Issue 4379: Part_combine_iterator: simplify context substitution

Part_combine_iterator: simplify context substitution.  There are no regression test differences, but I'm concerned that there might be a reason for the way this was written that the tests do not exercise.

Music_iterator::substitute_outlet: do nothing if the new outlet is
identical to the old (a sanity test)

[Keith, 2015-05-11 05:04:15]

This and the MMrest patch look sensible,
but it is a good idea to test them on some orchestral music from mutopiaproject.
 I tried them on a piece I have and saw no output differences.

https://codereview.appspot.com/238070043/diff/1/lily/part-combine-iterator.cc
File lily/part-combine-iterator.cc (left):

https://codereview.appspot.com/238070043/diff/1/lily/part-combine-iterator.cc...
lily/part-combine-iterator.cc:210: }
The use of loops to go through cases is HanWenn's style, but this is mysterious.

[Dan Eble, 2015-05-13 01:02:28]

On 2015/05/11 05:04:15, Keith wrote:
> This and the MMrest patch look sensible,
> but it is a good idea to test them on some orchestral music from
mutopiaproject.
>  I tried them on a piece I have and saw no output differences.

First I tried a score requiring 2.14 which would not compile after running
convert-ly.  Then I tried a few scores requiring 2.18 with horn, bassoon, and
string combining, and I did not notice any problems due to these changes.  (I
did find a problem caused by older changes which I reported to the bug list and
will address when I have time.)