/srv/irclogs.ubuntu.com/2014/01/13/#ubuntu-desktop.txt

ochosirobert_ancell: hm, i implemented the blanking now and it has exactly the effect i was afraid of.. the session that gets started does indeed inherit the blank settings00:33
ochosi(this is the patch, only weighs a few lines and applies against gtk-greeter trunk: http://dpaste.com/1548435/ )00:33
ochosii also added another call to reset the settings to defaults, but somehow that doesn't work :/00:34
robert_ancellochosi, does the session not apply the setting when it starts?00:34
robert_ancellThis is XFCE right?00:34
ochosiyeah00:34
ochosixfce doesn't do anything about these values unfortunately :/00:35
robert_ancellso it just uses the X defaults?00:35
ochosiso the greeter will have to reset whatever is default00:35
ochosiyup00:35
robert_ancellyeah, I guess the greeter can read it on startup and clear it00:35
ochosiwell if you look at the patch, that's what i tried00:35
ochosii added some debug output locally, the values do get read successfully00:35
robert_ancellochosi, did you call an XSync or equivalent?00:36
ochosihm, i didn't00:36
robert_ancellSince the greeter is quitting it might not have actually sent the request00:36
ochosioh00:36
robert_ancellthere's a gdk call for it00:36
ochosisince the simple call worked at the beginning, i presumed it would at the end as well...00:36
ochosigood catch, i'll see whether that helps00:36
ochosiXFlush (display); ?00:37
ochosirobert_ancell: somehow that doesn't seem to work either, tried both XSync and XFlush now00:44
robert_ancellhmm, X is notoriously hard to diagnose :(00:44
=== duflu_ is now known as duflu
=== doko_ is now known as doko
pittiGood morning05:54
=== jodh` is now known as jodh
Laneymorning09:04
Laneyhappy monday!09:04
Laneyor something like that09:04
seb128good morning desktopers09:05
seb128Laney, hey, happy monday! had a good w.e?09:05
Laneyhey seb128, it was very nice thank you - we went away to the peak district for some hiking09:07
Laneyyou?09:07
seb128mine was good as well09:09
seb128some tennis, some shopping, some video gaming09:09
seb128relaxing in between, and spa yesterday afternoon09:09
Laneysounds pleasant indeed09:11
pittihey Laney, bonjour seb12809:13
seb128pitti, lut, ça va ? bon w.e ?09:13
pittiseb128: oui, merci ! we did some badminton and a nice long walk, and played Wizard with some friends until late night on Saturday :)09:14
seb128wizard? I guess that's a board game?09:16
Laneysounds a bit like contract whist09:17
pittia card game09:17
Laneyplayed some of that over the weekend :-)09:17
pittihttp://en.wikipedia.org/wiki/Wizard_%28card_game%2909:17
Laney(hey pitti!)09:17
pittiit's a bit similar to Skat, but with simpler rules, and more interesting strategies09:17
seb128sounds like fun in any case ;-)09:18
Laneyhmm, two random reboots this morning09:36
Laneysuspecting disk issues09:36
seb128Laney, so, about those sounds setting in accountsservice ... you think we should make accountsservice replace gsettings as the configuration backend for those settings?10:46
seb128Laney, I've to admit it makes me uncomfortable, I've much less trust in accountsservice than gsettings to be a nice API to use and for stability/performance as well10:46
LaneyIt seems like the way to go for user properties which need to be accessed by parts of the 'system'10:47
seb128desrt, larsu: do you have an opinion on the topic? the context is the "ringtone" config for example10:48
seb128we had it in gsettings10:48
seb128but the greeter needs it10:48
seb128so we need it to accountsservice10:48
seb128I was suggesting "syncing" it there by some way but letting gsettings the primary storage and have user apps still use that10:48
seb128Laney suggests replace gsettings by accountsservice and have apps talk to accountsservice instead10:48
desrt...10:49
desrtdoes the phone really have users?10:49
seb128the tablet has10:49
desrtlike what do we do?  ringtone depending on who is selected at the login screen?10:49
seb128yes10:49
desrtthat seems pretty random10:49
