#include "lyric-phrasing-engraver.hh"
#include "note-head.hh"
#include "translator-group.hh"
-#include "side-position-interface.hh"
#include "spanner.hh"
-#include "paper-def.hh"
+
String get_context_id(Translator_group * ancestor, const char * type);
String trim_suffix(String &id);
ADD_THIS_TRANSLATOR (Lyric_phrasing_engraver);
+/*
+ TODO: this code is too hairy, and does things that should be in the
+ backend. Fixme.
+*/
+
+/*
+ TODO:
+
+ shared lyrics should be vertically centered:
+
+
+
+> About lyrics, it happens that there are common words for many bars, like
+> for a refrain say. When there is an even number of lyrics lines, I do not
+> know how to force the positioning of the common lyric line in the plain
+> middle of the others, because this is in between lines. Not a big matter,
+> but it would be a bit nicer if this was doable.
+
+ */
+
/*
We find start and end of phrases, and align lyrics accordingly.
Also, lyrics at start of melismata should be left aligned.
*/
}
-Voice_alist_entry *
+void
+Lyric_phrasing_engraver::finalize ()
+{
+ /*
+ but do need to unprotect alist_, since Engravers are gc'd now.
+ */
+
+ voice_alist_ = SCM_EOL;
+}
+
+
+Syllable_group *
Lyric_phrasing_engraver::lookup_context_id(const String &context_id)
{
SCM key = ly_str02scm(context_id.ch_C());
}
}
// ( ( alist_entry . old_entry ) . previous_entry )
- SCM val = gh_cons(gh_cons(Voice_alist_entry::make_entry (), SCM_BOOL_F),
- Voice_alist_entry::make_entry ());
+ SCM val = gh_cons(gh_cons(Syllable_group::make_entry (), SCM_BOOL_F),
+ Syllable_group::make_entry ());
voice_alist_ = scm_acons(key, val, voice_alist_);
return unsmob_voice_entry (gh_caar(val));
void
Lyric_phrasing_engraver::record_notehead(const String &context_id,
- Score_element * notehead)
+ Grob * notehead)
{
- Voice_alist_entry * v = lookup_context_id(context_id);
+ Syllable_group * v = lookup_context_id(context_id);
v->set_notehead(notehead);
if(!any_notehead_l_)
any_notehead_l_ = notehead;
}
void
-Lyric_phrasing_engraver::record_lyric(const String &context_id, Score_element * lyric)
+Lyric_phrasing_engraver::record_lyric(const String &context_id, Grob * lyric)
{
- Voice_alist_entry * v = lookup_context_id(context_id);
+ Syllable_group * v = lookup_context_id(context_id);
v->add_lyric(lyric);
}
void
-Lyric_phrasing_engraver::record_extender(const String &context_id, Score_element * extender)
+Lyric_phrasing_engraver::record_extender(const String &context_id, Grob * extender)
{
SCM key = ly_str02scm(context_id.ch_C());
if( ! gh_null_p(voice_alist_) ) {
// ( key . ( (alist_entry . old_entry) . previous_entry) )
SCM previous_scm = gh_cddr(s);
if(previous_scm != SCM_EOL) {
- Voice_alist_entry * v = unsmob_voice_entry(previous_scm);
+ Syllable_group * v = unsmob_voice_entry(previous_scm);
v->add_extender(extender);
}
}
void
Lyric_phrasing_engraver::record_melisma(const String &context_id)
{
- Voice_alist_entry * v = lookup_context_id(context_id);
+ Syllable_group * v = lookup_context_id(context_id);
v->set_melisma();
}
void
-Lyric_phrasing_engraver::acknowledge_element(Score_element_info i)
+Lyric_phrasing_engraver::acknowledge_grob(Grob_info i)
{
SCM p = get_property("automaticPhrasing");
if(!to_boolean(p))
return;
- Score_element *h = i.elem_l_;
+ Grob *h = i.elem_l_;
if (Note_head::has_interface(h)) {
/* caught a note head ... do something with it */
/* ... but not if it's a grace note ... */
- bool grace= to_boolean (i.elem_l_->get_elt_property ("grace"));
+ bool grace= to_boolean (i.elem_l_->get_grob_property ("grace"));
SCM wg = get_property ("weAreGraceContext");
bool wgb = to_boolean (wg);
if (grace != wgb)
/* what's its Voice context name? */
String voice_context_id = get_context_id(i.origin_trans_l_->daddy_trans_l_, "Voice");
record_notehead(voice_context_id, h);
+
+ /* is it in a melisma ? */
+ if(to_boolean(i.origin_trans_l_->get_property("melismaEngraverBusy"))) {
+ record_melisma(voice_context_id);
+ }
return;
}
+
/* now try for a lyric */
if (h->has_interface (ly_symbol2scm ("lyric-syllable-interface"))) {
- /* what's its LyricVoice context name? */
- String lyric_voice_context_id =
- get_context_id(i.origin_trans_l_->daddy_trans_l_, "LyricVoice");
- record_lyric(trim_suffix(lyric_voice_context_id), h);
- return;
- }
- /* finally for a melisma */
- if(h->has_interface (ly_symbol2scm ("melisma-interface"))) {
- String voice_context_id = get_context_id(i.origin_trans_l_->daddy_trans_l_, "Voice");
- record_melisma(voice_context_id);
+ /* what's its LyricsVoice context name? */
+ String voice_context_id;
+ SCM voice_context_scm = i.origin_trans_l_->get_property("associatedVoice");
+ if (gh_string_p (voice_context_scm)) {
+ voice_context_id = ly_scm2string(voice_context_scm);
+ }
+ else {
+ voice_context_id = get_context_id(i.origin_trans_l_->daddy_trans_l_, "LyricsVoice");
+ voice_context_id = trim_suffix(voice_context_id);
+ }
+ record_lyric(voice_context_id, h);
return;
}
- /* How about catching any extender items and then if we have a melisma,
- set the RIGHT item of the extender spanner to the melismatic note in
- the corresponding context (if any).
