Change \retrograde to deal with ties and repeat chords/notes

Change \retrograde to deal with ties and repeat chords/notes


Consists of commits:

Remove warning about ties in \retrograde


Change retrograde-music function to deal with ties


Let \retrograde expand repeat chords/notes

Since those lose their point of reference, expanding them
prior to retrograding keeps the results sane.

[thomasmorley651, 2016-08-16 06:23:17]

Didn't dive into the code in depth. From my (limited) testings, I'd say it's
very nice, thus:
LGTM

A nitpick:

https://codereview.appspot.com/302470043/diff/1/scm/modal-transforms.scm
File scm/modal-transforms.scm (right):

https://codereview.appspot.com/302470043/diff/1/scm/modal-transforms.scm#newc...
scm/modal-transforms.scm:188: ;; See also LSR #105.
There is no LSR-snippet 105. Must have been deleted somewhen in the past.
Actually the entire LSR doesn't mention retrograde at all nowadays.

[dak, 2016-08-16 07:53:16]

https://codereview.appspot.com/302470043/diff/1/scm/modal-transforms.scm
File scm/modal-transforms.scm (right):

https://codereview.appspot.com/302470043/diff/1/scm/modal-transforms.scm#newc...
scm/modal-transforms.scm:188: ;; See also LSR #105.
On 2016/08/16 06:23:17, thomasmorley651 wrote:
> There is no LSR-snippet 105. Must have been deleted somewhen in the past.
> Actually the entire LSR doesn't mention retrograde at all nowadays.

Done.

Namely, removed that line.  Will not reupload a new review just for that,
however.

[dak, 2016-08-16 07:54:50]

On 2016/08/16 06:23:17, thomasmorley651 wrote:
> Didn't dive into the code in depth. From my (limited) testings, I'd say it's
> very nice, thus:

I'm really bad at creating test cases.  Do you have any spares left over from
your (limited) testings that would make for a regtest?

[thomasmorley651, 2016-08-16 08:53:53]

On 2016/08/16 07:54:50, dak wrote:
> On 2016/08/16 06:23:17, thomasmorley651 wrote:
> > Didn't dive into the code in depth. From my (limited) testings, I'd say it's
> > very nice, thus:
> 
> I'm really bad at creating test cases.  Do you have any spares left over from
> your (limited) testings that would make for a regtest?

Nope. Though, I felt challenged and did some more testings and found some
issues.

With this example:

m = \relative g' { <g~ b>4 g8\glissando f <g b>4~ g8\glissando f }

\score {
  \new Voice { \m \retrograde  \m }
  \layout { ragged-right = ##f }
}

One Tie is omitted (without warning), glissandi not printed.


For a regtest perhaps:

motif = 
\relative { 
  \override TextSpanner.bound-details.left.text = "motif"
  <c' e>4~\startTextSpan  c16\< d^( e f~ f4:32)\! 
  << 
    \context Voice = "voice" { <g~ b>4 g8\glissando f\stopTextSpan } 
    \\ 
    { c2 } 
  >> 
}

\new Voice = "voice" { 
  \motif  
  \override TextSpanner.bound-details.left.text = "retrograde motif" 
  \retrograde  \motif 
}

Ofcourse the glissando isn't printed yet, as well. Interestingly a warning about
unterminated tie is issued, though.

[dak, 2016-08-16 14:24:50]

Regtest and fixes for glissandi, \<, \> and docs

[thomasmorley651, 2016-08-16 20:41:48]

On 2016/08/16 14:24:50, dak wrote:
> Regtest and fixes for glissandi, \<, \> and docs

Wow.
You've put up another load of work.

Some of the limitations you've mentioned yourself here:
http://lists.gnu.org/archive/html/lilypond-user/2016-08/msg00247.html
still persist. Though, tackling them would be beyond the topic of this patch:
dealing with ties.

Thus, pretty nice.
LGTM