/irc-logs / w3c / #webapps / 2009-04-13 / end
Options:
- # Session Start: Mon Apr 13 00:00:00 2009
- # Session Ident: #webapps
- # [00:00] * Joins: Lachy (Lachlan@93.158.28.130)
- # [00:00] * Quits: Lachy (Lachlan@93.158.28.130) (Client exited)
- # [00:06] * Joins: Lachy (Lachlan@93.158.28.130)
- # [01:27] * Quits: Lachy (Lachlan@93.158.28.130) (Quit: This computer has gone to sleep)
- # [04:20] * Quits: shepazu (schepers@128.30.52.30) (Client exited)
- # [04:20] * Joins: shepazu (schepers@128.30.52.30)
- # [04:23] * Quits: nico (nico@133.27.247.181) (Ping timeout)
- # [04:27] * Joins: nico (nico@133.27.247.181)
- # [04:41] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [08:48] * Joins: Lachy (Lachlan@93.158.8.176)
- # [09:27] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Tomorrow to fresh woods, and pastures new.)
- # [09:39] * Quits: Lachy (Lachlan@93.158.8.176) (Quit: Leaving)
- # [09:50] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [12:36] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Tomorrow to fresh woods, and pastures new.)
- # [12:49] * Joins: ArtB (d0309a43@128.30.52.43)
- # [13:00] * Quits: heycam (cam@210.84.37.90) (Ping timeout)
- # [13:06] * Joins: heycam (cam@124.168.119.97)
- # [13:19] * Quits: smaug (chatzilla@82.181.146.100) (Ping timeout)
- # [13:30] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [13:35] * Joins: smaug (chatzilla@88.192.219.38)
- # [13:58] <ArtB> Marcos, what is the security risk in http://lists.w3.org/Archives/Public/public-webapps/2009AprJun/0018.html ?
- # [13:59] <Marcos> Artb, "The problem in a nutshell is that by allowing multiple signatures,
- # [13:59] <Marcos> >> which is something we want to do, we create a situation in which not
- # [13:59] <Marcos> >> all of a signed widget's files are covered by the signature"
- # [14:01] <Marcos> This might be ok, so long as you cannot sneak additional XML elements into the top signature
- # [14:01] <ArtB> how could that be done in practice?
- # [14:01] <ArtB> that is, how can someone tamper with the sig files?
- # [14:02] <Marcos> e.g., (not itself signed) <signature> .... <script> evil = xhr()...</script> </signature>
- # [14:02] <Marcos> then, at runtime, you use XHR to read the signature
- # [14:02] <Marcos> extract the script
- # [14:02] <Marcos> I know, it's all a bit silly
- # [14:03] <Marcos> One can effectively change the top level signature because it is not itself signed
- # [14:04] <ArtB> When I see "It's not a big hole and the attacks require some "assistance" fromdevelopers, but unless there's a reason not to it should be pretty easyto close. "
- # [14:04] <ArtB> I don't get the sense this is critical to address
- # [14:05] <Marcos> I agree. I think we need to push Mark a bit more on it.
- # [14:05] <Marcos> I'm not too phased by it, which is why I did not put the text into the spec.
- # [14:06] <Marcos> However, if we can see it's a pretty easy security hole to close, then we should close it
- # [14:06] <ArtB> do you have any feedback from the people that would need to implement the requirement you proposed?
- # [14:06] <Marcos> no
- # [14:07] <Marcos> apart from Arve, saying, "WTF is this?" :)
- # [14:07] <ArtB> I wonder about an impl where things get unpacked such that this would not be an easy req to address
- # [14:07] <Marcos> I don't follow
- # [14:08] <Marcos> ok, I kinda do follow
- # [14:08] <Marcos> visualizing where such a thing could be a prob
- # [14:08] <ArtB> doesn't this say that at runtime, the UA must make sure the app never opens a sig file?
- # [14:08] <Marcos> exactly
- # [14:09] <ArtB> Yet the sig files may have been unpacked and placed in the FS somewhere outside of the zip
- # [14:09] <Marcos> that is also correct
- # [14:09] <ArtB> what if an impl wanted to delete the sig files after the zip is installed
- # [14:10] <Marcos> why would it want to do that?
- # [14:10] <ArtB> s/is installed/is verified/
- # [14:10] <ArtB> space
- # [14:11] <Marcos> seems like a strange thing to do. the imp would have to be pretty sure that another app could not then go in and modify the files
- # [14:11] <ArtB> but I guess if the sig files were deleted after initial verification this wouldn't be a prob :)
- # [14:12] <Marcos> that's true, but does not seem like a very smart thing to do. Seems better that if the UA detects that files have been changed or modified, it should revalidate
- # [14:12] <ArtB> is it your expectaion that a signed widget must be verified every time the widget is run?
