-
Simon Marlow authored
Extend the support for running arbitrary commands as tests. - generic_command is now called run_command - the stdin for the command can be redirected in the same way as for a normal test. - the command's stdout and stderr outputs are compared against <testname>.stdout and <testname>.stderr as for a normal test - running a command and ignoring the output altogether (just the exit code is checked) is provided by calling run_command_ignore_output instead of run_command.
ecb2cccb