IRC log of dig on 2011-01-06
Timestamps are in UTC.
- 12:36:14 [DIGlogger]
- DIGlogger (~dig-logge@groups.csail.mit.edu) has joined #dig
- 12:36:14 [jordan.freenode.net]
- topic is: Decentralized Information Group @ MIT http://dig.csail.mit.edu/
- 12:36:14 [jordan.freenode.net]
- Users on #dig: DIGlogger RalphS melvster IvanHerman mhausenblas timbl presbrey jambo webr3 danbri nunnun amy sandro mattl IPaparrizos ericP gbot30
- 12:38:34 [junaidnaseer]
- junaidnaseer (~junaid@pc222.phh.uni-kiel.de) has joined #dig
- 12:41:23 [junaidnaseer]
- junaidnaseer has quit (Client Quit)
- 12:57:10 [kennyluck]
- kennyluck (~kennyluck@2001:200:1c0:3602:225:ff:fe4d:f8c7) has joined #dig
- 13:39:20 [junaidnaseer]
- junaidnaseer (~junaid@pc222.phh.uni-kiel.de) has joined #dig
- 13:53:00 [kennyluck]
- but tabulator doesn't try to build panes on top of W3C widgets. lol
- 14:03:08 [nunnun]
- nunnun is now known as nunnun_away
- 14:06:58 [nunnun_away]
- nunnun_away is now known as nunnun
- 14:18:57 [webr3]
- kennyluck, nothing really tries to build on top of w3c widgets afaik, I've been nudging and pushing Marcos et al to try and get widget support in to the main browser context, and so that all extensions are deployed using web tech stack, just packaged/signed and cached
- 14:19:28 [webr3]
- some are v positive about it, but some of the browsers are not so keen on loosing their niche extension markets
- 14:23:02 [oshani]
- oshani (~oshani@124.43.46.240) has joined #dig
- 14:52:43 [marisol]
- marisol (~marisol@pool-151-203-91-62.bos.east.verizon.net) has joined #dig
- 15:36:49 [oshani]
- oshani has quit (Quit: Mama nidi!)
- 15:39:31 [oshani]
- oshani (~oshani@124.43.46.240) has joined #dig
- 16:02:17 [betehess]
- betehess (~betehess@31-35-254.wireless.csail.mit.edu) has joined #dig
- 16:09:23 [nunnun]
- nunnun has left #dig
- 16:09:44 [nunnun]
- nunnun (~nunnun@irc.v6.nunnun.jp) has joined #dig
- 16:11:00 [IvanHerman]
- IvanHerman has quit (Quit: bye guys)
- 16:31:38 [oshani_]
- oshani_ (~oshani@124.43.42.170) has joined #dig
- 16:34:04 [oshani]
- oshani has quit (Ping timeout: 260 seconds)
- 16:34:05 [oshani_]
- oshani_ is now known as oshani
- 16:39:49 [oshani]
- oshani has quit (Quit: Mama nidi!)
- 16:51:53 [oshani]
- oshani (~oshani@124.43.42.170) has joined #dig
- 17:53:30 [oshani]
- oshani has quit (Quit: Mama nidi!)
- 17:56:55 [oshani]
- oshani (~oshani@124.43.42.170) has joined #dig
- 18:05:55 [mhausenblas]
- mhausenblas has quit (Quit: mhausenblas)
- 18:07:50 [marisol]
- marisol is now known as mari-brb
- 18:08:42 [timbl]
- I have wonderde about tha
- 18:08:42 [timbl]
- tabulatro not using widget packaging
- 18:26:28 [timbl]
- timbl has quit (Read error: Connection reset by peer)
- 19:06:45 [timbl]
- timbl (~timbl@pool-96-237-236-230.bstnma.fios.verizon.net) has joined #dig
- 19:18:11 [oshani]
- oshani has quit (Quit: Mama nidi!)
- 19:46:37 [mari-brb]
- mari-brb is now known as mari-mami
- 20:14:09 [timbl_]
- timbl_ (~timbl@pool-96-237-236-230.bstnma.fios.verizon.net) has joined #dig
- 20:17:58 [timbl]
- timbl has quit (Ping timeout: 246 seconds)
- 20:17:58 [timbl_]
- timbl_ is now known as timbl
- 20:28:23 [melvster]
- melvster has quit (Remote host closed the connection)
- 20:28:42 [melvster]
- melvster (~melvster@p579F958C.dip.t-dialin.net) has joined #dig
- 20:43:51 [junaidnaseer]
- junaidnaseer has quit (Quit: Leaving.)
