Russ Brown wrote:
On Fri, 03 Feb 2006 17:05:26 +0100
Christian Boos <[EMAIL PROTECTED]> wrote:

Russ Brown wrote:
On Fri, 03 Feb 2006 15:59:33 +0100
Christian Boos <[EMAIL PROTECTED]> wrote:
...
Simple workflows can be configured directly in the trac.ini file.
It is possible to write more complex workflows as plugins.

See http://projects.edgewall.com/trac/wiki/WorkFlow

Thanks, I'll take a look. Does this require any changes to the trac
database that would make it incompatible with the trunk codebase?
I'm wondering if I'll be able to run this in parallel with our
existing installation for testing and evaluation purposes
There's no database upgrade.

However, if you add new resolutions or status *values*, and assign
them to some tickets, I'm not sure if the "trunk" can cope with
these ... So it's probably safer to experiment with a full copy of
your production environment.


I've just tried to do this but have come up against a problem that has
stumped me a little, though I'm sure there's a simple solution (this
will just be down to a little unfamiliarity with how the Python package
system works).

How do I run two different codebases of trac on the same machine at the
same time? I've copied the database (no problem), and I've set up a
different virtual host (again, no problem) but in the vhost
configuration the only way in which I can point apache at the trac
codebase is via the 'PythonHandler' directive, which from what I can
tell looks in /usr/lib/pythonx.x/site-packages, and since I already
have an installation of trac there that I don't want to mess with, I'm
not sure of how to go about getting the second instance working
independently, since it includes other trac packages via absolute
package references (trac.xxx), so simply putting this other instance of
trac into, say, trac2 doesn't work.


I'm certain that I probably just have to do something like set up
another python package path for this new virtual host to look at, but
how this is done isn't obvious to me.

Any hints

I recommend using the standalone tracd for doing your tests. It's simpler to setup, you can restart it anytime you want, you don't risk to disturb your production server.

However, if you insist on doing this in another virtual host, you should use the
PythonInterpreter  directive, see for example:

http://projects.edgewall.com/trac/ticket/2713

which suggests adding documentation for this topic (btw, documentation can always be
contributed directly, as it's in a Wiki ...)

-- Christian
_______________________________________________
Trac mailing list
[email protected]
http://lists.edgewall.com/mailman/listinfo/trac

Reply via email to