Let the parser interpret alist variables

Let the parser interpret alist variables

After

x.y = { c d e f }

this allows using

\x.y

instead of #(assoc-ref x 'y) in most places where the latter would work.


Also contains commit:

Create nested_property function

[thomasmorley651, 2016-03-13 23:39:29]

Even from description the consequences are not clear to me.
Can you give a little verbose example?

[lemzwerg, 2016-03-14 06:46:51]

LGTM.

https://codereview.appspot.com/290560043/diff/1/lily/nested-property.cc
File lily/nested-property.cc (right):

https://codereview.appspot.com/290560043/diff/1/lily/nested-property.cc#newco...
lily/nested-property.cc:137: for (;scm_is_pair (prop_path); prop_path = scm_cdr
(prop_path))
s/;scm/; scm/

[dak, 2016-03-14 08:32:05]

https://codereview.appspot.com/290560043/diff/1/lily/nested-property.cc
File lily/nested-property.cc (right):

https://codereview.appspot.com/290560043/diff/1/lily/nested-property.cc#newco...
lily/nested-property.cc:137: for (;scm_is_pair (prop_path); prop_path = scm_cdr
(prop_path))
On 2016/03/14 06:46:51, lemzwerg wrote:
> s/;scm/; scm/

Done.  Not uploading a separate review just for this, however.

[dak, 2016-03-14 08:49:43]

On 2016/03/13 23:39:29, thomasmorley651 wrote:
> Even from description the consequences are not clear to me.
> Can you give a little verbose example?

Basically, you can now use x.y for defining and \x.y for using a variable like
you could previously use x_y and \x_y but the value is really stored in an
alist.

That's not all that exciting, except that in connection with issue 4798 this
will also permit violin.1 or violin.#x which some people might consider almost
what they want.

However, the slightly nicer looking assignment

    violin1 = ...

is going to be ruled out by issue 4800 since I could not make the corresponding
\violin1 work.

It's sort of funny why \violin1 could not be made to work: I first let an alist
like \violin always grab the next expression and use it as an index.  This
caused \key c \major to stop working since \major has _exactly_ the form of an
alist indexed by a number.  So the indexing operation must be made explicit (dot
or comma), and in order to make it explicit for the assignment as well, issue
4800.

As a corollary, \minor.2 is equivalent to #-1/2 after this issue and issue 4798.