Hi Kristian,
Hi Will,
just a status update on my progress with the testing.
Basically I have now a working setup for my first goal,
which is to run a low-level test of the Synth code alone.
My test now invokes the calculation of 1 second of sound,
measures the running time (but throws the calculated samples away).
My setup is rather primitive and brute, but has the benefit of being simple:
The test is triggered via CLI; at this point existing sound is muted and
basically a regular shutdown is initiated, which causes the regular audio
backend to cease calculating sound. At this point, we can jump into a
test function, process it synchronously and print the result.
On return, the CLI thread and afterwards the application will shut down.
Please recall: what I have in mind is to build a test suite with a test runner.
Such a test runner would then for each test case launch Yoshimi, inject a CLI
script to load the necessary preset, and then trigger the test run, wait for
completion and evaluate the result and compare to a stored baseline.
The next step would be to dump calculated raw sound data into a file.
-- Hermann
PS: again, work-in-progress code is here
https://github.com/Ichthyostega/yoshimi.git
Branch "testing"
_______________________________________________
Yoshimi-devel mailing list
Yoshimi-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/yoshimi-devel
