OK
Pages: 1 2 3 4 5
Rating:
9.72
Thread Listing » AWN Applets
[Announcement] Standalone launchers with a separate taskmanager
#18947 by moonbeam (1.0000) posted on 11:13am Saturday, January 12th, 2008
( no avatar )
[EDIT]
[Q]



Note that the description in this has been superseded in some areas by later posts. It's a good idea to browse the whole thread.
[Q]



PLEASE NOTE dev work has moved back to trunk. Please replace references to testing etc with awn-extras (trunk).
[Q]



Thanks
[Q]



[/EDIT]
[Q]




The Last few days I've been working on a implementation of standalone launchers with a separate taskmanager.
[Q]



The work is incomplete but it is getting to the point of being usable. For those who are interested in trying it out, at the moment it requires desktop-agnostic, vala and my awn-extras branch (EDIT) OR the recommended option is awn-extras-testing and awn-core-testing. And probably some other things I'm neglecting to mention. My suggestion would be to drop by the irc channel if you want to get it up and running.
[Q]



A few comments on the current code state:
[Q]



-it is in development. Not remotely stable. Too many things to mention basically. The taskmand is fairly stable the launchers are not.
[Q]



Implementation details.
-taskmand is a dbus daemon. All it does is sits there and talks to the launcher applets and starts a new one if necessary. It does a few other things but that is about it. The taskmanager daemon is meant to be very simple. As of this writing the source file is 389 lines of C code. I anticipate with configuration support and the possibility of adding 1 or 2 additional dbus interfaces that it might climb ot 500 lines. It is designed so that it is possible to swap in different launchers.. as desired.
[Q]



-launchers... This is were the intelligence is meant to be. In essence a launcher is meant to be a minimal taskmanager. In essence it's up to the dev of the launcher to make it do as little or as much as they like. The launcher I am writing are meant to be relatively full features with the ability to manage more than one associated task in a single icon (all the email client windows under one icon, all the terminal windows etc) or to run in one icon per window mode. The launcher code is still in a state of flux... but it is coming along nice. The launchers will also have a mode where the taskmand for those who want to use a non-awn taskmanager.
[Q]



-Once the configuration backend is done for taskmand it will be possible to choose what launcher applet is used. This will allow the creation of different launchers (say experimental stacks style) as long as they follow the specs on how the taskmand functions. There is a document in the top level of my branch (in the Documents dir) that is _relatively_ up to date. The most up to date documentation of course is the source code :-)
[Q]



-Advantages:
[Q]



-This is(will be) a dropin replacement for the existing awn taskmanager/launcher. Just disable the awn taskmanager and startup this one.
[Q]



-Individually customizable behaviour (if the "launcher" applet allows.
[Q]



-Ability to solve the issue the of icons flashing in and out of the bar... and moving around. Most of that issue has just went away on it's own by the nature of the design.
[Q]



-stability of design.
-The core management is done by taskmand (approx 400 lines of C code) running as it's own separate process. If it does crash it should restart when needed through the dbus services process.
-Each launcher is its own applet and therefore it's own process.
-If a launcher or taskmand acts up it should not effect awn itself (see separate processes).
-flexibility (see the ability to use different launchers). Also note that launchers as I'm using the term here can mean something a bit more than is normally indicated by the term (see the document).

Some Issues:
-note the requirements mentioned above... Vala is needed for the launcher applet. There is nothing stopping development of launchers in python or C that can interact with taskmand though.
-awn-manager. awn-manager has this unfortunate tendency to think that it is the sole ruler of the awn-configuration. Which causes some bad interactions... In the pursuit of KISS taskmand and the launcher start and stop applets by modifications of applets_list key... they let awn do the real work. There is an addition to Desktop-Agnostics Configuration API that allows the use of advisory locking on configuration keys. Ultimately I hope this will filter down and any processes accessing/changing keys that might be touched/changed by other processes will use them when necessary.
[Q]



Anyway, I'm hoping that this will get a look over and receive some input. I'm particularly interested in any thoughts on taskmand behaviour and the dbus interface - as the launchers themselves are meant to be replaceable. I am mulling the possible need for a couple more interfaces... simplicity is desired.
[Q]



My apologies for any grammar spelling errors in this... I've got to do a couple things now. But I will come back and proofread this later.
[Q]

--
http://moon-shiny.blogspot.com/

*Edited at 6:36pm, 04/29/08
Score: 0.86 points
Any particular reason?
#18948 by asomething (0.9744) posted on 11:24am Saturday, January 12th, 2008
Sounds great... Even more of a reason to get desktop-agnostic merged with awn-core-testing.
[Q]

Score: 0 points
Any particular reason?
#18950 by malept (1.0000) posted on 11:46am Saturday, January 12th, 2008
( no avatar )
> Sounds great... Even more of a reason to get desktop-agnostic merged with awn-core-testing.
[Q]



IIRC, we need to get njpatel to sign off on it first.
[Q]
Score: 0 points
Any particular reason?
#18955 by moonbeam (1.0000) posted on 12:10pm Saturday, January 12th, 2008
( no avatar )
malept said:
IIRC, we need to get njpatel to sign off on it first.
[Q]



I really, really, really want the config stuff to be merged. It's a PITA doing applets otherwise - the other option is gconf only (not an option in my opinion) or liberal use of #ifdef #else (which obviously means that this should be abstracted away by a lower level library :-) ).
[Q]


