=== chaotic_ is now known as chaotic [03:10] tedg: Hey [03:10] Nafai: Good Evening [03:10] Should app indicators have any problems with a menu constructed by GtkBuilder that has sub-menus? [03:11] My initial conversion of gnome-bluetooth has produced interesting results: http://www.travishartwell.net/gnome-bluetooth.png [03:11] If you aren't familiar, the original version has sub-menus off of each of the devices (such as TMobile G1) [03:12] tedg ^^ [03:12] :) [03:14] Nafai: In theory there shouldn't be any issue. But I'm not sure if bratsche's GTK parsing code handles submenus. [03:15] Nafai: It shouldn't make a difference whether it's GtkBuilder or normal GtkMenu. [03:15] I figured that shouldn't be the issue, but I wanted to give the full details [03:17] I need to look at how the menus are built, perhaps there is something else I need to call, but I thought I'd run it by you as well [03:19] Nafai: Does it work other than the submenu? [03:20] I'm not sure if I added support for submenus. If not I can do that tomorrow probably. [03:20] bratsche: yes, the items that have actions associated with them do work [03:20] Oh, one thing I noticed, those items that are insensitive are not styled any differently than sensitive items [03:21] For example, in that screenshot, "Bluetooth: On" and "Devices" should be insensitive [03:22] Nafai: Hmm, both of those should work. [03:22] Nafai: You should file bugs on indicator-application for things that don't work so we don't forget them. [03:23] Now that I think of it, I had the same issue with brasero [03:23] Sure thing [03:23] I'll file bugs on the potential sub-menu parsing issue and the insensitive thing [03:24] (either tonight or first thing in the morning) [03:24] glad I'm helping to find bugs :) [03:24] Yup, that's the fun of running alpha code :) [03:24] Nafai: Can you assign the submenu parsing bug to me once you file it? [03:26] Sure, what's your username on launchpad? [03:27] You're Cody Russell, right? [03:27] :) [03:28] (I'm just as guilty, not using an easy nick -> name like tedg or jono or rickspencer3) :) [03:30] Nafai: Yes, but he fixed it on LP too ;) http://launchpad.net/~bratsche [03:30] :) [03:31] * Nafai files bugs [03:31] Should I assign the insensitive one to anyone? [03:32] You can assign that one to me too probably. If it turns out to be more of a tedg thing then I'll reassign it. [03:32] ok [03:33] This one has it as the #3 item: https://bugs.launchpad.net/indicator-application/+bug/513924 [03:33] Launchpad bug 513924 in indicator-application "Tooltips, formatting, and unsensitive items (affects: 1)" [Undecided,New] [03:33] Should I file a separate one? [03:34] The submenu one as well. [03:35] Right [03:35] But since the insensitive one is already reported, should I just comment on that one saying "Me too"? [03:36] tedg: https://bugs.launchpad.net/indicator-application/+bug/513924 [03:36] Launchpad bug 513924 in indicator-application "Tooltips, formatting, and unsensitive items (affects: 1)" [Low,In progress] [03:36] tedg: Have any comments on #1 and #2? [03:38] bratsche: I'll comment on the bug :) [03:38] * bratsche grabs a bag of popcorn [03:51] Ok, bug submitted #519625 [03:51] Plz school me on better bug reporting, kind of out of the habit. [03:52] is there a particular debug process if notify osd does not show up notifications? [04:01] nigel_nb: Check to see if notify-send works. [04:01] tedg, it does not work [04:02] nigel_nb: Then probably look at the log in ~/.cache and see if there is anything getting put in there. [04:02] tedg, its bug 518524 [04:02] Launchpad bug 518524 in notify-osd (Ubuntu) "Pidgins notifications doesn't show anymore (affects: 1)" [Low,Incomplete] https://launchpad.net/bugs/518524 [04:02] what should I be asking? [04:03] is there an apport hook for notifications? [04:03] nigel_nb: Hmm, I'm not sure what else to look for. That is odd. I don't believe there is an