/irc-logs / w3c / #webapps / 2009-05-27 / end
Options:
- # Session Start: Wed May 27 00:00:00 2009
- # Session Ident: #webapps
- # [00:00] * Joins: Marcos (Marcos@84.215.160.79)
- # [00:02] * Quits: aroben (aroben@17.203.12.32) (Ping timeout)
- # [00:07] * Joins: aroben (aroben@17.244.10.207)
- # [00:08] * Joins: aroben_ (aroben@17.151.119.212)
- # [00:10] * Quits: aroben (aroben@17.244.10.207) (Ping timeout)
- # [00:56] * Quits: Marcos (Marcos@84.215.160.79) (Quit: Marcos)
- # [01:08] * Joins: annevk (opera@94.210.210.44)
- # [01:10] * Quits: aroben_ (aroben@17.151.119.212) (Ping timeout)
- # [01:11] * Joins: aroben (aroben@17.246.18.233)
- # [01:15] * Quits: heycam (cam@124.168.66.131) (Quit: bye)
- # [01:47] * Quits: annevk (opera@94.210.210.44) (Quit: annevk)
- # [01:54] * Joins: heycam (cam@130.194.72.84)
- # [02:05] * Quits: aroben (aroben@17.246.18.233) (Connection reset by peer)
- # [03:51] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [05:04] * Joins: karl (karlcow@128.30.54.58)
- # [07:15] * Joins: phenny (phenny@80.68.92.65)
- # [07:33] * Quits: heycam (cam@130.194.72.84) (Ping timeout)
- # [07:51] * Joins: heycam (cam@130.194.220.215)
- # [08:24] * Joins: annevk (opera@94.210.210.44)
- # [09:30] * Quits: heycam (cam@130.194.220.215) (Quit: bye)
- # [09:32] * Joins: arve (arve@213.236.208.22)
- # [09:38] * Joins: shepazu (schepers@128.30.52.30)
- # [09:53] * Joins: heycam (cam@124.168.66.131)
- # [09:53] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [09:54] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [10:15] * Joins: tlr (tlr@128.30.52.30)
- # [10:33] * Joins: Marcos (Marcos@213.236.208.247)
- # [10:54] * Quits: Marcos (Marcos@213.236.208.247) (Quit: Marcos)
- # [10:56] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Tomorrow to fresh woods, and pastures new.)
- # [11:15] * Joins: Marcos (Marcos@213.236.208.22)
- # [11:29] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [12:42] * Joins: ArtB (d0309a43@128.30.52.43)
- # [12:43] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [12:49] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [13:16] <arve> oh, the widget uri scheme discussion got a whole lot harder
- # [13:17] <arve> there is one case where we can't avoid exposing it externally
- # [13:17] <arve> cross-document messaging
- # [13:18] <tlr> postMessage?
- # [13:18] <arve> yes
- # [13:18] <tlr> postMessage only needs an origin, not a URI
- # [13:19] <arve> the origin attr is a URI, iirc
- # [13:19] <tlr> no
- # [13:19] <tlr> origin can be a URI *or* a non-URI string.
- # [13:19] <tlr> it's the latter for things like sandboxed iframes.
- # [13:19] <tlr> or script from data: URIs
- # [13:20] <arve> The origin attribute represents, in server-sent events and cross-document messaging, the origin of the document that sent the message (typically the scheme, hostname, and port of the document, but not its path or fragment identifier).
- # [13:20] <tlr> of course, the other question is how one widget gets hold of another widget's window object in the first place
- # [13:20] <tlr> *typically*
- # [13:20] <arve> tlr: not between widgets
- # [13:20] <arve> between web pages and widgets
- # [13:20] <arve> <iframe> inside a widget
- # [13:20] <tlr> right
- # [13:21] <tlr> in which case things should work out perfectly nicely
- # [13:21] <anne> .origin would just be the empty string
- # [13:21] <tlr> ???
- # [13:21] <tlr> no
- # [13:21] <anne> for widgets, sure
- # [13:21] <tlr> no
- # [13:21] <tlr> for widgets, it would be a random string
- # [13:21] <tlr> that's the point
- # [13:21] <anne> unique id becomes empty string per spec
- # [13:21] <tlr> whoops, where's that?
- # [13:21] <arve> would you ever want to know that the origin was a widget?
- # [13:22] <arve> and not some other, undiscoverable origin
- # [13:22] <tlr> step back, more explicit
- # [13:22] <anne> oh sorry
- # [13:22] <anne> it's no longer the empty string
- # [13:22] <anne> it's "null" now
- # [13:23] <tlr> anne, ref?
- # [13:23] <anne> http://www.whatwg.org/specs/web-apps/current-work/#ascii-serialization-of-an-origin
- # [13:23] <tlr> thx
- # [13:23] * tlr woah, safari 4 beta doesn't behave that nicely on really large documents
- # [13:24] <arve> anne mind pointing to the multipage version?
