Add uprop_get_hst()

2 files changed, 56 insertions, 0 deletions

diff --git a/gen/data-files b/gen/data-files
index 156e4d5..1499eea 100755
--- a/gen/data-files
+++ b/gen/data-files
@@ -23,6 +23,7 @@ extracted/DerivedEastAsianWidth
 extracted/DerivedLineBreak
 extracted/DerivedNumericType
 extracted/DerivedNumericValues
+HangulSyllableType
 PropList
 Scripts
 SpecialCasing
diff --git a/gen/prop/hst b/gen/prop/hst
new file mode 100755
index 0000000..f1af5df
--- /dev/null
+++ b/gen/prop/hst
@@ -0,0 +1,55 @@
+#!/bin/sh
+
+set -e
+cd "${0%/*}/../.."
+exec >lib/unicode/prop/uprop_get_hst.c
+
+gawk '
+BEGIN {
+	FS = "( *#.*| +; +)"
+
+	print "/* This file is autogenerated by gen/prop/hst; DO NOT EDIT. */"
+	print ""
+	print "#include \"_bsearch.h\""
+	print "#include \"rune.h\""
+	print "#include \"unicode/prop.h\""
+	print ""
+}
+
+/^[^#]/ {
+	n = split($1, a, /\.\./)
+	lo = strtonum("0X" a[1])
+	hi = strtonum("0X" a[n])
+
+	for (i = lo; i <= hi; i++) {
+		gsub(/^; /, "", $2)
+		props[i] = "HST_" toupper($2)
+	}
+}
+
+END {
+	print "static const struct {"
+	print "\trune lo, hi;"
+	print "\tenum uprop_hst val;"
+	print "} lookup[] = {"
+
+	for (i = 0x1100; i <= 0x10FFFF; i++) {
+		if (!props[i])
+			continue
+		lo = i
+		while (props[lo] == props[i + 1])
+			i++
+		printf "\t{RUNE_C(0x%06X), RUNE_C(0x%06X), %s},\n", lo, i, props[i]
+	}
+
+	print "};"
+	print ""
+	print "_MLIB_DEFINE_BSEARCH(enum uprop_hst, lookup, HST_NA)"
+	print ""
+	print "enum uprop_hst"
+	print "uprop_get_hst(rune ch)"
+	print "{"
+	print "\treturn ch < RUNE_C(0x1100) ? HST_NA : mlib_lookup(ch);"
+	print "}"
+}
+' data/HangulSyllableType | sed 's/\s*$//'