Adds beam collision avoidance to auto-beaming

Adds beam collision avoidance to auto-beaming

[MikeSol, 2011-03-26 21:28:27]

I think this is a better approach to getting beam collision resolution out of
auto-beams.

It only effects 2 regtests:

repeat-percent.ly (the beamed C-Eb)
collision-dots-move.ly (there's only one beam & it moves)

I'd like to get this in 2.14.1 if possible in some form.

Cheers,
Mike

[hanwenn, 2011-03-26 21:46:16]

On Sat, Mar 26, 2011 at 6:28 PM,  <mtsolo@gmail.com> wrote:
> Reviewers: ,
>
> Message:
> I think this is a better approach to getting beam collision resolution
> out of auto-beams.

Quick comment: I'd say that timestep-count as a property is not
necessary.  The 'width' of an autobeam could also be garnered by
looking at its bounds, eg beam->get_bound(LEFT)-> get_column().  From
the column you could either look at rank or when.

A completely different approach now comes to mind: there is no strict
need to do all this processing 'on-line'.  You could also have the
engraver collect a list of grobs + timing-ranges in a list.  In
finalize(), you could sort the list, and decide which grobs to add
where.

-- 
Han-Wen Nienhuys - hanwen@xs4all.nl - http://www.xs4all.nl/~hanwen

[mike_apollinemike.com, 2011-03-26 21:54:44]

On Mar 26, 2011, at 5:46 PM, Han-Wen Nienhuys wrote:

> On Sat, Mar 26, 2011 at 6:28 PM,  <mtsolo@gmail.com> wrote:
>> Reviewers: ,
>> 
>> Message:
>> I think this is a better approach to getting beam collision resolution
>> out of auto-beams.
> 
> Quick comment: I'd say that timestep-count as a property is not
> necessary.  The 'width' of an autobeam could also be garnered by
> looking at its bounds, eg beam->get_bound(LEFT)-> get_column().  From
> the column you could either look at rank or when.
> 
> A completely different approach now comes to mind: there is no strict
> need to do all this processing 'on-line'.  You could also have the
> engraver collect a list of grobs + timing-ranges in a list.  In
> finalize(), you could sort the list, and decide which grobs to add
> where.


This sounds like a great idea.  Do you have time to write something along those
lines?

Cheers,
Mike

[hanwenn, 2011-03-26 21:56:25]

On Sat, Mar 26, 2011 at 6:54 PM, mike@apollinemike.com
<mike@apollinemike.com> wrote:

>> Quick comment: I'd say that timestep-count as a property is not
>> necessary.  The 'width' of an autobeam could also be garnered by
>> looking at its bounds, eg beam->get_bound(LEFT)-> get_column().  From
>> the column you could either look at rank or when.
>>
>> A completely different approach now comes to mind: there is no strict
>> need to do all this processing 'on-line'.  You could also have the
>> engraver collect a list of grobs + timing-ranges in a list.  In
>> finalize(), you could sort the list, and decide which grobs to add
>> where.

> This sounds like a great idea.  Do you have time to write something along
those lines?

Unfortunately, no.

-- 
Han-Wen Nienhuys - hanwen@xs4all.nl - http://www.xs4all.nl/~hanwen