Score: 0 points
Any particular reason?
#18959 by gilir (0.9999) posted on 3:39pm Saturday, January 12th, 2008
I'm glad that is now available for testing :)
[Q]



moonbeam said:
-launchers... This is were the intelligence is meant to be. In essence a launcher is meant to be a minimal taskmanager. In essence it's up to the dev of the launcher to make it do as little or as much as they like. The launcher I am writing are meant to be relatively full features with the ability to manage more than one associated task in a single icon (all the email client windows under one icon, all the terminal windows etc) or to run in one icon per window mode. The launcher code is still in a state of flux... but it is coming along nice. The launchers will also have a mode where the taskmand for those who want to use a non-awn taskmanager.
[Q]



Just to confirm that I'm not dreaming : you should be able in the future to group multiple windows of a type in one icon ? (like 3 windows of nautilus under only one icon of nautilus) ? It's sound that one of my most wanted features in AWN :)
[Q]



asomething said:
Sounds great... Even more of a reason to get desktop-agnostic merged with awn-core-testing.
[Q]



Yes, this is a good reason (1 more :))
[Q]



malept said:
IIRC, we need to get njpatel to sign off on it first.
[Q]



We can give it a try in testing and waiting feedback of Neil after. All changes I included in awn-core-testing are now in trunk. So we can begin some real testing :)
[Q]




moonbeam said:
I really, really, really want the config stuff to be merged. It's a PITA doing applets otherwise - the other option is gconf only (not an option in my opinion) or liberal use of #ifdef #else (which obviously means that this should be abstracted away by a lower level library :-) ).
[Q]



A good example (I think) it's compiz. You can change the backend of the configuration (gconf, flat file and kconf). I don't know if it's possible in the DA branch.
[Q]

Score: 0 points
Any particular reason?
#18963 by moonbeam (1.0000) posted on 4:37pm Saturday, January 12th, 2008
( no avatar )
gilir said:
Just to confirm that I'm not dreaming : you should be able in the future to group multiple windows of a type in one icon ? (like 3 windows of nautilus under only one icon of nautilus) ? It's sound that one of my most wanted features in AWN :)
[Q]



Not the future. Implemented today. It doesn't look pretty yet... but it's there. Automatic grouping anyway, I am planning on allowing drag and drop in case it doesn't figure out something should be grouped.
[Q]





--
http://moon-shiny.blogspot.com/

*Edited at 4:38pm, 01/12/08
Score: 0 points
Any particular reason?
#18965 by moonbeam (1.0000) posted on 5:13pm Saturday, January 12th, 2008
( no avatar )
gilir said:
A good example (I think) it's compiz. You can change the backend of the configuration (gconf, flat file and kconf). I don't know if it's possible in the DA branch.
[Q]



It's a build time configuration option... I think in compiz you can change at runtime. Basically you build with whatever backend (basically gconf or flat file) and the rest is more or less transparent... applets have no idea what the backend is if they use the awnconfigclient.
[Q]




Score: 0 points
Any particular reason?
#18966 by gilir (0.9999) posted on 5:22pm Saturday, January 12th, 2008
moonbeam said:
Not the future. Implemented today. It doesn't look pretty yet... but it's there. Automatic grouping anyway, I am planning on allowing drag and drop in case it doesn't figure out something should be grouped.
[Q]



Hummm, I must test this :)
[Q]



moonbeam said:
It's a build time configuration option... I think in compiz you can change at runtime. Basically you build with whatever backend (basically gconf or flat file) and the rest is more or less transparent... applets have no idea what the backend is if they use the awnconfigclient.
[Q]



Yes, it's runtime for compiz. The problem for build time is that you can only have one configuration flag when you install it :/. So, I need to investigate more to make package with multiple flag :)
[Q]

Score: 0 points
Any particular reason?
#18967 by malept (1.0000) posted on 6:16pm Saturday, January 12th, 2008
( no avatar )
gilir said:
Yes, it's runtime for compiz. The problem for build time is that you can only have one configuration flag when you install it :/. So, I need to investigate more to make package with multiple flag :)
[Q]



The reason that the DA config option has to be at build-time is because Awn proper is written in C :)
[Q]



gilir, on my TODO list is to split out config and other desktop-specific APIs into their own libraries - I'm thinking libawn-config and libawn-desktop. Would that help you with your packaging of the new DA features?
[Q]

--
http://blogger.malept.com/

*Edited at 6:17pm, 01/12/08
Score: 0 points
Any particular reason?
#18968 by gilir (0.9999) posted on 6:53pm Saturday, January 12th, 2008
malept said:
gilir, on my TODO list is to split out config and other desktop-specific APIs into their own libraries - I'm thinking libawn-config and libawn-desktop. Would that help you with your packaging of the new DA features?
[Q]



I don't think the location will have an impact on packaging. My problem is more the build system and the capacity of AWN to have multiple support.
[Q]



Technically I can have only one group of configuration flag to build package, like :
[Q]



--with-desktop=gnome --with-gconf=yes
or
--with-desktop=agnostic --with-gconf=no
[Q]



Each line need a proper package, so to have this 2 configurations, I need to build 2 differents package. On a distribution point of view, you don't make 2 packages, you choice the flags.
[Q]



To be perfect, in a perfect world ( :) ), a system like this will be fine :
--enable-gnome-support --enable-agnostic-support --enable-gconf --enable-ini
[Q]



And after, you decide of the support and the config in the configuration, because of course you can't have 2 config system running at the same time. But you need to adapt the build system and the way AWN use configuration and support.
[Q]



Score: 0 points
Any particular reason?
#18970 by malept (1.0000) posted on 7:33pm Saturday, January 12th, 2008
( no avatar )
Hmmm...I've never coded g_module support before :) I need to think about how that's going to work.
[Q]
Score: 0 points
Any particular reason?
#18974 by fde (0.1733) posted on 2:59am Sunday, January 13th, 2008
Hi! Everything sounds pretty nice! However, do you think it should be possible to choose a treshold in the number of windows of the same type under which the tasks are NOT grouped in a single icon?
[Q]



