=== m_conley is now known as m_conley_away [04:39] Bo. The dash isn't a feature-complete replacement for gnome-do :) [05:56] sb levelclear -level clientcrap,crap,joins,parts,quits,nicks,clientnotice === _LibertyZero is now known as LibertyZero [08:00] hey there everybody [08:01] Howdie MacSlow! Good weekend? [08:01] hey RAOF... yeah... spent most of it on twisting country-roads... leaving tire-rubber everywhere :) [08:02] Rev head! :P [08:02] RAOF, hell yeah! [08:12] good morning [08:17] good morning [08:24] good mornings [08:25] after sometimes the app-menu "loose" the window and only display "File" with just one Close option [08:32] zniavre, something similar to https://bugs.launchpad.net/unity/+bug/718926 ? [08:32] Ubuntu bug 718926 in bamf (Ubuntu) "Some apps don't integrate to appmenu after having their windows closed" [High,Triaged] [08:35] Davidc_3, thank you [08:36] zniavre, :) [08:40] there is no workaround or solution i guess we hav to wait for update so [09:17] Cimi: good morning. do you know how if it's possible to modify the gtkrc for the ambiance theme to make one style have fully transparent background ? [09:28] nerochiaro, i take the power to answer if you don't mind > https://wiki.ubuntu.com/DesktopTeam/RgbaGtkWithPPA/ [09:28] nerochiaro: what style? [09:28] ooops sorry [09:29] Cimi: zniavre: panel style in /usr/share/themes/Ambiance/gtk-2.0/apps/gnome-panel.rc [09:30] nerochiaro: you want a transparent gnome panel using the gnome panel's transparency? [09:31] nerochiaro: then simply comment the bg_pixmap line [09:31] Cimi: it's for unity-2d. we embed the indicators and i want them to be transparent. and i saw that they use the panel color [09:31] er, the panel style [09:31] unity-2d panel does not follow the gtkrc i believe [09:32] zniavre: the indicators follow it, the are embedded gtk widgets [09:33] a ok [09:33] * nerochiaro has stiff fingers this morning, can't type [09:33] nerochiaro: is there a ppa I could test for unity 2d playing with this stuff? [09:35] Cimi: yes, let me have a look [09:35] nerochiaro: ok so link me to the bugreport, PPA [09:35] I'll play with this [09:35] maybe later today or tomorrow [09:37] seb128: are there any high prio bugs you'd like to assign me [09:38] I've got one dbusmenu bug which is taking a while to figure out what's going wrong, and I think I should be doing something more productive [09:38] Cimi: https://launchpad.net/~unity-2d-team/+archive/unity-2d-daily [09:39] nerochiaro: thx [09:39] Cimi: I don't have a bug number yet, but all I really need is to have the indicators in the panel be transparent and show whatever is below them (in this case the unity-2d panel itself) [09:39] Cimi: thanks to you [09:40] yw [10:15] hmmm if a desktop file has spaces in the Name field, my indicator only shows the first word. Is this a known bug? === API is now known as Guest76134 === Guest76134 is now known as apinheiro [10:49] klattimer, hey, well there is quite some bugs open, just claim any which isn't assigned yet [11:02] seb128: from what list of bugs? [11:02] like I mean, where can I find a selection of appropriate bugs? [11:07] klattimer, https://launchpad.net/ubuntu/+source/indicator-datetime/+bugs [11:07] ? [11:08] I think I've been through most of those, mterry also has some branches [11:08] haven't seen this list today though so I'll have a poke [11:08] klattimer, well, things to work on [11:08] bug #691953 [11:08] Launchpad bug 691953 in indicator-datetime (Ubuntu) "Uninstalling Evolution removes calendar from clock applet" [Low,Triaged] https://launchpad.net/bugs/691953 [11:08] let me know if there's anything more important though [11:09] bug #726053 is that's still an issue [11:09] Launchpad bug 726053 in indicator-datetime (Ubuntu) "Resuming from suspend should trigger a time update" [Low,Confirmed] https://launchpad.net/bugs/726053 [11:09] bug #743394 [11:09] Launchpad bug 743394 in indicator-datetime (Ubuntu) "inicator-datetime dosen't show corretly the time when changing from 24 hours format to 12 hours" [Undecided,New] https://launchpad.net/bugs/743394 [11:10] #726053 I can't reproduce on my hp110 it sleeps most of the time and is up to date [11:10] hadn't seen that last one I'll check it out [11:11] uninstalling evo problem could be solved by at least showing the calendar but not hooking up events to it I think === daker_ is now known as daker [11:11] klattimer, well the comments suggest the calendar widget is hidden when the evolution binary is not on disk [11:12] yeah that's the case [11:12] klattimer, some users replace it by an empty file to workaround that [11:12] getting the calendar even empty should still be useful even if evo is not installed? [11:12] I'm looking at the code now to see how to work around it ;) [11:12] klattimer, bug #730476 if you have any clue as well [11:12] Launchpad bug 730476 in indicator-applet (Ubuntu) "Part of the time appears cut off when displaying the day of the week and date" [Undecided,Confirmed] https://launchpad.net/bugs/730476 [11:12] it's not happening there but several users have it and have screenshot showing the issue [11:13] so there is an issue [11:13] that looks like a 2d issue [11:16] klattimer, not sure, dpm has it and he's using 3d I think [11:16] oh? [11:16] hmm [11:16] might be a translation bug in working out the size [11:16] I'll look into it [11:17] thanks [11:17] I think I've got a nice fix for the evo bug btw [11:17] klattimer, I think you have enough with those ;-) [11:17] I'll push it and offer a merge request [11:17] ok [11:17] :) [11:17] great [11:17] btw the 2 second bug is fixed [11:17] now we need ted to merge the stack of fixes mterry and you got [11:17] but we need a merge on that one too as it's now sub 1 second [11:17] yeah, seems to work fine there as well [11:18] I _really_ want to fix https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/740980 but my testing brings up nothing to work with, I'll ask ted if he can look later [11:18] Ubuntu bug 740980 in indicator-datetime (Ubuntu) "Date with event isn't highlighted in the calendar" [Undecided,In progress] [11:20] klattimer, heya seb128, on unity-3d it's fine, the problem is only on the classic desktop (no effects) session [11:20] klattimer, so you are right it's a 2d bug [11:20] k [11:20] good to know [11:20] is there an easy way to switch to unity2d? [11:21] I only get it when my 3d borks [11:24] klattimer, well, dpm is using GNOME, so you can pick the classic session in gdm for that [11:24] k [11:24] klattimer, or you can run gnome-panel in your unity session if that's specific to it [11:25] seb128: https://code.launchpad.net/~karl-qdh/indicator-datetime/alwaysshowcalendar/+merge/55116 [11:25] Ubuntu bug 55116 in Launchpad itself "new bug mail seems to take a very long time to process (dup-of: 54369)" [Undecided,New] [11:25] Ubuntu bug 54369 in Launchpad itself "New bugs filed using the email interface should bypass the mail digest" [Low,Triaged] [11:26] klattimer, seems fine to me, and one less bug on the list ;-) === MacSlow is now known as MacSlow|lunch [12:28] smspillaz: ping [12:32] didrocks, hmm, https://bugs.launchpad.net/unity/+bug/721322 is still 'fix committed', but afaik it was included in 2 weeks ago's release, right? [12:32] Ubuntu bug 721322 in unity (Ubuntu) "Panel service needs to know coordinates of the indicators" [Undecided,Fix committed] [12:34] rodrigo_: it wasn't target to a milestone [12:34] didrocks, ah, ok, so can I just mark it as fix released? [12:34] rodrigo_: yeah :) [12:34] ok [12:34] thanks :) === clarita_ is now known as clarita [12:40] Hi Ayatana folks, I'm looking for people willing to test a Unity Books Lens, I'm in need of bug reports. Is anyone interested? [12:46] Is Python gi.repository.AppIndicator broken? I'm trying to move from appindicator to the pygi version, but get this error: [12:46] TypeError: GObject.