Re: lilypond ./ChangeLog input/test/to-xml.ly lily/...

[Juergen Reuter]

On Thu, 28 Nov 2002, Han-Wen Nienhuys wrote:

> ...
> > ! /* creats */      "Ligature_engraver",
>
> This fix  is incorrect. Ligature_engraver does not follow protocol. It
> does
>
>       Spanner *
>       Ligature_engraver::create_ligature_spanner ()
>       {
>         return new Spanner (SCM_EOL);
>       }
>
>

Ok, I am not that C++ guru.  This method should actually be something that
in e.g. Java would be called an abstract method: Ligature_engraver just
defines the interface of this method, but leaves the implementation to its
subclasses (but it must be declared in Ligature_engraver itself, since it
is virtually envoked from the process_music() code of Ligature_engraver).

> If the engraver doesn't create anything (which is not the case, it
> seems), then the /*creats */ entry should simply be empty.

Most probably.  But if I recall correctly, there was some problem with
acknowledge_grob() not being called in some cases.  Hence, when putting
the creats entry empty, one should double check that

http://lilypond.org/development/Documentation/user/out-www/lilypond/White-mensural-ligatures.html#White%20mensural%20ligatures

still shows ligatures (as the website of 1.7.8 currently does, while in
1.7.6, ligatures did not show).

Greetings,
Juergen