Create \= command for setting spanner-id

Create \= command for setting spanner-id

Also contains commits:

NR: document \= in slur sections


Make some regtests use \= instead of their own definitions

[thomasmorley651, 2015-10-04 00:45:36]

LGTM

Some nitpicks:

https://codereview.appspot.com/268050045/diff/1/Documentation/notation/expres...
File Documentation/notation/expressive.itely (right):

https://codereview.appspot.com/268050045/diff/1/Documentation/notation/expres...
Documentation/notation/expressive.itely:724: them with @code{\=} followed by an
identifying number or string.
If we explain the needed label may be number or string, why not give an example
for each in the code below.

https://codereview.appspot.com/268050045/diff/1/Documentation/notation/expres...
Documentation/notation/expressive.itely:728: <c~ f\=1( g\=2( >2 <c e\=1) a\=2) >
Am I correct this example does only work with the fix for issue 4625 applied?

[dak, 2015-10-04 03:58:31]

https://codereview.appspot.com/268050045/diff/1/Documentation/notation/expres...
File Documentation/notation/expressive.itely (right):

https://codereview.appspot.com/268050045/diff/1/Documentation/notation/expres...
Documentation/notation/expressive.itely:724: them with @code{\=} followed by an
identifying number or string.
On 2015/10/04 00:45:35, thomasmorley651 wrote:
> If we explain the needed label may be number or string, why not give an
example
> for each in the code below.

Because the documentation needs to be complete but the examples show the
intended practice?  spanner-id is a string and there does not seem to be a point
not to admit strings with \= but I don't expect anything but numbers to be in
use at Voice-level.  when non-Voice level slurs are supported by discriminating
using spanner-id, I expect something like a regexp or the first letter to be the
distinguishing feature.  Granted, that may make digits (which don't combine with
letters) not even the best idea as base.  The main advantage is that they save
you from inventing a name for something ad-hoc.

https://codereview.appspot.com/268050045/diff/1/Documentation/notation/expres...
Documentation/notation/expressive.itely:728: <c~ f\=1( g\=2( >2 <c e\=1) a\=2) >
On 2015/10/04 00:45:35, thomasmorley651 wrote:
> Am I correct this example does only work with the fix for issue 4625 applied?

Yes.  I saw no point in avoiding 4625 here as it likely will be the most common
reason to use \= and it has no bikesheddable elements that could still require
changing this example.  It's conceivable to assign in-chord slurs in sequence by
default so that no slur id is required.  But that would be a rather complex
extension.  So this example is likely to remain relevant for a while.

[Trevor Daniels, 2015-10-08 12:01:14]

LGTM