__init__() takes exactly 0 arguments (3 given) [12:46] when i try to create a AppIndicator.Indicator() [12:52] jderose, I got around this by using AppIndicator.Indicator.new() [12:52] artfwo: hmm, i tried that, but got same error.... do you pass any arguments to AppIndicator.Indicator.new()? [12:53] jderose, this code snippet just worked for me http://pastebin.ubuntu.com/586427/ [12:53] artfwo: if you could paste.ubuntu me a working example, i would be forever in your debt :) [12:53] hehe [12:54] * jderose tests [12:57] artfwo: works... a similar looking error is fouling me up further down the lane, mistook for same error when i tried this before. thanks! [12:57] jderose, yw [13:23] hi guys, is the self-implemented StatusNotifierItem the only way for pure Qt app to have indicator for now? except using C library, ofc [13:44] dbarth, you here? [13:48] dbarth, I have a merge proposal for the alt+f2 thing [13:49] dbarth, just wanted to ask if I should relate this to the dash bug or create a new one [13:49] dbarth, lets talk later [13:58] Cimi: if i want to get the gnome-panel itself to be transparent, the only thing i need to do is to remove bg_pixmap[NORMAL] from /usr/share/themes/Ambiance/gtk-2.0/apps/gnome-panel.rc ? === m_conley_away is now known as m_conley [14:00] nerochiaro: i GUESS SO [14:00] ops [14:00] Cimi: just tried, does not work [14:01] nerochiaro: I mean, you then need to right click on the panel [14:01] and enable transparency [14:03] Cimi: ah, it works. do you know what does the panel do when i enable transparency ? [14:06] Cimi: ah, looks like it just load the background manually and set it as background pixmap for the panel window === MacSlow|lunch is now known as MacSlow [14:37] this is a place to report scrollbar overlay errors please ? [14:48] i think that you can report bugs here https://bugs.launchpad.net/ayatana-scrollbar/ [15:09] didrocks: are you doing reviews today? [15:10] jcastro: yeah, I want to finish some work first, but I'll [15:10] I just can't do all of them :) [15:19] tedg, gwibber replies to yourself is fixed in trunk :) [15:20] kenvandine, Woot! That's service! [15:21] mterry: hi, i'm getting some dupes like yours https://bugs.launchpad.net/ubuntu/+source/unity-place-applications/+bug/744261 today, which is an old (popular!) bug that Should Be Fixed (TM) by the package version listed in the bug report. I'm wondering if you got this while running the old unity-place-applications but having upgraded to the newest package? [15:21] Ubuntu bug 744261 in unity-place-applications (Ubuntu) "unity-applications-daemon crashed with SIGABRT in raise() (dup-of: 738884)" [Medium,New] [15:21] Ubuntu bug 738884 in unity-place-applications (Ubuntu) "unity-applications-daemon crash in unity_package_searcher_get_random_apps()" [High,Fix released] [15:21] i fixed the annoying facebook api returning a dict for likes data now [15:21] kamstrup, no, this is fresh all the way [15:21] * kenvandine hates sudden changes to results [15:22] kamstrup, I rebooted since the last upgradd [15:22] mterry: the case is that the retrace matches exactly the code from the 0.2.40 packages and *not* that of 0.2.42... [15:23] * mterry shrugs at kamstrup. 