34 | <Philip`> | All the typos? |
| 00:34 | <Philip`> | Oh, types |
| 00:40 | <erlehmann> | in there any UA doing webforms 2.0 ? |
| 00:40 | <erlehmann> | is |
| 00:41 | <erlehmann> | or rather, parts of it ? |
| 00:41 | <Hixie> | opera 9 does most of it |
| 00:41 | <Hixie> | safari has had bits of it for a while |
| 00:41 | <Hixie> | other than that not much |
| 00:42 | <Hixie> | well, autocomplete is in most of them, and the stuff from wf1 is everywhere, of course |
| 00:43 | <erlehmann> | opera is nice :) |
| 00:47 | <Philip`> | erlehmann: All browsers are nice, in their own special ways :-) |
| 00:48 | <erlehmann> | Philip`: while i dislike opera for not being free, few others have done so well regarding standards |
| 01:01 | <erlehmann> | IE has web forms ? i am shocked and apalled. |
| 01:16 | <Hixie> | well it was a proprietary extension that we co-opted |
| 01:16 | <Hixie> | the spec source contains the text ="">^(?:</ |
| 01:16 | <Hixie> | that's... wrong |
| 01:20 | <erlehmann> | how far go the regular expressions ? |
| 01:20 | <Hixie> | ? |
| 01:20 | <erlehmann> | bah, common grammar mistake (i'm not really awake) |
| 01:21 | <erlehmann> | can i use only basic or also extended regexes ? |
| 01:21 | <Hixie> | in what? |
| 01:22 | <Hixie> | oh for pattern=""? |
| 01:22 | <Hixie> | you can use whatever ECMA262 defines |
| 01:22 | <erlehmann> | thx. i'll look it up. |
| 01:26 | <erlehmann> | the spec apparently doesn't say anything about that |
| 01:27 | <erlehmann> | but for the record, wikipedia says ECMA262 has something derived from perl regexes which aren't really POSIX |
| 01:31 | <erlehmann> | oops, my fault, i had read only the second edition - the third edition has regexes in 15.10 |
| 01:32 | <othermaciej> | ECMA-262 does define the regexp syntax |
| 01:36 | <erlehmann> | othermaciej: as i said, i got hold of an old PDF that was missing aforementioned section |
| 01:37 | <othermaciej> | all righty |
| 01:38 | <erlehmann> | sadly, that old pdf is the first search result in google and not even hosted by ECMA |
| 11:53 | <annevk2> | Is the Chromium team planning to replace certain parts of WebKit over time? E.g. http://code.google.com/p/chromium/issues/list?can=2&q=label:newhttp |
| 11:55 | <fakeolliej> | annevk2: webkit doesn't provide a network stack |
| 11:56 | <fakeolliej> | annevk2: it expects the platform to provide those, just as it expects the platform port to provide the front end API |
| 12:01 | <annevk2> | ah, so it's similar to Skia? ta |
| 12:02 | <fakeolliej> | annevk2: chrome effectively is its own platform |
| 12:02 | <fakeolliej> | annevk2: a la Qt |
| 12:02 | <fakeolliej> | annevk2: i have a dream that one day they may contribute to the engine itself |
| 12:03 | <fakeolliej> | annevk2: but my experience with adobe and nokia has lowered my expectation for such things :-/ |
| 12:23 | <annevk2> | CSS Namespaces Module in CR, a test suite for it is ready and I was never forced to work on either. That I dislike namespaces is such a myth :) |
| 12:28 | <hsivonen> | how did the WebKit XBL2 & MathML SoC projects go? |
| 12:30 | <annevk2> | if they were a success there would probably have been a blog post... |
| 12:31 | <annevk2> | afaik the Web Forms 2.0 project only landed autofocus, bit disappointing overall |
| 12:33 | <hsivonen> | where did Chrome's old http come from? |
| 12:35 | <gsnedders> | hsivonen: It's new as well |
