=== maclin1 is now known as maclin
pittiGood morning04:29
hikikohi :-)05:34
didrocksgood morning06:47
pittibonjour didrocks !06:48
didrocksbonjour pitti :)06:49
seb128good morning desktopers06:50
seb128hey pitti, wie gehts?06:50
seb128re didrocks06:50
didrocksre seb12806:50
pittihey seb128! gut, danke! und Dir?06:50
seb128pitti, auch gut, danke06:50
seb128a bit tired this morning though06:51
pittij'ai me levé à 5:30, c'est trop tôt -- mais, beacoup de temps calm dans le matin :)06:51
pittilike for unscrewing debhelper before everyone else starts uploading again :)06:51
seb128hehe, I saw that on -changes06:53
pittibut I didn't manage to break devel, just -proposed; so no fun for xenial users06:53
* pitti gives britney a cookie06:54
hikikoHey pitti didrocks seb12807:07
seb128hey hikiko! how are you?07:07
hikikoI am fine :-) and you?07:07
didrockshey hikiko07:08
hikikoBonjour Didier07:08
seb128I'm good thanks07:15
didrocksgood morning Trevinho08:54
seb128hey Trevinho08:55
seb128how are you?08:55
willcookemorning all08:58
willcookelarsu, I'm having trouble with borders on GeditCloseButton not doing what they are told, specifically radius.  But.. it looks like I can set the border on the GtkImage of the close button and that works better.  Is that OK?09:01
seb128hey willcooke09:01
seb128how are you?09:01
didrocksmorning willcooke09:01
seb128willcooke, larsu was around until like 10 minutes ago, he went to the gym09:02
willcookeaint no thing, it can wait09:02
seb128willcooke, did you copy what the gedit-adwaita is doing?09:02
seb128I didn't pay attention to the borders much but that looked better to me that the non themed ones09:02
seb128well the ones in the sidebar documents list at least09:02
seb128you might be speaking of tabs or something else09:02
willcookeseb128, yeah got that all sorted, just trying to match the theme as closely as possible from the old one specifically when you hover over a close button you get a border09:03
seb128hey Laney! how are you? getting over the cold?09:04
willcookeseb128, hrm, but good point - adawita is doing what I want actually09:04
willcookeseb128, also I've finished copying the bits over from Adwaita gedit to our theme, just need to get these close buttons done and we're good to go09:05
seb128let me know when you have that up09:05
willcookewill do09:05
seb128I can test it there and then start on the landing09:05
Trevinhoseb128: so, so...09:06
TrevinhoHey didrocks09:06
seb128Trevinho, sooooo09:06
Laneyseb128: it is more of a cough now09:07
Laneysooooooooooo I hope that is progress?!?!?!09:07
didrocksLaney: happy that you start feeling better! :)09:07
willcookeLaney, can you remind me what, if anything, we decided about libinput?09:09
Laneywillcooke: that we thought tjaalton was handling it...09:10
Laneyor would09:10
willcookeah yes, and he said something along the lines of "it's pretty much necessary for the future, so we should try and get it in for 16.04"09:11
willcooke$customer asking about whether libinput would be in 16.0409:11
didrocks$customer is a customer with some dollars? :)09:12
LaneyI'm sure that either way the OEM build would be able to have it09:14
* Laney is using libinput driver already09:14
tjaaltonxserver-xorg-input-libinput is in the archive, but in universe as nothing depends on it yet09:14
Laneyhey tjaalton09:14
willcookeah cool, thanks tjaalton09:14
tjaaltononce the unity input config capplet supports it we can include it via xserver-input-all09:14
willcookeand then it would be available to anything that wanted to use it?09:15
seb128tjaalton, it replaces other drivers right?09:15
seb128are Debian/Fedora using it?09:15
willcookeTrevinho, or maybe seb128 can you comment on input config capplet ^^ ?09:15
seb128would we loose any feature by switching?09:15
tjaaltonseb128: fedora is09:15
tjaaltonit doesn't support wacom yet, but that's coming in libinput itself09:16
tjaaltonbut it would replace evdev and synaptics as a start09:16
seb128willcooke, well basically it means updating u-s-d/u-c-c code by importing what GNOME did, our versions are quite different so it's non trivial but still doable09:16
seb128tjaalton, how risky is the change?09:17
seb128like is likely to make trackpads not work correctly?09:17
tjaaltonseb128: I think it's an overall improvement09:17
seb128or edge scrolling or some things work differently?09:17
tjaaltonall work09:17
seb128do you have any cycle to backport the libinput conversion in u-s-d/u-c-c?09:17
