[13:58] <kgunn> if folks want to join...
[13:59] <kgunn> https://plus.google.com/hangouts/_/hoaevent/AP36tYdeoIQjgglMiG2l--2zSdJl12RcGP4SVmbnUJlCuSjsDyWw5w?authuser=0
[14:15] <didrocks> kgunn: you were breaking up
[14:54] <jdstrand> rsalveti: ah, I just pinged Chris to join this session, so maybe you can ask your questions about hwaccel here :)
[14:54] <didrocks> jdstrand: rsalveti: https://plus.google.com/hangouts/_/hoaevent/AP36tYdv1-3CbM3iyNnFODizI2urwkZIi7l-Z_2EdBSK31bgzTt0pA
[14:55] <didrocks> bfiller: ^
[14:55] <bfiller> didrocks: thanks
[14:55] <bfiller> oSoMoN: ^^^
[14:56] <rsalveti> hm, got a conflict with core-1, but will try to join for a while
[14:56] <jdstrand> I too have a conflict
[14:56] <jdstrand> I've asked chris to join (who is a better representative for this anyway)
[15:00] <alex-abreu> what's the hg url?
[15:01] <didrocks> alex-abreu: https://plus.google.com/hangouts/_/hoaevent/AP36tYdv1-3CbM3iyNnFODizI2urwkZIi7l-Z_2EdBSK31bgzTt0pA
[15:01] <sergiusens> bfiller, your live
[15:01] <alex-abreu> didrocks, thx
[15:02] <didrocks> yw
[15:03] <dbarth> hey chrisccoulson, can you share the branch with the ua work?
[15:04] <chrisccoulson> https://code.launchpad.net/~chrisccoulson/oxide/network-callbacks
[15:04] <dbarth> sweet!
[15:09] <alesage> question about the possibility of getting a webdriver going: https://bugs.launchpad.net/oxide/+bug/1285750
[15:09] <udsbotu> Launchpad bug 1285750 in Oxide "Provide a web driver for user acceptance testing" [Medium,New] - Assigned to Alexandre Abreu (abreu-alexandre)
[15:09] <alesage> does the existing chromium driver help us?
[15:09] <bfiller> ppa install instructions here:  http://pastebin.ubuntu.com/7084806/
[15:10] <alex-abreu> alesage, not really we need a specific build
[15:14] <daker> Are we going to have a developer tools with Oxide ?
[15:14] <alex-abreu> daker, yes part of the plan
[15:14] <alex-abreu> daker, there is a bug for that, I am checking that in between other things
[15:14] <dbarth> didrocks: here is the ffe for now: https://bugs.launchpad.net/libunity-webapps/+bug/1290535
[15:14] <daker> for 14.04 or 14.10 ?
[15:14] <udsbotu> Launchpad bug 1290535 in webbrowser-app (Ubuntu) "[FFE] Webapps support for the new Oxide container" [Undecided,New]
[15:14] <alex-abreu> daker, ideally 14.04
[15:15] <daker> ok
[15:15] <bfiller> bugs here: https://bugs.launchpad.net/webbrowser-app/+bugs?field.tag=oxide
[15:15] <didrocks> daker: thx!
[15:15] <rickspencer3> QUESTION: sorry if I missed it, but is it useful if I try the browser on my desktop?
[15:15] <elopio> here is a bug reported for web driver:
[15:15] <elopio> https://bugs.launchpad.net/oxide/+bug/1285750
[15:15] <udsbotu> Launchpad bug 1285750 in Oxide "Provide a web driver for user acceptance testing" [Medium,New] - Assigned to Alexandre Abreu (abreu-alexandre)
[15:15] <dbarth> rickspencer3: it is, but you should be able to try that via the webapp container soon
[15:15] <bfiller> and here: https://bugs.launchpad.net/oxide
[15:16] <dbarth> rickspencer3: as webapps should run on top of an oxide container
[15:16] <rickspencer3> dbarth, ok
[15:16] <dbarth> and it works quite well; running youtube and google calendar with it everyday
[15:16] <rickspencer3> if it would help for me to install a ppa and try testing now, I'm happy to do that
[15:17] <elopio> dbarth is doing the hangout from oxide ;)
[15:17] <rickspencer3> nice!
[15:17] <chrisccoulson> i think that was a joke ;)
[15:17] <rickspencer3> bfiller, can you give me instructions for installing the web browser?
[15:17] <chrisccoulson> i'm fairly sure that wouldn't work yet :)
[15:17] <rickspencer3> I'll totally switch to it as my primary
[15:17] <bfiller> rickspencer3:  http://pastebin.ubuntu.com/7084806/
[15:17] <dbarth> elopio: not yet
[15:17] <rickspencer3> bfiller, that seems within my skills and abilities ;)
[15:18]  * rickspencer3 installs 
[15:18] <bfiller> rickspencer3: :)
[15:18] <dbarth> rickspencer3: just make sure to remove the ppa from your source list once installed
[15:18] <bfiller> rickspencer3: just make sure to remove that ppa from your sources list after, as it has a lot of other cruft that could mess up your desktop
[15:18] <bfiller> rickspencer3: we're working on moving the oxide stuff to it's own ppa but not there yet
[15:19] <rickspencer3> thanks dbarth and bfiller
[15:19] <rickspencer3> good tip
[15:20] <rickspencer3> bfiller, dbarth just ppa-purge?