seb128well, if it's not ringtone you can think wallpaper10:49
seb128or locale10:49
seb128or keyboard layout10:49
LaneyIt'll look up the 'active' user from logind10:49
chrisccoulsonhello desktoppers!10:50
Laneyand play that guy's ringtone10:50
seb128chrisccoulson, hey, how are you?10:50
* desrt makes a face10:50
chrisccoulsonseb128, yeah, not too bad thanks. how are you?10:50
desrtseb128: keep using accountsservice, i guess10:50
seb128chrisccoulson, I'm good, thanks10:50
* mpt imagines scrolling through the user list while the phone is ringing10:50
Laneyit's not about that really10:50
desrtseb128: this is the future... we can shove all kinds of crack in there now10:50
seb128desrt, "keep"? we currently use gsettings10:50
desrtwe can even have keys that have default values which contain '<' characters10:50
desrtTECHNOLOGY!!10:50
seb128desrt, so you would port desktop apps to talk to accountsservice rather than use gsettings?10:50
LaneyI would only do it for things that need to be there10:51
Laneygsettings is nicer to work with10:51
desrtseb128: it seems like there are maybe a half-dozen or dozen examples of these type of settings that need to be available from the login screen10:51
seb128Laney, what is the API like for a desktop app to get a value and key changes update with accountsservice? I've the feeling that's a lot less nice that the 1 line with gsettings-qt10:51
desrtrandom apps settings are obviously not this case10:51
Laneyseb128: well, indeed, we wrote a nice library for gsettings didn't we?10:51
Laneyit would be more than 1 line without that ...10:52
seb128Laney, right ... so we write a nice one for accountsservice? ;-)10:52
Laneythat'd be good10:52
desrtthis is spinning wildly out of control10:52
seb128that change just feels wrong to me10:52
desrtwe might have encrypted homedirs on the tablet?10:53
seb128I've trust in gsettings to be good (e.g robust and easy to use), I don't have the same trust in accountsservice10:53
mptdesrt, seb128: Last week we (charles, ted, and I) deprioritized all the indicator-related work based on who is selected, on the grounds that on Touch it’s pretty hard for an account to be selected without prompting you to log in.10:53
desrtseb128: accountsservice doesn't call abort()..... yet ;)10:53
seb128desrt, think convergence, we need to design for the desktop usecase, so yes, we mighgt have encrypted userdirs10:53
seb128mpt, well, if you use your device as an user and lock the screen that user should probably be the default one you are prompted to unlock if you turn the device back on?10:54
seb128"turn back on" being "resume from suspend", not a full power down/up cycle10:55
Laneyso what would you do?10:55
mptseb128, true.10:55
desrtseb128: does my desktop ring when i'm not logged in?10:55
Laneyyou want the telephony service to read accountsservice10:56
seb128desrt, if you have a SIM card, I guess so10:56
Laneyu-s-s to write to gsettings10:56
Laneyand ??? to mirror it?10:56
seb128Laney, is telephony-service a system or user service?10:56
desrtLaney: stop right there10:56
desrtyou're about to violate The Rule10:56
Laneydesrt: I'm asking seb128 what his design is10:57
Laneyseb128: It runs as the user currently because multi user isn't worked out yet, I guess10:57
seb128well, the design we have on the desktop with the background is10:57
seb128- the gsettings key is the canonical config10:57
desrtdesktop background is a great example of a violation of the rule :(10:57
seb128- nautilus/g-s-d monitor the key and write to a-s when it changes10:57
LaneyI don't think nautilus is a good example10:57
Laneyit's some weird local hack we did10:57
seb128- all the apps/config UI writes to gsettings10:57
seb128desrt, what rule?10:58
desrtseb128: data lives in one place10:58
desrt'sync' is bad10:58
Laneysync 'smells' to me10:58
desrti like sleep10:59
seb128desrt, yeah, what do you do up at 6am...10:59
desrtseb128: i started writing patches and i couldn't stop11:00
desrtnow that i stopped to talk on IRC i realise that i'm actually very tired11:00