- # [13:24] <anne> could be that http://www.whatwg.org/specs/web-apps/current-work/#unicode-serialization-of-an-origin applies here but the result is the same
- # [13:24] <anne> section 6.4
- # [13:25] <tlr> http://www.whatwg.org/specs/web-apps/current-work/multipage/browsers.html#origin
- # [13:26] <tlr> http://www.whatwg.org/specs/web-apps/current-work/multipage/comms.html#posting-messages
- # [13:26] <tlr> anne, the serialization isn't relevant for postMessage
- # [13:26] <tlr> as postMessage refers to the same origin definition
- # [13:26] <tlr> http://www.whatwg.org/specs/web-apps/current-work/multipage/browsers.html#same-origin
- # [13:27] <tlr> If A and B are both opaque identifiers, and their value is equal, then return true.
- # [13:27] <tlr> therefore, as long as the iframe follows the pattern of just calling postMessage back with the .origin attribute it was handed in the first place, all should be fine
- # [13:28] <arve> does 'null' qualify as an opaque identifier in this case?
- # [13:28] <tlr> null is what you get when you serialize
- # [13:28] <tlr> it isn't the opaque identifier itself
- # [13:28] <tlr> so the one thing that possibly needs clarification is whether, when ".origin" is passed into postMessage, an opaque identifier survives
- # [13:28] <tlr> That's a genuine HTML5 spec question, and needs solving there.
- # [13:29] <arve> tlr: the origin then is a guid
- # [13:30] <tlr> arve, yes
- # [13:30] <tlr> my point is that it doesn't need to be tied to the widget URI scheme
- # [13:30] <tlr> (one of the reasons why I don't like that tying is that this would be a URI scheme for which you can never write down an absolute URI -- violates principle of least surprise)
- # [13:31] <arve> except if you want to have interaction between web and widget, and authoratively at least know that something is a widget
- # [13:31] <tlr> in that case, you'd probably want to know a bit more about the widget
- # [13:31] <tlr> like, from whom it came
- # [13:32] <tlr> Adam Barth tells me that Mozilla is playing with chrome extensions and doing interesting things with putting public key fingerprints into origins.
- # [13:32] <tlr> I think he'll send a note to public-webapps about that soon.
- # [13:32] <arve> tlr: chrome extensions as in "Compatible with Google chrome"?
- # [13:33] <tlr> as in Mozilla chrome extensions
- # [13:33] <arve> or as in "unspecified extensions for the browser chrome, a la jetpack"
- # [13:33] <tlr> don't know about compatible with Google chrome, and had parsed this as unrelated to it
- # [13:33] <anne> tlr, are you saying no to me again?
- # [13:33] <tlr> I understood this to be the latter
- # [13:33] <tlr> anne, on what this time?
- # [13:33] <tlr> ;-)
- # [13:33] * anne is getting a bit annoyed
- # [13:33] * tlr chuckles
- # [13:34] <arve> what I really don't get with this origin bit is
- # [13:34] <anne> you did read the postMessage definition right?
- # [13:34] <tlr> yes
- # [13:34] <anne> and how it creates an event and sets the origin attribute?
- # [13:34] <arve> the origin attribute is specified as a DOMString, then it _will_ be null for comparison purposes
- # [13:34] <arve> (which is anne's point)
- # [13:35] <anne> specifically: "the origin attribute must be set to the Unicode serialization of the origin of the script that invoked the method"
- # [13:35] <tlr> argh
- # [13:35] <tlr> overlooked that piece, and suspect that it's a bug in html5
- # [13:36] <anne> you know, maybe I'll say no now
- # [13:36] <tlr> ;-)
- # [13:36] <tlr> look, the reason for targetOrigin is to make sure that the target window of a postMessage hasn't been navigated elsewhere
- # [13:37] <anne> yup
- # [13:37] <tlr> for the origin attribute on the MessageEvent, you've got two purposes:
- # [13:37] <tlr> 1. feed into targetOrigin when responding. In this case, it's fine to just pass through an opaque origin
- # [13:38] <tlr> 2. compare to some string that you know of. In this case, you'll actually want to do the serialization
- # [13:38] <tlr> so, what I'm saying is that origin is cast to DOMString too early in the spec
- # [13:38] <tlr> since you break the ability to pass messages back to anything that had a synthetic origin
- # [13:38] <anne> actually, you want do that the other way around
- # [13:38] <tlr> why?
- # [13:39] <anne> you're not going to reply before knowing if you trust them
- # [13:39] <tlr> depends on the use case
- # [13:39] <tlr> you might very well be exposing a public API
- # [13:39] <anne> in that case you'd use "*"
- # [13:39] <tlr> not necessarily
- # [13:39] <anne> why not?