Also, do you think it is possible to represent the tasks by a thumbnail of the corresponding window, updated at real-time (similarly as what is widely used in Compiz-Fusion)? Then, if you pass the treshold, they could appear as the icon of the correponding application.
[Q]



*Edited at 12:06pm, 01/13/08
Score: 0 points
Any particular reason?
#18975 by fde (0.1733) posted on 5:10am Sunday, January 13th, 2008
In fact, wouldn't that be even nicer if multiple windows of a type were grouped in a STACK?
[Q]



*Edited at 8:55am, 01/13/08
Score: 0 points
Any particular reason?
#18983 by malept (1.0000) posted on 11:36am Sunday, January 13th, 2008
( no avatar )
I hate stacks. I also hate that people want to turn Awn into an exact clone of the OSX bar. However, I'm just a (future) user. Moonbeam may implement it as an option.
[Q]
Score: 0 points
Any particular reason?
#18984 by moonbeam (1.0000) posted on 11:42am Sunday, January 13th, 2008
( no avatar )
malept said:
I hate stacks. I also hate that people want to turn Awn into an exact clone of the OSX bar. However, I'm just a (future) user. Moonbeam may implement it as an option.
[Q]



maybe, maybe not.
[Q]



I"m actually hoping that a few "launcher" implementations will appear and that someone else will do a nice, complete swiss army knife version :-)
[Q]




Score: 0 points
Any particular reason?
#18985 by asomething (0.9744) posted on 11:47am Sunday, January 13th, 2008
FYI:
[Q]



revno: 166
committer: Julien Lavergne <seagle@justice>
branch nick: awn-core-testing
timestamp: Sun 2008-01-13 20:20:41 +0100
message:
* Merge with the desktop-agnostic branch (Let's malept update the whiteboard ;))
* Debian dir update
- configure flag : --with-desktop=gnome --with-gconf
- add vala 0.1.6~svn as build-depends (need futur upload in PPA)
[Q]



Testing now....
[Q]

Score: 0 points
Any particular reason?
#18986 by fde (0.1733) posted on 11:54am Sunday, January 13th, 2008
malept said:
I hate stacks. I also hate that people want to turn Awn into an exact clone of the OSX bar. However, I'm just a (future) user. Moonbeam may implement it as an option.
[Q]



Ok, fine. In fact I never used those stacks, and I absolutely don't want my desktop to look like OSX. However, that would be much more convenient (and beautiful) to see the list of grouped tasks as a set of thumbnails instead of a textual list of their titles.
[Q]



EDIT: Moreover, I think - but I'm not sure - that OSX does NOT use those stacks to represent automatically grouped similar tasks.
[Q]





*Edited at 12:00pm, 01/13/08
Score: 0 points
Any particular reason?
#18991 by moonbeam (1.0000) posted on 12:19pm Sunday, January 13th, 2008
( no avatar )
fde said:
Ok, fine. In fact I never used those stacks, and I absolutely don't want my desktop to look like OSX. However, that would be much more convenient (and beautiful) to see the list of grouped tasks as a set of thumbnails instead of a textual list of their titles.
[Q]



I'm expect a transparent stack type presentation will eventually show up if this ends up being widely accepted (and possible if it isn't :-) )
[Q]



The thumbnails can definitely be done in a variety of ways... though I think the compiz style ones might require compiz (though it may just use the xrender extension)... I haven't seen the equivalent of live thumbnail windows in anything other than compiz so.... dunno. It probably would be cool to see... though I'd probably turn it off. Another issue with that feature is that it tends to be of variable use one you've scaled down to the size of an awn icon... sometimes it's obvious what the scaled window is, other times it's much more dubious. Maybe someone will get around to trying it out :-)
[Q]



At this point in time I'll be focusing on the basic functionality...
[Q]



-basic configurability.
-stackability, unstackability,
-left, right, middle clicks giving useful options.
-drag and drop of one icon into another (the stackability thing).
[Q]



etc.
[Q]

--
http://moon-shiny.blogspot.com/

*Edited at 12:25pm, 01/13/08
Score: 0 points
Any particular reason?
#18993 by fde (0.1733) posted on 2:05pm Sunday, January 13th, 2008
moonbeam said:
At this point in time I'll be focusing on the basic functionality...
[Q]



That's a wise way of doing things!
[Q]



Score: 0 points
Any particular reason?
#19020 by asomething (0.9744) posted on 9:09pm Monday, January 14th, 2008
@gilir
[Q]



I see that you packaged a Hardy deb of vala-0.1.6~svn20070114.2-1 on the testing PPA. Any plans to do one for Gutsy as well?
[Q]

Score: 0 points
Any particular reason?
#19025 by gilir (0.9999) posted on 11:29am Tuesday, January 15th, 2008
asomething said:
I see that you packaged a Hardy deb of vala-0.1.6~svn20070114.2-1 on the testing PPA. Any plans to do one for Gutsy as well?
[Q]



I upload it now, I hope it will build for gutsy :/
[Q]

Score: 0 points
Any particular reason?
#19029 by asomething (0.9744) posted on 12:45pm Tuesday, January 15th, 2008
gilir said:
I upload it now, I hope it will build for gutsy :/
[Q]



Looks like it failed... =(
[Q]



* Source Package: vala
* Version: 0.1.6~svn20070114.3-1
* Architecture: i386
* Archive: awn-testing PPA
* Component: main
* State: Failed to build
* Duration: a minute
* Build Log: http://launchpadlibrarian.net/11359991/buildlog_ubuntu-gutsy-i386.vala_0.1.6%7Esvn20070114.3-1_FAILEDTOBUILD.txt.gz
* Builder: https://launchpad.net/+builds/thallium
* Source: not available
[Q]

