We need to be able to determine which files were deleted between two
revisions. (In this case it is stored procedure files and we want to be
able to figure out which stored procedures can be dropped from the
database.) However, the method we use currently is way too expensive.
Currently we do an "svn update" of a working directory to an old
revision. We then do "svn update" again to the current revision and we
can parse the results to look for those that start with "D" to determine
which files have been deleted. We then have to parse it futher to only
find the paths that we know to contain stored procedures.
Unfortunantly, each "svn update" is massive and takes several minutes to
complete which is kind of a pain.
Since we do not need the actual contents it would be nice if we could
calulate this all on the server side and just get a list of files
changed. We tried using "svn diff -r OLDREV:NEWREV http://svn/trunk".
However, the results of "svn diff" are less than ideal - it would be
perfect if it supported the notion of the GNU diff --brief option.
Rather the only way to use "snv diff" would be to parse the diff result
for each file return and see which one return all "-"'s (meaning all
lines were deleted.)
Perhaps there is another way to do this?
Ray
Received on Wed Jun 30 03:47:11 2004