apport hook. [04:04] tedg, I'll ask for the log then [04:05] tedg, is the current version of notify-send 0.9.24? [04:05] because in karmic its only 0.4.5 [04:07] nigel_nb: No, they're versioned differently. notify-send comes from the libnotify package not the notify-osd one. [04:10] ah [04:11] tedg, so, is there any info that I can get that would be helpful (besides log) [04:11] nigel_nb: Perhaps video card/driver. I mean, the only thing I could think of at that point would be X bugs. [04:12] tedg, will do, thanks :) [04:37] tedg, did you see Nafai's mail? [04:37] jono: No [04:37] tedg, he found a couple of bugs [04:38] he wasn't on the list because I talked to him here :) [04:38] he knows about the bugs [04:38] jono: Yes, we discussed them in the backlog :) [04:38] ahhh sweet [04:38] cool :) [04:39] jono: Were you eating dinner again? Haven't I told you need to give up eating to keep up!?!?! :) [04:39] tedg, LOL [04:39] my mistake tedg :) [04:39] hey mccann :) [04:39] yo jono [04:39] tedg, I have asked for the log and info about video card + xorg log :) [04:40] mccann, hows tricks? [04:40] nigel_nb: Cool, hopefully something comes of that. [04:40] tedg, :) [04:40] jono: not bad. trying to smoke a cigar and not doing very well [04:41] mccann, lol [04:41] wrong end? :) [04:41] uneven burn and i didn't have a cutter so i had to bite it and came out bad [04:41] aha, gotcha :) [04:44] tedg, for issues about notify-osd not working/intermittent, adding the log files would be standard work flow then? [04:44] nigel_nb: I think so. We should probably figure out an apport hook for that.... [04:44] * tedg is scared of Pythons [04:44] heading for bed y'all, later [04:45] 'night Nafai [04:45] tedg: I'm scared of the seas than the pythons [04:45] er, the Cs :) [04:45] tedg, I'm not sure if I can, but I can kill some time trying to tame it ;) [04:46] i'll let you know if i come up with something :) [04:46] nigel_nb: Cool, that'd be great! [05:22] mccann: Don't accidentally swallow the tip! [05:44] bratsche: ick [05:45] Yum? [06:10] seg|ars: hi.. when switching themes in gwibber , i think you should force the font colors as well.. ex: with Ubuntu gwibber theme , if you are using dark system theme ,[where the system fonts are white] gwibber becomes unreadable [06:11] that's on my todo list. I'm going to fix the font colors in the ubuntu theme [06:12] neat.. [06:12] most of the themes for gwibber use system colors throughout, the ubuntu one is (I think) the only one that has statically colored backgrounds [06:12] it's a pretty trivial fix. I just have to look at what colors are used in the adium version and apply those to the css in the gwibber one [06:17] hmm , even the gwilouche,flat becomes kinda unreadable[seem to use system font color] , but Ubuntu theme is the worse [06:19] ugh you are right [06:20] I should do more testing with a dark gtk theme [06:25] seg|ars: should i reopen Bug #473515 ? or... [06:25] Launchpad bug 473515 in ubuntu (and 1 other project) "Dark Theme results in white text on white background for updates (affects: 6)" [Undecided,Invalid] https://launchpad.net/bugs/473515 [06:26] no, it's fine. I've added it to my todo list [06:26] ok.. thanks [06:26] thanks for bringing it to my attention [07:12] jono, are you around? could use some help with notify :) [07:12] I'm writing (trying to write) an apport hook for notify-osd... any suggestions as to what all info should be attached besides the notify-log [07:14] nigel_nb, I am about to head to bed [07:14] you mean info in the bubble? [07:15] jono, yep [07:15] apport hook is a cool idea [07:15] maybe summarize the problem [07:15] yeah. I was triaging bugs earlier and thought it might help. [07:15] I'm still learning though. right now, its set to only include the notify-osd.log from cache [07:17] anything more you'd like me to add? [07:18] nigel_nb, maybe: [07:18] App Name crashed: summarize problem [07:18] sweet, pal :-) [07:18] jono, I'll play around and let you know what happens :) [07:18] sweet! [07:18] ok, I am heading to shower [07:18] later! [07:19] good nite :) === chaotic_ is now known as chaotic [10:22] Hi there. [10:22] I just finished to develop an appindicate integration for Getting Things GNOME [10:23] I get a weird Segfault when the status is set the second time [10:23] any chance to understand why? === MacSlow is now known as MacSlow|lunch === MacSlow|lunch is now known as MacSlow [13:57] Good morning [13:57] invernizzi: Still around? [13:57] yes [13:58] Good morning [13:58] Figure out the segfaulting issue? [13:59] not really [13:59] hrm [13:59] I found an open bug for that [13:59] so it happens when you call app_indicator_set_status? [13:59] yes [14:00] bzr branch lp:gtg, activate and deactivate the "notification area" plugin [14:00] * Nafai looks [14:00] I've got some code where I've done it multiple times without issue, I wonder what's different... [14:01] HI jpetersen, how's your porting stuff going? [14:01] well, maybe I'm doing something wrong. That's why I asked [14:03] tedg is a great person to help :) [14:03] python, hrm. I wonder if it is a bindings issue [14:03] invernizzi, Just run gtg in the root of the checkout? [14:04] or ./scripts/debug.sh [14:04] either one [14:04] weird, it can't find configobj [14:04] I'll install the deps [14:05] sure [14:05] I've been doing too much C the last couple weeks :) [14:06] :-) [14:06] okay, I can reproduce, let me put it in a debugger to find out [14:07] Nafai, hi i am just working on gnome-settings-daemon [14:08] this has been fun so far, I'm starting my 4th conversion this morning [14:12] just a sec, installing more debug symbols so I can get a full stacktrace [14:12] take your time [14:19] hrm, I'm not familiar with the insides of this, but I have got a full C stack trace [14:19] Did you say there was a bug this might apply to? [14:19] sure [14:19] i'll find it [14:19] thanks [14:20] perhaps I'll post this stack trace to it [14:20] tedg: Who do we bug about the Python bindings? [14:20] Nafai: Probably the best person is kenvandine [14:20] ok [14:20] kenvandine: Hey! :) [14:20] :) [14:21] https://bugs.edge.launchpad.net/indicator-application/+bug/510169 [14:21] Launchpad bug 510169 in indicator-application "Segfault when setting status second time (affects: 2)" [Undecided,New] [14:21] Nafai: He doesn't hate Python, yet. I'm trying to teach him. [14:21] why would one hate Python? [14:21] If the alternative is C? :) [14:21] is it possible to have multiple app indicators in one application? [14:22] that would be required for gnome-settings-daemon (screen, keyboard, a11y) [14:22] jpetersen: Should be, but I haven't tested it. [14:22] jpetersen: You might find bugs. [14:22] invernizzi: Well, this stack trace is different [14:22] jpetersen: The key would be to ensure they all have different paths. [14:23] wow even included a test case! [14:23] Nafai, maybe because it's a differente application? [14:23] http://gist.github.com/300339 [14:23] invernizzi: I would guess, if it were the same bug it would still end up at the same place at the top [14:24] right. But the description of the bug is really the same, so maybe they're related [14:24] Right. [14:25] kenvandine: should invernizzi attach this stack trace and repro steps as well to that bug? [14:25] tedg, yes it seems there is a problem that in app_indicator_init the second app indicator is registered with dbus_g_connection_register_g_object at the same path (/org/ayatana/NotificationItem) [14:26] tedg, that seems to fail [14:26] jpetersen: just checking, you are giving the new ones unique names? [14:26] jpetersen: Hmm, okay. So there is probably needed a new "_new" function that can take a custom path. [14:27] Nafai, yes but the path seems to be always the same [14:27] ok, just double checking [14:27] in the example included in that bug i