0.2.40 was released a while ago. I've updated and rebooted many times since then [15:24] I mean 0.2.42 [15:24] mterry: i know, this is really odd! [15:25] mterry, ok, sorry I didn't see kamstrup was asking the same thing [15:25] kamstrup, so yeah it seems it's still happening [15:26] seb128: the thing that throws me off is that the retrace doesn't match the code in the package [15:26] the retrace matches the 0.2.40 code [15:27] seb128, mterry: no wait... this one looks good actually... [15:27] * kamstrup is maybe still in weekend mode... [15:28] \o/ === bcurtiswx_ is now known as bcurtiswx [15:33] mterry, seb128: phew, found the root cause... an undocumented exception from Xapian. Open Source 1 : Bad Documentation 0 [15:33] hehe [15:33] kamstrup, btw bug #739469 the description speaks about the application place not about the dash [15:33] Launchpad bug 739469 in unity-place-applications (Ubuntu) "Dash search unavoidably returns offensive results" [Undecided,New] https://launchpad.net/bugs/739469 [15:38] cyphermox, Are you confident that bug 741385 is about the ubuntu-mono icon cache? [15:38] Launchpad bug 741385 in ubuntu-mono (Ubuntu) "icons in notification area are broken" [Undecided,Confirmed] https://launchpad.net/bugs/741385 [15:38] cyphermox, Just trying to ensure I shouldn't be looking at it. [15:39] tedg, very confident but I wanted to try reverting to the old libappindicator, to see if the cache updates in that case [15:39] tedg, I doubt it does, but I want to make sure [15:40] tedg, ^^ I'm testing this right now [15:40] cyphermox, Cool, thanks for looking into it. [15:43] is sladen working on fixing the icon cache issue? [15:44] seb128, I pinged him, sent an email, and there's a bug open ;) [15:44] tedg, seems to fail to create the cache still [15:45] cyphermox, It's okay, as long as I can blame design, I'm good ;-) [15:47] cyphermox: haven't looked into the icon cache issue. It doesn't appear to affect the actual installation [15:47] sladen, right, it's true it didn't break install (although I guess it should have it fail) [15:49] sladen, problem is, this makes classic gnome setups look pretty bad (and also affects GDM afaict, the accessibility icon shows as a broken icon) [15:49] cyphermox: do you understand the problem space? (I don't, because I haven't looked into it yet) [15:50] cyphermox: is it a hook that we're calling/not calling from the post-inst? Or is it a piece of existing infrastructure taking exception to the format of one of the particular files [15:50] sladen, more or less. the directory gets scanned to update a cache of icons, but part of writing this yields to values that don't pass a validation that's done right before overwriting the old cache [15:50] sladen, option two, I believe ;) [15:51] it has issues when update-icon-caches / gtk-update-icon-cache gets run [16:12] bug 744086 [16:12] hm, no bot here? [16:13] "indicator-weather assert failure: *** glibc detected *** /usr/bin/python: realloc(): invalid old size: 0x00000000020c0420 ***" [16:13] because of libdbusmenu-gtk [16:14] jaytaoko, I just assigned a few nux crashers to you [16:14] one of them looks like it might be a mesa bug we're triggering [16:14] can you take a look when you get some time? [16:15] lamalex: send me a link please [16:16] jaytaoko, https://bugs.launchpad.net/nux/+bug/740077 [16:16] Ubuntu bug 740077 in nux (Ubuntu) "compiz crashed with SIGSEGV in __memcpy_ssse3_back()" [Medium,Confirmed] [16:16] https://bugs.launchpad.net/nux/+bug/740757 [16:16] Ubuntu bug 740757 in nux (Ubuntu) "compiz crashed with SIGSEGV in __libc_free()" [Medium,Confirmed] [16:16] https://bugs.launchpad.net/nux/+bug/740784 [16:16] Ubuntu bug 740784 in nux (Ubuntu) "compiz crashed with SIGSEGV in nux::GraphicsDisplay::GrabPointer()" [Medium,Confirmed] [16:23] lamalex: thanks [16:31] Hehe, jono is in Unity http://img24.imageshack.us/img24/8177/captureswj.png [16:33] didrocks, is this valid? https://bugs.launchpad.net/unity/+bug/740919 [16:33] Ubuntu bug 740919 in unity-place-applications (Ubuntu) "search for both localized and non-localized names/descriptions" [Undecided,New] [16:33] lamalex: not sure it makes sense to look for the english name [16:37] didrocks, it does [16:38] it depends on the keyboard layout, you'd