Some improvements in utf8tabs.py

4 files changed, 17 insertions, 19 deletions

diff --git a/examples/gauss1.c b/examples/gauss1.c
index 75d5567e..da6d73f4 100644
--- a/examples/gauss1.c
+++ b/examples/gauss1.c
@@ -1,7 +1,6 @@
 #include <time.h>
 #include <math.h>
 
-#define STC_IMPLEMENT
 #include <stc/crandom.h>
 #include <stc/cstr.h>
 
diff --git a/examples/shape.c b/examples/shape.c
index 051e1c63..e5ac0e58 100644
--- a/examples/shape.c
+++ b/examples/shape.c
@@ -4,7 +4,7 @@
 #include <stdio.h>
 #include <stc/ccommon.h>
 
-#define c_dyn_ptr(T, s) \
+#define c_dyn_cast(T, s) \
     (&T##_api == (s)->api ? (T*)(s) : (T*)0)
 
 #define c_vtable(Api, T) \
@@ -64,7 +64,7 @@ Triangle* Triangle_new(Point a, Point b, Point c)
 
 static void Triangle_draw(const Shape* shape)
 {
-    const Triangle* self = c_dyn_ptr(Triangle, shape);
+    const Triangle* self = c_dyn_cast(Triangle, shape);
     printf("Triangle : (%g,%g), (%g,%g), (%g,%g)\n",
            self->p[0].x, self->p[0].y,
            self->p[1].x, self->p[1].y,
@@ -103,7 +103,7 @@ void Polygon_addPoint(Polygon* self, Point p)
 
 static void Polygon_drop(Shape* shape)
 {
-    Polygon* self = c_dyn_ptr(Polygon, shape);
+    Polygon* self = c_dyn_cast(Polygon, shape);
     printf("poly destructed\n");
     PointVec_drop(&self->points);
     Shape_drop(shape);
@@ -111,7 +111,7 @@ static void Polygon_drop(Shape* shape)
 
 static void Polygon_draw(const Shape* shape)
 {
-    const Polygon* self = c_dyn_ptr(Polygon, shape);
+    const Polygon* self = c_dyn_cast(Polygon, shape);
     printf("Polygon  :");
     c_foreach (i, PointVec, self->points)
         printf(" (%g,%g)", i.ref->x, i.ref->y);
diff --git a/src/utf8code.c b/src/utf8code.c
index 4b55d2be..0cdcd8b6 100644
--- a/src/utf8code.c
+++ b/src/utf8code.c
@@ -105,7 +105,7 @@ int utf8_icmp_n(size_t u8max, const char* s1, const size_t n1,
         do { utf8_decode(&d2, (uint8_t)s2[j2++]); } while (d2.state);
         int c = utf8_tolower(d1.codep) - utf8_tolower(d2.codep);
         if (c || !s2[j2])
-            return c;        
+            return c;
     }
     return (j2 < n2) - (j1 < n1);
 }
diff --git a/src/utf8tabs.py b/src/utf8tabs.py
index 55be3f99..241f44af 100644
--- a/src/utf8tabs.py
+++ b/src/utf8tabs.py
@@ -60,7 +60,7 @@ def make_table(caselist):
         offset = b - a
 
         if abs(diff_a) > 2 or a - prev_a != diff_a or b - prev_b != diff_b or prev_offs != offset:
-            if  j > 0 and start_a not in [0xAB70, 0x13F8]:
+            if  j > 0 and start_a not in [0xAB70, 0x13F8]: # BUG in CaseFolding.txt V14
                 table.append([start_a, prev_a, prev_b, start_name])
             if j < n_1:
                 diff_a = caselist[j+1][0] - a
@@ -103,7 +103,7 @@ def print_index_table(name, indtab):
 def make_inverse_ind(table):
     inv = []
     for i in range(len(table)):
-        if table[i][2] not in [ord('i'), ord('s')]: # remove 'i'. 's' upcase mappings.
+        if table[i][2] not in [ord('i'), ord('s')]: # ignore 'i'. 's' upcase mappings.
             inv.append(i)
     # sort by mapped value table[:][2] (= inv) of the first element in each range entry
     inv.sort(key=lambda i: table[i][2] - (table[i][1] - table[i][0]))
@@ -124,25 +124,24 @@ def main():
     print('#include <stdint.h>\n')
     print('struct CaseMapping { uint16_t c0, c1, m1; };\n')
 
-    casemappings = compile_table('lowcase') # casefold
-    lowcase      = compile_table('lowcase', 'Lu') # unicode lowercase
+    casemappings = compile_table('lowcase') # Casefolding.txt
+    upcase       = compile_table('lowcase', 'Lu') # UnicodeData.txt uppercase
 
-    lowcase_ind = []
-    for v in lowcase:
-        try:
-            lowcase_ind.append(casemappings.index(v))
-        except:
-            lowcase_ind.append(len(casemappings))
+    # merge in additional Lu => Ll mappings from UnicodeData.txt
+    for v in upcase:
+        if v not in casemappings:
             casemappings.append(v)
 
-    casefold_len = len(casemappings)
+    # sort casemappings by uppercase values:
     casemappings.sort(key=lambda x: x[0])
-
     print_table('casemappings', casemappings, style=1)
+    
+    # index list sorted by mapped lowercase values:
     upcase_ind = make_inverse_ind(casemappings)
-    print('enum { casefold_len = %d };' % casefold_len)
+    print('enum { casefold_len = %d };' % len(casemappings))
     print_index_table('upcase_ind', upcase_ind)
 
+
 ########### main:
 
 if __name__ == "__main__":