#! /bin/sh -f

if make Scripts.txt
then	true
else	echo Could not acquire Unicode data file Scripts.txt
	exit 1
fi

if [ ! -f categors.sed ]
then	if make categors.sed
	then	true
	else	echo Could not build categories transformation script
		exit 1
	fi
fi


CC=${CC-cc}

(
cat <</eoc
#include <stdio.h>

long range_low;
long range_high = -2;
char * range_name = "";
char * range_category = "";

void
flush ()
{
	if (range_high >= 0) {
		printf ("	{0x%04X, 0x%04X, \"%s\", \"%s\"},\n", range_low, range_high, range_name, range_category);
	}
}

#define consider_sub_category

void
range (low, high, name, category)
	long low;
	long high;
	char * name;
	char * category;
{
	if (low == range_high + 1
		&& strcmp (name, range_name) == 0
#ifdef consider_sub_category
		&& strcmp (category, range_category) == 0
#else
		&& category [0] == range_category [0]
#endif
		) {
		range_high = high;
	} else {
		flush ();
		range_low = low;
		range_high = high;
		range_name = name;
		range_category = category;
	}
}

int
main () {
/eoc


# transform
#00F8..01BA    ; LATIN # L& [195] LATIN SMALL LETTER O WITH STROKE..LATIN SMALL LETTER EZH WITH TAIL
#01BB          ; LATIN # Lo       LATIN LETTER TWO WITH STROKE

LC_CTYPE=C egrep '^[^#]' Scripts.txt |
sed	-e '/^#/ d' \
	-e 's/^\([0-9A-F][0-9A-F]*\)\.\.\([0-9A-F][0-9A-F]*\) *; *\([^ ]*\) *# *\([^ ]*\).*/\1	{0x\1, 0x\2, "\3", "\4"},/' \
	-e 's/^\([0-9A-F][0-9A-F]*\) *; *\([^ ]*\) *# *\([^ ]*\).*/\1	{0x\1, 0x\1, "\2", "\3"},/' \
	-e 's,^\(....\)	,00\1	,' \
	-e 's,^\(.....\)	,0\1	,' \
| sort | sed -e 's,^.*	,	,' -e 's,_, ,g' |
sed -f categors.sed |
sed -e "s/{/range (/" -e "s/},/);/"


cat <</eoc
	flush ();
	return 0;
}
/eoc
) > mkscript.c

if $CC -o mkscript.exe mkscript.c
then	./mkscript.exe > scripts.t
	rm -f mkscript.c mkscript.exe
else	exit 1
fi
