2 7957cbd9 2021-01-27 op * Copyright (c) 2021 Omar Polo <op@omarpolo.com>
4 7957cbd9 2021-01-27 op * Permission to use, copy, modify, and distribute this software for any
5 7957cbd9 2021-01-27 op * purpose with or without fee is hereby granted, provided that the above
6 7957cbd9 2021-01-27 op * copyright notice and this permission notice appear in all copies.
8 7957cbd9 2021-01-27 op * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 7957cbd9 2021-01-27 op * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 7957cbd9 2021-01-27 op * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 7957cbd9 2021-01-27 op * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 7957cbd9 2021-01-27 op * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 7957cbd9 2021-01-27 op * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 7957cbd9 2021-01-27 op * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17 7957cbd9 2021-01-27 op #include <stddef.h>
18 7957cbd9 2021-01-27 op #include <stdint.h>
19 7957cbd9 2021-01-27 op #include <string.h>
21 7957cbd9 2021-01-27 op #include "gmid.h"
23 7957cbd9 2021-01-27 op #define BASE 36
24 7957cbd9 2021-01-27 op #define TMIN 1
25 7957cbd9 2021-01-27 op #define TMAX 26
26 7957cbd9 2021-01-27 op #define SKEW 38
27 7957cbd9 2021-01-27 op #define DAMP 700
28 7957cbd9 2021-01-27 op #define IBIAS 72
29 7957cbd9 2021-01-27 op #define IN 128
32 7957cbd9 2021-01-27 op adapt(int delta, int numpoints, int firsttime)
36 7957cbd9 2021-01-27 op if (firsttime)
37 7957cbd9 2021-01-27 op delta = delta / DAMP;
39 7957cbd9 2021-01-27 op delta = delta / 2;
41 7957cbd9 2021-01-27 op delta += (delta / numpoints);
44 7957cbd9 2021-01-27 op while (delta > ((BASE - TMIN) * TMAX) / 2) {
45 7957cbd9 2021-01-27 op delta = delta / (BASE - TMIN);
48 7957cbd9 2021-01-27 op return k + (((BASE - TMIN + 1) * delta) / (delta + SKEW));
51 7957cbd9 2021-01-27 op static const char *
52 35cf19e3 2021-01-28 op copy_label(const char *s, char *out, size_t len)
54 7957cbd9 2021-01-27 op char *end, *t;
57 7957cbd9 2021-01-27 op end = strchr(s, '\0');
62 7957cbd9 2021-01-27 op for (t = end; t >= s; --t)
63 7957cbd9 2021-01-27 op if (*t == '-')
69 7957cbd9 2021-01-27 op for (; s < t; ++s, ++out) {
78 7957cbd9 2021-01-27 op static unsigned int
79 7957cbd9 2021-01-27 op digit_value(char c)
81 7957cbd9 2021-01-27 op if ('A' <= c && c <= 'Z')
82 7957cbd9 2021-01-27 op return c - 'A';
84 7957cbd9 2021-01-27 op if ('a' <= c && c <= 'z')
85 7957cbd9 2021-01-27 op return c - 'a';
87 7957cbd9 2021-01-27 op if ('0' <= c && c <= '9')
88 7957cbd9 2021-01-27 op return 26 + c - '0';
94 4a3ab609 2021-01-29 op insert(char *out, size_t len, int codepoint, size_t i, const char **err)
99 a2fd8013 2021-01-29 op if (codepoint <= 0x7F) {
100 a2fd8013 2021-01-29 op *err = "puny: invalid decoded character (ASCII range)";
102 a2fd8013 2021-01-29 op } else if (codepoint <= 0x7FF) {
104 a2fd8013 2021-01-29 op } else if (codepoint <= 0xFFFF) {
106 a2fd8013 2021-01-29 op } else if (codepoint <= 0x10FFFF) {
109 a2fd8013 2021-01-29 op *err = "puny: invalid decoded character";
113 a2fd8013 2021-01-29 op if ((t = utf8_nth(out, i)) == NULL) {
114 a2fd8013 2021-01-29 op *err = "puny: invalid insert position";
118 a2fd8013 2021-01-29 op if (t + l >= out + len) {
119 a2fd8013 2021-01-29 op *err = "puny: insert would overflow";
123 7957cbd9 2021-01-27 op memmove(t + l, t, strlen(t));
