Re: Sets spanner length as a spanner property. (issue4641085)

[address@hidden]

On Jun 30, 2011, at 5:20 PM, Han-Wen Nienhuys wrote:

> On Thu, Jun 30, 2011 at 11:21 AM,  <address@hidden> wrote:
>> Reviewers: ,
>> 
>> Message:
>> Hey all,
>> 
>> This patch simplifies a lot of code duping I'm doing to get spanner
>> lengths in various scheme functions.  Please lemme know if it's OK!
> 
> Why is length not computed as a callback?  You should avoid
> set_property() as much as possible, since the order in which the code
> calling set_property() is not defined. Hence, readers of the length
> property cannot be sure that it has been set.
> 

There's no issue with computing this as a separate callback save the fact that 
it'd result in a code dup with the normalized endpoints callback.

> What calculations are simplified?  Can you include examples?  I am
> sure that the lilypond code uses spanner length in a ~100 places or
> so.  If you are starting to use a property for length, your code
> should have a 100 places changed.
> 

This makes it easier to do manipulations on a line-breaking spanner whose 
different pieces are not supposed to be copies of each other.

> I think you scheme code will be easier to follow if you just define a
> function that computes the length, if given a spanner.
> 

I have a version of this written in Scheme that I use in my graphical notation. 
 I'll throw up a new patch with this in the not-too-distant future.  But, 
again, it is a code dup from the normalized endpoints callback, which can 
calculate both spanner length and normalized endpoints in one fell swoop (at 
the expense of using set_property).

Cheers,
MS