Score: 0 points
Any particular reason?
#19031 by gilir (0.9999) posted on 12:49pm Tuesday, January 15th, 2008
asomething said:
Looks like it failed... =(
[Q]



It need 0.1.5 to compile SVN, so I send 0.1.5 backport to the repo to test it :) Just wait a bit more, I'm on it :)
[Q]

Score: 0.86 points
Any particular reason?
#19041 by plunn (1.0000) posted on 4:04pm Tuesday, January 15th, 2008
Well, it works great
[Q]



Moonbeams branch and the 0.7.6 version
[Q]



[Q]



Put the long start command within Sessions and logout/login...
[Q]



Thanks !
[Q]



*Edited at 4:05pm, 01/15/08
Score: 0 points
Any particular reason?
#19042 by moonbeam (1.0000) posted on 4:19pm Tuesday, January 15th, 2008
( no avatar )
plunn said:
Moonbeams branch and the 0.7.6 version
[Q]



At this point in time the version in testing is more up to date. I will be pushing the changes back into my branch in a bit.
[Q]



With the advent of awn-extras-testing I'll probably be working in that branches mostly (unless I am working on something that is extremely broken). What I will probably end up doing is merging awn-extras-testing into my branch. And keep my branch in sync with that.
[Q]



So be warned... my branch will end up being a close reflection of testing and where it isn't expect the parts that differ to be very "interesting"
[Q]


Score: 0 points
Any particular reason?
#19048 by davim (0.9996) posted on 5:26pm Tuesday, January 15th, 2008
moonbeam said:
At this point in time the version in testing is more up to date. I will be pushing the changes back into my branch in a bit.
[Q]



are all branches merged in the testing branch??!
[Q]

Score: 0 points
Any particular reason?
#19049 by moonbeam (1.0000) posted on 5:35pm Tuesday, January 15th, 2008
( no avatar )
davim said:
are all branches merged in the testing branch??!
[Q]



I believe more or less everything is in there at the moment.
[Q]



So the testing branches truly fall into the interesting category at the moment.
[Q]





Score: 0 points
Any particular reason?
#19050 by davim (0.9996) posted on 5:55pm Tuesday, January 15th, 2008
moonbeam said:
I believe more or less everything is in there at the moment.
[Q]



cool :) so when this is ready every thing will be in the trunk :) but what about the rewrite that neil started?? will we have a 0.2.2 when the testing branch goes to trunk?? are there any planed dates???
[Q]

Score: 0 points
Any particular reason?
#19052 by moonbeam (1.0000) posted on 6:18pm Tuesday, January 15th, 2008
( no avatar )
davim said:
cool :) so when this is ready every thing will be in the trunk :) but what about the rewrite that neil started?? will we have a 0.2.2 when the testing branch goes to trunk?? are there any planed dates???
[Q]



Don't know for sure on any of those questions :-) I know some of us hope that this stuff will get merged into trunk. And hopefully it'll get some more testing.
[Q]




Score: 0 points
Any particular reason?
#19056 by asomething (0.9744) posted on 8:34pm Tuesday, January 15th, 2008
gilir said:
It need 0.1.5 to compile SVN, so I send 0.1.5 backport to the repo to test it :) Just wait a bit more, I'm on it :)
[Q]



It's there and working now. Thanks a lot!
[Q]



I'm not running hardy yet. I'm willing to deal with awn breaking for testing purposes, but not my whole system... =)
[Q]

Score: 0 points
Any particular reason?
#19062 by plunn (1.0000) posted on 2:14pm Wednesday, January 16th, 2008
asomething said:
I'm not running hardy yet. I'm willing to deal with awn breaking for testing purposes, but not my whole system... =)
[Q]



Hardy is a nice guy for the moment.... :-)
[Q]



moonbeam said:
At this point in time the version in testing is more up to date. I will be pushing the changes back into my branch in a bit.
[Q]



Ok installed testing branch and it also works.
[Q]



The challenge...it would be nice to move gnome applets
in the same way.
[Q]





*Edited at 2:15pm, 01/16/08
Score: 0 points
Any particular reason?
#19064 by moonbeam (1.0000) posted on 8:13pm Wednesday, January 16th, 2008
( no avatar )
I thought I'd take a few minutes and write up some (possibly incomplete) instructions on getting things up and running.
[Q]



I am skipping over all the stuff on meeting the normal dependencies etc... if you don't know what I'm talking about then it probably is best to stop here.
[Q]



1) Install awn-core-testing. (Desktop Agnostic should be fine also but DA has been merged into testing so I'll assume testing will be used).
[Q]



a) If you are using Ubunt hardy or gutsy and want binaries then go https://launchpad.net/~awn-testing/+archive. install the avant-window-navigator-packages and awn-extras-testing packages for your particular configuration. If this is an option for you I would strongly recommending trying this first.
[Q]



b) If you are in a different environment than hardy or gutsy (or you just like installing from source).
[Q]



- First thing you need to do install val svn packages. There are some available for hardy and gutsy are the url mentioned in (a). Basically at this point you need a version of vala > 0.1.5. If your version is less than the 0.1.6 (svn) you need will need to upgrade: http://live.gnome.org/Vala#head-89a00120bd3c8f5338067fc923df27c73684b3f8. Note that if you are upgrading vala by building you will need to do it one version increment at a time. Example if you have 0.1.12 you will need to build/install 0.1.13 then 0.1.14 then 0.1.15 then svn.
-Then you can build and install avant-core-testing and awn-extras-testing as normal. https://code.launchpad.net/~awn-testing/
[Q]



-assuming all has went with the builds there should be a (assuming you did not use change the --prefix) awn.vapi file located in /usr/local/share/vala/vapi/. Assuming there is then there will also be taskmand file in /usr/local/bin.
[Q]






