zdb -DD
shows exactly the amount of the sources (~512MB) with a reference count of 2, for about 1/2 of all ordinary files in /usr/src/12-STABLE, that would fit.The data volume of aPretty sure SVN does one transfer to fetch the file into the pristine metadata area, and then copies that to the 'working' directory.
svn checkout
unfortunately points in the other direction, which disappoints me since it violates one of the basic principles of CS: not to add redundancy where it brings no benefit. There are many people with limited bandwidth and data volume, not only in the emerging countries. svn fetch
might be the way to go. You get a copy of the files, but no way to query the change history (or even figure out what exact version of the repository you have a copy of), and if you make local changes there's no way to easily see what change you made or commit that change (but only a select group are allow to do that anyway).OK, for just browsing/reading the sources that's an option. But to work on the sources, I need all that meta-data. Maybe the doubled volume comes fro this history meta-data? I'll see what the svn users tell me in reply to my bug report.For those who have a limited bandwidth, and how don't really care about the 'history' of the files,svn fetch
might be the way to go. You get a copy of the files, but no way to query the change history (or even figure out what exact version of the repository you have a copy of), and if you make local changes there's no way to easily see what change you made or commit that change (but only a select group are allow to do that anyway).