- 20:56:51 [junaidnaseer]
- junaidnaseer (~junaid@pc222.phh.uni-kiel.de) has joined #dig
- 21:03:20 [junaidnaseer]
- junaidnaseer has quit (Quit: Leaving.)
- 21:20:30 [RalphS]
- RalphS has quit (Quit: outah here ...)
- 21:21:43 [presbrey]
- I think tabulator will be implemented in chromium a lot more easily if it used local storage objects
- 21:25:49 [presbrey]
- chrome uses process separation for tabs and extensions and requires sendMessage and onMessage API between each context
- 21:29:37 [junaidnaseer]
- junaidnaseer (~junaid@pc222.phh.uni-kiel.de) has joined #dig
- 21:31:39 [timbl]
- timbl has quit (Remote host closed the connection)
- 21:31:47 [timbl]
- timbl (~timbl@pool-96-237-236-230.bstnma.fios.verizon.net) has joined #dig
- 21:32:34 [timbl]
- So it can't share a rdf store?
- 21:32:44 [junaidnaseer]
- junaidnaseer has quit (Client Quit)
- 21:33:04 [timbl]
- process separation for tabs
- 21:33:29 [timbl]
- Bu the tabulator experience is kinda predicated on accumukating more and more data in the whole session.
- 21:33:49 [timbl]
- I suupose an alternative is to put the tabulator store in a sep process running HTTP
- 21:38:21 [mari-mami]
- mari-mami has quit (Quit: mari-mami)
- 21:40:12 [presbrey]
- I'm not even sure content tabs can access the extension's local storage
- 21:40:22 [presbrey]
- its quite secure!
- 21:40:46 [presbrey]
- s/its/looks/
- 21:43:17 [presbrey]
- been playing with it a bit today
- 21:43:23 [presbrey]
- http://code.google.com/chrome/extensions/content_scripts.html#pi -- programmatic injection
- 21:44:05 [presbrey]
- they are advertising sending script around as strings
- 21:44:21 [presbrey]
- so no shared DOM access across all extension instances
- 21:46:10 [presbrey]
- http://code.google.com/chrome/extensions/content_scripts.html#host-page-communication -- host page communication
- 21:46:28 [presbrey]
- looks like quite a bit of hacking is needed
- 21:47:12 [presbrey]
- I committed a chrome extension shim
- 21:47:19 [presbrey]
- thought it might be an easy wrapping job
- 21:47:36 [presbrey]
- but *sigh*
- 21:50:06 [webr3]
- can it leverage workers?
- 21:50:26 [webr3]
- each webworker can be shared, so it's a central process that can be used by several tabs, ideal for dbs/storage
- 21:51:06 [webr3]
- workers etc have no access to dom though, thus no access to rdf/xml or rdfa..
- 22:00:20 [jambo]
- presbrey: does a chrome extension not have a single background page even though it's running in many tabs?
- 22:01:57 [jambo]
- i would think that you could keep the store in the background page and have a write a version of the store that actually just forwards messages off to the background page
- 22:03:09 [jambo]
- you're right though that it would not simply be a wrapper in that case
- 22:16:41 [presbrey]
- jambo: it does have a single background page
- 22:17:03 [presbrey]
- but that single page is in a separate process from all others (tabs, extensions, etc)
- 22:17:10 [jambo]
- sure
- 22:17:28 [presbrey]
- one quickie for you
- 22:17:48 [jambo]
- it would work though if you just used a different sort of store, right?
- 22:17:50 [presbrey]
- I notice lots of the libraries start by getting a reference to tabulator, var tabulator = Components...
- 22:18:04 [presbrey]
- where is that component first defined, xpcom.js?
- 22:18:52 [jambo]
- yeah, its where wrappedJSObject gets set.
- 22:19:11 [jambo]
- looking at hg, its line 32
- 22:20:12 [presbrey]
- next step I think is to backtrack from Chrome to Firefox 4 support
- 22:20:14 [jambo]
- remember that if you try to hang values created in a browser context off of that JS object, when that browser window closes, those values will get cleaned up and disappear
- 22:21:03 [presbrey]
- not trying to do anything new, loading in FF4, one of the var tabulator=Components[]... is the first to fail
- 22:21:13 [jambo]
- hrm
- 22:21:20 [jambo]
- where is it
- 22:21:35 [presbrey]
- DIGlogger, pointer?
