ROSTop (Re: [ROS Wiki] Update of "rqt/Plugins" by DanLazewatsky

692 views
Skip to first unread message

Isaac I. Saito

unread,
May 2, 2013, 10:24:26 PM5/2/13
to Dan Lazewatsky, ros-s...@googlegroups.com
Dan,

I saw your ROSTop rqt plugin. Looks nice! Thanks for letting us know
by adding it on wiki!

Do you mind if I put it into our rqt tool suite?? Its feature seems
generic enough and many users will appreciate it. If you don't, I will
import the pkg into rqt_common_plugins metapkg and it'll be released
as binary pkg. Good thing is that I'll take a co-maintainership
(double the eyes) while you'll remain the author.

--
Isaac

On Thu, May 2, 2013 at 12:28 PM, Moin <mo...@willowgarage.com> wrote:
> Dear Wiki user,
>
> You have subscribed to a wiki page or wiki category on "ROS Wiki" for change notification.
>
> The "rqt/Plugins" page has been changed by DanLazewatsky:
> http://www.ros.org/wiki/rqt/Plugins?action=diff&rev1=45&rev2=46
>
> ||[[kobuki_desktop|Kobuki Desktop]]||Python, C++||Software for simulating [[kobuki|Kobuki]]||
> ||[[http://www.ros.org/wiki/marble_plugin|Marble]]||C++||shows a geo position on an map. Marble is used as a widget base. The geo position has to be of the sensor_msgs::SatNavFix type. Made by Tobias Bär||
> ||[[rosgui_paramedit|ParamEdit]]||C++||List and edit parameters on the parameter server (similar functionality to command-line tool `rosparam`).||
> + ||[[rostop_gui|ROSTop]]||Python||A process monitoring utility similar to Ubuntu's System Monitor, which only shows info about ROS processes.||
> ||[[https://github.com/mayacakmak/pr2_hri/tree/master/pr2_pbd_gui/src|PR2 PBD GUI]]||Python||Predefined text to speech command I/F for `PR2` made by Maya Cakmak.||
> ||[[reconfigure_gui|ReconfigureGUI]]||Python||Tool to interface with nodes that have [[dynamic_reconfigure|dynamic_reconfigure]] enabled||
> ||[[sr_visualization|Shadow Robot Visualization stack]]||-||-||

Dan Lazewatsky

unread,
May 3, 2013, 11:38:05 AM5/3/13
to Isaac I. Saito, ros-s...@googlegroups.com
Hi Isaac -
You're definitely welcome to add it to the rqt tool suite; I'm glad the plugin can be useful to others. One thing to note about packing though - psutil, which is the library that gets all of the process info, is stuck at a very old version in the Precise deb which is missing a feature that it needs. Anything newer than Precise should be fine though. It's up to date in pip, but according to Tully, apt is preferred on Ubuntu. 

Let me know if there's anything else you need me to do, and I'm of course always open to bug reports and pull requests.

-Dan 

Isaac I. Saito

unread,
May 14, 2013, 12:44:40 AM5/14/13
to ros-s...@googlegroups.com
Dan,

thanks for the response. I've added the changes I think necessary.
https://github.com/130s/rqt_top

Changes are as follows:

- added rosdep.yaml
- catkinized
- changed pkg name to conform to the custom of existing rqt plugins
- added me as a maintainer

Considering Precise is LTS for ROS, I guess we need to work around the
psutil issue. I'm sure that rosdep.yaml file I added is NOT sufficient
or might not even be functioning properly (I added it following
William's suggestion in the rosdistro thread [0], but so far I'm not
yet sure how that file is supposed to be written..).

Looks like it runs fine on my Quantal-Groovy machine after "rosdep
install rqt_top".

We might want to change the pkg name to a little more general name
like rqt_procmonitor?

Once all look good, I'll pull it to rqt_common_plugins metapackage and
release it.

