A testing script for archive.
It tests the following commands:
archive stat
archive ls archive ls -l
archive mkdir dir1 archive mkdir dir1/dir2 (dir1 doesn't exist) archive mkdir -p dir1/dir2 (dir1 doesn't exist) archive mkdir -m MODE dir1 archive mkdir dir1 dir2 archive mkdir -p -m MODE dir1/dir2 dir1/dir3 (dir1 doesn't exist)
All the puts use -C to put them into the test directory. archive put -C path file1 archive put -C path file1 file2 archive put -C path dir1/file1 (dir1 doesn't exist) archive put -C path -p dir1/file1 (dir1 doesn't exist) archive put -C path -t tarfile.tar file1 file2 archive put -C path -M file1 (fails for local.pm.test !) archive put -C path fil* (test globbing) archive put -C path -t tarfile.tar -M fil*
It does not test:
-retry options -asf host options -C options -n (not implemented in archive) -f force options (implemented?)
There is an inherent difficulty in testing archive ls along with the other commands. We need to use those commands to create somthing for archive ls to look at (to test archive ls), and we need to use archive ls to look at the results of archive mkdir and put, to test mkdir and put commands. We essentially ignore this difficulty. It would take a fairly special circumstance to make everything look as if it worked when it didn't. The problems are more likely to arise when trying to diagnose a failure of TestArchive. Is the failure due to a failure in the archive get,put,etc or due to a failure of archive ls? The user will have to resort to non-pst commands to sort that out.