hanjoosten
commented
5 years ago A way that this could work is to create this code as an .adl script in the test environment. Then at least we know that the code is OK wrt the parser, typechecker and generator.
Then there are two more things to solve:
- [ ] Ensure that the contents of that .adl file and the contents shown in the markdown are the same
- [ ] Implement a testing strategy for the gui of the prototype itself. This is a project in its own, I fear.
In the documentation, a little script is provided to verify the workings of Ampersand after the installation process is complete. However, it happens that new Ampersands are published on the master branch that do not produce working prototypes for the code provided in the documentation. We need a procedure that allows a user to ALWAYS establish whether or not his installation works. How to do that?