Re: Feature request: revert in another context

[David Kastrup]

Simon Albrecht <address@hidden> writes:

> On 28.11.2016 23:13, David Kastrup wrote:
>> Simon Albrecht <address@hidden> writes:
>>
>>> On 28.11.2016 17:27, David Kastrup wrote:
>>>> Simon Albrecht <address@hidden> writes:
>>>>
>>>>> \version "2.19.49"
>>>>> %{
>>>>> It would be formidable if in such a case one wouldn’t need
>>>>> to look up the default stencil procedure, but could use either
>>>>> \undo or \revert.
>>>>> Is this a valid/sensible feature request?
>>>>> %}
>>>>> \score {
>>>>>     \new PianoStaff \with {
>>>>>       \omit SystemStartBrace
>>>>>       \accepts GrandStaff
>>>>>     } <<
>>>>>       \new GrandStaff \with {
>>>>>         % those don’t work
>>>>>         %\undo\omit SystemStartBrace
>>>>>         %\revert SystemStartBrace.stencil
>>>>>         % this does
>>>>>         \override SystemStartBrace.stencil =
>>>>> #ly:system-start-delimiter::print
>>>>>       } <<
>>>>>         \new Staff { 1 }
>>>>>         \new Staff { 1 }
>>>>>       >>
>>>>>       \new Staff { 1 }
>>>>>     >>
>>>>> }
>>>> The context modification for GrandStaff does not have access to the
>>>> unmodified definition of the enclosing PianoStaff (it doesn't even have
>>>> access to the unmodified definition of GrandStaff and does not know that
>>>> it will get applied to a GrandStaff) and even if it did, how should it
>>>> guess that you want to undo PianoStaff settings rather than Score
>>>> settings?
>>> OK, I’ll take that as a no. Or is it just that it would be complicated
>>> to implement a behaviour matching my naïve expectation?
>> It would be complicated to give a sensible definition of your naïve
>> expectation.
>>
>> If you wrote
>>
>>      \undo \omit GrandStaff.SystemStartBrace
>>
>> or the equivalent
>>
>>      \revert GrandStaff.SystemStartBrace.stencil
>>
>> in the music itself, you could not expect it to revert the settings of
>> PianoStaff .  So why do you expect it do do this in the context mod ?
>
> Put another way: I want to prevent the ‘trickling down’, or inheriting
> of the property from the parent context. Maybe there should be a
> separate syntax for that?

That does not even make sense.  Either the context has a property of its
own or it inherits the property.  There is no third option.

-- 
David Kastrup