Re: Allows glissandi between chords (issue4442082)

[Carl Sorensen]

On 4/28/11 6:35 AM, "Carl Sorensen" <address@hidden> wrote:

> On 4/27/11 5:00 PM, "address@hidden" <address@hidden> wrote:
> 
>> On 2011/04/26 17:11:02, c_sorensen_byu.edu wrote:
>>> On 4/26/11 11:00 AM, "address@hidden"
>> <mailto:address@hidden> wrote:
>> 
>>>> On Apr 26, 2011, at 7:55 AM, mailto:address@hidden wrote:
>>>> 
>>>>> On Apr 26, 2011, at 7:30 AM, mailto:address@hidden
>> wrote:
>>>>> 
>>>>> The old behavior can be achieved via:
>>>>> 
>>>>> \set glissandoMap = #'((0 . 0))
>>>>> 
>>>>> What would be the appropriate .ly file in ly/ to put this in?
>>>> 
>>>> Alternatively, we could write a convert-ly rule that changes all
>> \glissando to
>>>> \set glissandoMap = #'((0 . 0)) \glissando \unset glissandoMap.
>>>> 
>>>> I think that the majority of users who write glissandi between
>> chords will
>>>> want full-chord glissandi.  So, it seems that #'() should be the
>> default value
>>>> for glissandoMap (as it is in the proposed patch).
>>>> 
>>> If all notes should be part of the glissando, then I think we should
>> *not*
>>> write a convert-ly rule.  Rather, we should make sure that the new
>> setting
>>> is noted in changes, and we should make sure the new behavior is
>> documented.
>> 
>>> Thanks,
>> 
>>> Carl
>> 
>> 
>> 
>> Hey Carl et al,
>> 
>> Before I push this patch, are there any further questions/concerns
>> regarding it?
>> Specifically:
>> 
>> (a)  Are people all right with glissando'ed chords being default
>> behavior (given how ties work in lilypond, this seems to make sense)?
> 
> I'm fine with it, but I'm not a good reference -- my experience with music
> is too limited.
> 
>> 
>> (b)  With respect to my previous question about internal-properties, is
>> it ok to have glissando-index in define-grob-properties.scm and to put
>> it in the glissando-interface?  If not, where should I put it?
> 
> Personally, I think it should be in define-grob-properties.scm as an
> internal property, and not in an interface.  There are lots of examples of
> properties like that, as far as I can see.  For example, least-squares-dy,
> maybe-loose, positioning-done, important-column-ranks, pure-Y-extent.  In my
> mind, we don't have interfaces to things that the user couldn't set
> properly, and glissando-index is one of those.  We can read without an
> interface, we just can't set without an interface.
> 
> But I could easily be wrong on this.

In fact, I was wrong on this.  Please ignore all my comments on the
interface.  Mike's approach is correct.

Thanks,

Carl