dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley#ifndef lint
0c27b3fe77ac1d5094ba3521e8142d9e7973133fMark Andrewsstatic char *rcsid = "$Id: selectiveencode.c,v 1.1 2003/06/04 00:27:07 marka Exp $";
40f53fa8d9c6a4fc38c0014495e7a42b08f52481David Lawrence#endif
0c27b3fe77ac1d5094ba3521e8142d9e7973133fMark Andrews
0c27b3fe77ac1d5094ba3521e8142d9e7973133fMark Andrews/*
0c27b3fe77ac1d5094ba3521e8142d9e7973133fMark Andrews * Copyright (c) 2000,2002 Japan Network Information Center.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * All rights reserved.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *
70e5a7403f0e0a3bd292b8287c5fed5772c15270Automatic Updater * By using this file, you agree to the terms and conditions set forth bellow.
ab023a65562e62b85a824509d829b6fad87e00b1Rob Austein *
ab023a65562e62b85a824509d829b6fad87e00b1Rob Austein * LICENSE TERMS AND CONDITIONS
9c3531d72aeaad6c5f01efe6a1c82023e1379e4dDavid Lawrence *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * The following License Terms and Conditions apply, unless a different
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * license is obtained from Japan Network Information Center ("JPNIC"),
364a82f7c25b62967678027043425201a5e5171aBob Halley * a Japanese association, Kokusai-Kougyou-Kanda Bldg 6F, 2-3-4 Uchi-Kanda,
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * Chiyoda-ku, Tokyo 101-0047, Japan.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * 1. Use, Modification and Redistribution (including distribution of any
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * modified or derived work) in source and/or binary forms is permitted
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * under this License Terms and Conditions.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * 2. Redistribution of source code must retain the copyright notices as they
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * appear in each source code file, this License Terms and Conditions.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * 3. Redistribution in binary form must reproduce the Copyright Notice,
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * this License Terms and Conditions, in the documentation and/or other
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * materials provided with the distribution. For the purposes of binary
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * distribution the "Copyright Notice" refers to the following language:
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * "Copyright (c) 2000-2002 Japan Network Information Center. All rights reserved."
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * 4. The name of JPNIC may not be used to endorse or promote products
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * derived from this Software without specific prior written approval of
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * JPNIC.
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * 5. Disclaimer/Limitation of Liability: THIS SOFTWARE IS PROVIDED BY JPNIC
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL JPNIC BE LIABLE
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff * ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley */
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include <config.h>
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include <stddef.h>
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include <idn/assert.h>
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include <idn/logmacro.h>
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include <idn/result.h>
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include <idn/utf8.h>
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff#include <idn/debug.h>
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley#include "selectiveencode.h"
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halleystatic int is_domain_delimiter(char c);
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halleystatic char *find_nonascii(const char *s);
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halleyidn_result_t
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halleyidn_selectiveencode_findregion(const char *s,
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley char **startp, char **endp)
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley{
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff char *non_ascii;
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley char *start, *end;
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley assert(s != NULL && startp != NULL && endp != NULL);
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley TRACE(("idn_selectiveencode_findregion(s=\"%s\")\n",
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley idn__debug_xstring(s, 20)));
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley /*
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley * Scan the specified string looking for non-ascii character.
065273f52a2da0caa45123c70c6eeef4a7be7337Bob Halley */
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff if ((non_ascii = find_nonascii(s)) == NULL)
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley return (idn_notfound);
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley
7a9ded7b5fdaf79dae5469f8560bf4a237f8fbefBob Halley /*
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * Non-ascii character found.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * Determine the region to encode.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley */
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley /*
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * First, we scan backwards to find the beginning of the region
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * that should be converted.
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff */
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley start = non_ascii;
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley while (start > s) {
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley char *prev = idn_utf8_findfirstbyte(start - 1, s);
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley if (is_domain_delimiter(*prev))
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley break; /* Found */
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley start = prev;
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley }
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *startp = start;
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley /*
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley * Next we scan forwards looking for the end of the region.
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley */
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley end = non_ascii + idn_utf8_mblen(non_ascii);
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley while (!is_domain_delimiter(*end))
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff end += idn_utf8_mblen(end);
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley *endp = end;
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley return (idn_success);
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley}
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halleystatic int
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halleyis_domain_delimiter(char c) {
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley return ((unsigned char)c < 0x80 &&
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley !('A' <= c && c <= 'Z') &&
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley !('a' <= c && c <= 'z') &&
3ddd814a97de1d152ba0913c592d6e6dc83d38a6Michael Graff !('0' <= c && c <= '9') &&
d176cbca5b6e1f77345809944365a82d3550c4bcBob Halley c != '-' && c != '.');
d176cbca5b6e1f77345809944365a82d3550c4bcBob Halley}
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halleystatic char *
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halleyfind_nonascii(const char *s) {
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley while (*s != '\0' && (unsigned char)*s < 0x80)
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley s++;
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley if (*s == '\0')
d176cbca5b6e1f77345809944365a82d3550c4bcBob Halley return (NULL);
40f53fa8d9c6a4fc38c0014495e7a42b08f52481David Lawrence else
d176cbca5b6e1f77345809944365a82d3550c4bcBob Halley return ((char *)s);
dbef4f7fb77908f76dc5e4d9e20c184388d4610aBob Halley}
31d3e8d2fec874c271ace9666c14c32a4dbaf328David Lawrence