- 22:21:35 [presbrey]
- See http://dig.csail.mit.edu/irc/dig/2011-01-06#T22-21-35
- 22:22:49 [jambo]
- maybe something changed in ff4 so that the xpcom component cannot register?
- 22:23:39 [presbrey]
- that seems likely
- 22:23:57 [jambo]
- i would assume that getting null out of Cc.foo means the component didn't register
- 22:24:13 [jambo]
- are you just flat out plugging the code in?
- 22:24:37 [presbrey]
- mostly
- 22:24:46 [jambo]
- okay
- 22:24:49 [presbrey]
- I updated install.rdf and chrome.manifest a bit I think
- 22:24:57 [jambo]
- I couldn't find the reference in the log, do you know which file it is
- 22:25:03 [presbrey]
- yeah I'm looking
- 22:25:11 [jambo]
- OK
- 22:25:16 [presbrey]
- I have to reproduce it as diglogger doesn't have our pastes from yesterday
- 22:25:20 [jambo]
- ah
- 22:25:32 [jambo]
- alright, I have to go for a bit, but I'll be checking back
- 22:48:03 [presbrey]
- browserInit.js line 2 and statuswidget.js line 3
- 22:48:13 [presbrey]
- so yes I think the xpcom registration is failing
- 22:48:20 [jambo]
- okay.
- 22:48:36 [presbrey]
- https://developer.mozilla.org/en/XPCOM/XPCOM_changes_in_Gecko_2.0
- 22:48:44 [jambo]
- iw as just about to post that exact link
- 22:49:02 [presbrey]
- haha
- 22:49:33 [presbrey]
- how does the xpcom component work with the overlay? why are both needed?
- 22:49:33 [jambo]
- looks pretty simple
- 22:50:36 [jambo]
- stuff like icons appearing in the location bar, items that appear in the menu bar, etc, have to be loaded through the overlay of the browser.
- 22:50:54 [jambo]
- there used to be more in the overlay, but that _should_ be all gone now.
- 22:51:27 [jambo]
- i think oshani had put a lot of code for the policy stuff
- 22:51:30 [jambo]
- directly in the overlay
- 22:52:02 [presbrey]
- I see multiple implementations of the <div id="outline"> body replacement
- 22:52:12 [presbrey]
- are there multiple paths for setting up the replacement document?
- 22:53:10 [jambo]
- if its in xpcom components, i think that may have been copied to create an n3 converter rather than with inheritance for dubious reasons
- 22:53:33 [jambo]
- otherwise, i'm going to guess it might be related to the items in the location bar, which you would have to talk to kennyluck about
- 22:54:05 [jambo]
- it's been a while, though. I did think that I had cleaned all of that up in my changes, but I could be wrong :(
- 22:54:23 [presbrey]
- cool thanks
- 22:54:49 [presbrey]
- earlier I was trying to understand which setup paths would be important for a chrome extension
- 22:55:08 [presbrey]
- where the component registration might be abstracted away
- 22:55:47 [presbrey]
- and which part setup the outline interface
- 22:56:18 [presbrey]
- thinking it would be easy to tweak the addons once the outline was in place
- 22:56:22 [jambo]
- the biggest pain is that there isnt a very nice abstraction barrier between the outline code and the store
- 22:56:33 [presbrey]
- and kb
- 22:57:02 [jambo]
- they flow together a bit too much to just say, "well, create a kb over here, and create an outline over here, and just wire them together" unfortunately
- 22:57:19 [jambo]
- that's still hanging around from when it was assumed that everything would just be running in a single page anyway
- 23:00:07 [webr3]
- tomayac may be good to talk to, he's def made google extensions working w/ rdf(a)
- 23:00:45 [presbrey]
- cool w/ shared storage and cross-site xhr?
- 23:01:12 [webr3]
- unsure about shared storage, def w/ cross site xhr
- 23:01:47 [webr3]
- see: http://blog.tomayac.de/
- 23:02:06 [webr3]
- even the source of his extensions may be good to look at as reference
- 23:04:40 [webr3]
- DIGlogger, pointer?
- 23:04:40 [webr3]
- See http://dig.csail.mit.edu/irc/dig/2011-01-06#T23-04-40