desrti think i'm going to go grab a few hours :)11:00
Laneytata11:00
seb128;-)11:01
desrtone comment first:11:01
desrtthe only sane way to handle the sort of thing that you're proposing, in my opinion, is to do it from inside of dconf itself11:01
seb128desrt, Laney: ok, you win, let's drop gsettings and use a-s then, I just fear that a-s is going to suck to use for app writers compared to gsettings11:01
desrtthe refactoring dconf got a cycle or two ago makes it pretty easy to support some kind of a plugin system inside the writer service that could do this sort of thing11:01
desrtie: have a mapping of dconf paths to account service properties11:02
seb128that would be nice11:02
desrtthen it's not really a 'sync' exactly11:02
Laneyinteresting11:02
Laneyan AS backend?11:02
desrtbecause when you write it it's actually writing to both places straight away11:02
desrtLaney: as a plugin to dconf11:02
ali1234and what if the user's desktop doesn't use dconf to store settings?11:02
desrti'm probably not going to think this is a good idea once i've had some sleep :)11:03
desrtso there's that....11:03
seb128ali1234, what is the user uses iOS?11:03
ali1234iOS doesn';t run on ubuntu, stop with the irrelevant comments11:03
desrtta.11:03
seb128ali1234, not sure the question makes sense, we are talking about a stack we write and we are using dconf11:03
ali1234i'll make this very clear for you: what if the uses uses xubuntu11:03
seb128does xubuntu wants a phone edition and reflect session settings on their greeter?11:04
ali1234no and yes11:04
seb128in which case they should use dconf11:04
seb128or figure out another way to do what they want11:04
ali1234we already have it btw11:04
Laneyneither AS nor this dconf idea I don't really understand would be ubuntu specific11:04
seb128Laney, they might not want to use dconf apparently11:05
seb128or AS11:05
ali1234we currently acieve it using AS11:05
seb128ali1234, do you use AS as the primary storage? (e.g is your desktop side reading there)11:05
ali1234no11:05
seb128or do you write to AS in addition to the "normal config"?11:05
ali1234yes11:05
seb128desrt, night btw ;-)11:05
ali1234xfce stores all settings in xfconf11:06
seb128ok, so same we do for the background image atm11:06
ali1234the background selector thing is patched to also write to AS11:06
seb128ok11:06
seb128that's a design we don't want11:06
seb128because then you need to teach every single app that want to change the background to write to AS11:06
seb128and duplicate that code tons of times11:07
ali1234well if you replace it with a design that uses dconf, xfce will not be able to use it, and we will have to keep the bad design11:07
seb128well, you could patch xfconf to do the same thing11:07
seb128whatever our solution is, it's not going to talk to xfconf anyway11:08
ali1234true11:08
ali1234of course that means the xfconf has to know the meaning of the schema11:09
ali1234same goes for dconf11:09
ali1234you still have to write a patch for every program that sets the background, but all the patches go into dconf instead of each individual program, and they still have to be kept in sync with the programs, but the developers of the programs won't see the patches11:10
seb128Laney, I need some time to think about it, I don't like much having to "sync", but I like less replacing gsettings by a-s, especially for something as important as the ringtone (if something bugs in that stacks it might mean you don't receive calls anymore on your phone, which sucks)11:11
seb128ali1234, why would you have to write patch? atm in ubuntu-desktop we just patch nautilus, it already monitors the gsettings config to update the rendering, we just added a "write to a-s" to the function it calls11:12
LaneyI think you should try and get some evidence for your lack of trust11:13
LaneyI don't share it and I don't know where it comes from11:13
Laneyrelatedly, I was kind of surprised to see no fallback code in telephony-service11:13
Laneydon't know what happens if the ringtone gets deleted or something11:14
