mknames

#! /bin/sh -f

test -f UnicodeData.txt || make UnicodeData.txt
test -f NameAliases.txt || make NameAliases.txt


LC_ALL=C
export LC_ALL

scanunames() {
# modify UnicodeData.txt according to NameAliases.txt
sed	-e "/;abbreviation/ d" \
	-e "s,^\([0-9A-F]*\);\([^;]*\).*,/^\1;/ s/.*/\1;\2;;;;;;;;;;;;;/," \
	-e t -e d NameAliases.txt > mknames.sed

# scan UnicodeData.txt:
#	0153;LATIN SMALL LIGATURE OE;...
#	0000;<control>;Cc;0;BN;;;;;N;NULL;...
# generate lines:
#	{0x0153, "LATIN SMALL LIGATURE OE"},
#	{0x0000, "NULL"},
# and give precedence to NameAliases.txt:
sed	-f mknames.sed UnicodeData.txt |
sed	-e 's/^\([^;]*\);<control>;[^;]*;[^;]*;[^;]*;[^;]*;[^;]*;[^;]*;[^;]*;[^;]*;\([^;]*\);.*/\1;\2;/' \
	-e t \
	-e '/^[^;]*;</ d'
}

if false
then	scanunames |
	sed	-e 's/^\([^;][^;][^;][^;];\)/00\1/' \
		-e 's/^\([^;][^;][^;][^;][^;];\)/0\1/' \
		-e 's/^\([^;][^;]\)\([^;][^;]\)\([^;][^;]\);\([^;]*\);.*/  "\\x\1\\x\2\\x\3 \4",/' \
	> charnames.t
else	scanunames |
	sed	-e 's/^\([^;]*\);\([^;]*\);.*/\1 \2/' \
	> charnames.txt
fi

rm -f mknames.sed