[0] https://github.com/ros/rosdistro/pull/789#issuecomment-17348070

--
Isaac
> --
> --
> To unsubscribe from this group, send email to
> ros-sig-rqt...@googlegroups.com
> Google group at http://groups.google.com/group/ros-sig-rqt?hl=en
> rqt: http://ros.org/wiki/rqt
>
>

Dan Lazewatsky

unread,
May 14, 2013, 4:24:47 PM5/14/13
to ros-s...@googlegroups.com
Answered inline -

On Mon, May 13, 2013 at 11:44 PM, Isaac I. Saito <isa...@willowgarage.com> wrote:
Dan,

thanks for the response. I've added the changes I think necessary.
https://github.com/130s/rqt_top

Changes are as follows:

- added rosdep.yaml
- catkinized
- changed pkg name to conform to the custom of existing rqt plugins
- added me as a maintainer

Looks good. Tried it out in my VM, and everything seems fine.
 

Considering Precise is LTS for ROS, I guess we need to work around the
psutil issue. I'm sure that rosdep.yaml file I added is NOT sufficient
or might not even be functioning properly (I added it following
William's suggestion in the rosdistro thread [0], but so far I'm not
yet sure how that file is supposed to be written..).

I think there needs to be some special handling in rosdep.yaml for for precise. As it stands with the rosdep.yaml in there, it would install a version of psutil that's too old. Since, as I understand it, debs can't have pip dependencies, we will need to do something a little trickier. The first thing that comes to mind is to make a branch for precise, and pull in psutil as a submodule so it has its own version of psutil isolated what apt installs. No idea if this is a good idea though.
 

Looks like it runs fine on my Quantal-Groovy machine after "rosdep
install rqt_top".

We might want to change the pkg name to a little more general name
like rqt_procmonitor?

Definitely needs to be renamed. As it's named right now, running scripts/rqt_top confuses qt_gui_main():

qt_gui_main() found multiple plugins matching "rqt_top"
rqt_top/rqt_top.rostop_plugin.RosTop
rqt_topic/rqt_topic.topic.Topic

Something like rqt_procmonitor or rqt_process_monitor is fine with me.

Dirk Thomas

unread,
May 14, 2013, 4:40:28 PM5/14/13
to ros-s...@googlegroups.com
On 14.05.2013 13:24, Dan Lazewatsky wrote:
>
> On Mon, May 13, 2013 at 11:44 PM, Isaac I. Saito <isa...@willowgarage.com <mailto:isa...@willowgarage.com>> wrote:
>
> Considering Precise is LTS for ROS, I guess we need to work around the
> psutil issue. I'm sure that rosdep.yaml file I added is NOT sufficient
> or might not even be functioning properly (I added it following
> William's suggestion in the rosdistro thread [0], but so far I'm not
> yet sure how that file is supposed to be written..).
>
>
> I think there needs to be some special handling in rosdep.yaml for for precise. As it stands with the rosdep.yaml in there, it would install a version of psutil that's too old. Since, as I understand
> it, debs can't have pip dependencies, we will need to do something a little trickier. The first thing that comes to mind is to make a branch for precise, and pull in psutil as a submodule so it has
> its own version of psutil isolated what apt installs. No idea if this is a good idea though.

There can't be a separate version for precise.
Just embed the module if necessary (under a folder "external" or similar to express that it is external code).
Than the plugin can try importing the existing system installed version and if the version is to old fall back to the module embedded in the package.


> Looks like it runs fine on my Quantal-Groovy machine after "rosdep
> install rqt_top".
>
> We might want to change the pkg name to a little more general name
> like rqt_procmonitor?
>
>
> Definitely needs to be renamed. As it's named right now, running scripts/rqt_top confuses qt_gui_main():
>
> qt_gui_main() found multiple plugins matching "rqt_top"
> rqt_top/rqt_top.rostop_plugin.RosTop
> rqt_topic/rqt_topic.topic.Topic

This is not "confusing" qt_gui_main.
The argument you passed for stanting a specific plugin was just not specific enough.

rqt_top is a pretty good name imho.
What should be considered for renaming is "rostop" as the users might expect a command line tool like all other ros*.

