Lifted straight out of an old post on the devel list:


<code>
running multiple Rosegarden builds: msg#00002

Subject: running multiple Rosegarden builds
List-id: <rosegarden-devel.lists.sourceforge.net>

I know that the subject of how to install and run different builds of
Rosegarden at once comes up occasionally.  It's not hard to do; here's
what I do.

Now that I actually have more than about ten gig of disk space in my
development machine, I'm finally keeping more than one Rosegarden
source tree active, corresponding to more than one CVS branch.  At the
moment in fact I have ten build trees, plus the version of Rosegarden
that came with the distro (Studio to Go!) still installed and runnable
in /usr.

What I do is keep the source trees under /opt/rosegarden-build as
separate directories (1.0, head, glasgow_pitchtracker etc) with each
one configured to a separate target directory
under /opt/rosegarden-install:

  ./configure --prefix=/opt/rosegarden-install/glasgow_pitchtracker
  scons configure prefix=/opt/rosegarden-install/head
  scons configure debug=0 prefix=/opt/rosegarden-install/head_NODEBUG

etc.

Then I have a script that you can find in scripts/cc_run that runs a
particular build.  I have the HEAD version of this aliased to the shell
command "run".  It takes an argument for the build name, or guesses it
from the name of the current directory (handy if you're in a build
directory).  The -d argument starts it in gdb.  Pretty simple.

  run glasgow_pitchtracker
  run -d head
  run -d # when in the head build directory
  run 1.0 --existingsequencer # extra args passed to the RG process
  rosegarden # runs my packaged install in /usr/bin

etc.

There's nothing clever or inventive about this and the script doesn't
really do anything.  As much as anything, this is just a reminder that
running multiple builds is very easy so long as you remember always to
set the install prefix to something a long way away from the default.


Chris
</code>