2)if everything checks out. do a
[Q]



killall -HUP gconfd-2
[Q]



and
[Q]



sudo ldconfig
[Q]



-restart awn.
[Q]



-start awn-manager. add a standalone-launcher applet. Remove the awn taskmanager/launcher. Chances are this will cause awn to crash. If awn does not crash then close it. Start awn again.
[Q]



-Drag and drop a desktop file into the standalone-launcher. You can have as many standalone-launchers configured as you like.. but if you don't have at least one you will (at this time) need to start taskmand manually. If you have at least one launcher taskmand will get started automatically.
[Q]



And you're done.
[Q]



Misc notes:
[Q]



1) Just drag and drop the desktop file you want into the launcher.
2) To change the icon drag and drop the icon you want used into the launcher.
3) In the course of managing tasks taskmand will start a number of launchers in "anonymous" mode. At this time these will show up in awn-manager. They will get cleaned up when taskmand starts and stops.
4) awn-manager and taskmand do not always coexist peacefully. Sometimes awn will crasy ( or needed to be restarted) after making modifications with awn-manager.
5) configuration options can be found in gconf (or the appropriate ini if you build with desktop agnostic options). The dirs are
/apps/avant-window-navigator/applets/taskmand
/apps/avant-window-navigator/applets/standalone-launcher
6) default configuration is that the launchers will attempt to group related windows together. The popup is ugly at the moment... but it will be made shinier eventually. It is fairly good (but not perfect) at grouping at this time.
7) left click will open start a new task if there is none open, otherwise it will minimize, restore.
8) Middle click opens a new task regardless of others being open.
9) Right click does nothing yet.
10) No effects are implemented yet. you will get the normal hover effect happening though.
[Q]



Any other comments/questions? Just post here.
[Q]






[Q]

--
http://moon-shiny.blogspot.com/

*Edited at 8:18pm, 01/16/08
Score: 0 points
Any particular reason?
#19065 by plunn (1.0000) posted on 12:04am Thursday, January 17th, 2008
Just a memory note.
[Q]



Dependencies: flex, bison, gtk-doc-tools, libvala-dev, libvala0, valac added
[Q]



Vala (svn) broken for hardy, compile error
EDIT not broken...valac package installed and svn compiles
[Q]



awn-core-testing, compile/install OK
[Q]



Start not OK... :-)
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6e056b0 (LWP 5011)]
0xb73b4540 in pthread_mutex_lock () from /lib/tls/i686/cmov/libpthread.so.0

EDIT
awn.vapi file located in /usr/local/share/vala/vapi/
seems to be the problem. empty
[Q]



Also tested with ppa packages.
[Q]



*Edited at 12:38am, 01/17/08
Score: 0 points
Any particular reason?
#19066 by moonbeam (1.0000) posted on 5:53am Thursday, January 17th, 2008
( no avatar )
plunn said:
Start not OK... :-)
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6e056b0 (LWP 5011)]
0xb73b4540 in pthread_mutex_lock () from /lib/tls/i686/cmov/libpthread.so.0

EDIT
awn.vapi file located in /usr/local/share/vala/vapi/
seems to be the problem. empty
[Q]



That would be a problem if it was empty but only if awn was getting to the point of starting a launcher (and then it _should_ only crash the process). I might be wrong but I suspect this is a awn-core issue that you are seeing. It might be an idea to post this issue in the awn-core-testing thread also.
[Q]



Of note there was someone on irc yesterday having a very similar problem that was using hardy also.
[Q]



Score: 0 points
Any particular reason?
#19067 by plunn (1.0000) posted on 6:53am Thursday, January 17th, 2008
moonbeam said:
That would be a problem if it was empty but only if awn was getting to the point of starting a launcher (and then it _should_ only crash the process). I might be wrong but I suspect this is a awn-core issue that you are seeing. It might be an idea to post this issue in the awn-core-testing thread also.
[Q]



OK, well it ended in a total breakage for AWN...
[Q]



Impossible to run any version, uninstalled from source and also packages and reinstalled stable version. Seg faults...
[Q]



When and from what is awn.vapi created ? seems to be the key.
[Q]





*Edited at 6:53am, 01/17/08
Score: 0 points
Any particular reason?
#19068 by moonbeam (1.0000) posted on 7:01am Thursday, January 17th, 2008
( no avatar )
plunn said:
Impossible to run any version, uninstalled from source and also packages and reinstalled stable version. Seg faults...
[Q]



Are you able to successfully build other programs? The hardy user yesterday eventually determined the build environment on the system was borked.
[Q]



Basically awn.vapi just defines the awn vala api. It is used when compiling awn vala applets. It is not used at runtime in any way. I could be wrong but I think that the awan.vapi issue is a symptom of something else.
[Q]



Could you try building a different program from source (one that is not critical) and see what the result is... one of the test programs the user tried yesterday was affinity.
[Q]



If other programs are building and starting ok I'd suggest going into the awn-testing src trees and 'sudo make uninstall', then 'sudo ldconfig', then install the trunk version.
[Q]


Score: 0 points
Any particular reason?
#19077 by gilir (0.9999) posted on 10:10am Thursday, January 17th, 2008
@plunn
[Q]



vala bindings are in a separate package : vala-awn-testing
[Q]



Applets have problems to build on PPA, but all should be in place to compile them yourself using vala and awn PPA packages. I know it's working for sid, I'll test a bit more on hardy.
[Q]

Score: 0 points
Any particular reason?
#19078 by plunn (1.0000) posted on 10:14am Thursday, January 17th, 2008
Well, Vala is broken with latest svn update (compile errors) but there can be a dependency mismatch between your repo and Vala for the moment.
[Q]



gcc was broken yesterday not today (moonbeams question).
[Q]