- # [14:12] <Marcos> no, but it's probably a good idea to verify then every so often
- # [14:12] <ArtB> why?
- # [14:13] <Marcos> s/then/them.... in case they were changed via other means
- # [14:13] <Marcos> e.g., a widget somehow found another security hole
- # [14:13] <Marcos> and changed bank.wgt's content by adding a script
- # [14:14] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Tomorrow to fresh woods, and pastures new.)
- # [14:14] <Marcos> engines should protect their widgets. it would be like one web page changing the cookies of another web page... or worst
- # [14:15] <Marcos> i don't think sigs should be seen as one offs
- # [14:15] <ArtB> so you expect the UA to allow widgets to share file space with other widgets?
- # [14:15] <Marcos> yes, so long as they are careful to not allow those widgets to be tempered with
- # [14:15] <Marcos> s/those widgets/the content of those widgets/
- # [14:16] <ArtB> seems like a UA should effectively create a separate sandbox for each widget
- # [14:17] <Marcos> right
- # [14:18] <Marcos> and the UA could use the sig as the way of sandboxing
- # [14:19] <Marcos> I gotta go.. can we pick this up tomorrow
- # [14:19] <Marcos> ?
- # [14:20] <ArtB> l8r
- # [14:22] <Marcos> I need to read Frederick's responses again
- # [14:22] <Marcos> lets put in on the agenda for this week for discussion
- # [14:22] <Marcos> is that ok?
- # [14:22] <ArtB> I'll responde to your latest email; would prefer to close this before the meeting
- # [14:22] <ArtB> we got a *ton* of other higher priority things for this week's call
- # [14:23] <Marcos> ok
- # [14:23] <Marcos> sounds good
- # [14:58] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [15:00] * Joins: taf2 (taf2@65.210.82.235)
- # [15:37] <Marcos> Artb, lets talk priorities :)
- # [15:49] <ArtB> hi Marcos
- # [15:49] <Marcos> hi
- # [15:50] <ArtB> Let's start with P+C; OK?
- # [15:50] <Marcos> ok, sounds good
- # [15:50] * ArtB goes to Inbox ...
- # [15:50] * Marcos loads up spec
- # [15:51] * ArtB looks for darobin :-(
- # [15:51] <ArtB> <access> is one open issue
- # [15:51] <ArtB> Robin proposed http://www.w3.org/mid/24A3E8BD-03B4-43FB-88D0-45600872E25F@berjon.com
- # [15:51] <ArtB> I presume you can live with Robin's proposal, right?
- # [15:52] <Marcos> I can... I will work with him on it over the next 3 weeks. He was finishing off some other client projects last week. He will now be 100% full time on this
- # [15:52] <Marcos> this = <access>
- # [15:53] <ArtB> how close is Robin's proposal to what's in the P&C ED?
- # [15:53] <Marcos> give me a minute to re-read his proposal...
- # [15:53] <Marcos> just so I'm clear... it's been a while
- # [15:54] <ArtB> I think this is an area where new related work will be done in the context of the security work that follows-on from last December's workshop
- # [15:54] <Marcos> artb, it's fairly aligned
- # [15:56] <Marcos> like robin says, it just needs to be written in "spec-ese"
- # [15:59] * Marcos checks open issues DB
- # [16:00] <ArtB> Do you know if BONDI has any inputs on the <access> model?
- # [16:00] <Marcos> I can't comment on that
- # [16:00] <Marcos> because I don't know
- # [16:01] <ArtB> I was just wondering about the probability of any substantial new inputs
- # [16:02] <ArtB> I'll put this on the April 16 widgets agenda
- # [16:02] <Marcos> it is highly unlikely
- # [16:02] <Marcos> they will propose something new
- # [16:02] <ArtB> do you see this "something new" coming as a proposal for the Security WG?
- # [16:02] <ArtB> I think this is what Thomas and Nick mean by "API Bindings" in ...
- # [16:03] <ArtB> http://www.w3.org/2008/security-ws/report
- # [16:03] <ArtB> s/API Bindings/API Patterns/
- # [16:03] <ArtB> but I'm not sure
- # [16:06] <Marcos> having a look
- # [16:06] <Marcos> I don't think there will be anything new coming
- # [16:07] <Marcos> I think our APIs and Bondi's ones align
- # [16:08] <ArtB> Seems like we want an extremely simple model for v1
- # [16:09] <ArtB> I'm pretty sure TLR and the new Security WG are going to spend a lot of time in this area ...
- # [16:09] <Marcos> well, they can do that for 2.0
- # [16:09] <ArtB> but that WG won't even start until late summer at best
- # [16:09] <ArtB> bingo!
