I'm running a project which includes 2 instances of D2RQ, which is a (third-party) Tomcat application that I'm compiling myself with ant
using the supplied build.xml
.
Currently, I have a copy of the D2RQ code outside of my main development instances; when I need to update either of the D2RQ instances, I change the web.xml
file (just to set the name of the relevant .ttl
mapping file), compile the code, and deploy to Tomcat.
Are there any industry best practices as to how this should be managed in SVN? Other members of my team (the majority of whom don't currently use D2RQ) would like a straightforward way of storing the code. There are several options (in all cases I would obviously store both .ttl
files in the repo):
- store a complete copy of D2RQ code for each instance needed in SVN
- store only 1 copy of the the D2RQ instance, with very good documentation so that any of the other developers in my team are able to update if necessary (i.e. so they know they have to edit the
web.xml
before compiling D2RQ, etc) - store only the
web.xml
file (or, 1 copy for each of the D2RQ instances) in SVN. Perhaps this would lend itself to storing the remainder of the D2RQ code in a vendor branch? - modify the D2RQ
web.xml
to the extent that I can create a build script (either in maven or ant) that will always build both D2RQ instances when the overall project is built, rather than needing to compile them separately from each other (and from the overall application). The downside to this is that if a new version of D2RQ which modifies theweb.xml
, I'd have slightly more work to update my build script.