/irc-logs / w3c / #webapps / 2009-03-12 / end
Options:
- # Session Start: Thu Mar 12 00:00:00 2009
- # Session Ident: #webapps
- # [00:19] * Quits: aroben (aroben@71.58.77.15) (Connection reset by peer)
- # [00:34] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [00:45] * Quits: Hixie (ianh@129.241.93.37) (Quit: trying a new configuration)
- # [00:45] * Joins: Hixie (ianh@129.241.93.37)
- # [02:23] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [02:23] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [06:15] * Quits: heycam (cam@130.194.72.84) (Quit: bye)
- # [07:14] * Joins: heycam (cam@124.168.80.126)
- # [07:50] * Quits: gsnedders (gsnedders@86.136.52.180) (Quit: gsnedders)
- # [08:05] * Quits: timeless (timeless@65.75.195.122) (Ping timeout)
- # [08:33] * Joins: gsnedders (gsnedders@86.136.52.180)
- # [09:32] * Joins: annevk (opera@213.13.106.1)
- # [09:32] * Quits: gsnedders (gsnedders@86.136.52.180) (Quit: gsnedders)
- # [09:38] * Joins: tlr (tlr@128.30.52.30)
- # [10:39] * Joins: arve (arve@213.236.208.22)
- # [10:43] * Quits: arve (arve@213.236.208.22) (Quit: Ex-Chat)
- # [10:45] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [10:57] * Joins: arve (arve@213.236.208.22)
- # [11:16] <Marcos> Annevk, in CORS, Cross Domain Requests are per resource, right? that is, the example.com/ says allow foo.com, but that does not mean that foo.com can also read example.com/a
- # [11:17] <annevk> yes
- # [11:17] <Marcos> ok, cool. I'm just wondering how we can use CORS in widgets
- # [11:17] <Marcos> given that widgets have no origin
- # [11:18] <annevk> that makes things tricky, yeah
- # [11:20] <Marcos> Annevk, what if we used widget's ID as the origin?
- # [11:20] <Marcos> <widget id="http://microsoft.com"> ?
- # [11:21] <annevk> that seems kind of tricky as there's no way to guarantee actual uniqueness on that attribute...
- # [11:22] <Marcos> So, in CORS, the request origin is guaranteed?
- # [11:22] <Marcos> or in a browser, that is
- # [11:23] <annevk> a page served over HTTP always has an origin, yes
- # [11:23] <Dashiva> What about about:blank? Can't it use the same technique as that
- # [11:23] <annevk> (there's still ways to do attacks, but way more difficult than with widgets)
- # [11:23] <Marcos> no, what I mean is that the origin can be forged
- # [11:24] <annevk> really, how would you forge it?
- # [11:24] <annevk> Dashiva, what would that mean for widgets?
- # [11:24] <Dashiva> That every widget has its own unique origin
- # [11:24] <Marcos> Lets say I write my own little HTTP client, then when I send the request to a site, I pretend my origin is http://microsoft.com
- # [11:26] <Marcos> Dashiva: what would that look like?
- # [11:26] <Dashiva> Dunno. Maybe widget:hash(zipfile) or something? Just an idea.
- # [11:29] <annevk> Marcos, there's no issue in that scenario because user credentials are not involved and you can only access your own intranet and not that of the user
- # [11:29] <annevk> Marcos, and if user credentials are involved, again, it is only the user credentials of you
- # [11:30] <Marcos> Dashiva: yeah, that's kinda what I'm trying to get to. The problem is that widgets can be dynamically generated, so on the server you don't want to keep a list of every single widget that has been generated and then share all that info with other websites to allow CORS.
- # [11:30] <annevk> the simple answer here is to not use widgets :p
- # [11:31] <Marcos> hehe
- # [11:31] * Marcos resigns as Widget spec editor, quits Opera, and becomes a farmer
- # [11:31] <Dashiva> Bring out the cryptography, I guess
- # [11:32] <Dashiva> Domains work because domains are unique. There's nothing unique assigned to a widget, is there?
- # [11:33] <Marcos> nope
- # [11:33] * Marcos will be back in a bit
- # [11:33] <Dashiva> Some kind of crypto signature seems like it could do the trick, but getting that into CORS...
- # [11:38] * Joins: ArtB (d0309a43@128.30.52.43)
- # [11:48] * Quits: Lachy (Lachlan@85.196.122.246) (Quit: This computer has gone to sleep)
- # [12:06] * Joins: Lachy (Lachlan@213.236.208.22)
- # [12:32] <Marcos> CORS would not care, from my understanding, because it would just say access-control-allow-origin: widget://as4da4ads44dsa
- # [12:33] <annevk> yeah, that could work in theory
- # [12:34] <annevk> i still sort of wonder if we should not merge widgets and html5 so we don't have disadvantages of both
- # [12:38] <Marcos> annevk, tell me more (and where is Coimbra? :))
- # [12:40] <annevk> somewhere in portugal
- # [12:40] <annevk> well, HTML5 has the manifest thing, has most of the APIs, implemenations have arisen that allow "HTML5 apps" to be used standalone
- # [12:41] <annevk> there are some system integration APIs e.g. registerProtocolHandler and navigator.onLine
- # [12:41] <Marcos> The only problem (if it is a problem) is that the HTML5 manifest does not define content types for cached resources
- # [12:42] <annevk> they're fetched over HTTP so you get the content type from there
- # [12:43] <annevk> (and there's nothing stopping anyone from shipping a device that has several "html5 apps" cached locally from the start)
- # [12:43] <Marcos> but with widgets the resources are in the package, so the manifest format for widgets needs to state the type for when it is fed into the browsing context (when sniffing is not available).
