It is related to both. Conceptually, I would see the following. There would be a CVS resolve. The resolver would take the necessary credentials + the project name for the cvs repo. When I execute a resolve, it would make a call to CVS to retrieve each dependency artifact, then download it to the cache. I know this is possible as there are cvs web viewers out there that allow HTTP access to download from a project. For those with permissions to write to CVS, there would also be part of a script that would retrieve the dependencies from a user cache to somewhere on the user's file system, then execute an install from the file system to the cvs resolver. This part I am not sure could be done. Ill have to look more at how CVS and ssh could provide this. ________________________________
From: Michael Feinberg [mailto:[email protected]] Sent: Tue 4/27/2010 6:20 PM To: [email protected] Subject: RE: CVS Resolver Michael, We have written a repository resolver that pulls data out of the build system (which acts as a repository) and it severed us well. Is your question cvs- or resolver- related?
