I used to use SVN to store and version binary image files, and the
storage bloat of duplicating all of them in a checkout became unlivable.
I asked about fixing this problem in SVN at the time, and it went
nowhere. After trying numerous other tools and having nothing work
reliably I ended up writing a solution to my own problem myself
(https://github.com/robehickman/simple-http-file-sync).
My tool is very rudimentary, but in creating this I realized that the
problem of versioning binary files in an SVN-like way isn't very
difficult. Usefully diffing binary files isn't generally possible, and
at least for my own use even delta compressing them server side had no
value, as my own data set is primarily added to and the files are rarely
changed. Also hard drives are just so cheap now.
This problem has been in SVN forever and nobody seems that interested in
doing anything about it. Creating a new system is probably more productive.