seb128or is that something we should look at09:18
seb128I've that on my todolist but that list is long09:18
tjaaltonI'm working on lts-wily backports atm09:18
seb128I'm also trying to backport the new wacom code from g-s-d, same story09:18
tjaaltonthe u-s-d/u-s-c fork is for good?09:19
seb128well, for good ... at least for the coming LTS09:20
seb128I doubt it's relevant in unity809:20
tjaaltonno way to configure things?-)09:21
seb128in unity8?09:21
Laneythere's a new settings app there09:21
seb128let's keep that for another day09:21
seb128but yeah, there is ubuntu-system-settings09:21
seb128but it's limited, though there is ongoing work to make it fit for convergence09:21
seb128but that's not really important for the LTS/libinput discussion09:22
tjaaltonanyway, I think debian would be ripe to do the switch, need to ask other folks on the team09:22
tjaaltonthey don't have unity to worry about :)09:23
seb128GNOME was talking about dropping the non-libinput code09:23
seb128so Debian is eventually going to follow by just updating09:23
seb128tjaalton, what happens today if we start using libinput?09:31
seb128reading https://bugzilla.gnome.org/show_bug.cgi?id=740604 and looking to git, I'm not even sure we need u-s-d/u-c-c changes?09:31
ubot5Gnome bug 740604 in mouse "Avoid configuring xf86-input-libinput devices" [Normal,Resolved: fixed]09:31
seb128ah, it says they decided against the wrapper09:32
seb128so GNOME moved the handling in g-s09:32
tjaaltonso it seems09:32
tjaaltonI guess migrating now in xenial wouldn't be too dramatic09:33
tjaaltonbut at least all the forcepads I have need libinput, synaptics just doesn't work09:33
tjaaltonor needs too much pressure to move the cursor09:34
tjaaltonbut the kernel driver is lacking something so right-button emulation isn't working, making forcepads still somewhat crippled..09:34
tjaaltonbut that's not related to the migration09:34
pittihey Laney, how are you?09:35
* pitti missed you coming "in"09:36
pittiLaney: ignore the autopkgtest mass failure emails, dealt with (that was me breaking debhelper)09:36
seb128tjaalton, willcooke, I think we would need to update u-s-d/u-c-c to use the new api/property for libinput (not idea what changed but reading the GNOME bug it seems the way to change configs probably changed)09:40
willcookeand then for synaptics, we just keep using the old drivers?09:40
tjaaltonno libinput replaces the synaptics driver09:41
willcookeI R Confused.  So with libinput people with synaptics touchpads would have a bad time?09:42
tjaaltonlibinput replaces evdev and synaptics09:43
willcookeso people with forcepads would be happier with libinput09:43
willcookeand everyone else probably wouldn't notice09:43
Laneywould you remove those drivers?09:43
tjaaltonor my hw is just funky so that synaptics doesn't work well09:43
Laneyor have code to handle both?09:43
willcookehow do we turn "probably" in to something more concrete that we can rely on for an LTS?  Is it a case of spending more time looking in to it?09:44
tjaaltonno -evdev and -synaptics would still be around for some time, but probably gone before release09:44
tjaaltonmore folks using it. I know there have been issues with pointer acceleration code, but that's fixed by now I think09:45
seb128ok, so let's supposed we drop evdev and synaptics09:45
tjaaltonlibinput 1.2 will be released soon09:45
seb128and put libinput09:45
seb128what happens if we do no other change09:45
seb128if I go to u-c-c and change e.g edge scrolling09:45
seb128would that setting work?09:45
tjaaltonnot sure it would even show that09:46
seb128so we would regress functionnalities?09:46
tjaaltonon the config side yes09:46
tjaaltontwo-finger scroll still works :)09:47
tjaaltonat least here it just shows primary-button and double-click settings09:47
tjaaltonnothing else09:47
seb128so my touchpad has edge scrolling and I like that better than 2 fingers scrolling09:48
seb128with libinput I wouldn't be able to use edge scrolling anymore?09:48
tjaaltonit's easy to test, just install the driver09:48
tjaaltonyou can configure it with xinput09:49
Laneyoh, missed pitti's hello - hi!09:51
davmor2laney he was in stealth greeting mode09:51
seb128tjaalton, seems the current code does XChangeDeviceProperty() calls to change settings09:52
seb128        rc = XGetDeviceProperty (GDK_DISPLAY_XDISPLAY (gdk_display_get_default ()), xdevice,09:52
