-/*
- lyric-combine-music-iterator.cc -- implement Lyric_combine_music_iterator
-
+/*
+ new-lyric-combine-iterator.cc -- implement Lyric_combine_music_iterator
+
source file of the GNU LilyPond music typesetter
-
- (c) 1999--2003 Han-Wen Nienhuys <hanwen@cs.uu.nl>
-
- */
-#include "translator-group.hh"
-#include "lyric-combine-music.hh"
-#include "event.hh"
-#include "note-head.hh"
+ (c) 2004--2008 Han-Wen Nienhuys <hanwen@xs4all.nl>
+*/
+
+#include "dispatcher.hh"
+#include "global-context.hh"
#include "grob.hh"
+#include "input.hh"
+#include "international.hh"
+#include "listener.hh"
#include "music-iterator.hh"
+#include "music.hh"
+
+/*
+ This iterator is hairy. It tracks both lyric and melody contexts,
+ and has a complicated communication route, reading/writing
+ properties in both.
+
+ In the future, this should rather be done with
+
+ \interpretAsMelodyFor { MUSIC } { LYRICS LYRICS LYRICS }
+
+ This can run an interpret step on MUSIC, generating a stream. Then
+ the stream can be perused at leisure to apply durations to all of
+ the LYRICS.
+*/
class Lyric_combine_music_iterator : public Music_iterator
{
public:
- VIRTUAL_COPY_CONS (Music_iterator);
Lyric_combine_music_iterator ();
- Lyric_combine_music_iterator (Lyric_combine_music_iterator const&src);
- DECLARE_SCHEME_CALLBACK(constructor, ());
+ Lyric_combine_music_iterator (Lyric_combine_music_iterator const &src);
+ DECLARE_SCHEME_CALLBACK (constructor, ());
protected:
virtual void construct_children ();
virtual Moment pending_moment () const;
- virtual void do_quit();
+ virtual void do_quit ();
virtual void process (Moment);
- virtual Music_iterator *try_music_in_children (Music *) const;
-
+ virtual bool run_always ()const;
virtual bool ok () const;
virtual void derived_mark () const;
+ virtual void derived_substitute (Context *, Context *);
+ void set_music_context (Context *to);
private:
- bool get_busy_status ()const ;
- bool melisma_busy ();
-
+ bool start_new_syllable () const;
+ Context *find_voice ();
+ DECLARE_LISTENER (set_busy);
+ DECLARE_LISTENER (check_new_context);
+
+ bool music_found_;
+ Context *lyrics_context_;
+ Context *music_context_;
+ SCM lyricsto_voice_name_;
- Music_iterator * music_iter_;
- Music_iterator * lyric_iter_;
+ Moment busy_moment_;
+ Moment pending_grace_moment_;
+
+ Music_iterator *lyric_iter_;
};
+Lyric_combine_music_iterator::Lyric_combine_music_iterator ()
+{
+ music_found_ = false;
+ pending_grace_moment_.set_infinite (1);
+ lyric_iter_ = 0;
+ music_context_ = 0;
+ lyrics_context_ = 0;
+ busy_moment_.set_infinite (-1);
+}
/*
- Ugh, why static?
+ It's dubious whether we can ever make this fully work. Due to
+ associatedVoice switching, this routine may be triggered for
+ the wrong music_context_
*/
-Music *busy_req;
-Music *melisma_playing_req;
+IMPLEMENT_LISTENER (Lyric_combine_music_iterator, set_busy)
+void
+Lyric_combine_music_iterator::set_busy (SCM se)
+{
+ Stream_event *e = unsmob_stream_event (se);
-Lyric_combine_music_iterator::Lyric_combine_music_iterator ()
+ if ((e->in_event_class ("note-event") || e->in_event_class ("cluster-note-event"))
+ && music_context_)
+
+ busy_moment_ = max (music_context_->now_mom (),
+ busy_moment_);
+
+}
+
+void
+Lyric_combine_music_iterator::set_music_context (Context *to)
{
- if (!busy_req)
+ if (music_context_)
{
- busy_req
- = make_music_by_name (ly_symbol2scm ("BusyPlayingEvent"));
- melisma_playing_req
- = make_music_by_name (ly_symbol2scm ("MelismaPlayingEvent"));
+ music_context_->event_source ()->
+ remove_listener (GET_LISTENER (set_busy), ly_symbol2scm ("music-event"));
+ }
+
+ music_context_ = to;
+ if (to)
+ {
+ to->event_source ()->add_listener (GET_LISTENER (set_busy),
+ ly_symbol2scm ("music-event"));
}
-
- music_iter_ =0;
- lyric_iter_ =0;
+}
+
+bool
+Lyric_combine_music_iterator::start_new_syllable () const
+{
+ if (busy_moment_ < music_context_->now_mom ())
+ return false;
+
+ if (!lyrics_context_)
