I suspect most of the tests in develop test the wrong qasm.
The tests that also execute the backend, must test a qasm that is generated just before/after code generation.
A pass generating this was not present until now, so I added it.
It generates ’progname’_last.qasm.
See the new tests/test_multi_core.py in the multi_core branch for an example of its use.
I may be wrong with names etc. but the idea and the problem are clear, I hope.
So this means that we have to regenerate golden qasm outputs
when the qasm just before/after code generation is meant.
The current _scheduled.qasm is just before entering code generation or entering QX.
See the multi-core branch for the necessary changes.
I didn’t update the other tests.
The multi-core branch is ok, all tests pass there including the multi-core test itself.
I suspect most of the tests in develop test the wrong qasm. The tests that also execute the backend, must test a qasm that is generated just before/after code generation. A pass generating this was not present until now, so I added it. It generates ’progname’_last.qasm.
See the new tests/test_multi_core.py in the multi_core branch for an example of its use. I may be wrong with names etc. but the idea and the problem are clear, I hope.
So this means that we have to regenerate golden qasm outputs when the qasm just before/after code generation is meant. The current _scheduled.qasm is just before entering code generation or entering QX.
See the multi-core branch for the necessary changes. I didn’t update the other tests.
The multi-core branch is ok, all tests pass there including the multi-core test itself.