- This has the effect of finishing the extender under the last note
- of the melisma, instead of extending it to the next lyric.
-
- Problem: the extender request is thrown at the same moment as the next lyric,
- by which time we have already passed the last note of the melisma.
- However, the Lyric_phrasing_engraver remembers the last note, so just
- attach it to that, provided it was melismatic. If it was not melismatic,
- then ignore it and let the Extender_engraver take care of it (i.e. finish at next
- lyric).
- */
+
+ /* Catch any extender items and then if we have a melisma,
+ set the RIGHT item of the extender spanner to the melismatic note in
+ the corresponding context (if any).
+ This has the effect of finishing the extender under the last note
+ of the melisma, instead of extending it to the next lyric.
+
+ Problem: the extender request is thrown at the same moment as the next lyric,
+ by which time we have already passed the last note of the melisma.
+ However, the Lyric_phrasing_engraver remembers the last note, so just
+ attach it to that, provided it was melismatic. If it was not melismatic,
+ then ignore it and let the Extender_engraver take care of it (i.e. finish at next
+ lyric).
+ */
if(h->has_interface (ly_symbol2scm ("lyric-extender-interface"))) {
- String voice_context_id = get_context_id(i.origin_trans_l_->daddy_trans_l_, "LyricVoice");
+ String voice_context_id = get_context_id(i.origin_trans_l_->daddy_trans_l_, "LyricsVoice");
record_extender(trim_suffix(voice_context_id), h);
return;
}
}
-void Lyric_phrasing_engraver::process_acknowledged ()
+void Lyric_phrasing_engraver::create_grobs ()
{
/* iterate through entries in voice_alist_
for each, call set_lyric_align(alignment). Issue a warning if this returns false.
SCM v_entry = gh_cdar(v);
// ((current . oldflag) . previous)
if(!to_boolean(gh_cdar(v_entry))) { // not an old entry left over from a prior note ...
- Voice_alist_entry *entry = unsmob_voice_entry(gh_caar(v_entry));
+ Syllable_group *entry = unsmob_voice_entry(gh_caar(v_entry));
+
+ /*
+ TODO: give context for warning.
+ */
if(! entry->set_lyric_align(punc.ch_C(), any_notehead_l_))
warning (_ ("lyrics found without any matching notehead"));
warning (_ ("Huh? Melismatic note found to have associated lyrics."));
SCM previous_scm = gh_cdr(v_entry);
if(previous_scm != SCM_EOL) {
- Voice_alist_entry *previous = unsmob_voice_entry(previous_scm);
+ Syllable_group *previous = unsmob_voice_entry(previous_scm);
if (previous->lyric_count())
previous->adjust_melisma_align();
}
void
-Lyric_phrasing_engraver::do_pre_move_processing ()
+Lyric_phrasing_engraver::stop_translation_timestep ()
{
for(SCM v=voice_alist_; gh_pair_p(v); v = gh_cdr(v)) {
SCM entry_scm = gh_cdar(v);
// ((alist_entry . entry_is_old) . previous_entry)
- Voice_alist_entry * entry = unsmob_voice_entry(gh_caar(entry_scm));
+ Syllable_group * entry = unsmob_voice_entry(gh_caar(entry_scm));
// set previous_entry, set entry_is_old, and resave it to alist_
// but only change if this current was not old.
