Someone has a problem with RTT master and rock when ros is present. There is
indeed a line that blindly tries to load catkin.
Can we keep the pre-catkin behaviour of enabling catkin stuff only if a certain
CMake variable has been explicitly set ?
Sylvain
---------- Forwarded Message ----------
Subject: [Rock-dev] Unable to build rtt
Date: Wednesday, January 22, 2014, 10:42:32 AM
From: Malgorzata Goldhoorn <malgorzata [dot] goldhoorn [..] ...>
To: rock-dev [..] ...
Hi,
since yesterday I cannot compile RTT anymore. I have installed ROS hydro
via package system on my Ubuntu 13.04. For me, it seems that catkin
causes problems during the rock-build (see attached file).
Cheers,
Malgosia
Fwd: [Rock-dev] Unable to build rtt
On 01/23/2014 02:47 PM, Jonathan Bohren wrote:
> Then that's strange because I have catkin in my CMAKE_PREXIX_PATH, but
> when I go to the rtt directory and run the following, it builds without
> error
Do you also have the PYTHONPATH properly setup to point to the ROS stuff
? (I am asking since the log tells about python things not being available)
> So there's got to be something else going on here. What's different
> when someone builds it with the rock-ros bridge?
Nothing, really. The full environment can be seen within the build log.
Sylvain
Fwd: [Rock-dev] Unable to build rtt
On Wed, Jan 22, 2014 at 5:01 AM, Sylvain Joyeux <sylvain [dot] joyeux [..] ...>wrote:
> Someone has a problem with RTT master and rock when ros is present. There
> is
> indeed a line that blindly tries to load catkin.
>
Which line are you talking about? If it's the following, this shouldn't
cause any problems if you're building without catkin.
https://github.com/orocos-toolchain/rtt/blob/6b7d3c53aa9d2de1e70455ecf8a...
> Can we keep the pre-catkin behaviour of enabling catkin stuff only if a
> certain
> CMake variable has been explicitly set ?
>
The attached file looks like a build log for RTT, what exactly was the user
trying to compile?
-j
Fwd: [Rock-dev] Unable to build rtt
On Wednesday, January 22, 2014 11:43:22 AM Jonathan Bohren wrote:
> On Wed, Jan 22, 2014 at 5:01 AM, Sylvain Joyeux
<sylvain [dot] joyeux [..] ...>wrote:
> > Someone has a problem with RTT master and rock when ros is present. There
> > is
> > indeed a line that blindly tries to load catkin.
>
> Which line are you talking about? If it's the following, this shouldn't
> cause any problems if you're building without catkin.
> https://github.com/orocos-toolchain/rtt/blob/6b7d3c53aa9d2de1e70455ecf8a...
> 2decf0a8a/CMakeLists.txt#L183
Nope. This causes no problem if you are building without catkin being in the
CMAKE_PREFIX_PATH. It seems that if it is *and* you don't build with catkin
(which happens if you use Rock and the Rock/ROS bridge) it fails.
> > Can we keep the pre-catkin behaviour of enabling catkin stuff only if a
> > certain
> > CMake variable has been explicitly set ?
> The attached file looks like a build log for RTT, what exactly was the user
> trying to compile?
Well ... RTT.