Turns off TupletBracket collision avoidance if Script avoids slur

Turns off TupletBracket collision avoidance if Script avoids slur

[MikeSol, 2012-03-14 20:35:34]

Meh...not my best work, but it is a first step towards fixing this problem.  A
full solution would do tuplet avoidance for scripts in the same way they're done
for slurs.  Takers?

[dak, 2012-03-15 11:03:00]

On 2012/03/14 20:35:34, MikeSol wrote:
> Meh...not my best work, but it is a first step towards fixing this problem.

To me it looks more like an attempt to hide the problem's symptoms.  It does not
appear to have any relation with line-breaking which is the context in which the
problem manifested itself.

[mike_apollinemike.com, 2012-03-15 23:27:44]

On Mar 15, 2012, at 12:03 PM, dak@gnu.org wrote:

> On 2012/03/14 20:35:34, MikeSol wrote:
>> Meh...not my best work, but it is a first step towards fixing this
> problem.
> 
> To me it looks more like an attempt to hide the problem's symptoms.  It
> does not appear to have any relation with line-breaking which is the
> context in which the problem manifested itself.
> 

The problem is a circular dependency:

--) Tuplet bracket position calculation trys to avoid script
--) Tuplet bracket gets script's coordinates
--) Script has slur offset callback chained to it
--) Slur offset callback needs to know slur positions
--) Slur positions determined by encompass objects
--) Slur encompasses tuplet number
--) Tuplet number depends on tuplet bracket's position

I don't know where/how line breaking factors into it, but the above description
is definitely what's going on.  What I've proposed is a short, minimally
invasive workaround until a more thorough solution can be thought out.

Cheers,
MS

[mike_apollinemike.com, 2012-03-15 23:32:30]

On Mar 16, 2012, at 12:27 AM, mike@apollinemike.com wrote:

> On Mar 15, 2012, at 12:03 PM, dak@gnu.org wrote:
> 
>> On 2012/03/14 20:35:34, MikeSol wrote:
>>> Meh...not my best work, but it is a first step towards fixing this
>> problem.
>> 
>> To me it looks more like an attempt to hide the problem's symptoms.  It
>> does not appear to have any relation with line-breaking which is the
>> context in which the problem manifested itself.
>> 
> 
> The problem is a circular dependency:
> 
> --) Tuplet bracket position calculation trys to avoid script
> --) Tuplet bracket gets script's coordinates
> --) Script has slur offset callback chained to it
> --) Slur offset callback needs to know slur positions
> --) Slur positions determined by encompass objects
> --) Slur encompasses tuplet number
> --) Tuplet number depends on tuplet bracket's position
> 
> I don't know where/how line breaking factors into it, but the above
description is definitely what's going on.  What I've proposed is a short,
minimally invasive workaround until a more thorough solution can be thought out.

Sorry - I haven't had time to check the regtests yet, but I will either over the
weekend or next week - apparently some mangling is going on.
It'd delay 2.16, but this problem can be fully fleshed out over a few weeks.  Or
we could just revert the patch and then re-apply a better version after 2.16 is
out.  I'll have a better sense of the size of the job after I look into it a bit
more.

Cheers,
MS