Hi Jim, Have you used svn export? That is how I deploy from the repository to the development environment. The export command will push out only the files in your project and leave behind all of the svn meta directories and files. It will also omit any files that are not under revision control.
In addtion to your repository trunk you should organize your repository further by having branches and tags directories. A branch allows you to have 2 or more independent working revisions. This allows you to try out new features without disturbing your main line of development. A tag would be used to allow you to "roll back" to any given point in time (you wouldn't be using the rollback command, per se, just exporting from a different, earlier tag). Tags are often used to denote a particular release. On Tue, Feb 5, 2008 at 8:49 AM, Jim Hefferon <[EMAIL PROTECTED]> wrote: > Hello, > > Can I ask a best-practices question? I feel that what I'm doing is not > enough but I'm not sure what is the Right Thing. I know many folks on > this list do development of various kinds. > > I do a lot of web work. I have all my material, including dB table > generation and stocking, in a versioning system (it happens to be svn). > When I am good, I try to write unit tests, functional tests, and system > tests. I develop on a setup that is a different computer from but much > like the production system. > > The part that I don't get is how people bring the material out of the > svn archive. That is, suppose I get a bug report and I want to get all > the materials for version xxx onto my development machine (or to roll > back my production machine to version yyy). Is there a standard way to > do that, or at least a library of often-used and debugged routines? (I > work in Python.) > > I can't just make a lot of soft links from the svn tree to where I want > things to go for a number of reasons. For instance, one is that > Apache's suEXEC refuses soft links. Another is that the soft link tree > changes over time so I couldn't roll back to prior versions. Still > another is that I need to massage some of the files (say, doing a sed to > change the permissions's owner on some dB tables). > > I wrote a program that does the job for me but it is specialized to my > projects, obviously. It has a lot of subroutines that copy to a > directory all files whose names match a regular expression, for example, > and then I call those subroutines lots of times on the exact structure > of my tree. That seems suspiciously like I ought to be using a library, > where someone has carefully tested the routines (what if one of the > files is a soft link that points nowhere? that got me last week). Is > there such a thing but I missed it? > > Alternatively, maybe I'm doing everything all wrong (it has happened > before :-) ). I'd appreciate any tips. > > Thanks, > Jim > -- Robert Riggen Zend Certified Web Developer Essex Junction, VT 05452 (802) 578-6719 [EMAIL PROTECTED]