seb128                                 prop, 0, 2, False,09:52
seb128                                 XA_INTEGER, &act_type, &act_format, &nitems,09:52
seb128                                 &bytes_after, &data);09:52
seb128willcooke, so I guess it's currently not in our backlog, we don't have anyone understand the pro/con and the work needed09:53
seb128if that's something we should work on we need to create an item for it and investigate those09:53
seb128so not on the path to happen unless we decide it should09:54
seb128tjaalton, ^09:54
willcookethanks seb128 tjaalton09:57
willcookeyeah, sounds like there is a lot of work to be done before we know09:57
willcookemaybe I can find another team to look in to that09:57
seb128I'm unsure about "lot"09:57
willcookenon-trivial then :)09:58
seb128brb, restarting without evdev/synaptic drivers and with libinput one, just to see09:58
tjaaltonyou don't need to remove anything btw09:59
andyrockwillcooke: I need 5 more minutes09:59
seb128it might confuse u-c-c to think it can configure things where it can't09:59
willcookeandyrock, oki no worries - ping when you are ready and we can join the HO  ( <--- Trevinho, hikiko, seb128)10:00
andyrockwaiting for the lecture break but the the teacher is kind of excited maximizing a log likelihood :D10:01
Trevinhowillcooke: I'm in10:01
willcookeandyrock, :)10:02
hikikowillcooke, sec10:04
willcookeandyrock, we're in the HO and we will get started10:06
seb128willcooke, tjaalton, so installed libinput, removed evdev/synaptic, restarted session, my edge scrolling isn't working anymore10:07
seb128the touchpad section of the settings also vanished10:07
seb128no "tap to click"10:07
seb128or "edge scrolling"10:07
seb128so yeah, needs porting to the new api10:07
tjaaltonseb128: you don't need to remove anything, libinput just overrides the xorg.conf.d snippets evdev/synaptics provide10:08
seb128tjaalton, right, I just wanted to be a clean state to not confused the UI controls10:08
seb128to be on*10:08
seb128unsure what they check10:08
LaneyThese have to be updated to look at the libinput xinput properties no?10:09
andyrockwillcooke: looking for the link10:09
LaneyCurrently it looks for "Synaptics foo" for many foo10:10
Laneybut xinput list-props x shows a lot of "libinput foo"10:10
seb128Laney, right, that's basically what I was saying/confirming10:13
seb128needs code change10:13
seb128and GNOME didn't do those for us since they moved the handling to g-s10:13
Laneythey put it in gnome-screenshot?!?!?!?!?!?!?!?10:13
seb128libinput edge scrolling sucks :-(10:20
seb128I enabled it with "xinput set-prop 16 "libinput Scroll Method Enabled" 0 1 0"10:20
seb128but it's practically impossible to use on my latitude10:20
seb128dunno, I've to have my finger really close from the border and press more than before10:23
seb128and it only detects it as a scrolling half of the time10:23
seb128that is deprecated but it explained by libinput was not doing edgescrolling/how it conflicts with other features10:23
seb128so maybe they detection/logic is different from what older drivers are doing10:24
seb128in any case it's harder to use on my machine10:24
tjaaltoncould just be a bug10:25
tjaaltonseems fine on mine10:25
seb128well, it just confirms that it's not a trivial no-change swap10:25
seb128we are introducing new code and different behaviours10:25
seb128and different set of bugs10:25
seb128anyway, enough playing with that10:25
seb128it's enough to confirm that changing involves some wokr10:26
willcookethanks a lot seb128, tjaalton, Laney10:26
seb128going to let willcooke&other managers decide if it's high enough that we need to budget cycles for it10:26
seb128willcooke, yw!10:26
seb128brb, rebooting10:26
tjaaltonwell, perhaps keep evdev/synaptics around and make it easy to switch back10:27
Laneywho's going to do the u-c-c/u-s-d work?10:27
willcookeno one right now10:28
LaneyI mean in theory10:31
Laneyfeel like it's sort of being glossed over but that is tough10:32
seb128I missed some context I guess10:32
LaneyI tried edge and it seems okay here, hate that mode though :P10:32
willcookewell, I imagine someone in our team would work on it10:32
willcookeseb128, <Laney> who's going to do the u-c-c/u-s-d work?10:32
seb128Laney, that was my comment about "up to the managers to decide how much of a priority it is and what resources get allocated to it"10:34