Must "beat" this one...D
[Q]

Score: 0 points
Any particular reason?
#19079 by plunn (1.0000) posted on 10:32am Thursday, January 17th, 2008
Followup...
[Q]



Removed everything for compiling Vala, uninstalled all sources.
[Q]



Clean package install from testing PPA, AWN is back... :)
[Q]

Score: 0 points
Any particular reason?
#19080 by gilir (0.9999) posted on 10:34am Thursday, January 17th, 2008
plunn said:
Clean package install from testing PPA, AWN is back... :)
[Q]



A good news :)
[Q]

Score: 0 points
Any particular reason?
#19081 by plunn (1.0000) posted on 10:37am Thursday, January 17th, 2008
gilir said:
A good news :)
[Q]



The problem was this package...vala-awn-testing
[Q]



I started Vala compilation because ot this...and was "lost in the jungle"... :)
[Q]







Score: 0 points
Any particular reason?
#19082 by moonbeam (1.0000) posted on 11:22am Thursday, January 17th, 2008
( no avatar )
plunn said:
I started Vala compilation because ot this...and was "lost in the jungle"... :)
[Q]



Thanks for tracking it down. I suspect there was something else fundamentally wrong though as that just installs awn.vapi. Which only matters in the building of vala applets from source and the build should just fail.... needless to say I'm puzzled. Vala support does not effect the avant-window-navigator binary or its libraries in any way (I will confirm this).
[Q]



I suspect there is something more going on here than just that one file but I don't know what. Hopefully it won't happen again and we won't need to figure it out :-)
[Q]



Thanks for testing.
[Q]




Score: 0 points
Any particular reason?
#19083 by plunn (1.0000) posted on 11:31am Thursday, January 17th, 2008
moonbeam said:
Thanks for tracking it down. I suspect there was something else fundamentally wrong though as that just installs awn.vapi. Which only matters in the building of vala applets from source and the build should just fail.... needless to say I'm puzzled. Vala support does not effect the avant-window-navigator binary or its libraries in any way (I will confirm this).
[Q]



Well its more.. :-)
[Q]



I now lost my awn-manager, easiest way must be to compile it from
awn-core-testing. Mssing within packages and wrong version
within Hardy repo.
[Q]



I did a stupid install with awn source and directly got segfault.
Then I took the awn-manager within the source folder and it
was installed... But no standalone applet.
[Q]



It must have been mixed binarys.... :)
[Q]








Score: 0 points
Any particular reason?
#19087 by moonbeam (1.0000) posted on 1:32pm Thursday, January 17th, 2008
( no avatar )
plunn said:
It must have been mixed binarys.... :)
[Q]



I'm seeing some strange behaviours after doing a system update with testing... it's the segfault immediately issue others are seeing. Other people are building the newest rev without issue...
[Q]



I'm going to look into trying to trace down what is causing the issue. One interesting point is that the newest Desktop-Agnostic is function without issue so that tentatively eliminates a big swathe of potential causes of the bug.
[Q]



At this point I'm suspecting it's an update to a dependency that is the root cause.
[Q]


Score: 0 points
Any particular reason?
#19092 by moonbeam (1.0000) posted on 2:14pm Thursday, January 17th, 2008
( no avatar )
plunn said:
Well its more.. :-)
[Q]



what does this give you?
[Q]



pkg-config --modversion glib-2.0
[Q]
Score: 0 points
Any particular reason?
#19093 by plunn (1.0000) posted on 2:37pm Thursday, January 17th, 2008
moonbeam said:
pkg-config --modversion glib-2.0
[Q]



plun@dunder:~$ pkg-config --modversion glib-2.0
2.15.2
[Q]



Score: 0 points
Any particular reason?
#19095 by moonbeam (1.0000) posted on 5:25pm Thursday, January 17th, 2008
( no avatar )
Request.
[Q]



If anyone does build and is having awn crashes immediately please run
[Q]



ldd /path-to-awn/avant-window-navigator | grep glib
[Q]



and post the results in the awn-core-testing thread.
[Q]



I was having some build issues today and it turned out that both glib 1.2 and 2.x were being linked. I"d like to know if this is just something that is happening on my mutant system or if it is affecting a wider userbase.
[Q]
Score: 0 points
Any particular reason?
#19101 by plunn (1.0000) posted on 3:34am Friday, January 18th, 2008
Yup...
[Q]



plun@dunder:~/$ ldd /usr/bin/avant-window-navigator | grep glib
libdbus-glib-1.so.2 => /usr/lib/libdbus-glib-1.so.2 (0xb7c60000)
libglib-2.0.so.0 => /usr/lib/libglib-2.0.so.0 (0xb752b000)
libavahi-glib.so.1 => /usr/lib/libavahi-glib.so.1 (0xb7141000)
libglib-1.2.so.0 => /usr/lib/libglib-1.2.so.0 (0xb6fa0000)
[Q]




- Vala (svn) was fixed som now it compiles
[Q]



- awn-core and testing compiled/installed OK
[Q]




Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb6e296b0 (LWP 10100)]
0xb73d8540 in pthread_mutex_lock () from /lib/tls/i686/cmov/libpthread.so.0
[Q]



Hardy bug ?
[Q]

Score: 0 points
Any particular reason?
#19102 by moonbeam (1.0000) posted on 5:50am Friday, January 18th, 2008
( no avatar )
plunn said:
Hardy bug ?
[Q]



Well I'm using gentoo... so I'm beginning to think it might be a build system issue. Don't really know yet. I'll talk to our build system guru (malept) about it a bit more.
[Q]