+ return false;
+
+ if (!to_boolean (lyrics_context_->get_property ("ignoreMelismata")))
+ {
+ bool m = melisma_busy (music_context_);
+ if (m)
+ return false;
+ }
+
+ return true;
}
Moment
Lyric_combine_music_iterator::pending_moment () const
{
- Moment musnext = music_iter_->pending_moment ();
- return musnext;
+ Moment m;
+
+ m.set_infinite (1);
+
+ return m;
+}
+
+bool
+Lyric_combine_music_iterator::run_always () const
+{
+ return true;
}
bool
Lyric_combine_music_iterator::ok () const
{
- return music_iter_->ok ();
+ return lyric_iter_ && lyric_iter_->ok ();
}
void
-Lyric_combine_music_iterator::derived_mark()const
+Lyric_combine_music_iterator::derived_mark ()const
{
- if (music_iter_)
- scm_gc_mark (music_iter_->self_scm());
if (lyric_iter_)
- scm_gc_mark (lyric_iter_->self_scm());
+ scm_gc_mark (lyric_iter_->self_scm ());
+ if (lyrics_context_)
+ scm_gc_mark (lyrics_context_->self_scm ());
+ if (music_context_)
+ scm_gc_mark (music_context_->self_scm ());
+}
+
+void
+Lyric_combine_music_iterator::derived_substitute (Context *f, Context *t)
+{
+ if (lyric_iter_)
+ lyric_iter_->substitute_outlet (f, t);
+ if (lyrics_context_ && lyrics_context_ == f)
+ lyrics_context_ = t;
+ if (music_context_ && music_context_ == f)
+ set_music_context (t);
}
void
Lyric_combine_music_iterator::construct_children ()
{
- Lyric_combine_music const * m = dynamic_cast<Lyric_combine_music const*> (get_music ());
+ Music *m = unsmob_music (get_music ()->get_property ("element"));
+ lyric_iter_ = unsmob_iterator (get_iterator (m));
+ if (!lyric_iter_)
+ return;
+ lyrics_context_ = find_context_below (lyric_iter_->get_outlet (),
+ ly_symbol2scm ("Lyrics"), "");
+
+ if (!lyrics_context_)
+ {
+ m->origin ()->warning ("argument of \\lyricsto should contain Lyrics context");
+ }
- music_iter_ = unsmob_iterator (get_iterator (m->get_music ()));
- lyric_iter_ = unsmob_iterator (get_iterator (m->get_lyrics ()));
+ lyricsto_voice_name_ = get_music ()->get_property ("associated-context");
+
+ Context *voice = find_voice ();
+ if (voice)
+ set_music_context (voice);
+
+ /*
+ Wait for a Create_context event. If this isn't done, lyrics can be
+ delayed when voices are created implicitly.
+ */
+ Global_context *g = get_outlet ()->get_global_context ();
+ g->events_below ()->add_listener (GET_LISTENER (check_new_context), ly_symbol2scm ("CreateContext"));
+
+ /*
+ We do not create a Lyrics context, because the user might
+ create one with a different name, and then we will not find that
+ one.
+ */
}
-bool
-Lyric_combine_music_iterator::get_busy_status () const
+IMPLEMENT_LISTENER (Lyric_combine_music_iterator, check_new_context)
+void
+Lyric_combine_music_iterator::check_new_context (SCM sev)
{
- /*
- We have to use both the event and the busyGrobs queue. The
- busyGrobs queue doesn't contain any notes that have started this
- instant. */
- if (try_music (busy_req))
- return true;
+ if (!ok ())
+ return ;
- Translator_group * tr = music_iter_->report_to ();
+ // TODO: Check first if type=Voice and if id matches
+ Stream_event * ev = unsmob_stream_event (sev);
+ if (ev->get_property ("type") != ly_symbol2scm ("Voice"))
+ return ;
+
+ Context *voice = find_voice ();
- SCM grobs = tr->get_property ("busyGrobs");
- Moment now = tr->now_mom();
- for (; gh_pair_p (grobs); grobs = gh_cdr (grobs))
+ if (voice)
{
- SCM grob = gh_cdar (grobs);
- Moment end =*unsmob_moment (gh_caar (grobs));
-
-
- /*
- This is slightly ugh: we are now confunding the frontend
- (iterators) and the backend (note heads) */
- if (end > now
- && Note_head::has_interface (unsmob_grob (grob)))
- return true;
+ set_music_context (voice);
}
-
- return false;
}
+/*
+ Look for a suitable voice to align lyrics to.
-bool
-Lyric_combine_music_iterator::melisma_busy ()
+ Returns 0 if nothing should change; i.e., if we already listen to the
+ right voice, or if we don't yet listen to a voice but no appropriate
+ voice could be found.