don't see the problem [14:27] and can't repro it in ipython [14:27] different bug [14:29] ok [14:29] * invernizzi waves and goes to sleep, will report the bug tomorrow [14:31] thx invernizzi [14:33] tedg, maybe one could include the id in the path? [14:34] jpetersen: Hmm... that works for me. I like it. [14:34] something like /org/ayatana/NotificationItem/@id and /org/ayatana/NotificationItem/@id/Menu [14:34] jpetersen: Good idea, thanks! [14:35] Looks like vino is another one like brasero, their statusicon inherits from GtkStatusIcon. You know this would be easier to share code if they would have used composition instead of inheritance :) [14:52] Nafai: I have some tips from mpt for vino [14:53] Recommendation: Port menu directly. Change Preferences text to ?Show Remote Desktop menu:? ?Always?/?Whenever someone is connected?/?Never?. [14:53] Nafai: make sense? [14:53] Just a sec, let me look at the preferences again (just had it open) [14:54] Ah, instead of the current "Notification Area", etc [14:54] Yeah, makes sense [14:54] * jcastro nods [14:54] So behavior is identical, just tweak the preferences dialog :) [14:54] Nafai: what's the status of ken's reviews? [14:54] * jcastro would like to get some patches sent upstream today! [14:54] kenvandine: status? :) [14:55] soon... sorry! [14:56] np, just let me know if I can do anything [14:56] before lunch :) [14:56] what's your timezone btw? [14:56] EST [14:57] ok. I'm MST [15:02] hopefully if bratsche figures out the bug I reported, I'll have another two for kenvandine to look at tomorrow (vino and gnome-bluetooth) :) [15:05] kenvandine, let me know if you need some help reviewing [15:06] nah, i already started, will finish it up real soon [15:06] seb128, thx though :) [15:07] np [15:07] \o/ [15:13] can't wait until my first code is installed on someone's machine [15:13] from the repo [15:34] Nafai, building brasero :) [15:35] yay [15:39] jcastro: How does this look? http://www.travishartwell.net/vino-prefs.png [15:40] mpt: ^^^ [15:44] morning jono [15:45] Nafai, good [15:46] cool, thanks [15:46] (I know I said "menu:" rather than "Menu", but that would make it inconsistent with all the other headings, and it would be a bit of a diversion to change them all) [15:47] oh, heh, I didn't even notice the capitalization. [15:47] hey Nafai :) [15:48] looking forward to another jono webcast today :) [15:48] just think, during the last one I asked you how I could get a job with Canonical [15:51] Nafai, should be fun :) [15:59] Nafai, is there a way to test brasero without actually burning a disc? [15:59] kenvandine: heh, yeah. I just had it create ISOs [15:59] just a second [16:00] mind you, I don't have an optical drive in this laptop at all so I'm not sure if it presented me the options differently [16:00] ah [16:00] didn't know it would do that :) [16:00] thx [16:01] after you choose "create new data project", there is a drop down at the bottom, mine defaults to an image file, perhaps you need to change that [16:09] Nafai, ok the patch looks good [16:09] and it works... but i get a segfault in brasero [16:09] oh [16:09] if at the end of the burn you click around in the indicator [16:09] show and hide [16:09] then click cancel [16:10] let me take a look :) [16:10] thx [16:12] Hi, I want to add ApplicationIndicator support to my application. Can I run the indicator-applet under XFCE (it suppossed to support gnome applets)? [16:22] seb128, actually could you review the gnome-control-center patch? [16:22] i think you wanted that one :) [16:22] kenvandine, sure [16:22] bug 497857 [16:22] Launchpad bug 497857 in gnome-control-center (Ubuntu) (and 1 other project) "Support application indicators (affects: 1)" [Wishlist,Fix committed] https://launchpad.net/bugs/497857 [16:22] kenvandine, well, I worked on g-c-c quite a bit over years and I'm fine doing reviews [16:22] so I'm happy to review this one yes [16:23] ;-) [16:23] seb128, thx [16:23] for some reason I can't get it to show up as an indicator right now, it's falling back to a statusicon [16:25] is indicator-application-service running? [16:25] is gnome-user-share present in default install in lucid? it is reqd for bluetooth receive to work [16:25] seb128, Yeah, it is [16:26] i should ask in #ubuntu-desktop i guess [16:26] Nafai, try removing the applet and added it again maybe [16:26] weird, did a complete re-compile and it works now [16:26] must have been linking weird or something [16:26] tgpraveen12, no it's not [16:26] Nafai, ok [16:27] kenvandine: Okay repro'd, now investigating [16:27] cool [16:38] * Nafai wonders if this is a case of finalize already being called [16:38] got look up the semantics of that [16:43] kenvandine: So, a little C debugging advice, I can read the back trace: http://gist.github.com/300520 [16:43] but when I try to print the values in the memory locations in gdb (print *self, for example, to see what values are in that parameter), I'm getting "cannot access memory address at 0xf00" [16:44] I'm doing this by loading the program and a core dump, btw [16:48] stepping away for a short break, bbiab [17:00] GogglesGuy: I've never tried, but I don't see any reason it wouldn't work if they support gnome applets. [17:01] GogglesGuy: There was also a port of the messaging menu at some point to a native XFCE applet, it seems like that could be updated to work pretty easily. [17:01] kenvandine: Have you heard much about that recently? ^ [17:02] tedg: thanks. I'm trying to build the various packages now [17:02] Mark__T did it [17:50] back [17:54] ok, back to tracking down that segfault [17:56] Nafai: which app are you on? [17:56] * jcastro scrambles to get the bug #'s straight [17:57] jcastro, brasero, ken found a crasher :) [17:57] ah [17:57] but, otherwise, I'm working on vino [17:59] would it be safe to say that everything in the project is now blocking on ken? [17:59] :p [18:00] well, technically, gnome-bluetooth is blocked on bratsche :) [18:04] ah this is the nested menu bits right? [18:04] yes [18:07] hehe :) [18:07] Nafai, any progress on brasero? [18:07] what else is waiting for me to review? [18:08] seb128 is reviewing g-c-c [18:08] kenvandine: no, not yet, perhaps some help in debugging [18:08] I've got the backtrace, and it's in gtk code. but it can't find the code on my system because it is looking for it in /build/buildd [18:08] * kenvandine isn't great at that [18:08] kenvandine: g-p-m needs a look [18:08] ok [18:09] I did an apt-get src libgtk2.0-0 and added the appropriate source line in gdb, but that didn't seem to help [18:09] Nafai, which package? [18:09] you should install the -dbg packages for the libs [18:09] ah [18:09] yeah, have those too [18:10] you do? [18:10] deb http://ddebs.ubuntu.com lucid main restricted universe multiverse [18:10] in your sources? [18:10] libgtk2.0-0-dbg [18:10] you have that installed? [18:10] yeah [18:11] ok, figured it out [18:11] by the time this function is called, the ui manager is already NULL [18:12] wee. [18:13] probably because g_free(cobj->priv) has already been called in the finalizer [18:13] I would think the indicator should go away by this time [18:13] * Nafai reads the original code [18:21] * Nafai tries something [18:30] ok, interesting [18:31] in the tray icon version, you left click on it when the burn is complete, the icon and dialog disappears and it goes back to the main window [18:35] is activate the signal that gets called when an item is clicked? [18:37] kenvandine (or any other gtk hackers): around? [18:39] Here's what is happening. BraseroTrayIcon inherits from GtkStatusIcon. So when finalize is called for GtkStatusIcon from BraseroTrayIcon's finalize, the widget is destroyed, etc. [18:40] I'm not doing anything other than calling g_free on the private data structure of BraseroAppIndicator (which holds as a data member