Re: Updating translators

[Han-Wen Nienhuys]

Erik Sandberg schreef:
> 4. As a first demo, arpeggio-engraver has been rewritten to use these 
> mechanisms.

yes, see below.

>   B. Create a dummy listener that protects all incoming events, and unprotects 
> them at the end of each moment. The easiest way is probably to let 
> Score-level translators take care of this. This is the easiest and safest 
> solution, but also the most inefficient one.
>   C. Create some macro trickery in translators, so that incoming events are 
> remembered and protected automatically between try_music and 
> stop_translation_timestep.

go for either B or C. I'm not sure whether you can make B work, because 
it sounds as if there still might still be a small time unprotected time 
interval before the event reaches the protection engraver.

for protection, don't use

  event->protect()

rather use a derived_mark() and a vector<Event*>

>   D. Add an 'event property to Music, so that sent events are protected by 
> their music origins. This is the easiest solution, but I'm nervous about 
> memory errors that can happen when a different front-end is used (e.g., when 
> loading music streams directly from a file)

no, Music should not connect to stream events

> -bool
> -Arpeggio_engraver::try_music (Music *m)
> +IMPLEMENT_TRANSLATOR_LISTENER (Arpeggio_engraver, hear_arpeggio, 
> "ArpeggioEvent");
> +void Arpeggio_engraver::hear_arpeggio (Stream_event *ev)

can you reengineer this to have less redundancy? I would expect that you 
can get enough info from

  IMPLEMENT_TRANSLATOR_LISTENER (Arpeggio_engraver, arpeggio)


--

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

LilyPond Software Design
 -- Code for Music Notation
http://www.lilypond-design.com