+*/
+Context *
+Lyric_combine_music_iterator::find_voice ()
{
-#if 0
-
- Translator_group * trg = music_iter_->report_to();
- bool melisma_busy= to_boolean (trg->get_property ("melismaBusy"));
+ SCM voice_name = lyricsto_voice_name_;
+ SCM running = lyrics_context_
+ ? lyrics_context_->get_property ("associatedVoice")
+ : SCM_EOL;
- melisma_busy = melisma_busy || to_boolean (trg->get_property ("slurMelismaBusy"));
- melisma_busy = melisma_busy || to_boolean (trg->get_property ("tieMelismaBusy"));
- melisma_busy = melisma_busy || to_boolean (trg->get_property ("beamMelismaBusy"));
- return melisma_busy;
+ if (scm_is_string (running))
+ voice_name = running;
-#else
+ if (scm_is_string (voice_name)
+ && (!music_context_ || ly_scm2string (voice_name) != music_context_->id_string ()))
+ {
+ Context *t = get_outlet ();
+ while (t && t->get_parent_context ())
+ t = t->get_parent_context ();
- /*
- The above code is nicer since it doesn't rely on a special
- engraver to signal the melisma status. Unfortunately,
- music_iter_->report_to() might not be the context that sets the
- melisma properties, but rather a parent context.
- */
- return music_iter_->try_music (melisma_playing_req);
-#endif
-
+ string name = ly_scm2string (voice_name);
+ return find_context_below (t, ly_symbol2scm ("Voice"), name);
+ }
+
+ return 0;
}
void
-Lyric_combine_music_iterator::process (Moment m)
+Lyric_combine_music_iterator::process (Moment when)
{
- Moment my_next = music_iter_->pending_moment ();
- if (my_next > m)
- return;
+ (void) when;
- music_iter_->process (m);
+ /* see if associatedVoice has been changed */
+ Context *new_voice = find_voice ();
+ if (new_voice)
+ set_music_context (new_voice);
+
+ if (!music_context_)
+ return;
+
+ if (!music_context_->get_parent_context ())
+ {
+ /*
+ The melody has died.
+ We die too.
+ */
+ if (lyrics_context_)
+ lyrics_context_->unset_property (ly_symbol2scm ("associatedVoiceContext"));
+ lyric_iter_ = 0;
+ set_music_context (0);
+ }
+
- if (get_busy_status () && !melisma_busy () && lyric_iter_->ok ())
+ if (music_context_
+ && (start_new_syllable () ||
+ (busy_moment_ >= pending_grace_moment_))
+ && lyric_iter_->ok ())
{
- Moment m= lyric_iter_->pending_moment ();
+ Moment now = music_context_->now_mom ();
+ if (now.grace_part_)
+ {
+ pending_grace_moment_ = now;
+ pending_grace_moment_.grace_part_ = Rational (0);
+ return;
+ }
+ else
+ {
+ pending_grace_moment_.set_infinite (1);
+ }
+
+ Moment m = lyric_iter_->pending_moment ();
+ lyrics_context_->set_property (ly_symbol2scm ("associatedVoiceContext"),
+ music_context_->self_scm ());
lyric_iter_->process (m);
+
+ music_found_ = true;
}
+
+ new_voice = find_voice ();
+ if (new_voice)
+ set_music_context (new_voice);
}
void
Lyric_combine_music_iterator::do_quit ()
{
- if (music_iter_)
- music_iter_->quit();
- if (lyric_iter_)
- lyric_iter_->quit();
-
-}
-Lyric_combine_music_iterator::Lyric_combine_music_iterator (Lyric_combine_music_iterator const & src)
- : Music_iterator (src)
-{
- lyric_iter_ = 0;
- music_iter_ = 0;
+ if (!music_found_)
+ {
+ SCM voice_name = get_music ()->get_property ("associated-context");
- if (src.lyric_iter_)
- lyric_iter_ = src.lyric_iter_->clone ();
- if (src.music_iter_)
- music_iter_ = src.music_iter_->clone ();
+ string name;
+ if (scm_is_string (voice_name))
+ name = ly_scm2string (voice_name);
- if (lyric_iter_)
- scm_gc_unprotect_object (lyric_iter_->self_scm());
- if (music_iter_)
- scm_gc_unprotect_object (music_iter_->self_scm());
-}
+ get_music ()->origin ()->warning (_f ("cannot find Voice `%s'",
+ name.c_str ()) + "\n");
+ }
-Music_iterator*
-Lyric_combine_music_iterator::try_music_in_children (Music *m) const
-{
- Music_iterator * i = music_iter_->try_music (m);
- if (i)
- return i;
- else
- return lyric_iter_->try_music (m);
+ if (lyric_iter_)
+ lyric_iter_->quit ();
}
-
IMPLEMENT_CTOR_CALLBACK (Lyric_combine_music_iterator);
projects / lilypond.git / blobdiff