Laneyif we do no work then OEMs can still include it and preset their defaults in xorg.conf snippets btw, just users won't be able to configure it with a gui10:39
willcookeLaney, yeah good point.  If they do want it on in their images, I'm hoping that we can get them to commit to some of the work as well10:45
Laneywillcooke: yeah - the feature tests mean that we don't expose any *broken* UI10:46
Laney- g_bit_nth_lsf@Base 2.12.010:46
Laney- g_bit_nth_msf@Base 2.12.010:46
Laney- g_bit_storage@Base 2.12.010:46
Laney+#MISSING: 2.47.2-1# g_bit_nth_lsf@Base 2.12.010:46
Laney+#MISSING: 2.47.2-1# g_bit_nth_msf@Base 2.12.010:46
Laney+#MISSING: 2.47.2-1# g_bit_storage@Base 2.12.010:46
Laneythere's some weird shit going on around inline functions in this release10:47
larsuwillcooke: I guess it should be ok. in general we try to avoid setting things on widget names directly10:47
larsuand use classes instead10:47
larsumorning all that I have missed10:47
larsuhi Laney :)10:47
Laneybeen @ gym?10:47
larsuya. and going to lunch now10:47
seb128larsu, enjoy!10:47
Laneytough day10:47
* larsu will have to work more in the afternoon/evening today :)10:47
willcookelarsu, got it, thanks.  Had a look at how Adawaita was doing it and saw that was the correct way, and actually, I just got it sorted10:47
larsuwillcooke: very cool :)10:48
* larsu thinks willcooke is shaping up to be a good theme maintainer10:48
willcookeit seems though that -gtk-gradient borders can't have radiused edges10:48
larsuhm really?10:48
willcookeI'm verrrrry slow10:48
larsuif css can do it, gtk should be able to as well10:48
willcookeyeah, at least border-radius doesn't have an effect10:48
willcookemaybe I need to slice them to make it work10:48
willcookebut that's too hard for me10:49
larsuanyway - to late for my lunch date. see you in a bit10:49
willcookebackground-color: rgb(255,0,0); is the CSS equivalent of print "XXXXX I am here!"10:57
* willcooke <-- l33t hax0r10:57
Laney* { color: red; background-color: limegreen; } // IS THIS THING EVEN RUNNING?10:58
* Laney has done this kind of thing before10:59
Laneywhen fixing o-s issues back in the day11:00
seb128Trevinho, btw did you meant to ask about something earlier with your "so, so..."?11:02
Trevinhoseb128: no, as few mins before you asked how I was... :)11:03
seb128Trevinho, oh, sorry, I didn't read that as a reply, rather than as an opinion to a topic to discuss11:03
* seb128 hugs Trevinho11:03
seb128Trevinho, did you find a way to deal with the matching issue with nautilus? btw did you see my comment about gedit/are you looking at that one as well?11:10
willcookeseb128, I want to check a couple of bits of styling before commiting these changes.  Can I just add a "padding: 10px" (or whatever it was in to gtk-widgets.css for testing?11:11
willcookeoh, this is for the file dialog11:11
willcookenot the normal file browser, the other one11:12
seb128sure, why not?11:12
seb128the sidepane widget is shared between nautilus and file-selector11:12
seb128I'm unsure what you are asking?11:12
seb128https://code.launchpad.net/~willcooke/ubuntu-themes/fix1515557/+merge/278423 ?11:13
Trevinhoseb128: for nautilus https://code.launchpad.net/~3v1n0/bamf/safer-pid-desktop-registration/+merge/27846111:13
Trevinhoseb128: for gedit, I don't know... As I said yesterday we already use the _GTK_APPLICATION_ID (thanks to larsu's work) to use it as desktop-id... So not sure why it doesn't work11:13
Trevinhoseb128: I didn't see the issue btw (yet)11:14
willcookeseb128, ah, this is the one I meant:  https://bugzilla.gnome.org/show_bug.cgi?id=75858511:15
ubot5Gnome bug 758585 in file loading and saving "Open menu dropdown is 'compressed' & unintelligible with other themes" [Normal,Resolved: fixed]11:15
willcookeI'll just hack that fix in so I can test gedit11:16
willcookeand make sure I don't commit it11:16
seb128willcooke, you can also upgrade gedit, that fix landed in xenial yesterday11:16
willcookeah, even better, thanks seb12811:17
seb128Trevinho, with gedit the issue is that if I dnd gedit from the dash to the launcher I gedit a gedit.desktop added to my launcher11:20
seb128but if I start gedit from a command line I get11:21
seb128_GTK_APPLICATION_ID(UTF8_STRING) = "org.gnome.gedit"11:21
seb128which leads to a new icon in the launcher11:22
seb128same if I open a document from gedit11:22
Laneyseb128: could those theme tweaks go to the 3.18 branch too?11:24
Laneyoh they did11:24
Laneynice ;)11:24
