/*
* Copyright (C) 2012, 2013 Eric Biggers
*
- * This file is part of wimlib, a library for working with WIM files.
+ * This file is free software; you can redistribute it and/or modify it under
+ * the terms of the GNU Lesser General Public License as published by the Free
+ * Software Foundation; either version 3 of the License, or (at your option) any
+ * later version.
*
- * wimlib is free software; you can redistribute it and/or modify it under the
- * terms of the GNU General Public License as published by the Free
- * Software Foundation; either version 3 of the License, or (at your option)
- * any later version.
- *
- * wimlib is distributed in the hope that it will be useful, but WITHOUT ANY
- * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
- * A PARTICULAR PURPOSE. See the GNU General Public License for more
+ * This file is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
* details.
*
- * You should have received a copy of the GNU General Public License
- * along with wimlib; if not, see http://www.gnu.org/licenses/.
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this file; if not, see http://www.gnu.org/licenses/.
*/
#ifdef HAVE_CONFIG_H
#endif
#include "wimlib.h"
+#include "wimlib/assert.h"
#include "wimlib/encoding.h"
#include "wimlib/endianness.h"
#include "wimlib/error.h"
#include <stdlib.h>
#include <string.h>
+#ifdef HAVE_ALLOCA_H
+# include <alloca.h>
+#endif
+
bool wimlib_mbs_is_utf8 = !TCHAR_IS_UTF16LE;
/* List of iconv_t conversion descriptors for a specific character conversion.
struct iconv_list_head name = { \
.from_encoding = from, \
.to_encoding = to, \
- .list = LIST_HEAD_INIT(name.list), \
- .mutex = PTHREAD_MUTEX_INITIALIZER, \
}
static iconv_t *
errno = errno_save;
}
-/* Prevent printing an error message if a character conversion error occurs
- * while printing an error message. (This variable is not per-thread but it
- * doesn't matter too much since it's just the error messages.) */
-static bool error_message_being_printed = false;
-
#define DEFINE_CHAR_CONVERSION_FUNCTIONS(varname1, longname1, chartype1,\
varname2, longname2, chartype2,\
earlyreturn_on_utf8_locale, \
if (cd == NULL) \
return WIMLIB_ERR_ICONV_NOT_AVAILABLE; \
\
+ chartype2 *buf; \
+ size_t bufsize; \
+ bool buf_onheap; \
+ bufsize = (worst_case_len_expr) * sizeof(chartype2); \
/* Worst case length */ \
- chartype2 buf[worst_case_len_expr]; \
+ if (bufsize <= STACK_MAX) { \
+ buf = alloca(bufsize); \
+ buf_onheap = false; \
+ } else { \
+ buf = MALLOC(bufsize); \
+ if (!buf) \
+ return WIMLIB_ERR_NOMEM; \
+ buf_onheap = true; \
+ } \
+ \
char *inbuf = (char*)in; \
size_t inbytesleft = in_nbytes; \
char *outbuf = (char*)buf; \
- size_t outbytesleft = sizeof(buf); \
+ size_t outbytesleft = bufsize; \
size_t len; \
int ret; \
\
len = iconv(*cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft); \
if (len == (size_t)-1) { \
- if (!error_message_being_printed) { \
- error_message_being_printed = true; \
- err_msg; \
- error_message_being_printed = false; \
- } \
+ err_msg; \
ret = err_return; \
} else { \
- *out_nbytes_ret = sizeof(buf) - outbytesleft; \
+ *out_nbytes_ret = bufsize - outbytesleft; \
ret = 0; \
} \
put_iconv(cd); \
+ if (buf_onheap) \
+ FREE(buf); \
return ret; \
} \
\
\
len = iconv(*cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft); \
if (len == (size_t)-1) { \
- if (!error_message_being_printed) { \
- error_message_being_printed = true; \
- err_msg; \
- error_message_being_printed = false; \
- } \
+ err_msg; \
ret = err_return; \
} else { \
out[(LARGE_NUMBER-outbytesleft)/sizeof(chartype2)] = 0; \
return utf8_to_tstr(utf8str, strlen(utf8str), out, &out_nbytes);
}
+static void
+iconv_init(struct iconv_list_head *head)
+{
+ pthread_mutex_init(&head->mutex, NULL);
+ INIT_LIST_HEAD(&head->list);
+}
+
static void
iconv_cleanup(struct iconv_list_head *head)
{
}
}
+void
+iconv_global_init(void)
+{
+ iconv_init(&iconv_utf8_to_tstr);
+ iconv_init(&iconv_tstr_to_utf8);
+#if !TCHAR_IS_UTF16LE
+ iconv_init(&iconv_utf16le_to_tstr);
+ iconv_init(&iconv_tstr_to_utf16le);
+ iconv_init(&iconv_utf16le_to_utf8);
+ iconv_init(&iconv_utf8_to_utf16le);
+#endif
+}
+
void
iconv_global_cleanup(void)
{
#endif
}
-/* Upper case table --- Borrowed from NTFS-3g
+/* A table that maps from UCS-2 characters to their upper case equivalents.
