|
Packit |
b89d10 |
/**********************************************************************
|
|
Packit |
b89d10 |
regext.c - Oniguruma (regular expression library)
|
|
Packit |
b89d10 |
**********************************************************************/
|
|
Packit |
b89d10 |
/*-
|
|
Packit |
b89d10 |
* Copyright (c) 2002-2017 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>
|
|
Packit |
b89d10 |
* All rights reserved.
|
|
Packit |
b89d10 |
*
|
|
Packit |
b89d10 |
* Redistribution and use in source and binary forms, with or without
|
|
Packit |
b89d10 |
* modification, are permitted provided that the following conditions
|
|
Packit |
b89d10 |
* are met:
|
|
Packit |
b89d10 |
* 1. Redistributions of source code must retain the above copyright
|
|
Packit |
b89d10 |
* notice, this list of conditions and the following disclaimer.
|
|
Packit |
b89d10 |
* 2. Redistributions in binary form must reproduce the above copyright
|
|
Packit |
b89d10 |
* notice, this list of conditions and the following disclaimer in the
|
|
Packit |
b89d10 |
* documentation and/or other materials provided with the distribution.
|
|
Packit |
b89d10 |
*
|
|
Packit |
b89d10 |
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
Packit |
b89d10 |
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
Packit |
b89d10 |
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
Packit |
b89d10 |
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
Packit |
b89d10 |
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
Packit |
b89d10 |
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
Packit |
b89d10 |
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
Packit |
b89d10 |
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
Packit |
b89d10 |
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
Packit |
b89d10 |
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
Packit |
b89d10 |
* SUCH DAMAGE.
|
|
Packit |
b89d10 |
*/
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
#include "regint.h"
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static void
|
|
Packit |
b89d10 |
conv_ext0be32(const UChar* s, const UChar* end, UChar* conv)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
while (s < end) {
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
*conv++ = *s++;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static void
|
|
Packit |
b89d10 |
conv_ext0le32(const UChar* s, const UChar* end, UChar* conv)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
while (s < end) {
|
|
Packit |
b89d10 |
*conv++ = *s++;
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static void
|
|
Packit |
b89d10 |
conv_ext0be(const UChar* s, const UChar* end, UChar* conv)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
while (s < end) {
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
*conv++ = *s++;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static void
|
|
Packit |
b89d10 |
conv_ext0le(const UChar* s, const UChar* end, UChar* conv)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
while (s < end) {
|
|
Packit |
b89d10 |
*conv++ = *s++;
|
|
Packit |
b89d10 |
*conv++ = '\0';
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static void
|
|
Packit |
b89d10 |
conv_swap4bytes(const UChar* s, const UChar* end, UChar* conv)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
while (s < end) {
|
|
Packit |
b89d10 |
*conv++ = s[3];
|
|
Packit |
b89d10 |
*conv++ = s[2];
|
|
Packit |
b89d10 |
*conv++ = s[1];
|
|
Packit |
b89d10 |
*conv++ = s[0];
|
|
Packit |
b89d10 |
s += 4;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static void
|
|
Packit |
b89d10 |
conv_swap2bytes(const UChar* s, const UChar* end, UChar* conv)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
while (s < end) {
|
|
Packit |
b89d10 |
*conv++ = s[1];
|
|
Packit |
b89d10 |
*conv++ = s[0];
|
|
Packit |
b89d10 |
s += 2;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
static int
|
|
Packit |
b89d10 |
conv_encoding(OnigEncoding from, OnigEncoding to, const UChar* s, const UChar* end,
|
|
Packit |
b89d10 |
UChar** conv, UChar** conv_end)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
int len = (int )(end - s);
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
if (to == ONIG_ENCODING_UTF16_BE) {
|
|
Packit |
b89d10 |
if (from == ONIG_ENCODING_ASCII || from == ONIG_ENCODING_ISO_8859_1) {
|
|
Packit |
b89d10 |
