Re: Partial compilation (again)

[Urs Liska]

Am 21.11.2014 13:33, schrieb David Kastrup:
> Urs Liska <address@hidden> writes:
>
> > Am 21. November 2014 12:36:59 MEZ, schrieb David Kastrup <address@hidden>:
> >
> > > So I don't see this going anywhere fast.  To get a hook into cloning
> > > internal engraver states, you'd probably want to create a more
> > > object-oriented frame work for them.
> > Ok, I see. Thank you for the explanation.
> >
> > I'll try if I can get *somewhere* with an "external" approach as
> > outlined on -user.
> TeX, in contrast to LilyPond, works in a pipeline-like manner, producing
> output all the time.  In that case, a semi-external approach might work
> when you have a good idea what kind of input is subject to editing: you
> just fork off a copy of the executable when it gets to the place before
> the editing region.  And then you run that forked copy on the changed
> input from that point on.
>
> However, LilyPond first consumes all input and then starts processing,
> so the input file pointer into an unchanged part of the input is just
> not a useful indicator of LilyPond's progress and this kind of low-level
> approach does not look workable.

OK. I will try around anyway.
If I knew where a given system starts in terms of barnumbers I can try:

- setting suitable skipTypesettings commands one measure before and one 
measure after the range
- check for open spanners and optionally inject corresponding 
starts/ends in the respective "preroll" measure
- add manual breaks before and after the line
- compile this threeline document to three temporary files (with 
lilypond-book-preamble)
- replace the respective previous system's pdf with the new one.

setting system-count to 3 will result in the right document in any case, 
and in warnings when the system gets too full (so the recompilation 
doesn't work anymore).

Of course this isn't intended for the part of bringing a score to 
publication quality but rather for the stage when you're only dealing 
with the content. And I don't know how complicated things will become. 
But I think it is worth a try.

Urs