10.4 Finding the Shallowest Instance

Suppose you maintain local copies of sources from various projects, each with their own choice of directory organisation and source code management (SCM) tool. You need to periodically synchronize each project with its upstream tree. As the number local repositories grows, so does the work involved in maintaining synchronization. SCM utilities typically create some sort of administrative directory: .svn for Subversion, CVS for CVS, and so on. These directories can be used as a key to search for the bases of the project source trees. Suppose we have the following directory structure:

repo/project1/CVS
repo/gnu/project2/.svn
repo/gnu/project3/.svn
repo/gnu/project3/src/.svn
repo/gnu/project3/doc/.svn
repo/project4/.git

One would expect to update each of the projectX directories, but not their subdirectories (src, doc, etc.). To locate the project roots, we would need to find the least deeply nested directories containing an SCM-related subdirectory. The following command discovers those roots efficiently. It is efficient because it avoids searching subdirectories inside projects whose SCM directory we already found.

find repo/ \
-exec test -d {}/.svn \; -or \
-exec test -d {}/.git \; -or \
-exec test -d {}/CVS \; -print -prune

In this example, test is used to tell if we are currently examining a directory which appears to the a project’s root directory (because it has an SCM subdirectory). When we find a project root, there is no need to search inside it, and -prune makes sure that we descend no further.

For large, complex trees like the Linux kernel, this will prevent searching a large portion of the structure, saving a good deal of time.