Options:
- # Session Start: Tue Aug 16 00:00:00 2011
- # Session Ident: #css
- # [00:17] <fantasai> TabAtkins: So do you want me to resend those flex() comments to www-style for archiving?
- # [00:18] <fantasai> TabAtkins: Also, curl is available for Windows, too. And not available by default on many *nix systems
- # [00:19] * Quits: jdaggett (jdaggett@110.4.237.25) (Quit: jdaggett)
- # [00:19] <TabAtkins> fantasai: Yeah, go ahead.
- # [00:19] <TabAtkins> fantasai: Feel free to edit the wiki page accordingly.
- # [00:21] * Parts: stearns (anonymous@192.150.22.5)
- # [00:21] * Joins: stearns (anonymous@192.150.22.5)
- # [00:23] <fantasai> TabAtkins: Ok, I cut out the flex(<length>) comment since the old draft had it that way
- # [00:38] * Joins: homata (homata@58.158.182.50)
- # [00:39] * Joins: karl (karlcow@128.30.54.58)
- # [00:43] <fantasai> TabAtkins: What do you think of replacing attr(bgcolor, color, white) with attr(bgcolor as color, white)
- # [00:43] <fantasai> ?
- # [00:45] <TabAtkins> fantasai: I don't have a strong opinion. Seems fine to me. I know you're crusading for more examples of "function arguments use arbitrary CSS syntax". ^_^
- # [00:45] <fantasai> That and I can never remember which argument goes first otherwise. ^_^
- # [00:45] <TabAtkins> Good point. Yay for grammar!
- # [00:46] <TabAtkins> You mean between the type and default value, right?
- # [00:46] <fantasai> all of them really
- # [00:46] <fantasai> This way I think of it as a fallback list
- # [00:46] <TabAtkins> Well, it seems pretty natural to have the attr name first. But otherwise, yeah, this feels pretty good.
- # [00:46] <fantasai> first try "bgcolor's value (parsed as a color)" then try "white"
- # [00:46] <TabAtkins> Does anyone implement attr yet?
- # [00:46] <fantasai> in full? dunno
- # [00:47] <fantasai> but content: attr() is CSS2.1
- # [00:47] <TabAtkins> Oh, right, if all they do is accept an attr name, then this change is compatible.
- # [00:52] * Quits: karl (karlcow@128.30.54.58) (Quit: Freedom - to walk free and own no superior.)
- # [00:54] * Quits: tantek (tantek@70.36.139.219) (Quit: tantek)
- # [01:48] * Joins: arronei (arronei@131.107.0.72)
- # [01:49] * Quits: arronei_ (arronei@131.107.0.103) (Ping timeout)
- # [02:13] * Joins: tantek (tantek@159.63.23.38)
- # [02:15] <tantek> TabAtkins we could also use functions for the typecasting
- # [02:15] <tantek> e.g. attr(color(bgcolor))
- # [02:18] <tantek> and then simply define how color() parses for a <color> value the same as the 'color' property http://www.w3.org/TR/css3-color/#colorunits
- # [02:18] <tantek> that's consistent with say, the url() function
- # [02:18] <TabAtkins> tantek: That doesn't work generally - for example, you can't use url() to typecast to a url type.
- # [02:18] <tantek> which parses a string into a URL
- # [02:18] <tantek> hah - we both used the same example to prove different points - funny
- # [02:18] <TabAtkins> Heh.
- # [02:19] <TabAtkins> It would also mean that we can't ever introduce color() for anything else, since it's already defined as taking an arbitrary keyword representing an attribute.
- # [02:19] <tantek> yeah that would be annoying
- # [02:19] <tantek> oh wait
- # [02:19] <tantek> would prefer that color() parsed a *string*
- # [02:19] <tantek> rather than a reference to an attribute
- # [02:20] <tantek> hmm… so...
- # [02:20] <tantek> color(attr(bgcolor))
- # [02:20] <tantek> there, that's better
- # [02:20] <TabAtkins> attr(bgcolor as color) seems pretty clear too. ^_^
- # [02:20] <tantek> url(attr(background))
- # [02:21] <TabAtkins> I think that violates the 2.1 grammar.