- # [12:43] <annevk> try to get out of that box you're in :)
- # [12:43] <Marcos> seriously, annevk, I'm trying to.
- # [12:44] <annevk> ok, I'm saying we might not need packaging
- # [12:44] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
- # [12:45] <Marcos> I agree, the only need for packaging is for when widgets are acquired over a non http source
- # [12:46] <annevk> which is not a scenario I encounter often, if at all
- # [12:47] <annevk> and given how ever more likely it is that internet will be everywhere it seems sort of wrong to optimize for it
- # [12:47] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [12:47] * Joins: billyjackass (MikeSmith@mcclure.w3.org)
- # [12:48] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
- # [12:48] * billyjackass is now known as MikeSmith
- # [13:02] * Quits: annevk (opera@213.13.106.1) (Quit: annevk)
- # [13:08] <arve> «has most of the APIs» is not that true anymore
- # [13:08] <arve> ref. bondi
- # [13:08] * Quits: ArtB (d0309a43@128.30.52.43) (Quit: CGI:IRC (Ping timeout))
- # [13:09] * Joins: Art (d0309a43@128.30.52.43)
- # [13:09] * Quits: Art (d0309a43@128.30.52.43) (Quit: CGI:IRC)
- # [13:09] * Joins: ArtB (d0309a43@128.30.52.43)
- # [13:18] <MikeSmith> trackbot, status?
- # [13:18] * trackbot knows about the following 64 users: Olli, Joone, Josh, Lachlan, Frederick, Doug, Samuel, Michael(tm), Doug, Paul, David, Ian, Alexey, T.V., Arthur, Adele, Guido, Kris, Nikunj, Marc, Philip, Nick, Seung-Hun, Charles, Maciej, Marcos, Adrian, Adam, David, Carl, Carmelo, Paddy, Jean-Yves, Anne, Thomas, Arun, Jonathan, Chris, Soohong Daniel, Arve, Hallvord, Aaron, Bryan, Ian, Jere, Ian, Henry, Dave, Jonas, Mark, Rainer, Richard, Andrew, Katrin, Cameron,
- # [13:41] * Joins: timeless (timeless@65.75.195.122)
- # [13:53] <MikeSmith> Zakim, list
- # [14:04] * Quits: Marcos (Marcos@213.236.208.22) (Ping timeout)
- # [14:04] * Quits: arve (arve@213.236.208.22) (Ping timeout)
- # [14:04] * Joins: arve (arve@213.236.208.247)
- # [14:05] * Joins: Marcos (Marcos@213.236.208.247)
- # [14:07] * Joins: tlr (tlr@128.30.52.30)
- # [14:33] * Joins: aroben (aroben@71.58.77.15)
- # [15:45] * Quits: Marcos (Marcos@213.236.208.247) (Quit: Marcos)
- # [16:03] * Joins: Marcos (Marcos@213.236.208.22)
- # [16:04] * Quits: arve (arve@213.236.208.247) (Ping timeout)
- # [16:06] * Joins: arve (arve@213.236.208.22)
- # [16:27] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: This computer has gone to sleep)
- # [16:57] * Joins: Marcos_ (Marcos@213.236.208.22)
- # [16:57] * Quits: Marcos (Marcos@213.236.208.22) (Ping timeout)
- # [16:59] * Joins: anne (annevk@213.13.106.72)
- # [17:03] * Joins: gsnedders (gsnedders@86.136.52.180)
- # [17:04] * Quits: arve (arve@213.236.208.22) (Quit: Ex-Chat)
- # [17:06] * Joins: Lachy (Lachlan@85.196.122.246)
- # [17:07] * Quits: Marcos_ (Marcos@213.236.208.22) (Quit: Marcos_)
- # [17:11] * Joins: Marcos (Marcos@213.236.208.22)
- # [17:12] * gsnedders attacks Marcos
- # [17:12] <gsnedders> Speak now or forever hold your peace!
- # [17:13] <Marcos> AAAAHHHH!!!!
- # [17:13] <Marcos> ok ok...
- # [17:13] <Marcos> let me try the installer again
- # [17:13] <gsnedders> Yeah, I just tried installing it myself and I got some weird error
- # [17:13] <Marcos> do I need to d/l html5lib 0.10 ?
- # [17:14] <gsnedders> Get html5lib from SVN
- # [17:14] <gsnedders> It's quicker and more better.
- # [17:15] <Marcos> running installer for anolis...
