On Dec 20, 2011, at 06:08, Mojca Miklavec wrote: > I'm writing a script (also used as a cron job on the server) that > should work independent of whether or not a checkout has already been > done or not. I would like to do "repeated sparse checkout" as > explained below, but I'm not sure how to do it properly. (I have some > kind of a workaround for this problem, but I would like to know if > there is a more elegant solution to this.) > > Repository has the following structure: > - project/trunk (I would like to have it checked out) > - project/branches/xxx (I don't want them) > - project/tags/xxx (I would only like to have the latest one, but it > doesn't bother me too much if I don't delete older ones) > > The first time I can do the following: > latest=`svn list $URL/tags | grep beta | tail -1 | tr -d '/'` > svn co --depth=empty $URL > svn up project/trunk > svn up --set-depth empty project/tags # temporary workaround for a > bug in svn > svn up project/tags/$latest > > I could use "svn up" from that moment on, but since I would like the > script to work even if nothing has been checked out yet, I would like > to keep "svn co" in the script. > > The problem is that "svn co --depth=empty $URL" will delete all the > contents next time when I call it.
You should only be calling "svn co" once per working copy. Once you have a working copy, you should "svn up" it. > Is there any way to prevent > deleting existing contents within the scope of command line arguments? > A workaround is to use something like > if [ ! -d "project" ]; then > svn co --depth=empty $URL > fi > but I would be really happy if there was some command like: > > "please checkout $URL, but no need to fetch any files yet, in > particular don't fetch 'branches'; on the other hand please don't > delete 'trunk' and 'tags' if already present" I don't think that exists as a built-in command as such. But probably playing with the depth settings you can get what you want. I am not very familiar with these settings however so I cannot advise.