* Laney needed to pull11:24
seb128willcooke, Laney, https://bugs.launchpad.net/gtk/+bug/1519778 just for info (corresponding bugzilla bug filed as well)11:40
ubot5Ubuntu bug 1519778 in gtk+3.0 (Ubuntu) "Single click opening of selected items in file-selector is confusing" [High,New]11:40
willcookethx seb12811:41
willcookehrm, my theme changes are working.  Back to the drawing board...11:43
seb128not working?11:44
willcookemy selectors arent matching11:44
didrocksseb128: I guess will feels that working is wrong :)11:44
willcooke*are not11:44
seb128charles, hey, any chance you could have a look to https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1261191 ? it seems like a good one to fix for the LTS11:44
ubot5Ubuntu bug 1261191 in indicator-datetime (Ubuntu) "indicator-datetime list evolution calendar all day event a day before actual date of event" [High,New]11:44
didrocks"what, this is working, let's revert it" :)11:44
* Laney switches on bbc parliament12:12
TrevinhoLaney: what's happening?12:18
davmor2Sweet5hark: do you have a screenshot of what the icons in LO should look like on Ubuntu at all, I think I still see the gnome ones in kvm on the daily in live session on first start up but would like to be sure12:18
LaneyTrevinho: a bit of political theatre, some budget cuts to be announced12:19
Laneyseb128: thx for the bug12:19
* Laney screams12:23
Laneysun just started reflecting off a window opposite12:23
* Laney goes blind12:23
davmor2Laney: what is this sun of which you speak12:23
Laneyyou don't want it12:24
* didrocks would want some sun now :(12:26
* davmor2 doesn't believe Laney has it, you sure it is not a Police search helicopter light?12:26
desrtGood morning all12:40
desrttime to get my coffee, croissant, and a side of HIDEOUS ABI BREAKAGE12:41
Laneyhi desrt! :D12:42
desrtwill take a look :)12:43
Laneymaybe just a missing available annotation?12:43
* Laney wants a croissant12:43
desrtya.  visibility is the first thought that occurred to me too12:43
desrtsince I'm fairly sure the definitions are being properly emitted12:44
desrtbut seriously... the abi thing was third in that list for a good reason :)12:44
didrocksgood morning desrt!12:45
desrtmoin :)12:45
xnoxdesrt, what abi breakage?!12:45
didrocksxnox: a gliby thingy12:46
didrocksxnox: http://irclogs.ubuntu.com/2015/11/25/%23ubuntu-desktop.html#t10:4612:46
desrtinlining in glib has been a disaster for a while -- we had inline versions in headers and backup versions in the .c file in case the compiler that you were using to build your project which included the headers didn't suport inlining12:47
desrtwe gave up on that now since all compilers do inlining12:47
desrtso i cleaned up a lot of that mess, and tried to preserve the old versions for strict ABI compatibility12:47
desrtbut clearly failed :)12:47
=== hikiko is now known as hikiko|ln
* desrt escapes key-fob insanity13:44
* desrt dives into ABI insanity13:44
desrtour building has massive keyfobs that double as garage-door-openers and i'm tired of carrying mine around... and i've long suspected that the transmitter and keyfob aspects of it operate separately... so i took it apart today to figure out which part of it was responsible for each13:46
desrtoddly enough, the thing stopped working after i took it out of the case13:47
desrtturns out the nfc/rfid bit is the back panel of the case itself13:47
desrtLaney: visibility was indeed the problem.  want a new release today?13:49
desrt(i already pushed the fix if you want to cherrypick it in the meantime)13:49
=== hikiko|ln is now known as hikiko
Laneydesrt: probably a good idea, other distros might pick it up and not notice14:05
willcookelarsu, I'm going insane here.  My padding is being ignored for buttons in gedit, but everything else (like a test background colour) is being applied14:05
Laneythanks for fixing14:05
willcookeany ideas?14:05
willcookewhen I paste it in to the inspector it works14:07
seb128willcooke, what rule do you use?14:07
willcookedon;t understand "rule"14:07
willcookecss selector?14:08
seb128sorry, what lines do you add to the css14:09
seb128so we can try14:09
willcookeI'll pastebin it14:10
larsuwillcooke: you can see in the inspector where a value comes from14:10
larsuin "style properties"14:10
larsuit tells you file and line nr14:10
willcookelarsu, it's blank14:10
larsuthen your rule doesn't apply14:11
willcookebut the b/g colors *are* being applied14:11
