Issue 5275: Avoid duplicate rehearsal marks in regression tests

https://sourceforge.net/p/testlilyissues/issues/5275/

Having multiple marks with the same sequence number is something that
LilyPond might reasonably warn about in a future version.

I suppose LilyPond might reasonably warn about out-of-sequence marks,
but I haven't gone that far with this.  Let me know if you think I
should.

[Dan Eble, 2018-02-18 13:51:40]

Malte,
This patch changes a regression test you committed in January.  I think it still
covers what you intended to cover.

[Malte Meyn, 2018-02-20 16:28:21]

I see arguments both for and against such warnings.

• maybe CON, depending on the implementation: It should be possible to set the
first mark in a score to something different than 1 without a warning.
• CON: a user should be aware that they are responsible for correct numbering
(whatever that means) if they’re not using \mark \default because the only
reason I can imagine not to use \default is when one wants to have something
like out-of-sequence or multiple identical marks → warnings would disturb.
• PRO: multiple identical (number/formatted number, not markup) marks seem silly
to me.

https://codereview.appspot.com/334510043/diff/1/input/regression/rehearsal-ma...
File input/regression/rehearsal-mark-formatters.ly (right):

https://codereview.appspot.com/334510043/diff/1/input/regression/rehearsal-ma...
input/regression/rehearsal-mark-formatters.ly:7: The marks should read H, K, I,
K, 93, XCIV, 9, AB, CC, Dd, xcvi,
That 9 comes from \mark 111 with #format-mark-barnumbers and should be changed
to 7 because you deleted the first two measures.

[Dan Eble, 2018-02-20 17:52:47]

Fix expected value in test description