- Dirk

Dan Lazewatsky

unread,
May 14, 2013, 4:59:31 PM5/14/13
to ros-s...@googlegroups.com
On Tue, May 14, 2013 at 3:40 PM, Dirk Thomas <dth...@osrfoundation.org> wrote:
rqt_top is a pretty good name imho.
What should be considered for renaming is "rostop" as the users might expect a command line tool like all other ros*.

It's not really a command line tool though. The question is what is it best to be consistent with? ROS tools are ros*, but rqt_plugins are rqt_*.

Dirk Thomas

unread,
May 14, 2013, 5:01:01 PM5/14/13
to ros-s...@googlegroups.com
On 14.05.2013 13:59, Dan Lazewatsky wrote:
> On Tue, May 14, 2013 at 3:40 PM, Dirk Thomas <dth...@osrfoundation.org <mailto:dth...@osrfoundation.org>> wrote:
>
> rqt_top is a pretty good name imho.
> What should be considered for renaming is "rostop" as the users might expect a command line tool like all other ros*.
>
>
> It's not really a command line tool though. The question is what is it best to be consistent with? ROS tools are ros*, but rqt_plugins are rqt_*.

All rqt plugins start with "rqt" so that the user know to expect a GUI.

- Dirk

Dan Lazewatsky

unread,
May 14, 2013, 5:02:10 PM5/14/13
to ros-s...@googlegroups.com
Misread your comment. Reading it correctly, I agree.

On Tue, May 14, 2013 at 4:01 PM, Dirk Thomas <dth...@osrfoundation.org> wrote:
On 14.05.2013 13:59, Dan Lazewatsky wrote:

On Tue, May 14, 2013 at 3:40 PM, Dirk Thomas <dth...@osrfoundation.org <mailto:dthomas@osrfoundation.org>> wrote:

    rqt_top is a pretty good name imho.
    What should be considered for renaming is "rostop" as the users might expect a command line tool like all other ros*.


It's not really a command line tool though. The question is what is it best to be consistent with? ROS tools are ros*, but rqt_plugins are rqt_*.

All rqt plugins start with "rqt" so that the user know to expect a GUI.


- Dirk

--
--
To unsubscribe from this group, send email to ros-sig-rqt+unsubscribe@googlegroups.com

Aaron Blasdel

unread,
May 14, 2013, 7:23:30 PM5/14/13
to ros-s...@googlegroups.com
+1 for rqt_top

I'm not sure it is critical enough to warrant a script installed at the same level as rqt_console though. 
--

Daniel Stonier

unread,
May 15, 2013, 4:25:22 AM5/15/13
to ros-s...@googlegroups.com
On 15 May 2013 08:23, Aaron Blasdel <abla...@gmail.com> wrote:
+1 for rqt_top

I'm not sure it is critical enough to warrant a script installed at the same level as rqt_console though. 

+1 for a script in the global bin dir.
 

On May 15, 2013, at 5:59, Dan Lazewatsky <d...@lazewatsky.com> wrote:

On Tue, May 14, 2013 at 3:40 PM, Dirk Thomas <dth...@osrfoundation.org> wrote:
rqt_top is a pretty good name imho.
What should be considered for renaming is "rostop" as the users might expect a command line tool like all other ros*.

It's not really a command line tool though. The question is what is it best to be consistent with? ROS tools are ros*, but rqt_plugins are rqt_*.

--
--
To unsubscribe from this group, send email to ros-sig-rqt...@googlegroups.com
Google group at http://groups.google.com/group/ros-sig-rqt?hl=en
rqt: http://ros.org/wiki/rqt
 
 

--
--
To unsubscribe from this group, send email to ros-sig-rqt...@googlegroups.com
Google group at http://groups.google.com/group/ros-sig-rqt?hl=en
rqt: http://ros.org/wiki/rqt
 
 

Isaac Isao Saito

unread,
May 15, 2013, 5:13:10 AM5/15/13
to ros-s...@googlegroups.com
On Tue, May 14, 2013 at 3:40 PM, Dirk Thomas <dth...@osrfoundation.org> wrote:
> On 14.05.2013 13:24, Dan Lazewatsky wrote:
>
> Just embed the module if necessary (under a folder "external" or similar to
> express that it is external code).
> Than the plugin can try importing the existing system installed version and
> if the version is to old fall back to the module embedded in the package.