- # [13:40] <tlr> I might very well want to send a response to the precise guy from whom the question came, to not leak their private information to a third party
- # [13:40] <tlr> without caring who they are
- # [13:40] <tlr> being able to pass through origin even when it's an opaque identifier gives me that ability
- # [13:40] <anne> interesting point
- # [13:40] <tlr> casting to DOMstring earlier breaks it
- # [13:40] <tlr> therefore, spec bug
- # [13:40] * Quits: heycam (cam@124.168.66.131) (Ping timeout)
- # [13:40] <anne> not necessarily
- # [13:41] <anne> only if opaque id origins are considered relevant enough
- # [13:41] <tlr> (and yes, you're totally right about the "compare to some known URI" use case for having a string representation -- not disputing that)
- # [13:42] <anne> so basically you want .origin to become some kind of Origin object that stringifies to "null" or an actual origin
- # [13:43] <tlr> yes
- # [13:43] <anne> for quite a minor use case
- # [13:43] <anne> feel free to go for it :)
- # [13:43] <tlr> lol
- # [13:46] * Joins: heycam (cam@210.84.19.113)
- # [13:59] <tlr> http://lists.w3.org/Archives/Public/public-html/2009May/0478.html
- # [14:43] * Joins: taf2 (taf2@38.99.201.242)
- # [16:00] <ArtB> shepazu, how about capturing http://www.w3.org/mid/4A1C76D9.9040607@w3.org in the Widgets V2 UC + Reqs doc -> http://www.w3.org/2008/webapps/wiki/Widgets2_UC%26R
- # [16:23] * ArtB is now known as ArtB_
- # [17:17] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
- # [17:39] * Quits: arve (arve@213.236.208.22) (Client exited)
- # [17:42] * ArtB_ is now known as ArtB
- # [18:06] * Quits: annevk (opera@94.210.210.44) (Quit: annevk)
- # [18:08] <Marcos> Artb, so what was the next priority?
- # [18:09] * Marcos prints A&E and DigSig, and arms himself with a red pen!
- # [18:17] <ArtB> Marcos, A+E!
- # [18:17] <ArtB> let's get that puppy to LC as soon as we can
- # [18:17] * Marcos is on it!
- # [18:18] <ArtB> And don't worry about that A+E Editor named Arve -> just do the right thing :-)
- # [18:18] <Marcos> :)
- # [18:29] * Quits: Marcos (Marcos@213.236.208.22) (Quit: Marcos)
- # [18:51] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [19:02] * Joins: darobin (darobin@82.233.247.234)
- # [19:05] * Joins: aroben (aroben@17.246.18.201)
- # [19:29] * Joins: msJacobRossi (836b0053@128.30.52.43)
- # [19:31] * Quits: msJacobRossi (836b0053@128.30.52.43) (Quit: CGI:IRC)
- # [19:32] * Joins: jrossi (836b0053@128.30.52.43)
- # [19:40] * Quits: ArtB (d0309a43@128.30.52.43) (Quit: CGI:IRC (EOF))
- # [19:46] * Quits: darobin (darobin@82.233.247.234) (Quit: darobin)
- # [20:03] * Joins: annevk (opera@94.210.210.44)
- # [20:35] * Joins: shepazu (schepers@128.30.52.30)
- # [20:36] * Quits: aroben (aroben@17.246.18.201) (Quit: aroben)
- # [20:47] <jrossi> shepazu, are we meeting?
- # [20:48] <shepazu> jrossi: no, sorry, I'm at Google I/O
- # [20:48] <shepazu> next week, both chaals and I will be available
- # [20:48] <shepazu> and I'll update the document :(
- # [20:48] <jrossi> A couple of our other folks are there as well.
- # [20:48] <jrossi> Thanks!
- # [20:48] <shepazu> np
- # [20:48] <jrossi> (Travis says hi.)
- # [20:49] <shepazu> howdy, travis :)
- # [20:50] <jrossi> Good luck at Google I/O. We're going offline...
- # [20:54] * Quits: jrossi (836b0053@128.30.52.43) (Quit: CGI:IRC)
- # [20:57] * Quits: annevk (opera@94.210.210.44) (Quit: annevk)
- # [21:11] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
- # [21:14] * Joins: ArtB (d0309a43@128.30.52.43)
- # [21:21] * tlr is now known as tlr-off
- # [22:29] * Quits: ArtB (d0309a43@128.30.52.43) (Quit: CGI:IRC (EOF))
- # [22:32] * Joins: aroben (aroben@17.246.18.201)
- # [22:40] * Joins: arve (arve@84.202.133.45)
- # [23:31] * Joins: Marcos (Marcos@213.236.208.22)
- # Session Close: Thu May 28 00:00:00 2009
The end :)