Score: 0 points
Any particular reason?
#19103 by plunn (1.0000) posted on 6:52am Friday, January 18th, 2008
moonbeam said:
Well I'm using gentoo... so I'm beginning to think it might be a build system issue. Don't really know yet. I'll talk to our build system guru (malept) about it a bit more.
[Q]



Well again... this is just a composite challenge.
[Q]



Everything is broken with compiz-fusion running.
Metacity with composite, its OK.
[Q]



So back to the real thing with real 3D... :)
[Q]



EDIT
[Q]



Nope crashed again....really strange behaviors
[Q]



*Edited at 6:56am, 01/18/08
Score: 0 points
Any particular reason?
#19108 by malept (1.0000) posted on 9:59am Friday, January 18th, 2008
( no avatar )
plunn: what's the output of `grep AWN_LIBS Makefile` ?
[Q]
Score: 0 points
Any particular reason?
#19111 by plunn (1.0000) posted on 11:27am Friday, January 18th, 2008
malept said:
plunn: what's the output of `grep AWN_LIBS Makefile` ?
[Q]



plun@dunder:~/awn-extras-testing$ grep AWN_LIBS MakefileAWN_LIBS = -pthread -L/usr/local/lib -lawn -lgnome-desktop-2 -lgnomeui-2 -lSM -lICE -lstartup-notification-1 -lbonoboui-2 -lgnomevfs-2 -lgnomecanvas-2 -lgnome-2 -lpopt -lart_lgpl_2 -lgtk-x11-2.0 -lgdk-x11-2.0 -latk-1.0 -lgdk_pixbuf-2.0 -lm -lpangocairo-1.0 -lpango-1.0 -lcairo -lbonobo-2 -lbonobo-activation -lORBit-2 -lgconf-2 -lgthread-2.0 -lrt -lgmodule-2.0 -ldl -lgobject-2.0 -lglib-2.0 -lglib
[Q]



Score: 0 points
Any particular reason?
#19123 by moonbeam (1.0000) posted on 11:40am Saturday, January 19th, 2008
( no avatar )
For those interested. Vala Version 0.1.6 was released.
[Q]



Score: 0 points
Any particular reason?
#19174 by plunn (1.0000) posted on 5:55am Tuesday, January 22nd, 2008
Follow up...
[Q]



Today everything worked...
[Q]



Latest Vala, awn-extras-testing and awn-core-testing.
[Q]



No seg faults
[Q]



Just to figure out about the standalone launcher :)
[Q]

Score: 0 points
Any particular reason?
#19175 by moonbeam (1.0000) posted on 7:09am Tuesday, January 22nd, 2008
( no avatar )
plunn said:
Just to figure out about the standalone launcher :)
[Q]



Add a launcher and drag and drop the desktop file you want into it.
[Q]



In general don't be surprised if you need to restart awn after adding removing an applet using awn-manager (as I've said before it currently doesn't play nice with others)
[Q]
Score: 0 points
Any particular reason?
#19177 by plunn (1.0000) posted on 10:07am Tuesday, January 22nd, 2008
moonbeam said:
Add a launcher and drag and drop the desktop file you want into it.
[Q]



Yup it works....
[Q]



Did you remove stalonetray ?
[Q]





Score: 0 points
Any particular reason?
#19178 by moonbeam (1.0000) posted on 10:23am Tuesday, January 22nd, 2008
( no avatar )
plunn said:
Yup it works....
[Q]



Good. I hope that later today I'll commit a simple applet that allows placement of the tasks. Basically drop in in the applet list where you want it and taskmand will put new icons to the left or right of it. This will be an additional option over top of the current configuration method. I also intend that the applet will server to start up taskmand also... though that might not happen today.
[Q]



plunn said:
Did you remove stalonetray ?
[Q]



It's still in my branch. And I will be keeping it there. Though it doesn't really belong in the awn branches as such - I just put it in mine as a convenience.
[Q]



I will get around to syncing my branch with testing really soon (it is out of date relative to testing). RL has been keeping me occupied :-)
[Q]




Score: 0 points
Any particular reason?
#19179 by plunn (1.0000) posted on 10:27am Tuesday, January 22nd, 2008
Ok thanks... :)
[Q]

Score: 0 points
Any particular reason?
#19185 by asomething (0.9744) posted on 10:08pm Tuesday, January 22nd, 2008
Hey moonbeam,
[Q]



FYI: awn-extras-testing rev357 is failing to build. Erroring out on:
[Q]