I examined, and found that there are many psutil related files (I
paste `locate` output at #1). Here I'm not sure if these many files
are "embeddable" in a folder as I have no idea how to realize their
directory hierarchy esp. for dynamic library files (it maybe easy for
others. I just don't know how).

So I still tried rosdep but haven't been successful and opened a question
http://answers.ros.org/question/62756/install-rosdep-definition-locally-in-a-package/

Thanks for the help,
Isaac

#1 locate output on Quantal:
$ locate psutils
/usr/lib/psutils
/usr/lib/psutils/md68_0.ps
/usr/lib/psutils/md71_0.ps
/usr/lib/pyshared/python2.7/_psutil_linux.so
/usr/lib/pyshared/python2.7/_psutil_posix.so
/usr/lib/python2.7/dist-packages/_psutil_linux.so
/usr/lib/python2.7/dist-packages/_psutil_posix.so
/usr/lib/python2.7/dist-packages/psutil
/usr/lib/python2.7/dist-packages/psutil-0.6.1.egg-info
/usr/lib/python2.7/dist-packages/psutil/__init__.py
/usr/lib/python2.7/dist-packages/psutil/__init__.pyc
/usr/lib/python2.7/dist-packages/psutil/_common.py
/usr/lib/python2.7/dist-packages/psutil/_common.pyc
/usr/lib/python2.7/dist-packages/psutil/_compat.py
/usr/lib/python2.7/dist-packages/psutil/_compat.pyc
/usr/lib/python2.7/dist-packages/psutil/_psbsd.py
/usr/lib/python2.7/dist-packages/psutil/_psbsd.pyc
/usr/lib/python2.7/dist-packages/psutil/_pslinux.py
/usr/lib/python2.7/dist-packages/psutil/_pslinux.pyc
/usr/lib/python2.7/dist-packages/psutil/_psmswindows.py
/usr/lib/python2.7/dist-packages/psutil/_psmswindows.pyc
/usr/lib/python2.7/dist-packages/psutil/_psosx.py
/usr/lib/python2.7/dist-packages/psutil/_psosx.pyc
/usr/lib/python2.7/dist-packages/psutil/_psposix.py
/usr/lib/python2.7/dist-packages/psutil/_psposix.pyc
/usr/lib/python2.7/dist-packages/psutil/error.py
/usr/lib/python2.7/dist-packages/psutil/error.pyc
/usr/share/doc/psutils
/usr/share/doc/python-psutil
/usr/share/doc/psutils/README.gz
/usr/share/doc/psutils/changelog.Debian.gz
/usr/share/doc/psutils/copyright
/usr/share/doc/python-psutil/changelog.Debian.gz
/usr/share/doc/python-psutil/copyright
/usr/share/doc/python-psutil/examples
/usr/share/doc/python-psutil/examples/disk_usage.py
/usr/share/doc/python-psutil/examples/free.py
/usr/share/doc/python-psutil/examples/iotop.py
/usr/share/doc/python-psutil/examples/killall.py
/usr/share/doc/python-psutil/examples/meminfo.py
/usr/share/doc/python-psutil/examples/netstat.py
/usr/share/doc/python-psutil/examples/nettop.py
/usr/share/doc/python-psutil/examples/pmap.py
/usr/share/doc/python-psutil/examples/process_detail.py
/usr/share/doc/python-psutil/examples/top.py
/usr/share/doc/python-psutil/examples/who.py
/usr/share/lintian/overrides/psutils
/usr/share/pyshared/psutil
/usr/share/pyshared/psutil-0.6.1.egg-info
/usr/share/pyshared/psutil/__init__.py
/usr/share/pyshared/psutil/_common.py
/usr/share/pyshared/psutil/_compat.py
/usr/share/pyshared/psutil/_psbsd.py
/usr/share/pyshared/psutil/_pslinux.py
/usr/share/pyshared/psutil/_psmswindows.py
/usr/share/pyshared/psutil/_psosx.py
/usr/share/pyshared/psutil/_psposix.py
/usr/share/pyshared/psutil/error.py
/var/cache/apt/archives/python-psutil_0.6.1-1ubuntu1_amd64.deb
/var/lib/dpkg/info/psutils.list
/var/lib/dpkg/info/psutils.md5sums
/var/lib/dpkg/info/python-psutil.list
/var/lib/dpkg/info/python-psutil.md5sums
/var/lib/dpkg/info/python-psutil.postinst
/var/lib/dpkg/info/python-psutil.prerm

