Made performKeyToDegree much closer to inverse of performDegreeToKey

[triss]

See issue #1163

[triss]

I've made the patch now....

Unfortunately I've just noticed it has issues with some scales, specifically Scale.ajam.

Seems to work perfectly with the scales I'm more familiar with though. Scale.minor, major, dorian and all the other modes.

Would you mind showing me how to throw together a Unit test for this using your framework? I'm not likely to get the chance to fiddle with your unit testing library for a week or so.

[crucialfelix]

this example should be pretty easy to understand:

https://github.com/supercollider-quarks/CommonTests/blob/master/TestArray.sc

you just subclass UnitTest and write test_methods { }

that repository is out of date, its just an example because its easy to
link to github repos.

On Wed, Aug 6, 2014 at 4:02 PM, Tristan Strange notifications@github.com
wrote:

I've made the patch now....

Unfortunately I've just noticed it has issues with some scales,
specifically Scale.ajam which has 24 note per octave I think.

Seems to work perfectly for twelve notes per octave though. I can't fathom
why it's an issue when the number is bigger.

Would you mind showing me how to throw together a Unit test for this using
your framework? I'm not likely to get the chance to fiddle with your unit
testing library for a week or so.

—
Reply to this email directly or view it on GitHub
#1164 (comment)
.

[telephon]

any news on this? Do you have a simple test file?

[crucialfelix]

It sounds like this changes behavior, and this is a very musical method that is often used. We can't change it without unit tests that confirm that we won't have herds of angry musicians telling us that we embarrassed themselves in public (because they updated the night before a gig)

[jamshark70]

If this changes behavior, I'd share Felix's concern. I'd suggest taking the time to draw up a reasonably complete feature spec before merging anything. "Make it behave more like an inverse operation" is awfully vague and invites dispute... but if we have a spec, then there's something written down to explain to current and future users why it's this way and not another way.

SC development almost never begins with gathering requirements and speccing features... fine until something breaks.

It would be nice to do without the temporary array. This is basically a math op, which should avoid unnecessary load on the garbage collector.

[telephon]

comma?