Publicly distributable test cases go in the test/cases directory. Put private and/or confidential cases in test/privatecases, which Git ignores.
Within those directories, place your database files into the data subdirectory.
Test cases are JSON files named [something].json with defined keys:
- cmd_args: the arguments passed to pgdbf to generate the test output. If this is a string, it will be passed as a single argument to pgdbf. If you're attempting to pass more than one argument, this is probably not what you want. In that case, set
cmd_argsto a list of string arguments.- With one argument:
"cmd_args": "mytable.dbf" - With multiple arguments:
"cmd_args": ["-m", "memofile.fpt", "datafile.dbf"]
- With one argument:
- head: a string to be matched against the start of the test output
- length: the expected length of the test output
- md5: the expected MD5 hex digest of the test output
- tail: a string to be matched against the start of the test output
Unknown keys are ignored.
Public test cases are loaded and executed in alphabetical order, followed by private test cases.
Run the runtests.py program from within the test/ directory to execute all unit tests. Use the -p argument to select a different pgdbf executable to test (such as a freshly built on at src/pgdbf). By default, all test cases in test/cases and test/privatecases are executed in alphabetical order. If you specify one or more test cases as command line arguments, only those test cases will be executed.
Before each test case is executed, runtests.py CDs into the directory containing that test case. This way, relative paths to data files (like data/my-database.dbf) work.
Example:
./runtests.py ../src/pgdbf
will test a freshly built-but-not-installed binary with all test cases in test/cases and test/privatecases.
./runtests.py my-test-case.json
will test the first pgdbf executable in $PATH with just my-test-case.json.
Of course. :-)