- # [17:15] <Marcos> src/lxml/lxml.etree.c:133987: error: invalid application of 'sizeof' to incomplete type 'struct __pyx_obj_4lxml_5etree__ParserSchemaValidationContext'
- # [17:15] <Marcos> lipo: can't open input file: /var/tmp//cc5RVomh.out (No such file or directory)
- # [17:15] <Marcos> error: Setup script exited with error: command 'gcc' failed with exit status 1
- # [17:15] * gsnedders wants more error
- # [17:15] <gsnedders> Can you pastebin the entire output of the script?
- # [17:16] <Marcos> oh, hang, better email that
- # [17:16] * ArtB thinks gsnedders means mo' betta' :)
- # [17:16] <Marcos> rc/lxml/lxml.etree.c:169:31:src/lxml/lxml.etree.c:169:31: error: libxml/schematron.h: No such file or directory error: libxml/schematron.h: No such file or directory
- # [17:16] <Marcos> well, that's an obvious one :P
- # [17:17] <gsnedders> xmllint --help
- # [17:17] <gsnedders> What does that output?
- # [17:17] <Marcos> lots of -- options\
- # [17:17] <gsnedders> s/help/version/
- # [17:17] * gsnedders headdesks
- # [17:17] <Marcos> xmllint: using libxml version 20616
- # [17:17] <Marcos> compiled with: DTDValid FTP HTTP HTML C14N Catalog XPath XPointer XInclude Unicode Regexps Automata Schemas
- # [17:18] <gsnedders> Ah, you need a mo' betta' (thanks to ArtB) version of libxml
- # [17:18] <Marcos> mmmkay, is there a magic update command?
- # [17:19] <gsnedders> No
- # [17:19] <gsnedders> :P
- # [17:19] <Marcos> easy_install lxml :D
- # [17:19] <Marcos> or is that a different thing?
- # [17:19] <Marcos> bah, that crashed too
- # [17:20] <gsnedders> lxml is only a python wrapper around it
- # [17:20] <Marcos> ok, makes sense
- # [17:20] <gsnedders> macports is probably the best advice
- # [17:21] * Marcos tries http://www.explain.com.au/oss/libxml2xslt.html
- # [17:21] <gsnedders> Marcos: Those are still ancient versions!
- # [17:21] <Marcos> oh ok
- # [17:21] <gsnedders> I mean, they don't support XML 1.0 Fifth ed.!
- # [17:21] <gsnedders> :P
- # [17:22] * Marcos scratches his head
- # [17:22] <gsnedders> http://www.macports.org/ !
- # [17:22] * Marcos resumes processing
- # [17:23] <Marcos> ...installing...
- # [17:25] <gsnedders> Then sudo port sync; sudo port install libxml2 libxslt
- # [17:25] <Marcos> sudo: port: command not found :(
- # [17:26] <Marcos> udo port install libxml2 libxslt
- # [17:26] <Marcos> no port :(
- # [17:27] <gsnedders> You installed it yet?
- # [17:27] <Marcos> guess I should find the port directory?
- # [17:27] <gsnedders> echo $PATH
- # [17:27] <gsnedders> Or did you open the CLI before installing?
- # [17:27] <Marcos> I did... let me open new one
- # [17:27] <gsnedders> source ~/.profile ~/.bash_profile; echo $PATH
- # [17:28] <Marcos> that did it :D
- # [17:28] * Marcos watches the magic happen
- # [17:29] * Marcos remembers he had to do all this last time...
- # [17:29] * Marcos is sure he will bother gsnedders again in the future with exactly the same questions
- # [17:30] * gsnedders expects so
- # [17:30] * gsnedders grumbles
- # [17:31] <Marcos> ...building libxml2...
- # [17:33] * Marcos has become too accustomed to his idiotic iphone apps... if it don't wobble, then it's too hard
- # [17:34] * Marcos sudo python setup.py install 's anolis
- # [17:35] * Marcos wishes he was young again... when all this command line stuff made sense... don't worry, gsnedders, you'll get old like me one day when even clicking the mouse becomes a source of confusion.
- # [17:37] <Marcos> ...and we are back in business! Thanks gsnedders!!! you rock and roll!!!
- # [17:37] <gsnedders> Marcos: sudo make me a sandwich
- # [17:37] <Marcos> hehe
- # [17:37] <Dashiva> sudo tell me a story about the good old days
- # [17:50] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
- # [18:02] * Quits: sicking (chatzilla@63.245.220.241) (Client exited)
- # [18:04] * Joins: sicking (chatzilla@63.245.220.241)
- # [18:52] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [20:17] * Quits: Hixie (ianh@129.241.93.37) (Ping timeout)
- # [20:18] * Joins: Hixie (ianh@129.241.93.37)
- # [20:27] * Quits: anne (annevk@213.13.106.72) (Ping timeout)
- # [20:52] * Joins: anne (annevk@213.13.106.1)
- # [20:58] * Quits: ArtB (d0309a43@128.30.52.43) (Quit: CGI:IRC)
- # [22:28] * Quits: heycam (cam@124.168.80.126) (Quit: bye)
- # [23:30] * Joins: heycam (cam@130.194.72.84)
- # [23:50] * Quits: sicking (chatzilla@63.245.220.241) (Client exited)
- # Session Close: Fri Mar 13 00:00:01 2009
The end :)