Fix #687: Include MIDI swing script in default distribution

Fix #687: Include MIDI swing script in default distribution

This commit is based on Johannes Rohrer’s contribution,
which in turn stemmed in part from other contributor’s work:
https://sourceforge.net/p/testlilyissues/issues/687/?page=1&limit=25#c2cf

(I haven’t contacted Johannes to ask for authorization; I assume
we can safely distribute it under GPL like the rest of LilyPond.
Full credit is given in the source code, and his many comments
have all been preserved.)

Documentation and regtest added, largely inspired by Peter Chubb’s
articulate.ly script.

[thomasmorley651, 2019-03-10 20:02:07]

Hi Valentin,

up to now I didn't test, but read the code only.
Several small suggestions:

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly
File ly/swing.ly (right):

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode99
ly/swing.ly:99: (fold (lambda (x y) (and x y)) #t (map rational? lst))))
Why not
(every rational? lst)
?

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode102
ly/swing.ly:102: (if (< n 1) '() (append (integers-from-1-to (1- n)) (list n))))
Why not
(iota n 1 1)
?

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode112
ly/swing.ly:112: (ly:duration? (ly:music-property music 'duration))))
Mmh, strange check

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode119
ly/swing.ly:119: (factr (ly:make-moment (+ 1 (/ delta dur)))))
Why not
(1+ (/ delta dur))
?

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode163
ly/swing.ly:163: (#t
(else

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode229
ly/swing.ly:229: (integers-from-1-to (- n 1))))
(1- n)

[Valentin Villenave, 2019-03-10 22:58:05]

Thanks!  I actually hadn’t reviewed Johannes’ code in detail, but you make good
points.

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly
File ly/swing.ly (right):

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode99
ly/swing.ly:99: (fold (lambda (x y) (and x y)) #t (map rational? lst))))
On 2019/03/10 20:02:06, thomasmorley651 wrote:
> Why not
> (every rational? lst)
> ?

Sure.

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode102
ly/swing.ly:102: (if (< n 1) '() (append (integers-from-1-to (1- n)) (list n))))
On 2019/03/10 20:02:06, thomasmorley651 wrote:
> Why not
> (iota n 1 1)
> ?

Since this function is only called once, and with (- n 1), we don’t even need
anything other than iota.

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode112
ly/swing.ly:112: (ly:duration? (ly:music-property music 'duration))))
On 2019/03/10 20:02:06, thomasmorley651 wrote:
> Mmh, strange check

Indeed. But I can’t seem to make it work without it.

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode119
ly/swing.ly:119: (factr (ly:make-moment (+ 1 (/ delta dur)))))
On 2019/03/10 20:02:06, thomasmorley651 wrote:
> Why not
> (1+ (/ delta dur))
> ?

Done.

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode163
ly/swing.ly:163: (#t
On 2019/03/10 20:02:06, thomasmorley651 wrote:
> (else

Done.

https://codereview.appspot.com/572520044/diff/574520044/ly/swing.ly#newcode229
ly/swing.ly:229: (integers-from-1-to (- n 1))))
On 2019/03/10 20:02:06, thomasmorley651 wrote:
> (1- n)

No longer needed if using (iota n 1), anyway.

[Valentin Villenave, 2019-03-10 22:58:20]

More concise code with srfi-1 procedures

[dak, 2019-03-12 22:15:48]

https://codereview.appspot.com/572520044/diff/544550044/Documentation/notatio...
File Documentation/notation/input.itely (right):

https://codereview.appspot.com/572520044/diff/544550044/Documentation/notatio...
Documentation/notation/input.itely:246: \paper @{ @dots{} @}
I don't understand these diff lines.  Any idea where they are from?

https://codereview.appspot.com/572520044/diff/544550044/Documentation/notatio...
Documentation/notation/input.itely:3812: Grace notes are ignored and simply left
unaffected.
What's with tuplets?  When following actual practice, most written tuplets
should be rendered exactly.  Things get really murky when they span only part of
a usually swung unit.

Of course, I have no idea what swing.ly actually does with them currently.  It
should be documented what currently is there and put a warning when the code
does not yet make it do what would likely be wanted.

https://codereview.appspot.com/572520044/diff/544550044/ly/swing.ly
File ly/swing.ly (right):

https://codereview.appspot.com/572520044/diff/544550044/ly/swing.ly#newcode97
ly/swing.ly:97: #(define (rational-list? lst)
I don't think this does what you think it does.  Every number (except complex
one) counts as "rational" in Scheme terms.  Including (sin 2.2).

[Valentin Villenave, 2019-03-13 08:51:34]

On 2019/03/12 22:15:48, dak wrote:
>
https://codereview.appspot.com/572520044/diff/544550044/Documentation/notatio...
> Documentation/notation/input.itely:246: \paper @{ @dots{} @}
> I don't understand these diff lines.  Any idea where they are from?

I thought they were because of tabs in the original file, but on second look it
doesn’t seem to be the case.  Removed.

>
https://codereview.appspot.com/572520044/diff/544550044/Documentation/notatio...
> Documentation/notation/input.itely:3812: Grace notes are ignored and simply
left
> unaffected.
> What's with tuplets?

Left unaffected as well.  Added to the doc section.

> https://codereview.appspot.com/572520044/diff/544550044/ly/swing.ly#newcode97
> ly/swing.ly:97: #(define (rational-list? lst)
> I don't think this does what you think it does.  Every number (except complex
> one) counts as "rational" in Scheme terms.  Including (sin 2.2).

Yeah, that originated from Johannes’ code but we don’t really need that much
type checking.  I’m gonna go with (pair?) in the music-function arg signature,
much like what we already do with compoundMeter.

V.

[Valentin Villenave, 2019-03-13 08:51:52]

Further simplification; add CHANGES item.

[Valentin Villenave, 2019-03-18 18:59:11]

On 2019/03/13 08:51:52, Valentin Villenave wrote:
> Further simplification; add CHANGES item.

Pushed as
https://git.savannah.gnu.org/cgit/lilypond.git/commit/?id=464cf0d8111976c63fe...
Thanks!