Getting string events from the music stream

[Carl Sorensen]

I'm working on trying to make it so both the fretboard_engraver and the
tab_note_head engraver use the same algorithm to choose strings and frets
for notes.

In so doing, I'm looking at some code in the tab-note-heads engraver that
seems wrong to me.  But I'm sufficiently inexperienced with engravers that
I'd like to see if my thinking is correct.

Here is the code from lily/tab-note-heads-engraver.cc, with my questions
interspersed.  Line numbers are included for convenience if more information
is wanted.

 62 IMPLEMENT_TRANSLATOR_LISTENER (Tab_note_heads_engraver, note);
 63 void
 64 Tab_note_heads_engraver::listen_note (Stream_event *ev)
 65 {
 66   note_events_.push_back (ev);
 67 }
 68 
 69 IMPLEMENT_TRANSLATOR_LISTENER (Tab_note_heads_engraver, string_number);
 70 void 71 Tab_note_heads_engraver::listen_string_number (Stream_event *ev)
 72 {
 73   tabstring_events_.push_back (ev);
 74 }
 75 

OK, so here it seems we are setting up to listen to both note events and
string events; no problem here, it seems to make sense.  Now we move on to
process_music:



 76 void
 77 Tab_note_heads_engraver::process_music ()
 78 {
 79   vsize j = 0;
 80   for (vsize i = 0; i < note_events_.size (); i++)
 81     {
 82       SCM string_tunings = get_property ("stringTunings");
 83       int string_count = scm_ilength (string_tunings);
 84       bool high_string_one = to_boolean (get_property
("highStringOne"));
 85 
 86       Stream_event *event = note_events_[i];
 87 
 88       Stream_event *tabstring_event = 0;

We've grabbed a note_event; so far so good.

 89 
 90       for (SCM s = event->get_property ("articulations");
 91            !tabstring_event && scm_is_pair (s); s = scm_cdr (s))
 92         {
 93           Stream_event *art = unsmob_stream_event (scm_car (s));
 94 
 95           if (art->in_event_class ("string-number-event"))
 96             tabstring_event = art;
 97         }

Now we search through the articulations of the note_event to see if there's
a string event;  that makes sense.

 98 
 99       if (!tabstring_event && j < tabstring_events_.size ())
100         {
101           tabstring_event = tabstring_events_[j];
102           if (j + 1 < tabstring_events_.size ())
103             j++;
104         }

Now if we didn't find an articulation on the note event, we search through
the list of tabstring events to find a non-empty tabstring event.  Why are
we doing this second check?  Is there another way to get a tabstring event
besides adding a string as an articulation to a note?  Is this code
redundant, like I think it is?  Actually, it seems to me to be wrong,
because it loses the registration with the note event, because nothing makes
i stay the same as j.

105 
106       int string_number = 0;
107       if (tabstring_event)
108         string_number = scm_to_int (tabstring_event->get_property
("string-number"));
109 

Now that I have an event (either from the articulations list or from the
tabstring_events[] array), I can get the string number.  From here on out,
I'm fine with the code.

110       if (!string_number)

So is there something I'm missing about the code in lines 99-104.  Is it
redundant like I think it is?  In fact, can I eliminate the string-event
listener from the tab-note-head engraver?

In the automatice fret calculation routines used by the fretboard engraver,
all of the string information is extracted from the list of notes; the
string-number events are never used.

If anybody could help me see if my thoughts are correct, or identify the
errors in my thinking, I'd be grateful.

Thanks,

Carl