larsuwhat is it? Do you have a branch?14:11
larsuhehe I was about to complain about the non-clickiness of that first link ;)14:12
larsuthat's apps/gedit.css?14:12
* willcooke wonders if it could be some of the @xyz definitions being wrong for our theme14:22
larsuwillcooke: ya, that seems to be it indeed (remove the "theme_" prefix)14:23
willcookeheh, thanks larsu, sorry to bother you with that14:27
larsuno worries :)14:28
Laneypitti: do you know if we're allowed to use backports in cloud deployed things?14:28
Laneytrusty doesn't have a new enough lmdb, but I could backport it14:28
pittiLaney: I'm not sure that we even have a policy about what you can use14:28
Laneythat sounds great to me ;-)14:29
pittiLaney: indeed trusty is pretty strongly recommended, but as long as you keep stuff updated backports sohuld be fine14:29
LaneyI thought it was pretty tight14:29
Laneybut actually I found a page about dev managed stuff even in prodstack14:29
Laneyso if you get full control...14:29
* Laney goes hog wild14:29
pittiLaney: I need a newer twitter-bootstrap3 for debci as well14:29
Laneypitti: I saw something where you wget a deb from the archive14:30
Laneythat was great :P14:30
pittiLaney: yes, I manage autopkgtest myself, it's not CI managed14:30
pittiLaney: well, it's an utter hack, but *shrug*14:30
Laneypitti: this is prodstack, right?14:30
Laneyjust the workers in scaling?14:30
pittiLaney: the adt testbeds are in scaling; the workers themselves are on a PS instance14:31
Laneyyeah sorry, that's what I meant14:31
pittiLaney: what are you charming up, OOI?14:31
Laneyappstream extractor14:31
Laneyanother thing I'm not sure about is how to handle the cinder volume holding the archive mirror14:32
Laneyit'd be a pain to have to start this from fresh each time14:32
Laneyso maybe I can do it out of band14:32
pittiLaney: archive mirror?14:32
Laneyjuju doesn't seem to support this in the way I want yet14:32
Laneyyeah, need a local one14:32
pittiLaney: ah, you want to attach a separate storage to your instance with that data14:33
Laneyyes, it shouldn't be tied to the unit(?)14:33
Laneyotherwise you have to re-mirror from scratch each time14:33
pittiLaney: where "each time" is just "when you redeploy", no?14:33
pittiwhich should happen like twice a year14:33
pittiso I wouldn't sweat that too much14:33
pittiLaney: and the mirror that you get in the cloud is pretty fast already14:34
Laneydon't know14:34
Laneyit's still a multiple hour operation14:34
pittiLaney: TBH I'd just start with setting up debmirror as part of your charm14:34
pittiLaney: you can upgrade a charm, you don't always need to re-deploy from scratch14:34
LaneyI thought I might have the deploy script handle attaching a volume14:35
pittiwelll, that indeed sounds elegant14:35
Laneyi.e. do it out of band14:35
pittiLaney: I found that you can't charm quite a number of things anyway; you're going to need some deploy.sh script which invokes the various juju deploy, juju add-relation etc. commands, so you can do it there too14:36
* Laney shrugs, will see, maybe it does prove to not be required that often14:36
Laneyit would also be good to store the assets themselves in a resillient fashion14:36
Laneyso after a redeploy we can serve the latest run until we catch up again14:36
Laneymaybe I'm overthinking this for this early stage :)14:37
pittiLaney: so you can deploy your service14:37
pittiLaney: then use direct OpenStack commands to add the storage to your newly created machine14:37
Laneyyeah, that bit should be fine14:38
Laneyand I store the assets on this volume too14:38
pittimachine_no="$($(juju status myservice) | grep machine: | grep -o '[0-9]\+')"14:39
pitti    machine=juju-${env}-machine-${machine_no}14:39
pittiand then $machine is the thing to talk to with openstack commands line nova14:39
Trevinhodesrt: hey, do you know what's the best way to get a gfile uri to be transformed to the actual path? I.e. I'd like to get the actual path from "trash://"14:54
larsuTrevinho: g_file_get_path()14:59
larsuit returns NULL though for things that don't have a path14:59
Trevinholarsu: tried in python but I get nothing15:00
larsuincluding trash://, I think15:00
Trevinhothat's the thing15:00
Trevinhoif you don't initialize it with a path, it won't work15:00
larsutrash isn't really a path...15:00
* Trevinho wonders what nautilus uses15:00
larsugio api15:00