- # [02:21] <TabAtkins> Possible the core grammar.
- # [02:21] <tantek> functions can't call functions?
- # [02:21] <TabAtkins> url() can't, I think.
- # [02:21] <tantek> pretty sure core grammar allows for nested ()
- # [02:21] <TabAtkins> General functions definitely can, yes.
- # [02:21] <tantek> also pretty sure core grammar doesn't define specific functions
- # [02:22] <TabAtkins> http://www.w3.org/TR/CSS21/syndata.html
- # [02:22] <TabAtkins> It certainly does define url() specially. ^^
- # [02:22] <TabAtkins> Because url() has magical tokenization rules.
- # [02:22] <tantek> 2.1 does yes
- # [02:22] <TabAtkins> No, that's the core grammar.
- # [02:22] <TabAtkins> 2.1 grammar is in appendix G.
- # [02:24] <tantek> wow: "User agents may vary in how they handle invalid URIs or URIs that designate unavailable or inapplicable resources. "
- # [02:24] <tantek> ?!?
- # [02:24] <tantek> it's been too long since I edited this chapter
- # [02:24] <TabAtkins> That just means that, for example, CSS doesn't define how to display an invalid url.
- # [02:24] <tantek> like maybe 5-6 years?
- # [02:24] <tantek> "handle invalid URIs"
- # [02:25] <tantek> could be enough wiggle room to support attr()
- # [02:25] <tantek> or not - if implementations already try to load example.com/attr()
- # [02:25] <tantek> frustrating
- # [02:26] <TabAtkins> url(attr()) is an invalid url token already.
- # [02:30] <TabAtkins> (Unescaped parens are not allowed in the URI token.)
- # [02:36] <tantek> that's actually good then!
- # [02:36] <tantek> it means existing implementations would ignore url(attr(background)) as an illegal value
- # [02:36] <tantek> *should*
- # [02:36] <TabAtkins> If you think the direction you're going is good. ^_^
- # [02:36] <tantek> :)
- # [02:36] <tantek> well rather, it provides the opportunity should we decide to take it
- # [02:37] <tantek> I'm more brainstorming out loud here
- # [02:37] <tantek> "… as color" seems ok too
- # [02:37] <tantek> just trying to solicit consideration of options
- # [02:38] <TabAtkins> Ah, kk.
- # [02:38] <tantek> however regarding the original / existing syntax
- # [02:38] <tantek> attr(attrname, type, default)
- # [02:38] <tantek> I too have difficulties remembering the order
- # [02:38] <tantek> so yeah - we should improve on that
- # [03:05] * Quits: homata (homata@58.158.182.50) (Quit: Leaving...)
- # [03:18] * Joins: homata (homata@58.158.182.50)
- # [03:19] * Quits: stearns (anonymous@192.150.22.5) (Ping timeout)
- # [03:52] * Quits: nimbupani (Adium@24.18.47.160) (Client exited)
- # [04:15] * Quits: dbaron (dbaron@63.245.220.240) (Ping timeout)
- # [04:18] * Joins: stearns (anonymous@50.132.9.217)
- # [04:36] * Joins: nmccully (nmccully@192.150.22.5)
- # [04:37] * Parts: nmccully (nmccully@192.150.22.5)
- # [05:20] * Joins: dbaron (dbaron@173.228.28.25)
- # [05:26] <tantek> hey TabAtkins - if you're still there, I'm willing to add element(#ID) as an additional at-risk syntax for the nav-* directional properties to be consistent, but the existing #ID syntax should stay as that's already been interoperably implemented.
- # [05:26] <tantek> [CSS3-UI]
- # [05:27] <tantek> note that this would be making an exception for element(#ID), as any other features not currently implemented are being punted to CSS4-UI
- # [05:29] <TabAtkins> Who's already implemented it?