Dirk Thomas

unread,
May 15, 2013, 12:19:17 PM5/15/13
to ros-s...@googlegroups.com
On 15.05.2013 01:25, Daniel Stonier wrote:
>
> On 15 May 2013 08:23, Aaron Blasdel <abla...@gmail.com <mailto:abla...@gmail.com>> wrote:
>
> +1 for rqt_top
>
> I'm not sure it is critical enough to warrant a script installed at the same level as rqt_console though.
>
>
> +1 for a script in the global bin dir.

We should only place really critically important scripts into the global bin in order to not clutter it.
Therefore I would suggest keeping it in the package bin folder as it is for almost all other rqt plugins (except the ones migrated from rx since the users expect them to be on the path).

- Dirk

William Woodall

unread,
May 15, 2013, 1:40:57 PM5/15/13
to ros-s...@googlegroups.com
Just to play the devil's advocate, once you have more that one rqt_* program in PATH, it really doesn't matter how many rqt_* packages have scripts in PATH.

You might even consider it inconsistent to have some plugins as scripts in the PATH and other's not.




--
--
To unsubscribe from this group, send email to ros-sig-rqt+unsubscribe@googlegroups.com



--
William Woodall
ROS Development Team

Aaron Blasdel

unread,
May 15, 2013, 11:29:22 PM5/15/13
to ros-s...@googlegroups.com
+1 to no global bin copy.
If we start putting everything in the global it will be a problem.

Daniel Stonier

unread,
May 19, 2013, 7:52:28 PM5/19/13
to ros-s...@googlegroups.com
On 16 May 2013 12:29, Aaron Blasdel <abla...@gmail.com> wrote:
+1 to no global bin copy.
If we start putting everything in the global it will be a problem.


I'd like to respond to this, but I think this is the wrong thread. I'll move the discussion to the build sig.

Isaac Isao Saito

unread,
Jun 6, 2013, 7:53:45 PM6/6/13
to ros-s...@googlegroups.com
Dirk,

On Wed, May 15, 2013 at 5:40 AM, Dirk Thomas <dth...@osrfoundation.org> wrote:
>> Definitely needs to be renamed. As it's named right now, running
>> scripts/rqt_top confuses qt_gui_main():
>>
>> qt_gui_main() found multiple plugins matching "rqt_top"
>> rqt_top/rqt_top.rostop_plugin.RosTop
>> rqt_topic/rqt_topic.topic.Topic
>
>
> This is not "confusing" qt_gui_main.
> The argument you passed for stanting a specific plugin was just not specific enough.

Could you tell me how to pass specific info in rqt plugin pkg in
general so that the error above doesn't happen?

# I'm trying to add a patch to this pkg but at loss where to change.

Isaac

Dirk Thomas

unread,
Jun 6, 2013, 8:04:24 PM6/6/13
to ros-s...@googlegroups.com
rqt_top is simply not unique when looking at the list of available plugins.
rqt uses the passed string to match it against the list of available plugins (|.*string.*|).
Only if that matches only one plugin it is unambiguous - else rqt has no way to determine which oe the user wants.

