still doesnt work

toki/toki_sitelen

@@ -1,89 +1,5 @@
 #!/bin/sh
 
-toki_ucsur "$@" | while read -r codepoint; do
-	# normalize to U+000000
-	codepoint="$(printf '%s\n' "$codepoint" | sed 's/^U+//')"
-	codepoint="U+$( \
-		dd if=/dev/zero bs=1 count=$( \
-			printf '%s\n' "$codepoint" \
-				| wc -c \
-				| xargs printf '7 - %s\n' \
-				| bc \
-			) 2>/dev/null | tr '\0' 0)$codepoint"
+set -e
 
-	codepoint_bin="$(printf '%s\n' "$codepoint" \
-		| sed \
-			-e 's/^U+//' -e 's/0/0000/g' -e 's/1/0001/g' \
-			-e 's/2/0010/g' -e 's/3/0011/g' -e 's/4/0100/g' \
-			-e 's/5/0101/g' -e 's/6/0110/g' -e 's/7/0111/g' \
-			-e 's/8/1000/g' -e 's/9/1001/g' -e 's/A/1010/g' \
-			-e 's/B/1011/g' -e 's/C/1100/g' -e 's/D/1101/g' \
-			-e 's/E/1110/g' -e 's/F/1111/g')"
-
-	printf '%s\n' "$codepoint_bin" \
-		| dd bs=17 count=1 2>/dev/null \
-		| grep 1 2>/dev/null 1>&2 \
-			|| bytes=1
-	printf '%s\n' "$codepoint_bin" \
-		| dd bs=13 count=1 2>/dev/null \
-		| grep 1 2>/dev/null 1>&2\
-			|| bytes=2
-	printf '%s\n' "$codepoint_bin" \
-		| dd bs=8 count=1 2>/dev/null \
-		| grep 1 2>/dev/null 1>&2 \
-			&& bytes=4 \
-			|| bytes=3
-
-	# TODO: How to bring bin,oct, or hex to actual binary in POSIX?
-
-	utf8_bin="$(case $bytes in \
-	1) printf '0%s\n' "$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 8)" ;; \
-	2) printf '110%s10%s\n' "$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 12 | head -c 5)" \
-		"$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 7)" ;; \
-	3) printf '1110%s10%s10%s\n' "$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 17 | head -c 4)" \
-		"$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 13 | head -c 6)" \
-		"$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 7)" ;; \
-	4) printf '11110%s10%s10%s10%s\n' "$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 22 | head -c 3)" \
-		"$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 19 | head -c 6)" \
-		"$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 13 | head -c 6)" \
-		"$(printf '%s\n' "$codepoint_bin" \
-			| tail -c 7)" ;; \
-	esac)"
-
-	utf8_oct="$(dd if=/dev/zero bs=1 count=$( \
-			printf '%s\n' "$utf8_bin" \
-				| wc -c \
-				| xargs printf '34 - %s\n' \
-				| bc \
-			) 2>/dev/null \
-		| tr '\0' 0 \
-		| xargs printf "%s$utf8_bin\n" \
-		| sed 's/.../& /g' \
-		| sed \
-			-e 's/000/0/g' -e 's/001/1/g' -e 's/010/2/g' \
-			-e 's/011/3/g' -e 's/100/4/g' -e 's/101/5/g' \
-			-e 's/110/6/g' -e 's/111/7/g' \
-		| tr -d ' ')"
-
-	# a little fucky
-	utf8_hex="$(printf '%s\n' "$utf8_bin" \
-		| sed \
-			-e 's/0000/0/g' -e 's/0001/1/g' -e 's/0010/2/g' \
-			-e 's/0011/3/g' -e 's/0100/4/g' -e 's/0101/5/g' \
-			-e 's/0110/6/g' -e 's/0111/7/g' -e 's/1000/8/g' \
-			-e 's/1001/9/g' -e 's/1010/A/g' -e 's/1011/B/g' \
-			-e 's/1100/C/g' -e 's/1101/D/g' -e 's/1110/E/g' \
-			-e 's/1111/F/g')"
-
-	printf '%s\n' "$utf8_bin"
-	shift
-done
+toki_ucsur "$@" | utf8