- # [05:34] <TabAtkins> Eh, I'm heading home now (but continue talking if you want - I'll see the scrollback in the morning). I don't think it's a big deal, but I know that Webkit is having some problems trying to implement it as-written, because hashes get turned into colors at a fairly low level.
- # [05:42] <tantek> Opera has implemented it for years (since early 2000s), as has/did Tasman (which is apparently now shipping in MSTV set top boxes), and from my understanding, TV set top box browsers in general all have this now.
- # [06:02] * Joins: nimbupani (Adium@24.18.47.160)
- # [06:03] * Quits: dbaron (dbaron@173.228.28.25) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [07:03] * Quits: tantek (tantek@159.63.23.38) (Quit: tantek)
- # [09:12] * Joins: Ms2ger (Ms2ger@91.181.135.59)
- # [10:10] * Quits: nimbupani (Adium@24.18.47.160) (Ping timeout)
- # [10:17] * Joins: nimbupani (Adium@32.173.62.108)
- # [10:19] * Parts: nimbupani (Adium@32.173.62.108)
- # [11:17] * Joins: Martijnc (Martijnc@84.192.44.100)
- # [11:24] * Quits: homata (homata@58.158.182.50) (Quit: Leaving...)
- # [11:28] * Joins: lhnz (lhnz@188.223.83.48)
- # [11:50] <Bert> fantasai, the webmasters proposed Thursday for CSS Speech.
- # [11:50] * Quits: Martijnc (Martijnc@84.192.44.100) (Quit: Martijnc)
- # [11:51] * Joins: Martijnc (Martijnc@84.192.44.100)
- # [12:38] * Quits: Ms2ger (Ms2ger@91.181.135.59) (Ping timeout)
- # [12:52] * Joins: Ms2ger (Ms2ger@91.181.160.131)
- #
- # Session Start: Tue Aug 16 14:45:31 2011
- # Session Ident: #css
- # [14:45] * Now talking in #css
- # [17:34] * Quits: anne (annevk@83.85.115.123) (Client exited)
- # [17:34] * Joins: anne (annevk@83.85.115.123)
- # [17:36] * Joins: dbaron (dbaron@173.228.28.227)
- # [17:48] * Joins: nimbupani (Adium@24.18.47.160)
- # [18:54] * Joins: tantek (tantek@70.36.139.219)
- # [18:55] <fantasai> Bert: Ok, cool
- # [18:56] * Quits: tantek (tantek@70.36.139.219) (Connection reset by peer)
- # [18:57] * Quits: dbaron (dbaron@173.228.28.227) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [19:09] * Quits: stearns (anonymous@50.132.9.217) (Quit: stearns)
- # [19:17] * Quits: Ms2ger (Ms2ger@91.181.40.134) (Ping timeout)
- # [19:20] * Joins: Ms2ger (Ms2ger@91.181.40.134)
- # [19:27] * Joins: dbaron (dbaron@63.245.220.240)
- # [19:45] * Joins: stearns (anonymous@192.150.22.5)
- # [20:03] * Quits: dbaron (dbaron@63.245.220.240) (Ping timeout)
- # [20:05] * Joins: dbaron (dbaron@63.245.220.240)
- # [22:29] * Quits: Ms2ger (Ms2ger@91.181.40.134) (Quit: nn)
- # [22:52] * Quits: Martijnc (Martijnc@84.192.44.100) (Quit: Martijnc)
- # [22:54] * Joins: plinss_ (plinss@192.6.114.30)
- # [23:26] * Parts: nimbupani (Adium@24.18.47.160)
- # [23:26] * Quits: anne (annevk@83.85.115.123) (Client exited)
- # [23:27] * Joins: anne (annevk@83.85.115.123)
- # [23:40] * Quits: anne (annevk@83.85.115.123) (Client exited)
- # [23:42] <dbaron> fantasai, I'm having trouble understanding the property grammar for http://dev.w3.org/csswg/css3-writing-modes/#unicode-bidi given that it has one more ] than it does [
- # [23:43] * Joins: anne (annevk@83.85.115.123)
- # Session Close: Wed Aug 17 00:00:00 2011
The end :)