elwp@gmx.de writes:
I wonder how fast CVS is with so many small files.
I guess one has to spread the files over subdirs, e.g. as follows:
Nancy -> N/a/Nancy New York -> N/e/New York Nueremberg -> N/u/Nuremberg
or
Nancy -> N/Na/Nancy New York -> N/Ne/New York Nueremberg -> N/Nu/Nuremberg
And I doubt that the scripts that have to be develped to mirror the database to CVS on the server and to put the files back to the database on the client machine are much simpler than my programs, but you are welcome to write them, of course. :-)
But i expect the scripts to be trivial - as yours ;)
I'd simply write every successfull commit into a separate table (or an own database) from where you can dump to the CVS - all this could be done at the servers idle time.