*conv = (UChar* )xmalloc(len * 2);
|
|
Packit |
b89d10 |
CHECK_NULL_RETURN_MEMERR(*conv);
|
|
Packit |
b89d10 |
*conv_end = *conv + (len * 2);
|
|
Packit |
b89d10 |
conv_ext0be(s, end, *conv);
|
|
Packit |
b89d10 |
return 0;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else if (from == ONIG_ENCODING_UTF16_LE) {
|
|
Packit |
b89d10 |
swap16:
|
|
Packit |
b89d10 |
*conv = (UChar* )xmalloc(len);
|
|
Packit |
b89d10 |
CHECK_NULL_RETURN_MEMERR(*conv);
|
|
Packit |
b89d10 |
*conv_end = *conv + len;
|
|
Packit |
b89d10 |
conv_swap2bytes(s, end, *conv);
|
|
Packit |
b89d10 |
return 0;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else if (to == ONIG_ENCODING_UTF16_LE) {
|
|
Packit |
b89d10 |
if (from == ONIG_ENCODING_ASCII || from == ONIG_ENCODING_ISO_8859_1) {
|
|
Packit |
b89d10 |
*conv = (UChar* )xmalloc(len * 2);
|
|
Packit |
b89d10 |
CHECK_NULL_RETURN_MEMERR(*conv);
|
|
Packit |
b89d10 |
*conv_end = *conv + (len * 2);
|
|
Packit |
b89d10 |
conv_ext0le(s, end, *conv);
|
|
Packit |
b89d10 |
return 0;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else if (from == ONIG_ENCODING_UTF16_BE) {
|
|
Packit |
b89d10 |
goto swap16;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
if (to == ONIG_ENCODING_UTF32_BE) {
|
|
Packit |
b89d10 |
if (from == ONIG_ENCODING_ASCII || from == ONIG_ENCODING_ISO_8859_1) {
|
|
Packit |
b89d10 |
*conv = (UChar* )xmalloc(len * 4);
|
|
Packit |
b89d10 |
CHECK_NULL_RETURN_MEMERR(*conv);
|
|
Packit |
b89d10 |
*conv_end = *conv + (len * 4);
|
|
Packit |
b89d10 |
conv_ext0be32(s, end, *conv);
|
|
Packit |
b89d10 |
return 0;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else if (from == ONIG_ENCODING_UTF32_LE) {
|
|
Packit |
b89d10 |
swap32:
|
|
Packit |
b89d10 |
*conv = (UChar* )xmalloc(len);
|
|
Packit |
b89d10 |
CHECK_NULL_RETURN_MEMERR(*conv);
|
|
Packit |
b89d10 |
*conv_end = *conv + len;
|
|
Packit |
b89d10 |
conv_swap4bytes(s, end, *conv);
|
|
Packit |
b89d10 |
return 0;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else if (to == ONIG_ENCODING_UTF32_LE) {
|
|
Packit |
b89d10 |
if (from == ONIG_ENCODING_ASCII || from == ONIG_ENCODING_ISO_8859_1) {
|
|
Packit |
b89d10 |
*conv = (UChar* )xmalloc(len * 4);
|
|
Packit |
b89d10 |
CHECK_NULL_RETURN_MEMERR(*conv);
|
|
Packit |
b89d10 |
*conv_end = *conv + (len * 4);
|
|
Packit |
b89d10 |
conv_ext0le32(s, end, *conv);
|
|
Packit |
b89d10 |
return 0;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else if (from == ONIG_ENCODING_UTF32_BE) {
|
|
Packit |
b89d10 |
goto swap32;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
return ONIGERR_NOT_SUPPORTED_ENCODING_COMBINATION;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
extern int
|
|
Packit |
b89d10 |
onig_new_deluxe(regex_t** reg, const UChar* pattern, const UChar* pattern_end,
|
|
Packit |
b89d10 |
OnigCompileInfo* ci, OnigErrorInfo* einfo)
|
|
Packit |
b89d10 |
{
|
|
Packit |
b89d10 |
int r;
|
|
Packit |
b89d10 |
UChar *cpat, *cpat_end;
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
if (IS_NOT_NULL(einfo)) einfo->par = (UChar* )NULL;
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
if (ci->pattern_enc != ci->target_enc) {
|
|
Packit |
b89d10 |
r = conv_encoding(ci->pattern_enc, ci->target_enc, pattern, pattern_end,
|
|
Packit |
b89d10 |
&cpat, &cpat_end);
|
|
Packit |
b89d10 |
if (r != 0) return r;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
else {
|
|
Packit |
b89d10 |
cpat = (UChar* )pattern;
|
|
Packit |
b89d10 |
cpat_end = (UChar* )pattern_end;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
*reg = (regex_t* )xmalloc(sizeof(regex_t));
|
|
Packit |
b89d10 |
if (IS_NULL(*reg)) {
|
|
Packit |
b89d10 |
r = ONIGERR_MEMORY;
|
|
Packit |
b89d10 |
goto err2;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
r = onig_reg_init(*reg, ci->option, ci->case_fold_flag, ci->target_enc,
|
|
Packit |
b89d10 |
ci->syntax);
|
|
Packit |
b89d10 |
if (r != 0) goto err;
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
r = onig_compile(*reg, cpat, cpat_end, einfo);
|
|
Packit |
b89d10 |
if (r != 0) {
|
|
Packit |
b89d10 |
err:
|
|
Packit |
b89d10 |
onig_free(*reg);
|
|
Packit |
b89d10 |
*reg = NULL;
|
|
Packit |
b89d10 |
}
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
err2:
|
|
Packit |
b89d10 |
if (cpat != pattern) xfree(cpat);
|
|
Packit |
b89d10 |
|
|
Packit |
b89d10 |
return r;
|
|
Packit |
b89d10 |
}
|