Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
Hi Folks! I have a basic question about setting up repositories. This has likely been covered before but I can't find exactly what I am looking for in the archives, so here goes. I am wondering how to best setup our repositories. We are new to CVS. Repository setup seems simple enough at first blush, but these are not the usual programming projects (SAS code, in this case)....please bear with me as I give some background: We have many Clients, each with several Projects and a handful of offices that will house CVS servers. To keep it simple, we'll say one client deals with only one office and all their projects are on a single CVS server. Our working file structures (where the sandboxes live) are similar to the following, with separate Development and Production folders within a project: \\FileServer\Client1\C1Proj1\Dev\Code <\\FileServer\Client1\C1Proj1\Dev\Code> \\FileServer\Client1\C1Proj1\Prod\Code <\\FileServer\Client1\C1Proj1\Prod\Code> \\FileServer\Client1\C1Proj2\Dev\Code <\\FileServer\Client1\C1Proj2\Dev\Code> \\FileServer\Client1\C1Proj2\Prod\Code <\\FileServer\Client1\C1Proj2\Prod\Code> \\FileServer\Client2\C2Proj1\Dev\Code <\\FileServer\Client2\C2Proj1\Dev\Code> \\FileServer\Client2\C2Proj1\Prod\Code <\\FileServer\Client2\C2Proj1\Prod\Code> ..and so on. My current thinking is to have a repository for each Client and define modules that relate to each Project within the repository. The repository would look like this: \\CVSRepServer\CVSRep\Client1\CVSROOT <\\CVSRepServer\CVSRep\Client1\CVSROOT> \\CVSRepServer\CVSRep\Client1\C1Proj1\Code <\\CVSRepServer\CVSRep\Client1\C1Proj1\Code> \\CVSRepServer\CVSRep\Client1\C1Proj12\Code <\\CVSRepServer\CVSRep\Client1\C1Proj12\Code> \\CVSRepServer\CVSRep\Client2\C2Proj12\Code <\\CVSRepServer\CVSRep\Client2\C2Proj12\Code> I would set up regular modules in the modules file that would look like: C1Proj1 -d Code C1Proj1/Code C1Proj2 -d Code C1Proj2/Code C2Proj1 -d Code C2Proj1/Code This would allow our programmers to refer to the project by project name, even though they are not checking out a folder by that name (they will retrieve the folder /Code down into the /Dev or /Prod folder in the sandbox). For example: > cvs checkout C1Proj1 Does this make sense? Defining a repository for each individual Project would result in a very large number of repositories! We have tens of clients, and a single client may have several projects. Are there maintenance issues and limitations of this approach that I should be aware of? We are starting with the basic CVS functionality then will likely use additional tools like CVSMailer and ViewCVS at a later time. Thanks in advance! Tim Tim Williams SAS Systems Administrator PRA International Charlottesville, VA USA direct 434.951.3504 main 434.951.3000 fax 434.951.3001 mobile 804.338.1063 WilliamsTim at praintl.com