if(! to_boolean(gh_cdar(entry_scm))) {
- Voice_alist_entry * previous_entry = unsmob_voice_entry(gh_cdr(entry_scm));
+ Syllable_group * previous_entry = unsmob_voice_entry(gh_cdr(entry_scm));
previous_entry->copy(entry);
entry_scm = gh_cons(gh_cons(gh_caar(entry_scm), SCM_BOOL_T), gh_cdr(entry_scm));
voice_alist_ = scm_assoc_set_x(voice_alist_, gh_caar(v), entry_scm);
-/*=========================================================================================*/
-
-/** Voice_alist_entry is a class to be smobbed and entered as data in the association list
- member of the Lyric_phrasing_engraver class.
-*/
-
-Voice_alist_entry::Voice_alist_entry()
-{
- first_in_phrase_b_=true;
- melisma_b_ = false;
- clear();
-}
-
-void
-Voice_alist_entry::clear()
-{
- notehead_l_=0;
- lyric_list_.clear();
- longest_lyric_l_=0;
- shortest_lyric_l_=0;
- melisma_b_ = false;
-}
-
-void
-Voice_alist_entry::copy( Voice_alist_entry *from)
-{
- notehead_l_ = from->notehead_l_;
- lyric_list_ = from->lyric_list_;
- longest_lyric_l_ = from->longest_lyric_l_;
- shortest_lyric_l_ = from->shortest_lyric_l_;
- melisma_b_ = from->melisma_b_;
- alignment_i_ = from->alignment_i_;
- first_in_phrase_b_ = from->first_in_phrase_b_;
-}
-
-void
-Voice_alist_entry::set_first_in_phrase(bool f)
-{
- first_in_phrase_b_ = f;
-}
-
-void
-Voice_alist_entry::set_notehead(Score_element * notehead)
-{
- if(!notehead_l_) {
- /* there should only be a single notehead, so silently ignore any extras */
- notehead_l_=notehead;
- }
-}
-
-void
-Voice_alist_entry::add_lyric(Score_element * lyric)
-{
- lyric_list_.push(lyric);
- /* record longest and shortest lyrics */
- if( longest_lyric_l_ ) {
- if(lyric->extent(X_AXIS).length() > (longest_lyric_l_->extent(X_AXIS)).length())
- longest_lyric_l_ = lyric;
- if(lyric->extent(X_AXIS).length() < (shortest_lyric_l_->extent(X_AXIS)).length())
- shortest_lyric_l_ = lyric;
- }
- else
- longest_lyric_l_ = shortest_lyric_l_ = lyric;
-}
-
-void
-Voice_alist_entry::add_extender(Score_element * extender)
-{
- if(notehead_l_ && melisma_b_) {
- dynamic_cast<Spanner*>(extender)->set_bound (RIGHT, notehead_l_);
- // should the extender finish at the right of the last note of the melisma, or the left?
- // Comments in lyric-extender.hh say left, but right looks better to me. GP.
-
- // Left:
-// extender->set_elt_property("right-trim-amount", gh_double2scm(0.0));
-
- // Right:
- Real ss = extender->paper_l ()->get_var ("staffspace");
- extender->set_elt_property("right-trim-amount",
- gh_double2scm(-notehead_l_->extent(X_AXIS).length()/ss));
- }
-}
-
-void
-Voice_alist_entry::set_melisma()
-{
- melisma_b_ = true;
-}
-
-bool
-Voice_alist_entry::set_lyric_align(const char *punc, Score_element *default_notehead_l)
-{
- if(lyric_list_.size()==0) {
- // No lyrics: nothing to do.
- return true;
- }
-
- Score_element * lyric;
- alignment_i_ = appropriate_alignment(punc);
-
- // If there was no notehead in the matching voice context, use the first
- // notehead caught from any voice context (any port in a storm).
- if(!notehead_l_) {
- notehead_l_ = default_notehead_l;
- }
- Real translation = amount_to_translate();
- for(int l = 0; l < lyric_list_.size(); l++) {
- /** set the x alignment of each lyric
- */
- lyric = lyric_list_[l];
- lyric->set_elt_property("self-alignment-X", gh_int2scm(alignment_i_));
-
- // centre on notehead ... if we have one.
- if(notehead_l_) {
- /* set the parent of each lyric to the notehead,
- set the offset callback of each lyric to centered_on_parent,
- */
- lyric->set_parent(notehead_l_, X_AXIS);
- lyric->add_offset_callback (Side_position::centered_on_parent, X_AXIS);
- /* reference is on the right of the notehead; move it left half way, then centralise */
- lyric->translate_axis (translation-(notehead_l_->extent(X_AXIS)).center(), X_AXIS);
- }
- }
- return (notehead_l_);
-}
-
-Real
-Voice_alist_entry::amount_to_translate()
-{
- Real translate = 0.0;
- if(alignment_i_ != CENTER) {
- // right or left align ...
