Add uprop_is_ce() and uprop_is_comp_ex()

3 files changed, 29 insertions, 9 deletions

diff --git a/gen/data-files b/gen/data-files
index ac984b9..c648868 100755
--- a/gen/data-files
+++ b/gen/data-files
@@ -14,6 +14,7 @@ readonly PATHS='
 	BidiMirroring
 	Blocks
 	CaseFolding
+	CompositionExclusions
 	DerivedAge
 	DerivedCoreProperties
 	DerivedNormalizationProps
diff --git a/gen/prop/bool-props b/gen/prop/bool-props
index a79f858..010dbf8 100755
--- a/gen/prop/bool-props
+++ b/gen/prop/bool-props
@@ -62,12 +62,17 @@ extpict=Extended_Pictographic
 
 props4='
 cwkcf=Changes_When_NFKC_Casefolded
+comp_ex=Full_Composition_Exclusion
 '
 
 props5='
 bidi_m=Bidi_Mirrored
 '
 
+props6='
+ce=Composition_Exclusion
+'
+
 manual='
 idst=IDS_Trinary_Operator
 idsu=IDS_Unary_Operator
@@ -87,16 +92,18 @@ gen()
 	IFS='
 '
 	parallel-moreutils -- $(
-		for prop in $props1; do gen $prop PropList & done
-		for prop in $props2; do gen $prop DerivedCoreProperties & done
-		for prop in $props3; do gen $prop emoji-data & done
+		for prop in $props1; do gen $prop PropList                  & done
+		for prop in $props2; do gen $prop DerivedCoreProperties     & done
+		for prop in $props3; do gen $prop emoji-data                & done
 		for prop in $props4; do gen $prop DerivedNormalizationProps & done
-		for prop in $props5; do gen $prop DerivedBinaryProperties & done
+		for prop in $props5; do gen $prop DerivedBinaryProperties   & done
+		for prop in $props6; do gen $prop CompositionExclusions     & done
 	)
 )
 
 printf '[[_mlib_pure]] bool uprop_is_%s(rune);\n' \
-	$(printf '%s\n' $props1 $props2 $props3 $props4 $props5 | cut -d= -f1) \
+	$(printf '%s\n' $props1 $props2 $props3 $props4 $props5 $props6 \
+	  | cut -d= -f1) \
 | gawk '
 	/PROP PREDICATES END/ { no = 0 }
 	FILENAME != "-" && !no { print }
diff --git a/gen/prop/bool-props.py b/gen/prop/bool-props.py
index 62cfe74..5ac0287 100755
--- a/gen/prop/bool-props.py
+++ b/gen/prop/bool-props.py
@@ -8,17 +8,29 @@ from lib import *
 
 def parse(file: str) -> list[bool]:
 	xs = [False] * 0x110000
-	if sys.argv[1] == 'Indic_Conjunct_Break':
-		sys.argv[1] = 'InCB;'
+
+	sys.argv[1] = ({
+		'Composition_Exclusion': '',
+		'Indic_Conjunct_Break': 'InCB;',
+	}).get(sys.argv[1], sys.argv[1])
+
 	with open(file, 'r') as f:
 		for line in f.readlines():
 			if (
-				len(line) == 0
+				len(line.strip()) == 0
 				or line[0] == '#'
 				or sys.argv[1] not in line
 			):
 				continue
-			parts = [int(x, 16) for x in line.split(';')[0].strip().split('..')]
+			parts = [
+				int(x, 16) for x in (
+					line
+						.split('#')[0]
+						.split(';')[0]
+						.strip()
+						.split('..')
+				)
+			]
 			for i in range(parts[0], parts[len(parts) - 1] + 1):
 				xs[i] = True
 	return xs