Other formats for regression tests
Could we use audio? For example, I remember staccato working at a point and then stopped.
What about GIF (or equivalent moving picture) files?
I don't know if it would take longer (compiler, contributors, etc), but it might be useful sometimes.
Comments
Staccato does work for me, it doesn't for you?
In reply to Staccato does work for me, it by Jojo-Schmitz
It does now, yes (sorry, I should have clarified).
It still serves as example though.
Audio files are huge, and binary, and comparing them is far from an exact science. I have to think that is pretty much out of the question. We do have playback tests based on *MIDI*, however. Not a ton of them, but that seems to the way forward if we feel the need for more tests. Hard enough to get more tests written for the core notation features, however...
In reply to Audio files are huge, and by Marc Sabatella
I feel what we need is more people writing tests and maintaining tests. The visual tests for example are almost not used. Only Werner checks them during his developments. So anything that cannot be automated is not a good candidate because it needs even more eye balls.