127 7957cbd9 2021-01-27 op t[1] = ( codepoint & 0x3F) + 0x80;
128 35cf19e3 2021-01-28 op t[0] = ((codepoint >> 6) & 0x1F) + 0xC0;
131 7957cbd9 2021-01-27 op t[2] = ( codepoint & 0x3F) + 0x80;
132 35cf19e3 2021-01-28 op t[1] = ((codepoint >> 6) & 0x3F) + 0x80;
133 7957cbd9 2021-01-27 op t[0] = ((codepoint >> 12) & 0x0F) + 0xE0;
136 7957cbd9 2021-01-27 op t[3] = ( codepoint & 0x3F) + 0x80;
137 35cf19e3 2021-01-28 op t[2] = ((codepoint >> 6) & 0x3F) + 0x80;
138 7957cbd9 2021-01-27 op t[1] = ((codepoint >> 12) & 0x3F) + 0x80;
139 7957cbd9 2021-01-27 op t[0] = ((codepoint >> 18) & 0x07) + 0xF0;
146 a2fd8013 2021-01-29 op decode(const char *str, char *out, size_t len, const char **err)
150 7957cbd9 2021-01-27 op unsigned int oldi, bias, w, k, digit, t;
151 7957cbd9 2021-01-27 op unsigned int numpoints;
152 7957cbd9 2021-01-27 op const char *s;
154 44ee1bac 2021-01-27 op if (!starts_with(str, "xn--")) {
155 7957cbd9 2021-01-27 op strncpy(out, str, len);
159 7957cbd9 2021-01-27 op /* skip the xn-- */
162 7957cbd9 2021-01-27 op if (strchr(str, '-') != NULL) {
163 a2fd8013 2021-01-29 op if ((s = copy_label(str, out, len)) == NULL) {
164 a2fd8013 2021-01-29 op *err = "puny: invalid label";
167 7957cbd9 2021-01-27 op if (*s == '-')
172 7957cbd9 2021-01-27 op numpoints = strlen(out);
176 7957cbd9 2021-01-27 op bias = IBIAS;
178 7957cbd9 2021-01-27 op while (*s != '\0') {
182 7957cbd9 2021-01-27 op for (k = BASE; ; k += BASE) {
183 a2fd8013 2021-01-29 op if (*s == '\0') {
184 a2fd8013 2021-01-29 op *err = "puny: label truncated?";
187 7957cbd9 2021-01-27 op /* fail eventually? */
188 7957cbd9 2021-01-27 op digit = digit_value(*s);
191 7957cbd9 2021-01-27 op /* fail on overflow */
192 7957cbd9 2021-01-27 op i += digit * w;
194 7957cbd9 2021-01-27 op if (k <= bias)
196 7957cbd9 2021-01-27 op else if (k >= bias + TMAX)
199 7957cbd9 2021-01-27 op t = k - bias;
201 7957cbd9 2021-01-27 op if (digit < t)
203 7957cbd9 2021-01-27 op w *= (BASE - t);
206 7957cbd9 2021-01-27 op bias = adapt(i - oldi, numpoints+1, oldi == 0);
207 7957cbd9 2021-01-27 op n += i / (numpoints+1); /* fail on overflow */
208 7957cbd9 2021-01-27 op i = i % (numpoints+1);
210 4a3ab609 2021-01-29 op if (!insert(out, len, n, i, err))
219 0cd66754 2021-01-27 op static const char *
220 35cf19e3 2021-01-28 op end_of_label(const char *hostname)
222 7957cbd9 2021-01-27 op for (; *hostname != '\0' && *hostname != '.'; ++hostname)
224 7957cbd9 2021-01-27 op return hostname;
228 a2fd8013 2021-01-29 op puny_decode(const char *hostname, char *out, size_t len, const char **err)
230 35cf19e3 2021-01-28 op char label[LABEL_LEN];
231 7957cbd9 2021-01-27 op const char *s, *end;
234 7957cbd9 2021-01-27 op memset(out, 0, len);
235 44ee1bac 2021-01-27 op if (hostname == NULL)
238 7957cbd9 2021-01-27 op s = hostname;
240 415ac7a2 2021-01-28 op end = end_of_label(s);
242 a2fd8013 2021-01-29 op if (l >= sizeof(label)) {
243 a2fd8013 2021-01-29 op *err = "label too long";
247 35cf19e3 2021-01-28 op memcpy(label, s, l);
248 35cf19e3 2021-01-28 op label[l] = '\0';
250 a2fd8013 2021-01-29 op if (!decode(label, out, len, err))
253 7957cbd9 2021-01-27 op if (*end == '\0')
256 a2fd8013 2021-01-29 op if (strlcat(out, ".", len) >= len) {
257 a2fd8013 2021-01-29 op *err = "domain name too long";
261 7957cbd9 2021-01-27 op l = strlen(out);
262 a2fd8013 2021-01-29 op if (l >= len) {
263 a2fd8013 2021-01-29 op *err = "domain name too long";