This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

------------------------------------------------

To use directly from the package root directory, you must first
create the auto-generated files and compile any shared libraries
inplace via './setup.py build_ext --inplace'.

  Then you can run the tests via:
./testserver.py SERVER:/PATH [options] [flags] [testcode]

Useful options are:
all : run all standard tests (defaults to running none)
--rundeps : run dependencies even if not specifically requested on command line
--maketree : will (re)create the test tree. Ie it will create and populate
	     the directories /PATH/tmp and /PATH/tree
-o, --outfile : will save results into given file.
		results can then be examined using ./showresults.py FILE
--showflags : will show a list of available flags
-v : show test status as they are being run, instead of summary at end

------------------------------------------------

Usage:
Each test has a unique code.  It will also belong to one or more
groups which are indicated by flags. Codes and flags can be put on the
command line as is, or preceded by 'no'.  After options and the
SERVER:/PATH argument are removed, each test is compared to the
remaining args left to right.  Its status starts as DONTRUN, then is modified 
in turn by each argument code or flag that it matches

example1:
./testserver.py tiffin:/pynfstest --maketree link noblock getattr BLOCK1

This will (re)create the test tree, then run:
1. the test with code BLOCK1
2. every test that has the 'getattr' flag
3. every test that has the 'link' flag but does not have the 'block' flag
Before running each test, it will try to run any tests it depends on.  However,
the dependent tests must pass the same checks on flags and codes.

example2:
./testserver.py tiffin:/pynfstest --rundep link getattr noblock 

This will use the previously created test tree, and run:
1. every test that has the 'getattr' flag but does not have the 'block' flag
2. every test that has the 'link' flag but does not have the 'block' flag
Before running each test, it will make sure any test dependencies are run
first, regardless of their flags.

example3:
./testserver.py tiffin:/pynfstest --force all noblock getattr 

This will use the previously created test tree, and run:
1. every test that has the 'getattr' flag
2. every other test in the standard suite that does not have the 'block' flag
Each test will be run, regardless of its dependencies.

------------------------------------------------

The test suite will first check that the tmp directory is empty, and if not, 
try to clean it up.  This can be skipped by using --noinit.

Then 2 null calls are sent.

Then each test is run, each separated by a null call.

Finally, 2 null calls are sent, and the it tries to clean tmp.  This step can
be skipped using --nocleanup.
 
------------------------------------------------

Problems:

Get a warning message along the lines of:
RuntimeWarning: Python C API version mismatch for module gssapi: This Python has API version 1012, module gssapi has version 1011.

This imples the file lib/rpc/rpcsec/gssapi.so was compiled with header files
from a different version of python than is curently being run.
Delete gssapi.so and try `python2.3 setup.py build_ext --inplace' (or whatever
version (>=2.3) that you use).

If this gives an error similar to:
error: invalid Python installation: unable to open /usr/lib/python2.3/config/Makefile (No such file or directory)
and you installed from python-*.rpm, you need to also install python-devel-*.rpm.