Re: SystemStart changes after stop/startStaff - how to get it back

[Thomas Morley]

Hi Aaron,

Am Di., 14. Apr. 2020 um 06:20 Uhr schrieb Aaron Hill
<address@hidden>:
>
> I think I have addressed the original issue in this patched version of
> the System_start_delimiter_engraver.  This is implemented in Scheme, but
> the changes should be easily ported back to C++.
>
> The strategy is to acknowledge the end of StaffSymbols to be able to
> remove their corresponding entries in the internal structure.
> Additionally, the context in which the StaffSymbol correlates is
> recorded so a future StaffSymbol of the same context takes the same
> place.  This should keep things stable if staves stop and start at odd
> places.

many thanks for your work!

For now I tested your engraver with my workaround in
mid-staff-gap-04.ly (see my previous post in this thread).
All works fine.

Though as nice as my current workaround looks like, it's still that: a
workaround.
I'll have a closer look at your engraver. Now that the functionality
is available in scheme I may find a proper fix (or not ...)
First step as always: I need to understand the given code ...

>
> One thing I noticed was the C++ engraver acknowledges
> SystemStartDelimiters (of which it is the one creating them) and calls
> the Bracket_nesting_group::add_support function.  However, in my porting
> of the engraver to Scheme, I could not get this acknowledger to trigger.
>   Does LilyPond have code coverage tests that could confirm whether that
> logic is necessary?  In any event, I omitted it from my Scheme engraver.

Can't help here, sorry.


Best,
  Harm