On 2010-06-02 07:10, Brad Hards wrote:
> I've been using buildbot to test openchange (http://www.openchange.org) for a
> while, and it mostly works pretty well for us for client-side testing (against
> a running Exchange server).
> We're about to move into testing of the server side of openchange. Essentially
> I want to build the source, configure the server, start the server, wait a
> little while (or for a certain debug output line) and then run my client
> tests. When all the tests are done, kill the server and wait for the next
> Any suggestions, tips or tricks for how to best approach this?
Historically the answers to similar requests have been that BB is not a
build system and thus such things lies outside it's reponsibilities.
Also, having a build process that relies heavily on a single server (the
bb master) and cannot be used outside that system is not something I'd
I'd vote for having a single build that has it's own scripts/tools to
start/stop servers or clients on different computers.
Nevertheless, it might be possible for you to get BB to coordinate
different tasks on different servers, and others have suggested ways to
do that (and there will surely be more help if you get into specifics).
But, when/if doing so, you should bear in mind that it has not been a
design goal of buildbot to go down that route.
Now, there might be ways to extend buildbot in that direction, but I'm
not sure on how or why. Dustin (and the mail-list archives) knows more
about that than me. ;)