have to switch to native layout just to look for the app, if your layout is english [16:39] if I have an english layout selected, it's just easier to type "terminal" === 20QAADT57 is now known as mterry [16:44] didrocks, think i should add a design task? [16:44] or just say wontfix [16:44] lamalex: yeah, you should add a design task IMHO [16:45] are the overlay scrollbars in Natty now? [16:45] I don't see them [17:02] jono, only available from in PPA. [17:02] s/in/a/ [17:02] damn... [17:07] Davidc_3, there is already a ppa? [17:10] AndreaAzzarone, if you are talking about my Lens, no, it's not packaged (but the code is n the Unity Books Lens project onaunchpad) . I was answering Jono's question about scrollbars. [17:11] Davidc_3, yes i am talking about your Lens... [17:11] Davidc_3, i will try it soon! :) [17:12] AndreaAzzarone, thanks for your support :) [17:14] Davidc_3, no icon is shown it the lens... only text! [17:16] AndreaAzzarone, I know, it happens... Don't know if it's Unity's or Google's fault. [17:20] Davidc_3, ok... [17:24] AndreaAzzarone, Unity doesn't seem to like fetching many http icons at the same time. For now, reloading books-lens.py helps. :) [17:27] Davidc_3, whoa! [17:27] what lens is that? [17:28] jcastro, did you see http://img24.imageshack.us/img24/8177/captureswj.png ? [17:28] jono, Books Lens ( https://launchpad.net/unity-books-lens ). [17:28] Davidc_3, nice! which site does it query? [17:28] Davidc_3, want to work on a Severed Fifth lens? [17:28] ;-) [17:29] I really want to see an awesome Severed Fifth lens that demonstrates how lenses can be used for specific projects [17:29] jono, Google Books, but I'd like to add OpenLibrary and others in the future. [17:29] Davidc_3, nice! [17:30] jono, Severed Fifth... Well I'm not that much into metal ;) [17:30] Davidc_3, think of it more as a way of spreading Free Culture :-) [17:31] jono, good point, I'll look into it when I have something stable with the Books :) [17:32] thanks Davidc_3 :-) [17:32] jono, yw [17:48] where are the docs for making a lens? === alecu is now known as alecu-lunch [18:07] jono, https://wiki.ubuntu.com/Unity/Lenses , http://bazaar.launchpad.net/~unity-team/unity-place-sample/unity-place-python/files , http://bazaar.launchpad.net/~stefano-palazzo/askubuntu-lens/trunk/files [18:09] jono: Kamstrup and stefano-palazzo files are IMHO the best way to understand python lenses. As a proof : I'm an absolute python noob. [18:19] thanks Davidc_3 [18:42] lamalex: how are your tooltips over the launcer? Mine are sticking around way too long, driving me mad. [19:06] it's funny that the scrollbar overlay can work with gksu synaptic but app-menu not === daker is now known as daker_ [19:37] jcastro, no complaints [20:05] jcastro, http://ubuntuone.com/p/jtL/ [20:05] Service Temporarily Unavailable [20:05] sigh [20:06] u1 is undergoing maintenance [20:06] jcastro, it's a screenshot of the gwibber lense [20:06] :) === alecu-lunch is now known as alecu [21:11] guys, is there a way to hide top panel on natty? [21:14] For what reason? [21:14] have more free pixel on a netbook [21:17] We already got free pixels from the menus moving from windows to the panel, and from maximized window titles moving to it. [21:19] sure, and if i can set top panel to autohide number of free pixel grow up [21:19] every single pixel here is precious [21:21] there's not a way to do that using gconf-editor? [21:21] kenvandine, ping [21:22] nmarques, pong [21:24] kenvandine, I wanted to ask you something, are you aware if there are patches for indicator-sound for 2 fixes: 1. fix the double entries on the "choose playlist" and 2. fix the icons. I know updated software does has this fixes, but I can't really update it because of libnotify. And by the way awesome job on indicator-datetime (liboobs drop, yay)... [21:25] humm... you