Trevinhoyeah, I know.. I mean, what it uses to get the proper path to show15:01
larsuwhere does it show that?15:01
TrevinhoI mean, when you open trash:// in the location bar, it just open it as an ur and then fetches the children of that, or it maps to the real path to show?15:01
larsuTrevinho: it fetches the children with gio api15:09
larsuthere's no path on the file system mapping to that15:09
TrevinhoMh, I see..15:10
desrtTrevinho: what larsu said15:10
desrtTrevinho: imagine you have http:// uri15:10
TrevinhoWell the thing is that if you go in the trash, the you open a subfolder...15:10
desrtsome GFile's simply do not have a path15:10
Trevinhonautilus open the actual paths15:10
Trevinhothe location is not trash://subfolder15:10
desrtthat seems odd15:11
Trevinhoerr trash:///subfolder15:11
desrtcuriously, i also see the same15:11
desrtnautilus may have some special code to do this15:11
Trevinhoyeah, that's why I was wondering how to dothat...15:11
desrtpretty sure there is nothing on GIO API to support this15:11
desrt  standard::target-uri: file:///home/desrt/.var/lib/Trash/files/trashthis15:12
desrtso there you go15:12
larsuerrr where do you see this?15:12
Trevinhostandard::target-uri ?15:12
desrtthis is very strange, since the type of the thing is a directory15:13
desrti wonder why i did that.....15:13
larsuno sorry, I mean in nautilus15:14
desrti didn't do it.  it was hadess15:14
larsuI don't see the target-uri btw15:14
desrtsee https://bugzilla.gnome.org/show_bug.cgi?id=66779415:14
ubot5Gnome bug 667794 in trash backend "g_file_get_path() shouldn't return NULL" [Normal,Resolved: fixed]15:14
desrtso there you go15:16
desrtTrevinho: my official answer is "it didn't used to do that but now it does and i don't really understand the reason"15:16
Trevinhomh, ok :-D15:17
desrt"so that totem can play DVDs from the trash"15:17
larsuwtf? Seriously?15:17
desrtthat's what the commit says :)15:17
desrtLaney: new glib release is out15:20
desrthopefully this one isn't hideously broken15:21
Trevinhodesrt: so.... There's no really way to get me the actual trash path?15:21
desrtTrevinho: but there is!15:21
desrtexamine the standard::target-uri property, i guess15:21
Trevinhodesrt: a part from going trhough children..15:21
desrtyou can query it from g_file_query_info15:21
Trevinhodesrt: there's nothing for gvfs-info trash:///15:21
desrtit has to be on a file in the trash15:22
Trevinhodesrt: there's for gvfs-info trash:///trashed-child15:22
desrttrash:/// isn't an actual location15:22
desrtit is a combination of several locations (for example trash folders on removable media)15:22
Trevinhoyeah, but I wanted to get from trash:/// -> file:///home/marco/.local/share/Trash/files/15:22
desrtdifferent items can be in different actual locations15:22
desrtif you want to know the homedir one, the answer is simple15:22
desrtby the spec it's in XDG_DATA_HOME/Trash/files15:23
Trevinhodesrt: ok, that's what I wanted to use...15:23
Trevinhodesrt: but you defined data_home on ~/.var/lib ?15:23
larsuTrevinho: what are you doing?15:23
desrtTrevinho: i set my XDG_DATA_HOME to that15:23
Trevinholarsu: launcher icons <-> storage windows matching15:24
Trevinhodesrt: ok, then I'd go for that then15:24
desrtread http://standards.freedesktop.org/trash-spec/trashspec-1.0.html15:24
desrtbut consider removable media15:24
larsuTrevinho: shouldn't that matching be on the uri, and not on the path?15:24
Trevinhodesrt: yeah, that's the thig15:25
Trevinholarsu: in fact I was doing it using the URI...15:25
desrtjust use GFile...15:25
desrtyou do not want to reimplement the trash spec15:25
larsunautilus windows tell you on dbus what they are showing right now (as a URI)15:25
desrtsays someone who has implemented the trash spec15:25
Trevinholarsu: then... If you open nautilus trash and you go inside a trashed folder, nautilus sayus that you're in ~/.local/....15:25
larsuTrevinho: that's a bug15:26
desrtya.. this is a combination of a nautilus bug and hadess being insane15:26
Trevinholarsu: that's what I'm using, but it's not using the URI for trash sub-folders15:26
desrtone of those two situations ought to be rectified15:26
larsuit should though15:26
desrt(perhaps both)15:26
desrti can totally see the nautilus logic here15:26
Trevinhothat's why I asked... otherwise my coed would have worked perfectly without caring about the actual trash uri15:26
desrt"did the user click on something that has a symlink destination?  go to it."15:26