| 12:35 | <annevk2> | hsivonen, heh, good question :) |
| 12:38 | olliej | reappears |
| 12:38 | <olliej> | sorry this system wasn't set to autojoin whatwg |
| 12:38 | <olliej> | :D |
| 12:38 | <olliej> | hsivonen: cpst's project morphed into the general "make jsc faster" |
| 12:38 | <olliej> | rather than the pattern matching stuff |
| 12:38 | <olliej> | or tile matching |
| 12:38 | <olliej> | which is what frostbyte actually does |
| 12:39 | <olliej> | frostbyte == khtml's new js engine |
| 12:40 | <olliej> | the svg filters guy accomplished very little alas (which is not surprising given the level he was at -- i need to understand the us education system better i don't know what any of the crazy names mean) |
| 12:40 | <olliej> | keishi got the firebug api's going |
| 12:40 | <olliej> | even doing the work to jsc to make line number info get passed in appropriately |
| 12:40 | <olliej> | the xbl one didn't really get to completion |
| 12:41 | <olliej> | i think that's it |
| 12:46 | <annevk2> | olliej, wasn't there Web Forms 2.0 and MathML as well? |
| 12:47 | <olliej> | i don't think we had sufficient slots for the mathml person |
| 12:47 | <olliej> | and yes there was a webforms 2 guy |
| 12:47 | <olliej> | however based on my complete absence of rememberance i suspect he may have disappeared into the ether |
| 12:48 | <hsivonen> | olliej: :-( that doesn't sound good |
| 12:48 | <olliej> | hsivonen: yeah |
| 12:48 | <olliej> | i assume he failed |
| 12:48 | <olliej> | but i honestly don't know for sure |
| 12:48 | gsnedders | should get around to learn C++ so he can take part in two years time |
| 12:48 | <gsnedders> | :P |
| 12:48 | <gsnedders> | *learning |
| 12:48 | <olliej> | gsnedders: you cna work on the inspector |
| 12:48 | <olliej> | html+js+css |
| 12:49 | <gsnedders> | olliej: That doesn't help me learn C++ though :) |
| 12:49 | <hsivonen> | gsnedders: I had hello world -level C++ experience when I started my summer coding |
| 12:50 | gsnedders | notes the first item on his summer to-do list was ship Anolis 1.0 |
| 12:50 | <gsnedders> | That still hasn't happened. |
| 12:50 | <olliej> | hsivonen: you were GSoC type person? |
| 12:51 | <hsivonen> | olliej: I was in 2006, but it was not Google's but a thing imitating SoC |
| 12:51 | <olliej> | ah righto |
| 12:51 | gsnedders | is too young for GSoC |
| 12:52 | <hsivonen> | (I had much better payment terms that what one gets with SoC) |
| 12:52 | <olliej> | :D |
| 12:52 | <olliej> | $/LoC? ;D |
| 12:53 | <hsivonen> | no, checkpoints instead of all or nothing at the end |
| 12:54 | <olliej> | GSoC has a midway checkpoint |
| 12:54 | <gsnedders> | olliej: Like, "Does it do less than when you started?" |
| 12:54 | <hsivonen> | oh. is that a new thing? |
| 12:54 | <hsivonen> | (I thought they didn't have checkpoints in 2005) |
| 12:56 | <olliej> | hsivonen: maybe -- i haven't done anything in previous years |
| 13:15 | <gsnedders> | Lachy: You still want to get me on standardssuck.org for HTTP5? |
| 13:30 | <erlehmann> | about SVG in HTML - obviously <object> doesn't suffice, but why ? |
| 13:31 | <erlehmann> | gsnedders: standardssuck.org is true to its name and has only flash video :) |
| 13:32 | <gsnedders> | erlehmann: Actually, FLV is fully documented nowadays :P |
| 13:34 | <erlehmann> | gsnedders: Adobe Flash Player isn't |
| 13:35 | <gsnedders> | erlehmann: Sure, but you can just get the FLV and download that :P |