could either try to cherry pick those fixes... or back out the libnotify change [21:25] which is minimal... so would be the easiest [21:28] I'm going to take a closer look into it [21:28] ion, are you there? [21:28] probably backout the libnotify change [21:28] Yes, why? [21:29] just for an answer [21:29] if not, you may tell me that [21:30] I have no idea. [21:30] ok, thank you [21:33] kamstrup, i have a basic gwibber place :) [21:33] or lense [21:33] whatever the cool kids are calling it these days [21:34] kenvandine: haha, just awesome :-) [21:34] kenvandine: screenie? [21:34] one sec [21:46] http://ubuntuone.com/p/jtL/ [21:46] kamstrup, ^^ [21:47] kamstrup, i need the list renderer though :) [21:47] not very friendly just listing a bunch of names [21:48] tedg, ^^ [21:50] kenvandine, Cool! [21:50] kenvandine: indeed, if njpatel doesn't, do it tomorrow i'm gonna have a stab at it [21:52] and [21:52] http://ubuntuone.com/p/juQ/ [21:52] that is from the dash, so askubuntu and gwibber results :) [21:53] kamstrup, awesome [21:53] kenvandine, You can just hard code that search text for all users. [21:53] tedg, don't joke... i did that briefly while testing [21:53] "ted gould" [21:54] kamstrup, is there any way to have a text entry other than search? [21:54] * kenvandine is thinking it would be cool to have one for posting [21:55] kenvandine: there might... I think there's something about a key in the .place file to override the default... ask didrocks [21:55] kenvandine, Your training is coming along beautifully young jedi. [21:55] hehe [21:55] kamstrup, that would be cool! [21:56] kenvandine: it would be nice to be able to change the looking glass icon in the entry as well, but you can't do that currently [21:57] nice features for next cycle :) [21:57] kamstrup, as it is right now, the only thing i still need to make this useful is the list renderer [21:57] no pressure [21:58] kenvandine: indeed, let's see tomorrow, no promises that I can tackle the Nux beast :-) [21:59] kamstrup, in the list view... would it be possible to have more than one action available? [21:59] like selecting different part of the post would do different things? [21:59] kenvandine: not currently i'm afraid [22:00] like raise a gwibber posting dialog setup for reply or open gwibber with the message thread selected [22:00] bummer [22:00] ok [22:00] kamstrup, i'll settle for a basic list renderer then... that can display the name, avatar and plain text [22:00] :) [22:01] kenvandine: i think the deeper problem is that the Dash is mostly designed for the simple find-activate usage and not really any interactions more complex than that [22:01] yeah, understood [22:01] not for "working with data" [22:02] but I think this gwibber thing is really interesting because it's somewhat in the gray zone [22:02] kamstrup, should this be called a lense or place? [22:02] * kenvandine is confused by the current terminology [22:03] tedg: cia [22:03] o [22:04] kenvandine: I think it's places for N but lenses for O, so it depends on how forward looking you are :-) [22:04] very forward [22:04] i'll rename it to lense :) [22:04] Cimi, Howdy [22:04] tedg: so, here I am [22:04] kenvandine, You'll just get burned as it'll change three times ;-) [22:04] kamstrup, i want to get it in a ppa working with the natty versions of stuff before UDS [22:04] Cimi, There you are ;-) [22:05] lol [22:05] tedg, at least it isn't indicators [22:05] :) [22:05] tedg: where shall I start? [22:05] tedg, damn you're consistent! [22:05] kamstrup, Oh, oh, can they be "data indicators" ? [22:06] Cimi, I think you need to print out "whether it's floating" and the ref count in each case. The ref count can be one, but it's a floating ref. [22:06] tedg: each case of what? [22:06] Cimi, the ref and unref. [22:07] tedg: there's only one time