If it turns out that the other implementations arn't close enough to what is needed I'd like to know whether there are features which should be implemented (in the hack program that is). Does it need to treat stdout/stderr in a different way (we can capture both independently or let them through). The example in the TODO info imples that it was for logging stderr with stdout going direct to the normal fd but I might be mis-reading that.
If both stdout/stderr should be logged do we need to log them with different severity/errors? Would it be a good idea to have the code examine an environment variable to select options (to save having to build long command-line arguments for the wrapper)? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