unicode/Makefile

@@ -0,0 +1,2 @@
+utf8: utf8.c
+	$(CC) -I../ascii -g -o utf8 utf8.c

unicode/utf8.c

@@ -0,0 +1,105 @@
+#include <stdio.h> /* fprintf(3), getc(3), stderr, stdin, EOF */
+#include <string.h> /* memset(3) */
+#include "ascii.h" /* ASCII_HEXADECIMAL_DIGITS_LOWER,
+                    * ASCII_HEXADECIMAL_DIGITS_UPPER */
+
+static char *hex = ASCII_HEXADECIMAL_DIGITS_UPPER
+	ASCII_HEXADECIMAL_DIGITS_LOWER;
+
+void print_hexascii(unsigned char *hexes, int n){
+	if(n % 2 != 0)
+		return;
+	while(n --> 0){
+		putc((char)(((hex - strchr(hex, hexes[0])) << 4)
+			+ (hex - strchr(hex, hexes[1]))), stdout);
+		++hexes;
+	}
+}
+
+int main(int argc, char *argv[]){
+	int c;
+	int i;
+	int l; /* line counter */
+	char *n;
+	unsigned char utf32_hex[8]; /* nybbles */
+	long int utf32_lit;
+	unsigned char utf8_hex[8]; /* nybbles */
+	long int utf8_lit;
+
+	c = '\0';
+	i = 0;
+	l = 1;
+	while(c != EOF){
+		memset(utf32_hex, 0, sizeof utf32_hex);
+		memset(utf8_hex, '0', sizeof utf8_hex);
+		for(
+			i = 0, n = NULL, utf32_lit = 0, utf8_lit = 0;
+			(c = getc(stdin)) != '\n'
+			&& c != EOF
+			&& i < (sizeof utf32_hex) / (sizeof *utf32_hex);
+			++i
+		){
+			if(
+				(i == 0 && c != 'U')
+				|| (i == 1 && c != '+')
+				|| i > 10
+				|| (i > 1 && (n = strchr(hex, c))
+					== NULL)
+			){
+				fprintf(stderr, "%s: %s: Syntax error.\n",
+					argv[0], l);
+				while((c = getc(stdin)) != '\n' && c != EOF);
+				i = -1;
+				break;
+			}
+			if(n != NULL){
+				n -= 16;
+				utf32_hex[i - 2] = *(n -= 16 * (n - hex > 16));
+			}else if(i >= 2)
+				utf32_hex[i - 2] = c;
+		}
+		if(i == -1 || i < 3)
+			continue;
+		while(utf32_hex[7] == '\0'){ /* slow but easy */
+			for(i = 0; i < 7; ++i)
+				utf32_hex[i + 1] = utf32_hex[i];
+			utf32_hex[0] = '0';
+		}
+		/* this code is embarrassing */
+		for(i = 0; i < 8; ++i)
+			utf32_lit = (utf32_lit << 4)
+				+ strchr(hex, utf32_hex[i]) - hex;
+		if(utf32_lit < 128){
+			utf8_hex[7] = utf32_hex[7];
+			utf8_hex[6] = utf32_hex[6];
+			i = 6;
+			goto done;
+		}else{
+			utf8_hex[7] = hex[utf32_lit & 15];
+			utf8_hex[6] = hex[((utf32_lit >> 4) & 3) + 8];
+		}
+		if(utf32_lit < 2048){
+			utf8_hex[5] = hex[(utf32_lit >> 6) & 15];
+			utf8_hex[4] = hex[((utf32_lit >> 10) & 1) + 12];
+			i = 4;
+			goto done;
+		}else{
+			utf8_hex[5] = hex[(utf32_lit >> 6) & 15];
+			utf8_hex[4] = hex[((utf32_lit >> 10) & 3) + 8];
+		}
+		if(utf32_lit < 65536){
+			utf8_hex[3] = hex[(utf32_lit >> 12) & 15];
+			utf8_hex[2] = 14;
+			i = 2;
+			goto done;
+		}else{
+			utf8_hex[3] = hex[(utf32_lit >> 12) & 15];
+			utf8_hex[2] = hex[((utf32_lit >> 16) & 3) + 8];
+			utf8_hex[1] = hex[(utf32_lit >> 21) & 3];
+			utf8_hex[0] = hex[15];
+			i = 0;
+		}
+done:		print_hexascii(utf8_hex + i, 8 - i);
+		++l;
+	}
+}