[15:20] <daker> i use the ua-string override for some webapps
[15:21] <dbarth> rickspencer3: never really used ppa-purge much, but this should work, yes
[15:23] <rickspencer3> arg ... g+ won't let me go there in the browser :?
[15:24]  * rickspencer3 shakes fist at Google
[15:36] <oSoMoN> http://pastebin.ubuntu.com/7085057/
[15:37] <pmcgowan> Has there been any updates from the Mozilla folks on their success or lack thereof in getting websites fixed?
[15:37] <pmcgowan> Do we have a plan for how to avoid the offers for native app installs?
[15:38] <daker> they evangelize their UA, so they the websites that don't serve mobile content for their UA
[15:39] <pmcgowan> daker, right, and we somewhat benefit from that, we did talk to them directly
[15:39] <daker> pmcgowan: i believe oSoMoN and chrisccoulson had a meeting with them
[15:39] <pmcgowan> yes
[15:42] <pmcgowan> ah right, the toolkits
[15:42] <pmcgowan> beuno was going to help on that
[15:42] <pmcgowan> I think I would prefer to have our own pure user agent but some sort of blanket override possibility, as even the bastard string may not work
[15:43] <pmcgowan> the problem is of course you cannot go back
[15:44] <rickspencer3> it will help them win if Ubuntu becomes more common
[15:44] <rvr> The UA pattern to serve desktop/mobile content is a broken one
[15:47] <pmcgowan> rickspencer3, whats your thinking, go with a practical string or with a pure string that needs overrides everywhere?
[15:47] <rickspencer3> pmcgowan, I think we should start with the practical over the pure
[15:48] <rickspencer3> that's my $.02
[15:48] <rickspencer3> make the web work
[15:48] <rickspencer3> the trade off for the Ux is getting those darn "install my app" pages
[15:48] <pmcgowan> yeah thats the issue
[15:50] <rickspencer3> pmcgowan, yeah, I think we should accept that trade off for now
[15:50] <rickspencer3> imho
[15:50] <pmcgowan> yeah, no perfect answer, the web is busted
[15:53] <pmcgowan> its also hard to get statistics to show how often our browser is used if the string changes
[15:54] <pmcgowan> when will downloads work?
[15:56] <didrocks> bfiller: ok, let's wrap up maybe and switch to next session?
[15:57] <pmcgowan> thanks
[15:57] <didrocks> bfiller: https://plus.google.com/hangouts/_/hoaevent/AP36tYes4tHRU4-_4wV_CTOx5w50Gu0iCBJwKuk0vrxpHKJHImEvog for your next session
[15:59] <bfiller> renato: ^^^
[16:01] <bfiller> spec is here: https://docs.google.com/a/canonical.com/document/d/1qa0twAq3oRbzsIM5kkjsWm0D1cUP1IQrOyv4fgYIfIo/edit#
[16:04] <tedg> Hmm, wasn't in the room. Can someone repaste the doc?
[16:05] <kenvandine> https://docs.google.com/a/canonical.com/document/d/1qa0twAq3oRbzsIM5kkjsWm0D1cUP1IQrOyv4fgYIfIo/edit#
[16:05] <tedg> kenvandine, Thanks!
[16:08] <tedg> Why do we want them as one process? It seems like running the sync component and then having the memory free'd is a good thing?
[16:09] <mardy> tedg: +1
[16:09] <tedg> What is sync monitor watching? Could we just make it conditions of the upstart job?
[16:10] <mardy> tedg: online accounts events (account created, enabled, disabled, deleted)
[16:10] <mardy> tedg: maybe something else as well, I don't know
[16:10] <tedg> mardy, Those are all dbus signals, no?
[16:11] <bfiller> sync monitor watching UOA enable/disable events and changes to the address book service so it knows when to kick off the sync
[16:11] <mardy> tedg: yes
[16:11] <tedg> So yeah, we can use the upstart dbus bridge.
[16:12] <tedg> start on dbus SIGNAL=enable or dbus SIGNAL=disable
[16:14] <jono> hey all
[16:14] <tedg> Would be interesting if we could run syncevolution based on a push notification.
[16:15] <jono> sorry, just joined, do we have any ETAs when we can expose google calendar events to the calendar app?
[16:16] <jono> I think that will significantly ramp up calendar development
[16:16] <charles> +1
[16:16]  * tedg thinks jono just wants his team to show up to meetings
[16:16] <jono> lol
[16:17] <charles> having google calendar events show up in the datetime indicator automatically without a lot of manual user config work would be a big win
[16:18] <tedg> Yeah, I'm always surprised how many people don't know that feature in datetime exists.
[16:18] <tedg> Would love that on the desktop as well.
[16:18] <jono> tedg, because it never works :-)
[16:18] <tedg> jono, Heh, I use it constantly.