ali1234seb128: yes, that same patch is currently in xfdesktop11:14
ali1234point being if somebody changes the way nautilus handles the wallpaper, then the patch won't apply in ubuntu and somebody will fix it. but not if the patch lives in dconf: then nobody will notice until it breaks on an install11:17
seb128Laney, ok, fair enough, I guess we should give it a try ... the lack of trust gets from "direct read from a mmap <-> talking to a process over dbus" and from "nice api designed to be nice to app writers <-> raw dbus calls to a service"11:17
seb128gets->comes11:17
seb128reading from a mmap is as solid as you can get, dbus roundtrip with another process involved seems more likely to hit bugs11:18
seb128ali1234, well, configuration keys should never change, but yeah, if that happens somebody will need to update all clients11:20
seb128Laney, didrocks' new landing process is going to be nice for those changes btw, when we do a landing ask it's going to give us a ppa with the new telephony-service and u-s-s to test ;-)11:21
Laneyheh11:21
=== MacSlow is now known as MacSlow|lunch
seb128Laney, ok, I made my mind, +1 for using a-s ... if we add fallback code, to play the default sound in telephony-service, in case the config is empty/buggy11:50
Laneyseb128: righto11:51
* Laney just built telephony-service for armhf to test that11:52
Laneynot the default thing, haven't done that yet11:52
Laneydear phone, please kindly turn on, no love, iain11:53
Laneydecent, it works12:05
=== alan_g is now known as alan_g|afk
=== MacSlow|lunch is now known as MacSlow
=== alan_g|afk is now known as alan_g
=== alan_g is now known as alan_g|lunch
=== alan_g|lunch is now known as alan_g
=== ara is now known as Guest58013
seb128Laney, there is a webkit 2.3.2 -> 2.3.4 update to do, want to do it?14:20
seb128Laney, I'm just asking because there is a wishlist bug from adam-yorba about it, wondering if I should assign it ;-)14:20
Laneyseb128: will do at some point14:47
Laneytest building on all arches sucks though14:47
seb128Laney, ok, so I'm assigning the bug to you14:47
Laneysure14:47
seb128right, I was pondering just doing a dch -v update and pushing to a ppa just to see what happens14:47
seb128;-)14:47
mlankhorstsure, makes sense14:49
Laneyif you have a ppa which builds for ppc* and arm64 ;-)14:49
Laneythere will be some patch wrangling14:49
mlankhorsthm speaking of which, I'll need ppc64 and arm64 enabled on x-staging at some point14:50
=== mjohnson15_2 is now known as mjohnson15
=== mjohnson15_2 is now known as mjohnson15
=== vrruiz_ is now known as rvr
=== vrruiz_ is now known as rvr
=== alan_g is now known as alan_g|tea
sil2100seb128: hi! How do you stand with free time today? You think you'll have a moment for a preNEW review?15:47
seb128sil2100, you can assume that every day is busy but it doesn't hurt to ask ... what source this time?15:48
seb128we need a queue for preNEW like for sponsoring etc15:48
sil2100seb128: it's15:48
sil2100lp:account-plugin-evernote this time15:48
sil2100(wild newline)15:49
seb128sil2100, ok, let me have a look15:49
sil2100seb128: thank you! :)15:49
seb128you are lucky, I just caught me at a time where I was done with something and looking for the next thing to pick from my todolist ;-)15:49
sil2100hah, excellent timing then ;)15:50
=== alan_g|tea is now known as alan_g
seb128sil2100, there is a typo in the description15:52
seb128" This plugin enables developers to authenticate to their Evernote developer15:52
seb128 sanbox account, which allows syncing of Evernote notes across Ubuntu devices."15:52
seb128sanbox -> sandbox15:52
seb128sil2100, NEWS and README are empty, no need to list them to install (e.g drop debian/docs)15:53
sil2100seb128: hah, didn't see that typo - will fix those issues straight away15:56
seb128sil2100, do we know where the icon is coming from?15:57
sil2100seb128: I will make sure we know exactly, but from what I see it's our creation right now16:02
sil2100seb128: it's the Reminders app icon: https://launchpad.net/reminders-app16:03
