Add uprop_get_inpc()

2 files changed, 56 insertions, 0 deletions

diff --git a/gen/data-files b/gen/data-files
index 323b358..63e0e76 100755
--- a/gen/data-files
+++ b/gen/data-files
@@ -24,6 +24,7 @@ readonly PATHS='
 	extracted/DerivedNumericType
 	extracted/DerivedNumericValues
 	HangulSyllableType
+	IndicPositionalCategory
 	PropList
 	Scripts
 	SpecialCasing
diff --git a/gen/prop/inpc b/gen/prop/inpc
new file mode 100755
index 0000000..b4bd85c
--- /dev/null
+++ b/gen/prop/inpc
@@ -0,0 +1,55 @@
+#!/bin/sh
+
+set -e
+cd "${0%/*}/../.."
+exec >lib/unicode/prop/uprop_get_inpc.c
+
+gawk '
+BEGIN {
+	FS = "( *#.*| +; +)"
+
+	print "/* This file is autogenerated by gen/prop/inpc; 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] = "INPC_" toupper($2)
+	}
+}
+
+END {
+	print "static const struct {"
+	print "\trune lo, hi;"
+	print "\tenum uprop_inpc val;"
+	print "} lookup[] = {"
+
+	for (i = 0x900; i <= 0x10FFFF; i++) {
+		if (!props[i])
+			continue
+		lo = i
+		while (props[lo] == props[i + 1])
+			i++
+		printf "\t{RUNE_C(0x%06X), RUNE_C(0x%06X), %s},\n", lo, i, props[i]
+	}
+
+	print "};"
+	print ""
+	print "_MLIB_DEFINE_BSEARCH(enum uprop_inpc, lookup, INPC_NA)"
+	print ""
+	print "enum uprop_inpc"
+	print "uprop_get_inpc(rune ch)"
+	print "{"
+	print "\treturn ch < lookup[0].lo ? INPC_NA : mlib_lookup(ch);"
+	print "}"
+}
+' data/IndicPositionalCategory | sed 's/\s*$//'