[16:18]  * tedg setup the ical feed for vUDS
[16:18] <charles> tedg, it's because people don't know about syncevolution and its config is nontrivial
[16:18] <jono> dpm, can you throw me the hangout URL
[16:18] <charles> having it Just Work would be so much better
[16:18] <tedg> charles, Sure it needs to be part of configuring online accounts.
[16:18] <dpm> jono, https://plus.google.com/hangouts/_/hoaevent/AP36tYes4tHRU4-_4wV_CTOx5w50Gu0iCBJwKuk0vrxpHKJHImEvog
[16:20] <tedg> Wow, just saw jono on the feed.
[16:20] <tedg> Seems like a 2 min delay today on hangouts.
[16:20] <jono> tedg, I am everywhere
[16:20] <jono> check your closet
[16:20] <tedg> jono, I know you're in the closet
[16:21]  * tedg locks the door
[16:22] <tedg> jono, We should talk about the delay with Google though, it has made vUDS harder today.
[16:22] <jono> tedg, I am coming out of the closet
[16:22] <jono> tedg, delay
[16:22] <jono> ?
[16:22] <tedg> jono, Today the delay for the hangout feed is about 2 min, making HO/IRC sync harder.
[16:22] <jono> yikes
[16:23] <balloons> tedg, yea I agree it feels worse
[16:24] <Saviq> syncevo on MeeGo lets you set schedules
[16:24] <Saviq> i.e. every 10 minutes during the day, every hour overnight
[16:25] <Saviq> and that's per-provider, something like that would be relatively cheap, and could be modified to take WiFi vs. 3G into account
[16:29] <tedg> When you get there (fighting the delay) why do we want Qt binding for SyncEvo backends instead of extending at the EDS backend?
[16:32]  * mardy afk
[16:35] <Saviq> dpm, jono - it works with the steps from http://notyetthere.org/syncing-ubuntu-touch-with-owncloud-or-any-carddav-server/
[16:35] <Saviq> with the caveat that you need to tweak some of the parameters
[16:36] <Saviq> to make it use a calendar and not the contacts
[16:36] <jono> thanks Saviq
[16:36]  * Saviq is syncing the calendar for a while now
[16:37] <dpm> thanks!
[16:42] <pmcgowan> that model is not great for the phone?
[16:43] <pmcgowan> periodic polling that is by the syncevo daemon
[16:43] <pmcgowan> right, push
[16:45] <tedg> The problem with us doing it is that we need to then move the Google credientials for all the users to our server.
[16:45] <nxvl> and that's something most users won't accept
[16:46] <tedg> nxvl, Well, most users probably would accept it, but I don't want that responsibility :-)
[16:46] <nxvl> and why some people didn't trusted blackberry e-mail
[16:46] <nxvl> tedg: well, yeah, that as well
[16:47] <tedg> We probably don't want a library, and loading that.
[16:47] <tedg> It's hard to confine a loadable module.
[16:47] <tedg> It'd be better if we could run a process that then could use some IPC.
[16:48] <tedg> We don't want a trusted daemon loading untrusted code into its memory space.
[16:49] <tedg> I think the simple use-case is a Facebook type app that could put birthdays in the calendar.
[16:50] <tedg> Or allowing enterprises to put the corporate directory as a click package.
[16:50] <kenvandine> tedg, that would be a cool use case
[16:52] <tedg> Or Salesforce having all your business contacts as an addressbook.
[16:52] <tedg> bfiller, renato, jono, ping on the security issue above
[16:54] <jdstrand> it is impossible to confine a loaded module
[16:54] <jdstrand> said loaded module must be in a separate process to confine it
[16:54] <bfiller> tedg: not sure what the security issue is.. can you summarize?
[16:54] <tedg> bfiller, Don't use a lib, needs to be a separate process.
[16:55] <bfiller> tedg: it is already a seperte process (sync-monitor/sync evo)
[16:55] <tedg> bfiller, In the 3rd party case.
[16:55] <bfiller> ah
[16:55] <tedg> The comment makes sense to where I am in the HO feed :-)
[16:56]  * jdstrand has no context at alol
[16:56] <jdstrand> all
[16:56] <didrocks> bfiller: it's time to start wrapping up (there is a meeting about Qt 5.2 after that one)
[16:56]  * jdstrand is in another session
[16:56] <bfiller> ok
[16:56] <tedg> alol is my favorite state
[16:56] <didrocks> jdstrand: double yourself!
[16:56] <jdstrand> mine too, it just isn't the state I'm in :)
[16:56] <jdstrand> didrocks: man, I wish I could code that up :)
[16:57] <didrocks> jdstrand: I heard good things about go with multiple processes :p
[16:57]  * tedg is staticly compiled
[16:58] <dobey> tedg: you wish :)
[16:58] <zyga-uds> thanks
[16:58] <zyga-uds> very interesting session
[17:02] <jdstrand> bfiller, tedg: again completely lacking context but I will say it is ok to use a .so file *if* you fork, aa_change_profile(), then load. the new scopes architectureis doing something similar iirc
[17:04] <jdstrand> though there might be some issues there with shared memory. if going that route, I recommend talking to the security team for specifics
[17:05] <tedg> jdstrand, The problem in this case is that you'd want more than one.