when g_object_ref is called for the thumb [22:07] tedg: and everytime it's incrementing the ref counter [22:07] tedg: totally "place indicators" ;-) Just like the Laucnher indocator on the left and the panel indicator on the top :-) [22:07] Cimi, Is it floating before that ref? [22:08] kenvandine, sweet, I got 0.6.5.1 to work with libnotify 0.6.0 ;) [22:09] New Rule: When kenvandine is in the room nothing will be referred to as anything other than a type of indicator. [22:11] tedg: well, the thing is more complicated than it seems [22:11] tedg: also other pointers, like adjustment or the parent [22:12] are sometimes incremented and sometimes not [22:12] so frankly I'm not sure they are freed correctly [22:12] the thumb is just visible [22:12] anyway the g object is not floating [22:14] tedg: so your code of swap_thumb/adjustment/pager is not properly destroying references I guess [22:14] any hint? [22:15] code is here http://goo.gl/uepE6 [22:18] Cimi, It's not floating? You're inheriting from GtkWidget, no? [22:18] gtkwindow [22:20] Cimi, Yeah, so that always starts with a floating reference. Someone is ref'ing it before that function even sees it. [22:20] Cimi, It should be floating until someone does a ref_sink() [22:21] tedg: I am calling [22:21] os_scrollbar_swap_thumb (scrollbar, os_thumb_new()) [22:21] tedg: ^ [22:21] that means I'm kinda creating the pointer I guess [22:23] tedg: os_thumb_init is still not floating [22:23] Cimi, So could something in the thumb_init() function be ref'ing it? [22:23] tedg: at the begin [22:26] tedg: changing the type from GTK_TYPE_WINDOW to GTK_TYPE_WIDGET changed that into floating [22:26] tedg: but that's not a fix [22:29] Cimi, Is it a TOPLEVEL? Perhaps GTK is sinking toplevel windows? [22:29] tedg: popup [22:30] but init is before setting the type [22:32] Cimi, Huh, the gtk_window_new() function takes a type. Should you maybe be setting it in your g_object_new() ? [22:32] Cimi, Perhaps it defaults to TOPLEVEL? [22:36] tedg: even calling g_object_new with type GTK_WINDOW_POPOP fixes that [22:36] neither sorry [22:37] tedg: looking at the code [22:38] tedg: it's referencing it in gtk_window_init [22:38] Cimi, Huh, that's interesting. So all windows have their own reference then. [22:38] yes [22:38] Cimi, So that's where your extra is coming from. [22:39] Cimi, I'd leave your function the same, and unref after calling it. [22:39] tedg: unref after calling what? [22:39] So something like [22:39] bob = os_thumb_new() [22:39] os_scrollbar_swap_thumb (scrollbar, bob) [22:39] unref(bob) [22:40] tedg: unref in the swap thumb? [22:40] after sinking? [22:40] Cimi, Leave the ref and the unref in swap. That way you can see the count there. Makes it easier to debug. [22:40] tedg: line 611 [22:40] ? [22:40] in the nested {} [22:41] ? after settings callbacks? [22:41] Cimi, Don't change anything in os_scrollbar_swap_thumb() [22:41] tedg: to line 429? [22:41] *so [22:42] Cimi, Yes. Do you see why? [22:42] tedg: no :) [22:42] I didn't understand why there and not in the nested {} [22:43] Cimi, Because the _new() is making the ref in this case. And it means that _swap_thumb() is balanced in it's ref's and unref's. This make that function easier to debug. [22:43] ok [22:44] tedg: but if in notify_orientation_cb [22:44] priv->thumb is null [22:45] then I should not unref it, or it's safe to do it anyway? [22:45] Cimi, No you should protect against thumb being null. [22:46] tedg: fo, if priv->thumb != NULL ... unref ? [22:46] *so [22:46] Cimi, Yup [22:46] tedg: UDS session on reference counters and g object memory management pls :P [22:49] Cimi, UDS isn't about teaching it's about planning :-) [22:49] Not a conference. [22:49] tedg: I tried :D [22:50] cimi desktop summit session on ref counting gobject memory management would have made sense [22:50]