I am working on a large sized project that is about 1020 MB in size. This is because, apart from the code, we have other resources, like graphics, XML configurations, etc.
For binary data, SVN stores the entire copy on every new revision, not deltas. Thus - you get big repositories over time. .svn-base
stores the pristine state of the working copy - if you have a big working copy, you have the same size of pristine-data.
If there are many updates to your working copy, it might grow in size, even more than the checked out contents might suggest. Running svn cleanup
will free this space again.
So that svn revert
doesn't need to contact the server.
SVN actually stores another copy of the file locally. That's why the .svn
dirs sum up to be almost as the project code base itself.