Re: Regtest changes phase 1 (issue 6454121)

[k-ohara5a5a]

Sorry I didn't notice this before.

'relative-repeat.ly' was testing something different than you thought.


http://codereview.appspot.com/6454121/diff/12002/input/regression/relative-repeat.ly
File input/regression/relative-repeat.ly (left):

http://codereview.appspot.com/6454121/diff/12002/input/regression/relative-repeat.ly#oldcode2
input/regression/relative-repeat.ly:2: texidoc = "Relative mode for
repeats uses order of entry."
The point of the original regression test was that
  \relative c' {
   \repeat unfold 3 {f2 bes2}
   \alternative { {a1}{e}{b} } }
is very different from a textual unfolding of the repeats.
  \relative c' {f2 bes2 a1 f2 bes2 e f2 bes2 b }

Relative-pitch-entry, and default-durations follow the order of typed
input, not the order of printed output.

Barchecks and octave checks would make the test self-explanatory.  Maybe
there are enough other uses of relative mode with repeats that we could
leave this one out.

http://codereview.appspot.com/6454121/diff/12002/input/regression/relative-repeat.ly
File input/regression/relative-repeat.ly (right):

http://codereview.appspot.com/6454121/diff/12002/input/regression/relative-repeat.ly#newcode2
input/regression/relative-repeat.ly:2: texidoc = "Unfolded repeats take
alternative notes in the order
We don't need the replacement test, because 'repeat-unfold.ly' covers
that function.

http://codereview.appspot.com/6454121/