+
+ Molecule m = text2molecule (me, ly_car (text), p);
+
+ /*
+ TODO: look at padding?
+
+ Look ahead here for kern and raise.
+
+ (cols "foo" ((raise . 1) "bar"))
+ (cols "foo" ((bold (raise . 1)) "bar"))
+
+ When constructing the molecule for bar, all normal extra
+ properties found, such as bold, are used for the construction
+ of bar's molecule. But for kern or raise, it seems that we're
+ too late then, translating bar's molecule has no effect (or
+ maybe the effect of translating gets nullified when bar's
+ molecule is `added_to_edge' of the molecule for foo?)
+
+ So, while constructing foo's molecule, we look ahead for the
+ raise of bar. The HEAD of the description of bar may be a
+ single property, or a list, so we must check that too.
+ */
+
+ SCM next_p = SCM_EOL;
+ if (gh_pair_p (ly_car (text)))
+ next_p = scm_list_n (gh_call2 (f, sheet, ly_caar (text)), SCM_UNDEFINED);
+ SCM next_k = ly_assoc_chain (ly_symbol2scm ("kern"), next_p);
+ Real next_kern = kern[axis];
+ if (gh_pair_p (next_k) && gh_number_p (ly_cdr (next_k)))
+ next_kern = gh_scm2double (ly_cdr (next_k)) * staff_space;
+
+ SCM next_r = ly_assoc_chain (ly_symbol2scm ("raise"), next_p);
+ Real next_raise = 0;
+ if (gh_pair_p (next_r) && gh_number_p (ly_cdr (next_r)))
+ next_raise = gh_scm2double (ly_cdr (next_r)) * staff_space;
+
+ o[Y_AXIS] = next_raise;
+
+ if (!m.empty_b ())
+ {
+ m.translate (o);
+ if (mol.empty_b ())
+ mol = m;
+ else
+ {
+ if (axis == Y_AXIS && baseline_skip)
+ next_kern += baseline_skip - m.extent (Y_AXIS)[UP];
+ mol.add_at_edge (axis, axis == X_AXIS ? RIGHT : DOWN, m, next_kern);
+ }
+ }
+ text = ly_cdr (text);