- /* If length of longest lyric < 2 * length of shortest lyric,
- - centre longest lyric on notehead
- Otherwise
- - move so shortest lyric just reaches notehead centre
- */
- // FIXME: do we really know the lyric extent here? Some font sizing comes later?
- if((longest_lyric_l_->extent(X_AXIS)).length() <
- (shortest_lyric_l_->extent(X_AXIS)).length() * 2 )
- translate = alignment_i_*(longest_lyric_l_->extent(X_AXIS)).length()/2;
- else
- translate = alignment_i_*(shortest_lyric_l_->extent(X_AXIS)).length();
- }
- return translate;
-}
-
-
-/** determine what alignment we want.
- Rules: if first_in_phrase_b_ is set, then alignment is LEFT.
- otherwise if each syllable ends in punctuation, then alignment is RIGHT
- otherwise alignment is centre.
-*/
-int
-Voice_alist_entry::appropriate_alignment(const char *punc)
-{
- if(first_in_phrase_b_)
- return LEFT;
-
- Score_element * lyric;
- bool end_phrase = true;
- /* use a property to determine what constitutes punctuation */
-
- for(int l = 0; l < lyric_list_.size() && end_phrase; l++) {
- lyric = lyric_list_[l];
- SCM lyric_scm = lyric->get_elt_property("text");
- String lyric_str = gh_string_p(lyric_scm)?ly_scm2string(lyric_scm):"";
- char lastchar;
- if(lyric_str.length_i()>0) {
- lastchar = lyric_str[lyric_str.length_i()-1];
- /* If it doesn't end in punctuation then it ain't an end of phrase */
- if(! strchr(punc, lastchar)) {
- /* Special case: trailing space. Here examine the previous character and reverse the
- sense of the test (i.e. trailing space makes a break without punctuation, or
- suppresses a break with punctuation).
- This behaviour can be suppressed by including a space in the
- phrasingPunctuation property, in which case trailing space always means
- the same as punctuation.
-
- FIXME: The extra space throws alignment out a bit.
- */
- if(lastchar == ' ') {
- if(lyric_str.length_i()>1) {
- lastchar = lyric_str[lyric_str.length_i()-2];
- if(strchr(punc, lastchar))
- end_phrase=false;
- }
- }
- else
- end_phrase=false;
- }
- }
- }
- if(end_phrase)
- return RIGHT;
-
- return CENTER;
-}
-
-/** We don't know about the melisma until after the initial alignment work is done, so go
- back and fix the alignment when we DO know.
-*/
-void
-Voice_alist_entry::adjust_melisma_align()
-{
- if(notehead_l_) {
- // undo what we did before ...
- Real translation = -amount_to_translate();
- // melisma aligning:
- switch (alignment_i_) {
- // case LEFT: // that's all
- case CENTER: // move right so smallest lyric is left-aligned on notehead
- translation += (shortest_lyric_l_->extent(X_AXIS)).length()/2;
- break;
- case RIGHT: // move right so smallest lyric is left-aligned on notehead
- translation += (shortest_lyric_l_->extent(X_AXIS)).length();
- break;
- }
- for(int l = 0; l < lyric_list_.size(); l++) {
- lyric_list_[l]->translate_axis (translation, X_AXIS);
- }
- }
-}
-
-
-bool
-Voice_alist_entry::is_empty()
-{
- return lyric_list_.size()==0;
-}
-
-void
-Voice_alist_entry::next_lyric()
-{
- first_in_phrase_b_ = (alignment_i_ == RIGHT);
- clear();
-}
-
-/* SMOB */
-
-#include "ly-smobs.icc"
-
-SCM
-Voice_alist_entry::mark_smob (SCM)
-{
- return SCM_EOL;
-}
-
-int
-Voice_alist_entry::print_smob (SCM, SCM port, scm_print_state * )
-{
- scm_puts ("#<Voice_alist_entry>", port);
- return 1;
-}
-
-IMPLEMENT_UNSMOB(Voice_alist_entry, voice_entry);
-IMPLEMENT_SIMPLE_SMOBS(Voice_alist_entry);
-IMPLEMENT_DEFAULT_EQUAL_P(Voice_alist_entry);
-
-SCM
-Voice_alist_entry::make_entry ()
-{
- Voice_alist_entry *vi = new Voice_alist_entry;
- return vi->smobbed_self ();
-}
projects / lilypond.git / blobdiff