- # [16:09] <Marcos> exactly
- # [16:10] <Marcos> I'm happy to see we only have 4 outstanding issues
- # [16:10] <Marcos> in the issues DB
- # [16:10] * Joins: aroben (aroben@71.58.77.15)
- # [16:11] <ArtB> http://www.w3.org/2008/webapps/track/products/8
- # [16:12] <ArtB> what's the status of your I18N document, Marcos?
- # [16:13] <Marcos> Almost done
- # [16:13] <Marcos> hopefully finish today
- # [16:14] <Marcos> I bought myself some coke zero to help me with the job
- # [16:14] <Marcos> :)
- # [16:14] <Marcos> however, it's grown to 21 pages :(
- # [16:14] <Marcos> Hopefully the i18n WG will be able to help us out quickly
- # [16:16] <ArtB> maybe we need a simple albeit feature challenged model for v1 and a plan to add richness for v2
- # [16:17] <Marcos> ArtB: maybe we need to drop it all together
- # [16:17] <Marcos> then bring it back in v2
- # [16:17] <Marcos> it might be ok if we can get agreement quickly
- # [16:18] <Marcos> it's not that complex... there is just lots of ways of doing i18n and each has subtly different effects
- # [16:18] * Quits: smaug (chatzilla@88.192.219.38) (Client exited)
- # [16:19] * Joins: smaug (chatzilla@88.192.219.38)
- # [16:20] <ArtB> when do you expect to complete your proposal?
- # [16:21] <Marcos> hopefully tonight
- # [16:21] <Marcos> if not, tomorrow.
- # [16:22] <Marcos> parts 1,2,3 are done
- # [16:22] <Marcos> and could potentially be reviewed
- # [16:22] <ArtB> please ping me when done and I will review it
- # [16:22] <ArtB> and add it to the agenda
- # [16:22] <Marcos> ok great
- # [16:23] <ArtB> The next P+C open work item is the base folder and xml:base
- # [16:23] <Marcos> that's part of the i18n
- # [16:23] <Marcos> proposal
- # [16:24] <ArtB> ok; I don't rember that but I read the proposal on the 7th
- # [16:24] <Marcos> might have added it after
- # [16:25] <ArtB> does bcp47 sub-lang handling?
- # [16:26] <Marcos> yep
- # [16:26] <Marcos> bcp47 defines the algorithm
- # [16:27] <ArtB> right, sorry, that's what I was asking
- # [16:28] <Marcos> np, I'm having lunch, but have 1 eye on IRC
- # [16:28] <ArtB> are there other issues for P&C?
- # [16:28] * Marcos thinks...
- # [16:28] <Marcos> um, no. that's it for P&C... other stuff is just work that needs to be done
- # [16:29] <Marcos> i.e., once we have a solutions to issues, then they just need to be spec'ed
- # [16:30] <Marcos> that's not a lot of work
- # [16:30] <ArtB> Re A+E, do you have an ETA from Arve on his proposal to address the various Issues identified in the latest ED?
- # [16:31] <ArtB> Storage and preferences are separate ...
- # [16:32] <Marcos> bb in 5... we should ask the guy from access to take over that spec
- # [16:32] <Marcos> he is good
- # [16:32] <ArtB> Marcos, I need to leave for a bit too; TTYL
- # [16:33] <Marcos> np :)
- # [16:51] * Quits: ArtB (d0309a43@128.30.52.43) (Client exited)
- # [18:04] * Joins: ArtB (d0309a43@128.30.52.43)
- # [18:25] * Quits: taf2 (taf2@65.210.82.235) (Ping timeout)
- # [18:38] * Joins: taf2 (taf2@65.210.82.235)
- # [18:42] * Quits: taf2 (taf2@65.210.82.235) (Quit: taf2)
- # [18:46] * Joins: taf2 (taf2@65.210.82.235)
- # [19:27] * Quits: Marcos (Marcos@84.215.160.79) (Quit: Marcos)
- # [19:48] * Joins: Marcos (Marcos@84.215.160.79)
- # [21:03] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [21:07] * Quits: smaug (chatzilla@88.192.219.38) (Ping timeout)
- # [21:20] * Joins: shepazu_ (schepers@128.30.52.30)
- # [21:24] * Quits: shepazu (schepers@128.30.52.30) (Ping timeout)
- # [22:00] * Quits: shepazu_ (schepers@128.30.52.30) (Ping timeout)
- # [22:13] * Joins: shepazu (schepers@128.30.52.30)
- # [22:32] * Quits: ArtB (d0309a43@128.30.52.43) (Quit: CGI:IRC)
- # [23:54] * Quits: taf2 (taf2@65.210.82.235) (Quit: taf2)
- # [23:56] * Quits: aroben (aroben@71.58.77.15) (Connection reset by peer)
- # Session Close: Tue Apr 14 00:00:00 2009
The end :)