Feature #6456
Updated by osmith about 1 month ago
I've been debugging why osmotestconfig.py fails with my patch https://gerrit.osmocom.org/c/osmo-sgsn/+/36845, when it runs in "make distcheck".
The purpose of the script is to run Osmocom programs with multiple configs from doc/example, write the config back, and run the program again, to see if that works and if it still runs as I understand it. This is very useful and works well as long as there are no errors.
But there are several hurdles that make it hard to figure out what went wrong in case the program returns an error, which IMHO could be improved in osmotestconfig.py:
* It does not clearly state _where_ a failure occurs, instead multiple checks with multiple configs are done and a counter gets increased for each failure. In the end there is only something like "ERRORS: 2", but it's not clear at all where they come from. So at least it should clearly print where it failed.
* A lot of unrelated output is printed with each run, such as output of ps | grep ...
* No logs of the Osmocom program that failed are shown so we can't see where they failed. I'm not sure if they are stored anywhere... in my opinion it would be useful to just stop on the first error and print the output of the Osmocom program.
* It seems to just catch and ignore python stack traces, and increasing the error counter, instead of printing the stack traces