an AppIndicator) [18:40] Is there something I need to call on AppIndicator to tell it to "go away"? [18:40] tedg^^ perhaps? [18:41] or clicked [18:42] humm [18:42] because finalize has been called by this time, that's why we are getting segfaults because of invalid pointers [18:43] yeah [18:43] the examples on the wiki don't show anything about cleaning up [18:43] kind of a problem :) [18:44] Nafai: You need to do a g_object_unref [18:45] Of the AppIndicator pointer? [18:47] * Nafai tries [18:47] Nafai: Yes. [18:47] ok [18:49] i think you did [18:50] if (priv->indicator) { [18:50] should be [18:50] if (priv->indicator != NULL) { [18:50] i think [18:50] right tedg? [18:50] although... i wonder why finalize is getting called a second time [18:50] i guess missing a unref then [18:51] kenvandine: I'm not looking at the code, but I would say that you shouldn't unref a NULL pointer :) [18:51] yeah... but if (priv->indicator) [18:51] is not the same as [18:51] if (priv->indicator != NULL) [18:51] They're actually the same. [18:51] they are? [18:51] In some compilers. [18:51] hehe [18:51] ok [18:51] yeah, the unref didn't change things [18:51] TRUE = !0 [18:51] NULL == 0 [18:55] Sorry if I'm slightly unresponsive for a few minutes; my wife forgot a short paper for her college class at home and I have to re-type it for her because she didn't save it only printed it. :) [19:03] ok [19:03] back to this [19:05] Trying to figure out what's appropriate here [19:11] I've got a problem with actions used in a menu in AppIndicator. The custom labels aren't used, but the stock labels instead. [19:13] I've got no idea how to solve this, could anyone help? [19:13] The changeset of the support can be found at . [19:14] * Nafai looks [19:15] where is the menu constructed in that code? [19:18] Nafai: in ui.h, give me a sec, I'm looking up a link to the files [19:18] thanks [19:19] Nafai: Menu is 'icon-popup' in , the actions are defined in . [19:19] the menu is at the bottom of the ui.h file [19:21] * Nafai nods [19:21] I'm looking at the GtkActionEntry array, and it looks similar to the ones I've used before [19:22] I'm not sure what could be going on [19:23] tedg, bratsche? Perhaps one of you could look at qense's problems? [19:23] qense: btw, are you qense on digg.com? (I assume since it is an unusual nick) [19:23] Nafai: yes, that's what was bugging me, the [19:23] code looks perfectly normal [19:23] Nafai: tedg already had a look on it yesterday but couldn't find it [19:23] Nafai: I am qense on Digg, yes [19:24] Whenever I look at upcoming, digg tells me that you are the digg'r most like me [19:25] I'm nafai77 [19:25] Nafai: that name does sound familiar [19:25] qense: "The custom labels aren't used, but the stock labels instead." <-- that will always happen if you use stock. [19:25] Because stock items are not just a convenience for loading icons, they really are a matched pair. [19:26] bratsche: but according to the GTK+ documentation of GtkActionEntries/Array(What's it called again?) this shouldn't happen when the label is not NULL. [19:26] bratsche: http://library.gnome.org/devel/gtk/unstable/GtkActionGroup.html#GtkActionEntry [19:27] I'd have to trace through some code to see what the problem is. There's a possibility that this documentation is wrong. :/ [19:28] I also don't completely understand how the icon naming spec stuff works though. [19:29] bratsche: these action entries do work the way the documentation says it should when launching it next to a GtkStatusIcon, but not when sending it to AppInd via libdbusmenu. [19:29] However, I couldn't find anything in the code of AppInd or libdbusmenu that changed the menu or the actions. [19:29] Ah okay, then maybe it's a bug of mine. [19:30] Feel free to file bugs in Launchpad and assign to me. [19:30] Is there a document that describe the differences between the KStatusNotifier API and the added extensions? [19:30]