Minor Grumph.

[swestrup]

As part of the work crisis, I have several tasks to perform this weekend. Most are no big deal, like running some (semi-)automated tests and updating some documentation. What is not so easy is the part where I'm supposed to 'debug' any anomalies in the test results. Lets see if I can enumerate the reasons for the difficulty:

1. I have never run these tests before, so I don't know what good output looks like.
2. I have never used the software in a state when it wasn't broken, so I don't know what ITS correct behavior is.
3. I have only a sketchy knowledge of the various parts of the system and their interactions, making it hard to diagnose a problem.
4. None of the diagnostic parts of the program tools I have are documented.
5. We have no troubleshooting guide.

So, I am running tests I do not understand, on a platform with unresolved issues that have been known to break scripts in the past, with a system in a configuration that I have never seen before, without most of the documentation needed to understand what my command options are, and I am expected to debug any problems.

So 10 out of 10 for confidence in me to do the job, but minus several million for reasonable expectations. All I can really do is run all the various tests I can think of and post the results to the company engineers mailing list, in the hopes that someone else can shed some light on any irregularities that show up, or at least provide vague indications of where I can look and what tools are available for the diagnosis.