So you just need to pass a string which matches only a single plugin.

- Dirk

Isaac Isao Saito

unread,
Jun 6, 2013, 8:45:42 PM6/6/13
to ros-s...@googlegroups.com
Dirk,

sorry to bother you a lot, but then where in an rqt plugin to pass an
unambiguous string?
I assume you're not saying that I have to change the name of my pkg.

This is the branch I'm working on if this helps.
https://github.com/130s/rqt_top/tree/embed_psutil

Isaac

Dirk Thomas

unread,
Jun 6, 2013, 8:53:18 PM6/6/13
to ros-s...@googlegroups.com
You can print the list of available plugins with "rqt --list-plugins":
rqt_action.action_plugin.ActionPlugin
rqt_bag.bag.Bag
rqt_console.console.Console
rqt_dep.ros_pack_graph.RosPackGraph
rqt_graph.ros_graph.RosGraph
...

Your string must only match one of them and than it works - have you tried "rqt_top."?
You are just unlucky with "rqt_top" since there is also "rqt_topic".

- Dirk

Isaac Isao Saito

unread,
Jun 6, 2013, 9:03:31 PM6/6/13
to ros-s...@googlegroups.com
Ok, now I'm tempted to change the pkg name to avoid colliding with rqt_topic.

"--list-plugins" argument is what I've been kind of looking for and
somehow didn't notice. Thanks.

> Your string must only match one of them and than it works - have you tried "rqt_top."?

I haven't tried and won't try since I guess including dot in the pkg
name might not be a good idea...

Isaac

Dirk Thomas

unread,
Jun 6, 2013, 9:07:49 PM6/6/13
to ros-s...@googlegroups.com
On 06.06.2013 18:03, Isaac Isao Saito wrote:
> Ok, now I'm tempted to change the pkg name to avoid colliding with rqt_topic.
>
> "--list-plugins" argument is what I've been kind of looking for and
> somehow didn't notice. Thanks.
>
>> Your string must only match one of them and than it works - have you tried "rqt_top."?
>
> I haven't tried and won't try since I guess including dot in the pkg
> name might not be a good idea...

I am not suggesting to use a dot in the package name.
I am proposing to call "rqt -s rqt_top." which should be unique and run your plugin.
And that string "rqt_top." is what you could pass as the standalone argument in your script https://github.com/130s/rqt_top/blob/embed_psutil/scripts/rqt_rostop#L39

- Dirk

Isaac Isao Saito

unread,
Jun 6, 2013, 9:52:31 PM6/6/13
to ros-s...@googlegroups.com
Wow, good to know!
This time, however, changing script may not work; I got the following
error. Since with this particular package, I think I/we have no
problem in changing pkg name, I won't try hacking this. Thanks anyway,
a lot to learn for me.

{{{
130s@130s-serval:~/catkin_ws/src/rqt_common_plugins$ rosrun rqt_top rqt_top &
[2] 31013
130s@130s-serval:~/catkin_ws/src/rqt_common_plugins$ Traceback (most
recent call last):
File "/home/130s/catkin_ws/src/rqt_common_plugins/rqt_top/scripts/rqt_top",
line 34, in <module>
roslib.load_manifest(pkg)
File "/opt/ros/groovy/lib/python2.7/dist-packages/roslib/launcher.py",
line 62, in load_manifest
sys.path = _generate_python_path(package_name, _rospack) + sys.path
File "/opt/ros/groovy/lib/python2.7/dist-packages/roslib/launcher.py",
line 93, in _generate_python_path
m = rospack.get_manifest(pkg)
File "/usr/lib/pymodules/python2.7/rospkg/rospack.py", line 133, in
get_manifest
return self._load_manifest(name)
File "/usr/lib/pymodules/python2.7/rospkg/rospack.py", line 172, in
_load_manifest
retval = self._manifests[name] =
parse_manifest_file(self.get_path(name), self._manifest_name)
File "/usr/lib/pymodules/python2.7/rospkg/rospack.py", line 164, in get_path
raise ResourceNotFound(name, ros_paths=self._ros_paths)
rospkg.common.ResourceNotFound: rqt_top.
}}}

Isaac

Dirk Thomas

unread,
Jun 7, 2013, 2:07:08 AM6/7/13
to ros-s...@googlegroups.com
Isaac, I did not suggest so change the name your try to resolve as a package.
Only the argument "standalone" passed to rqt.