desrtwhen the logic should be:15:26
desrt"did the user click on something that IS A SYMLINK AND has a symlink destination?  go to it."15:26
desrtwhy hadess sets symlink destination on a directory?  no idea.15:27
Trevinhodesrt: so you'd say we should fix nautilus or..... just use a workaround in unity side?15:28
larsufix nautilus :)15:30
* larsu hides15:30
desrtTrevinho: i'm chatting with hadess.  i think we need to fix gvfs.15:30
desrtwe should also think about fixing nautilus while we're at it15:30
Trevinhodesrt: ok, thanks... Let me know what's the final decision...15:32
* Trevinho drives back to Florence15:32
seb128Trevinho, safe drive!15:32
Trevinhosnow allert :o15:32
Trevinhoseb128: thanks15:33
seb128tedg, hey, do you remember what is the deal with bug #1165420? did we remove indicators content from the hud on purpose? or was it a side effect of the gmenu transition?15:41
ubot5bug 1165420 in hud (Ubuntu) "Unable to access indicators from HUD" [High,Confirmed] https://launchpad.net/bugs/116542015:41
attentehi, does anyone know or have a debian/watch file for github releases?15:45
seb128attente, ibus has one15:52
seb128https://github.com/ibus/ibus/releases .*/ibus(?:-|_|_v|V)(\d\S*)\.(?:tar\.xz|txz|tar\.bz2|tbz2|tar\.gz|tgz)15:52
seb128unsure if it works though15:52
attenteseb128: thanks, i will try it15:54
xnoxchecking for XMIRMODULES... no16:09
xnoxchecking whether to build Xmir DDX... yes16:09
xnoxconfigure: error: Xmir build explicitly requested, but required modules not found.16:09
* xnox needs help16:09
seb128xnox, where is that?16:09
xnoxseb128, on s390x16:10
seb128robert_ancell would probably be your best bet but he's sleeping at this time16:10
xnoxi guess i should skip it, just like it is skipped on arm64, powerpc and ppc64el16:10
seb128maybe bregma or ChrisTownsend or tedg can help you16:10
seb128yeah, probably16:10
xnox#shame -> i already bootstrapped mir16:10
tedgseb128, it was removed because it failed in user testing. The plan was to bring it back with additional info, but it never got implemented.16:59
seb128tedg, ok, seems quite some users liked/would like that (me included), I was wondering if that was something to "fix" for the LTS17:00
seb128I though it was buggy17:00
seb128not disabled on purpose17:00
tedgSadly, no clue on Mir. Try #ubuntu-Mir17:00
seb128I would argue that the audiance that uses hud is probably not confused by indicators being listed in there17:01
tedgseb128, I am +1, I doubt there is anyone in design that remembers that user study anyway ;-)17:01
seb128would it be easy to re-enable? do you remember where that was changed?17:02
tedgseb128, it has been rewritten since...17:02
seb128oh, right :-/17:02
seb128tedg, thanks17:02
seb128so a valid wishlist but not likely trivial17:03
tedgBut we now have an indicator registry, so it should be easier now.17:03
seb128tedg, k, let's see if somebody can pick it up as LTS polish, I put it on the list17:07
didrocksgood night everyone!17:38
willcookecya didrocks17:39
didrockssee you!17:39
Trevinhodesrt: got anything from that discussion?17:55
desrtannoyingly, no17:55
Trevinhodesrt: mh, so... going for a workaround for now17:55
desrtTrevinho: we're still chatting about it18:03
desrtTrevinho: but sure... more workarounds are cool too18:03
Trevinhodesrt: well, I don't love them, but if we can't find a better solution upstream...18:04
desrtwe can18:04
desrtbut you haven't given much time18:04
Trevinhomaybe :-)18:04
TrevinhoI was wondering... In case of a subfolder trashed on an external device...18:05
Trevinhowhat's its uri?18:05
TrevinhoI mean, trash://media-id/folder or what?18:05
attenteseb128: could i submit a new package for upload to the archive?18:10
=== alan_g is now known as alan_g|EOD
Laneyglib uploaded to sync tomorrow, some first work on appstream charm pushed18:22
Sweet5harkwillcooke: fyi http://nabble.documentfoundation.org/Candidacy-Bjoern-Michaelsen-td4167393.html19:13
willcookeSweet5hark, good stuff!  Good luck19:14
* Sweet5hark starts kissing babies and buys a smear campaign on TV.19:15
ricotzSweet5hark, lol, thumbs up19:55
=== ralsina is now known as ralsina_
=== ralsina_ is now known as ralsina
willcookeI have vanquished by Gtk theme foes with a combination of brute force and ignorance20:48
willcookeI'll finish it tomorrow20:48
willcookeg'night all20:48
=== vrruiz_ is now known as rvr

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