repository - How can we improve our use of SVN? -

repository - How can we improve our use of SVN? -

i started in new job 2 years ago , svn repository process existing. process came because there never much thought set best forward-looking way leverage svn manage codebase. we're in situation need create alter before grow much more. we're looking improve way utilize svn suited our needs.

we have base of operations framework of our client code sits on top of managed in 1 repository. we have repository manage of our custom perl modules , repository manage of our skins base of operations framework. then have repositories each of our customers' business units.

there code across these repositories overlaps in /var/www, /usr/local, /opt directories , subdirectories.

up we've managed doing our changes in working area of server (/var/www, /opt, /usr) moving our changes , directory construction on working re-create in our home directory can perform our svn operations on files.

i've done experimentation checking out repositories on root of server have sudo svn operations. other concern if switch repository-for-customer-a , have create changes base of operations framework changes beingness tracked? included in svn stat of repository-for-customer-a? or if svn commit server commit files both repositories @ once?

managing svn way totally inefficient , liability because there files missed essential fix/change , when comes merging code trunk aren't making utilize of svn branching/merging it's lot more work our gatekeepers. there lot of overhead in our processes can eliminated here if can figure out how leverage svn way meant used.

i've come stack overflow because respect format , community here. advice appreciated!

your workflow improve lot if distinguish between source code , installed program. code should flow source location (that has no production relevance) installed location, in:

http://your-svn-server | checkout v /home/local-checkout | create install v /var/www /opt /usr

make sure code never moves in hierarchy, except through svn commit. especially, don't monkey-patch /var/www , re-create files checked-out directory. you'll never re-create all relevant changes.

this sounds lot of work, isn't. make install script takes little time run, when no compiler involved. on upside, defined deployment procedure helps in future when want migrate different box or install on new machine.

svn repository development-process