- Dirk

Isaac Isao Saito

unread,
Jun 7, 2013, 3:10:47 AM6/7/13
to ros-s...@googlegroups.com
Dirk, thanks for the care.
I finally managed to keep the pkg name rqt_top and get it run
standalone by the following script (though I'm not sure if this is
what you meant):

https://github.com/130s/rqt_top/blob/d43168e65cd61657fbb980e1ccb0344755664de2/scripts/rqt_top

Whole branch: https://github.com/130s/rqt_top/tree/embed_psutil

Isaac

Isaac Isao Saito

unread,
Jun 7, 2013, 4:30:12 AM6/7/13
to ros-s...@googlegroups.com
Dan,

back to psutil issue on Precise (or possibly earlier too), I've
managed to embed psutil into the pkg itself as a tentative workaround.
It's running both on Quantal and Precise that don't have python-psutil
installed.

Therefore, I moved the pkg into rqt_common_plugins metapackage.
https://github.com/ros-visualization/rqt_common_plugins/tree/groovy-devel/rqt_top

Things to consider:

1. If psutil's binary is installed somehow, that breaks functionality
on Precise https://github.com/ros-visualization/rqt_common_plugins/issues/89
2. How it's implemented is far from ideal.

For 1, dummest but easy workaround is to write document.

For 2, I need suggestion/patch. What I've done is I copy-pasted the
files that seemed necessary at least on Quantal (python-psutil 0.6.1).

Isaac

Dan Lazewatsky

unread,
Jun 10, 2013, 12:21:53 PM6/10/13
to ros-s...@googlegroups.com
Isaac -
Did you actually get it working with the embedded version of psutil? I just cloned the rqt_common_plugins metapackage and am having some issues. First, I see some errors with catkin_make. After a bunch of 

-- +++ processing catkin package: 'rqt_xxx'
-- ==> add_subdirectory(rqt_common_plugins/rqt_xxx)

I see:

-- Configuring incomplete, errors occurred!
Invoking "cmake" failed

Not sure if that's related, but when the plugin tries to import psutil, it can't find the _psutil_posix module. This module comes from a .so file which doesn't look like it is getting built. Unless I'm missing something, I think all of psutil needs to be there, and its install script needs to be run.

-Dan

Isaac Isao Saito

unread,
Jun 10, 2013, 1:37:32 PM6/10/13
to ros-s...@googlegroups.com
Dan,

sorry, I haven't pushed a couple of .so files (had to force-commit
since .so is in my .gitignore). Now you can pull the update and try to
build.

By the way,

> I see:
>
> -- Configuring incomplete, errors occurred!
> Invoking "cmake" failed

Did this cmake error look related to rqt_top? If not what is it about?

Isaac

Dan Lazewatsky

unread,
Jun 11, 2013, 4:27:52 PM6/11/13
to ros-s...@googlegroups.com
On Mon, Jun 10, 2013 at 12:37 PM, Isaac Isao Saito <13...@lateeye.net> wrote:
Dan,

sorry, I haven't pushed a couple of .so files (had to force-commit
since .so is in my .gitignore). Now you can pull the update and try to
build.

It's working now. I have a feeling though, that all of psutil should be included, and should be built during the package's build process. That may have been your eventual plan anyway, but I just wanted to mention it. 
 

By the way,

> I see:
>
> -- Configuring incomplete, errors occurred!
> Invoking "cmake" failed

Did this cmake error look related to rqt_top? If not what is it about?

It doesn't look related to rqt_top, but I couldn't get it to give me any more detailed info. Everything seems to work though.
Reply all
Reply to author
Forward
0 new messages