 
Making all in taskmand
make[3]: Entering directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand'
dbus-binding-tool --prefix=taskmand --mode=glib-server taskmand-infos.xml> server-bindings.h
dbus-binding-tool --prefix=taskmand --mode=glib-client taskmand-infos.xml> client-bindings.h
make all-recursive
make[4]: Entering directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand'
Making all in data
make[5]: Entering directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand/data'
make[5]: Leaving directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand/data'
make[5]: Entering directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand'
if gcc -DHAVE_CONFIG_H -I. -I. -I../.. -DLIBAWN_USE_GNOME -DORBIT2=1 -pthread -I/usr/local/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/gnome-desktop-2.0 -I/usr/include/libgnome-2.0 -I/usr/include/gnome-vfs-2.0 -I/usr/include/gnome-vfs-module-2.0 -I/usr/lib/gnome-vfs-2.0/include -I/usr/include/gconf/2 -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/freetype2 -I/usr/include/libpng12 -I/usr/include/libgnomeui-2.0 -I/usr/include/startup-notification-1.0 -I/usr/include/libart-2.0 -I/usr/include/gnome-keyring-1 -I/usr/include/libbonoboui-2.0 -I/usr/include/libgnomecanvas-2.0 -I/usr/include/orbit-2.0 -I/usr/include/libbonobo-2.0 -I/usr/include/bonobo-activation-2.0 -I/usr/include/libxml2 -I/usr/include/gail-1.0 -DDATADIR=\""/usr/local/share"\" -DLIBDIR=\""/usr/local/lib"\" -DGNOMELOCALEDIR=\""/usr/local/share/locale"\" -I/usr/include/dbus-1.0/ -I/usr/include/dbus-1.0 -I/usr/lib/dbus-1.0/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/libwnck-1.0 -I/usr/include/gtk-2.0 -I/usr/include/startup-notification-1.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/freetype2 -I/usr/include/libpng12 -I../ -g -O2 -MT taskmand.o -MD -MP -MF ".deps/taskmand.Tpo" -c -o taskmand.o taskmand.c; \
then mv -f ".deps/taskmand.Tpo" ".deps/taskmand.Po"; else rm -f ".deps/taskmand.Tpo"; exit 1; fi
In file included from taskmand.c:43:
server-bindings.h:165: error: ‘taskmand_launcher_position’ undeclared here (not in a function)
taskmand.c: In function ‘remove_from_response_tree’:
taskmand.c:80: warning: passing argument 2 of ‘g_tree_lookup’ makes pointer from integer without a cast
taskmand.c:88: warning: passing argument 2 of ‘g_tree_remove’ makes pointer from integer without a cast
taskmand.c: In function ‘add_to_response_list’:
taskmand.c:97: warning: passing argument 3 of ‘g_list_find_custom’ from incompatible pointer type
taskmand.c:100: warning: passing argument 2 of ‘g_tree_replace’ makes pointer from integer without a cast
taskmand.c: In function ‘check_response_list’:
taskmand.c:109: warning: passing argument 2 of ‘g_tree_lookup’ makes pointer from integer without a cast
taskmand.c: In function ‘taskmand_launcher_register’:
taskmand.c:148: warning: passing argument 3 of ‘g_list_find_custom’ from incompatible pointer type
taskmand.c: In function ‘taskmand_inform_task_ownership’:
taskmand.c:157: warning: passing argument 3 of ‘g_list_find_custom’ from incompatible pointer type
taskmand.c:162: warning: passing argument 2 of ‘g_list_find’ makes pointer from integer without a cast
taskmand.c:164: warning: passing argument 2 of ‘g_list_remove’ makes pointer from integer without a cast
taskmand.c:168: warning: passing argument 2 of ‘g_tree_remove’ makes pointer from integer without a cast
taskmand.c:177: warning: passing argument 2 of ‘g_list_find’ makes pointer from integer without a cast
taskmand.c:179: warning: passing argument 2 of ‘g_tree_lookup’ makes pointer from integer without a cast
taskmand.c:179: warning: comparison between pointer and integer
taskmand.c:182: warning: passing argument 2 of ‘g_list_remove’ makes pointer from integer without a cast
taskmand.c:184: warning: passing argument 2 of ‘g_tree_remove’ makes pointer from integer without a cast
taskmand.c: In function ‘taskmand_launcher_unregister’:
taskmand.c:209: warning: passing argument 3 of ‘g_list_find_custom’ from incompatible pointer type
taskmand.c: In function ‘launch_anonymous_launcher’:
taskmand.c:285: warning: passing argument 1 of ‘g_list_length’ from incompatible pointer type
taskmand.c: In function ‘_launcher_response_timeout’:
taskmand.c:322: warning: passing argument 2 of ‘g_tree_lookup’ makes pointer from integer without a cast
taskmand.c:322: warning: comparison between pointer and integer
taskmand.c:324: warning: passing argument 2 of ‘g_tree_replace’ makes pointer from integer without a cast
taskmand.c:324: warning: passing argument 3 of ‘g_tree_replace’ makes pointer from integer without a cast
taskmand.c:327: warning: passing argument 2 of ‘g_tree_remove’ makes pointer from integer without a cast
taskmand.c:328: warning: passing argument 2 of ‘g_list_find’ makes pointer from integer without a cast
taskmand.c: In function ‘_window_opened’:
taskmand.c:354: warning: passing argument 2 of ‘g_list_append’ makes pointer from integer without a cast
taskmand.c:355: warning: passing argument 3 of ‘g_timeout_add’ makes pointer from integer without a cast
taskmand.c:356: warning: passing argument 2 of ‘g_tree_replace’ makes pointer from integer without a cast
taskmand.c:356: warning: passing argument 3 of ‘g_tree_replace’ makes pointer from integer without a cast
taskmand.c: In function ‘_window_closed’:
taskmand.c:364: warning: passing argument 2 of ‘g_list_remove’ makes pointer from integer without a cast
make[5]: *** [taskmand.o] Error 1
make[5]: Leaving directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand'
make[4]: *** [all-recursive] Error 1
make[4]: Leaving directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand'
make[3]: *** [all] Error 2
make[3]: Leaving directory `/home/andrew/source-installs/development/awn-extras-testing-install/src/taskmand'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory `/home/andrew/source-installs/development/awn-extras-testing-install/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/andrew/source-installs/development/awn-extras-testing-install'
make: *** [all] Error 2
[Q]



Thanks!
[Q]

Score: 0 points
Any particular reason?
#19189 by moonbeam (1.0000) posted on 12:29am Wednesday, January 23rd, 2008
( no avatar )
asomething said:
FYI: awn-extras-testing rev357 is failing to build. Erroring out on:
[Q]



thanks
[Q]



revision 358 should fix that.
[Q]
Score: 0 points
Any particular reason?
Pages: 1 2 3 4 5
Thread Listing » AWN Applets » [Announcement] Standalone launchers with a separate taskmanager

Post A Reply:

Anonymous posting has been disabled. Please register first before attempting to post.
Powered by Metaforum ©2004-2010
Get your own Free AJAX Forum Messageboard by visiting the offical site.
.