sil2100seb128: should I add copyright for this as well? Since I saw that the reminders-app doesn't have it copyrighted under debian/copyright as well16:04
sil2100seb128: so I thought it's a convention16:05
seb128sil2100, I'm asking because our artwork is usually CC-BY-SA16:05
seb128sil2100, but I'm not picky about it16:05
seb128having it under GPL wfm16:05
seb128sil2100, otherwise it looks good for NEW, maybe just check with design that the license for the artwork is ok16:05
sil2100seb128: will do, thank you! Fixing what was pointed out and asking design in a moment16:06
seb128sil2100, great16:06
sil2100seb128: https://code.launchpad.net/~sil2100/account-plugin-evernote/more_minor_fixes/+merge/201450 <- in case you want to confirm the fixes ;)16:13
seb128sil2100, looks good, thanks16:25
=== Sweetsha1k is now known as Sweetshark
=== gatox is now known as gatox_lunch
GunnarHjbregma: ping?17:33
bregmaGunnarHj, pong17:33
GunnarHjbregma: Hi Stephen! Your remark yesterday on the ubuntu-doc list made me curious. Can you give me a pointer to some info on the planned changes for 14.04 on input method handling?17:33
bregmaGunnarHj, we have related blueprint https://blueprints.launchpad.net/unity/+spec/client-1404-unity7-hotkey-handling but nothing specific on IM stuff17:36
bregmathere were problems in 13.10 wih IM and keyboard selection conflicts17:36
GunnarHjbregma: Indeed there were, and I hope that most of it has been dealt with by now. But as regards IM there have also been discussions on reverting to how things worked in 13.04, or switch to fcitx. I thought that nothing of that will happen, but your list comment made me wonder. ;-)17:39
bregmawe do not plan to switch to fcitx in Ubuntu 14.04 LTS17:41
GunnarHjbregma: Will the current "Text Entry" UI be just about as it is now?17:41
bregmaGunnarHj, we have some IM changes in Unity pending upstream17:42
attentebregma, GunnarHj, i have a ppa that should fix some of the problems17:42
attentehttps://launchpad.net/~attente/+archive/gnome-key-grabber17:43
GunnarHjThanks, attente!17:43
GunnarHjattente: One thing I'm wondering about is the default keyboard shortcuts for switching input source. Currently they involve the "Super" key which does not work.17:44
attenteGunnarHj, they'll work with the ppa, but the super key still also activates simultaneously17:45
attenteso you'll see the launcher numbers the same way the super key is held down17:46
attenteand possibly the shortcut help screen17:46
GunnarHjattente: Wouldn't Alt+Shift be a better default?17:46
GunnarHjattente: Less confusing...17:47
=== gatox_lunch is now known as gatox
attenteGunnarHj, i'm not sure tbh... it seems like a lot of the non-latin guys want something like that, while the cjk guys prefer something like modifier+space17:48
GunnarHjattente: Ok... What's most important, of course, is that the default works.17:49
GunnarHjattente: Should I install all those four packages to test?17:49
attenteGunnarHj, right17:49
attenteGunnarHj, sure, if you're up for testing the ppa, by all means17:49
GunnarHjattente: Will do. Somehow I got myself involved in this mess (via ubuntu-docs). ;-)17:50
attenteGunnarHj, heh17:51
GunnarHjattente: I tested your PPA, and can confirm that Super+Space works with those packages.18:32
attenteGunnarHj, thanks18:33
xclaessedunno who to ask, but what's the plan regarding kdbus in ubuntu ?18:45
xclaesseas I understand it needs a bit of userspace in systemd18:45
xclaessewill ubuntu implement its own ?18:46
=== tkamppeter_ is now known as tkamppeter
DelemasAfter a 13.04 to 13.10 upgrade my server logs are getting spammed every 6 seconds by systemd-logind as per: http://pastebin.ca/2535595 This happens even when user isn't logged in. Anyone know why?19:14
=== Guest87464 is now known as ayan
DelemasNevermind I found the cause... Monitoring service run amok...19:47
=== mjohnson15_2 is now known as mjohnson15
=== thumper is now known as thumper-gym

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!