Add Indic Conjunct Break data

gen/wcwidth.py

@@ -72,6 +72,7 @@ marks = set(emoji_skin_tone_modifiers) | flag_codepoints
 not_assigned = set(range(0, sys.maxunicode))
 property_maps: dict[str, set[int]] = defaultdict(set)
 grapheme_segmentation_maps: dict[str, set[int]] = defaultdict(set)
+incb_map: dict[str, set[int]] = defaultdict(set)
 
 
 def parse_prop_list() -> None:
@@ -264,6 +265,15 @@ def parse_grapheme_segmentation() -> None:
     for line in get_data('ucd/auxiliary/GraphemeBreakProperty.txt'):
         chars, category = split_two(line)
         grapheme_segmentation_maps[category] |= chars
+    for line in get_data('ucd/DerivedCoreProperties.txt'):
+        spec, rest = line.split(';', 1)
+        category = rest.strip().split(' ', 1)[0].strip().rstrip(';')
+        chars = parse_range_spec(spec.strip())
+        if category == 'InCB':
+            # Most InCB chars also have a GBP categorization, but not all,
+            # there exist some InCB chars that do not have a GBP category
+            subcat = rest.strip().split(';')[1].strip().split()[0].strip()
+            incb_map[subcat] |= chars
 
 
 def get_ranges(items: list[int]) -> Generator[Union[int, tuple[int, int]], None, None]:
@@ -443,46 +453,63 @@ def gen_names() -> None:
             print(cp, *words, end=end, file=f)
 
 
+def gofmt(*files: str) -> None:
+    subprocess.check_call(['gofmt', '-w', '-s'] + list(files))
+
+
 def gen_grapheme_segmentation() -> None:
     with create_header('kitty/grapheme-segmentation-data.h') as p, open('tools/wcswidth/grapheme-segmentation-data.go', 'w') as gof:
         gp = partial(print, file=gof)
-        p('typedef enum GraphemeBreakProperty {')  # }
-        p('    GBP_None,')
         gp('package wcswidth\n\n')
-        gp('type GraphemeBreakProperty uint8\n')
-        gp('const (')  # )
-        gp('None = iota')  # )
-        for category in grapheme_segmentation_maps:
-            p(f'    GBP_{category},')
-            gp(f'    {category}')
-        p('} GraphemeBreakProperty;')
-        p('')
-        gp(')')
-        gp('')
-        p('static inline GraphemeBreakProperty')
-        p('grapheme_break_property(const char_type c) {')  # }
-        p('\tswitch(c) {')  # }
-        gp('func GraphemeBreakPropertyFor(code rune) GraphemeBreakProperty {')  # }
-        gp('\tswitch code {')  # }
-        for category, codepoints in grapheme_segmentation_maps.items():
-            p(f'\t\t // {category} ({len(codepoints)} codepoints ''{{''{')
-            gp(f'\t\t // {category} ({len(codepoints)} codepoints ''{{''{')
-            for spec in get_ranges(list(codepoints)):
-                write_case(spec, p)
-                p(f'\t\t\treturn GBP_{category};')
-                write_case(spec, gp, for_go=True)
-                gp(f'\t\t\treturn {category}')
-            p('\t\t // }}''}')
+        def enum(name: str, *items: str, prefix: str = '') -> None:
+            p(f'typedef enum {name} {{')  # }}
+            gp(f'type {name} uint8\n')
+            gp('const (')  # )
+            for i, x in enumerate(items):
+                x = prefix + x
+                p(f'\t{x},')
+                if i == 0:
+                    gp(f'{x} {name} = iota')
+                else:
+                    gp(x)
+            p(f'}} {name};')
+            gp(')')
             p('')
-            gp('\t\t // }}''}')
             gp('')
-        p('\t}')  # }
-        gp('\t}')  # }
-        p('\treturn GBP_None;')  # }
-        gp('\treturn None')  # }
-        p('}')
-        gp('}')
-    subprocess.check_call(['gofmt', '-w', '-s', gof.name])
+
+        enum('GraphemeBreakProperty', 'None', *grapheme_segmentation_maps, prefix='GBP_')
+        enum('IndicConjunctBreak', 'None', *incb_map, prefix='ICB_')
+
+        def get_cat(name: str, c_func_name: str, go_func_name: str, prefix: str, m: dict[str, set[int]]) -> None:
+            p(f'static inline {name}')
+            p(f'{c_func_name}(const char_type c) {{')  # }}
+            p('\tswitch(c) {')  # }
+            gp(f'func {go_func_name}(code rune) {name} {{')  # }}
+            gp('\tswitch code {')  # }
+            for category, codepoints in m.items():
+                p(f'\t\t // {category} ({len(codepoints)} codepoints ''{{''{')
+                gp(f'\t\t // {category} ({len(codepoints)} codepoints ''{{''{')
+                category = prefix + category
+                for spec in get_ranges(list(codepoints)):
+                    write_case(spec, p)
+                    p(f'\t\t\treturn {category};')
+                    write_case(spec, gp, for_go=True)
+                    gp(f'\t\t\treturn {category}')
+                p('\t\t // }}''}')
+                p('')
+                gp('\t\t // }}''}')
+                gp('')
+            p('\t}')  # }
+            gp('\t}')  # }
+            p(f'\treturn {prefix + "None"};')  # }
+            gp(f'\treturn {prefix + "None"}')  # }
+            p('}')
+            gp('}')
+        get_cat('GraphemeBreakProperty', 'grapheme_break_property', 'GraphemeBreakPropertyFor', 'GBP_', grapheme_segmentation_maps)
+        p('')
+        gp('')
+        get_cat('IndicConjunctBreak', 'indic_conjunct_break', 'IndicConjunctBreakFor', 'ICB_', incb_map)
+    gofmt(gof.name)
 
 
 def gen_wcwidth() -> None:
@@ -549,7 +576,7 @@ def gen_wcwidth() -> None:
         p(f'#define UNICODE_MINOR_VERSION {uv[1]}')
         p(f'#define UNICODE_PATCH_VERSION {uv[2]}')
         gop('var UnicodeDatabaseVersion [3]int = [3]int{' f'{uv[0]}, {uv[1]}, {uv[2]}' + '}')
-    subprocess.check_call(['gofmt', '-w', '-s', gof.name])
+    gofmt(gof.name)
 
 
 def gen_rowcolumn_diacritics() -> None:
@@ -595,7 +622,7 @@ def gen_rowcolumn_diacritics() -> None:
         p('\t}')
         p('\treturn 0;')
         p('}')
-    subprocess.check_call(['gofmt', '-w', '-s', go_file])
+    gofmt(go_file)
 
 
 def main(args: list[str]=sys.argv) -> None: