rhl-bthr
commented
4 years ago After #34, I felt all but qpipe were straightforward and are not bringing good coverage. Should we instead write our own test cases as we have done in the past?
To build a good testing module, we need to cover all type of DFA transition combinations on the parser, maybe also keeping in mind the optimizations done by p4c.
The type is dependent on - input state, output state, and transition.
An input state can be one or more of -
- Start state
- in-built / non in-built
- can / can-not do some processing before the transition (we know that can-not is buggy)
An output state can be one or more of -
- Accept state
- in-built / non-inbuilt
A transition can be -
- based on multiple fields from different headers (not supported)
- one field, different header
- one field, same header
- multiple fields, same header
- always
- default
I don't think these should be too many test cases to build, but I'm not sure.
The biggest impediment in formally announcing the tool is the lack of confidence in its correctness. The tool has not been tested extensively for all the backends that it supports with a variety of P4 programs. We can find more public domain P4 programs here for testing.
Possibly also need to craft custom P4 programs to test handling of corner cases.