| 13:36 | <gsnedders> | And SWF is open too |
| 13:36 | <erlehmann> | you bet how often i look around in the source code ... |
| 13:37 | <gsnedders> | erlehmann: ? |
| 13:38 | <erlehmann> | i meant it often comes to me that i look through the source of a web site to find the embedded FLV's url |
| 13:38 | <gsnedders> | erlehmann: ah |
| 13:38 | <gsnedders> | Well, get someone to implement <http://www.adobe.com/devnet/swf/pdf/swf_file_format_spec_v9.pdf>; :P |
| 13:39 | <olliej> | gsnedders: open source != standard :D |
| 13:39 | <gsnedders> | olliej: What open source impls are there? There is no open source impl. or standard! :P |
| 13:39 | <gsnedders> | Just a spec! :P |
| 13:40 | <olliej> | gsnedders: isn't there swfdec or some such? |
| 13:40 | <gsnedders> | olliej: Nothing useful in the real world :P |
| 13:40 | <olliej> | righto |
| 13:40 | olliej | would much rather flash just died |
| 13:40 | <gsnedders> | Partial impls, but no full impls |
| 13:40 | <gsnedders> | Me too, but it won't. |
| 13:40 | <olliej> | but i dont' see that happening in the near future |
| 13:40 | <olliej> | yup |
| 13:41 | <olliej> | like IE it is a bain on our existence |
| 13:41 | <olliej> | (not that gecko and webkit haven't left bits of cruft all over the web) |
| 13:41 | <erlehmann> | gsnedders: gnash ppl already said that this docs don't help |
| 13:41 | gsnedders | wants Web 2.0 :P |
| 13:42 | <erlehmann> | gsnedders: you are wrong here. this channel is the web 3.11 workgroup |
| 13:42 | <gsnedders> | erlehmann: Nah, we're just documenting Web 1.0 |
| 13:43 | <erlehmann> | also, embracing and extending it ;) |
| 13:44 | <gsnedders> | erlehmann: sssh. You don't mention that. The log watchers might see! |
| 13:45 | <erlehmann> | CAN I HAZ PLACEHOLDER ATTRIBUTE ? |
| 14:05 | <Lachy> | gsnedders, yes |
| 14:05 | <gsnedders> | Lachy: at TPAC? |
| 14:05 | <Lachy> | sure |
| 14:06 | <Lachy> | you can do it with either annevk2 or myself |
| 14:06 | <Lachy> | or marcos, if he'll be there |
| 15:50 | jgraham | wonders what stuff he needs to pack |
| 16:16 | Philip` | wonders how the canvas colour discussion has gone from "rgb("+r+","+g+","+b+")" being long and hard to type, to suggesting that people write { red: 0, green: 0.3, blue: 0.6, alpha: 1.0 } instead |
| 16:19 | <Philip`> | At least one of the people writing CSS3 Color clearly had a pro-HSL agenda, so I guess they wouldn't be happy at an API that was specifically for RGB(A) and non-extensible |
| 17:00 | <erlehmann> | ♖ ♘ ♗ ♕ ♔ ♗ ♘ ♖ |
| 17:00 | <erlehmann> | ♙ ♙ ♙ ♙ ♙ ♙ ♙ |
| 17:00 | <erlehmann> | ♙ |
| 17:00 | <erlehmann> | ♟ ♟ ♟ ♟ ♟ ♟ ♟ ♟ |
| 17:00 | <erlehmann> | ♜ ♞ ♝ ♛ ♚ ♝ ♞ ♜ |
| 17:11 | <jcranmer> | fixed-width chess pieces are hard to see well |
| 17:12 | <jcranmer> | besides, white's supposed to go first |
| 17:16 | <Philip`> | Depends on whether your IRC client has black-on-white text |
| 17:17 | <Philip`> | but then I think the king and queen would be the wrong way around |
| 17:17 | <gsnedders> | Philip`: yeah, they are |
| 17:20 | <Philip`> | Not that it makes any practical difference in the gameplay, but it's the principle that matters |
| 17:33 | <erlehmann> | mine has a) black on white and b) proper unicode support |
| 18:01 | <gsnedders> | Hixie: Do you have GT4? |
| 18:10 | <jcranmer> | erlehmann: I have white on black and proper unicode support |