On 8/7/2015 3:31 PM, Matt Olson wrote:
Does anyone use SVN (Subversion) to store their RPG Source?
I've been trying Git in... solo mode, for lack of a better term. I use
Git to be able to roll back changes, to know what changes were made
when, and what bits and pieces went into a given change. I use RDi, and
Git more or less plugs in. Because of the way Git tracks changes, it's
easiest to create an iProject and let Git track the working iProjects
directory, where the source members exist as stream files.
Because my colleagues use classic source members, I must as well, so the
'source of record' is still in source files and source members on the
IBM side. My various versions are kept in Git on my PC and also on the
IFS, where I push the changes should a colleague decide to join me.
What are the limitations / gotchas?
SVN isn't a turnkey solution to software lifecycle management on IBM i.
Let me qualify that. I am an RPG programmer. I deal with ILE RPG
source members which compile to *PGM and *SRVPGM objects. Git or SVN
will handle versioning the source members well enough, but it won't do
anything at all to help you keep track of
a) whether you remembered to compile the source
b) whether the source actually compiled without errors
c) whether this particular object was deployed to the production
environment with the rest of the related objects (CPF4131 - Level check).
Objects, plural, because there can be display files, data queues, job
descriptions and many other objects - some of which haven't any
associated source member for SVN or GIT to operate on.
I think there's a Subversion port that runs on IBM i, the idea being
that you check out and in via PDM user commands. If you can't install
that for some reason, you'll need a way to get the source member into a
stream file that Subversion natively understands. Although, you might
be able to map a directory and have SVN deal with the files in IFS
notation: /QSYS.LIB/BUCK.LIB/QRPGLESRC.FILE/DEMODATES.MBR
As an Amazon Associate we earn from qualifying purchases.