Tobias Reiher
Tobias Reiher
Removing Python 3.7 (#1145) did not solve the problem. Also, pytest-timeout did not stop the test after the defined duration. https://github.com/Componolit/RecordFlux/runs/8083850144?check_suite_focus=true
The issue still persists: - 2020-09-01: https://github.com/Componolit/RecordFlux/actions/runs/2974825503/jobs/4765968081 - 2020-09-07: https://github.com/Componolit/RecordFlux/actions/runs/3003269702/jobs/4821388388 - 2020-09-09: https://github.com/Componolit/RecordFlux/actions/runs/3018170749/jobs/4852823000 - 2020-09-15: https://github.com/Componolit/RecordFlux/actions/runs/3055954758/jobs/4929594509
One example would be the reserved field in IPv6: > Reserved 8-bit reserved field. Initialized to zero for > transmission; ignored on reception. There are also more complex cases like...
I think we should implement #700 where we decided to remove the need for an explicit definition of a final state.
There are several checks, where SPARK is able to show that these checks are unnecessary. The code generator does not yet contain the necessary information to come to the same...
It is now possible to integrate icontract with Hypothesis: [icontract-hypothesis](https://github.com/mristin/icontract-hypothesis). The most interesting part is the automatic inference of Hypothesis strategies based on contracts, i.e. test cases for a function...
Maybe. I would just keep this issue open until we find a solution. I think that is sufficient in this case.
It is not yet possible to set a sequence field by referring to a sequence field of another message. A workaround could be storing the sequence in a temporary variable.
The specification shouldn't be case-sensitive, but I also don't really like that `F2` and `f2` refers to the same thing. What about showing a warning when different notations are used...
That sounds even better. I think the verification should still be case-insensitive to avoid subsequent errors and therefore I have created a separate issue for enforcing "case preservation": #563.