+ * Index and array values are both CPU endian.
+ * Note: this is only an *approximation* of real UTF-16 case folding.
+ */
+static u16 upcase[65536];
+
+/*
+ * Initialize the 'upcase' table.
+ *
+ * Code modified from NTFS-3g, original copyright notices below:
*
* Copyright (c) 2000-2004 Anton Altaparmakov
* Copyright (c) 2002-2009 Szabolcs Szakacsits
* Copyright (c) 2008-2011 Jean-Pierre Andre
* Copyright (c) 2008 Bernhard Kaindl
*
- * License is GPLv2 or later.
- */
-
-/**
- * ntfs_upcase_table_build - build the default upcase table for NTFS
- * @uc: destination buffer where to store the built table
- * @uc_len: size of destination buffer in bytes
- *
- * ntfs_upcase_table_build() builds the default upcase table for NTFS and
- * stores it in the caller supplied buffer @uc of size @uc_len.
+ * License GPLv2 or later.
*
- * Note, @uc_len must be at least 128kiB in size or bad things will happen!
+ * The expanded table itself is the same as that used by Windows Vista for
+ * NTFS case insensitivity.
*/
-static void
-ntfs_upcase_table_build(utf16lechar *uc, u32 uc_len)
+void
+init_upcase(void)
{
/*
* This is the table as defined by Vista
* "Start" is inclusive and "End" is exclusive, every value has the
* value of "Add" added to it.
*/
- static int uc_run_table[][3] = { /* Start, End, Add */
+ static const u16 uc_run_table[][3] = { /* Start, End, Add */
{0x0061, 0x007b, -32}, {0x00e0, 0x00f7, -32}, {0x00f8, 0x00ff, -32},
{0x0256, 0x0258, -205}, {0x028a, 0x028c, -217}, {0x037b, 0x037e, 130},
{0x03ac, 0x03ad, -38}, {0x03ad, 0x03b0, -37}, {0x03b1, 0x03c2, -32},
{0x1fb3, 0x1fb4, 9}, {0x1fcc, 0x1fcd, -9}, {0x1fd0, 0x1fd2, 8},
{0x1fe0, 0x1fe2, 8}, {0x1fe5, 0x1fe6, 7}, {0x1ffc, 0x1ffd, -9},
{0x2170, 0x2180, -16}, {0x24d0, 0x24ea, -26}, {0x2c30, 0x2c5f, -48},
- {0x2d00, 0x2d26, -7264}, {0xff41, 0xff5b, -32}, {0}
+ {0x2d00, 0x2d26, -7264}, {0xff41, 0xff5b, -32},
};
/*
* "Start" is exclusive and "End" is inclusive, every second value is
* decremented by one.
*/
- static int uc_dup_table[][2] = { /* Start, End */
+ static const u16 uc_dup_table[][2] = { /* Start, End */
{0x0100, 0x012f}, {0x0132, 0x0137}, {0x0139, 0x0149}, {0x014a, 0x0178},
{0x0179, 0x017e}, {0x01a0, 0x01a6}, {0x01b3, 0x01b7}, {0x01cd, 0x01dd},
{0x01de, 0x01ef}, {0x01f4, 0x01f5}, {0x01f8, 0x01f9}, {0x01fa, 0x0220},
{0x048a, 0x04bf}, {0x04c1, 0x04c4}, {0x04c5, 0x04c8}, {0x04c9, 0x04ce},
{0x04ec, 0x04ed}, {0x04d0, 0x04eb}, {0x04ee, 0x04f5}, {0x04f6, 0x0513},
{0x1e00, 0x1e95}, {0x1ea0, 0x1ef9}, {0x2183, 0x2184}, {0x2c60, 0x2c61},
- {0x2c67, 0x2c6c}, {0x2c75, 0x2c76}, {0x2c80, 0x2ce3}, {0}
+ {0x2c67, 0x2c6c}, {0x2c75, 0x2c76}, {0x2c80, 0x2ce3},
};
/*
* Set the Unicode character at offset "Offset" to "Value". Note,
* "Value" is host endian.
*/
- static int uc_byte_table[][2] = { /* Offset, Value */
+ static const u16 uc_byte_table[][2] = { /* Offset, Value */
{0x00ff, 0x0178}, {0x0180, 0x0243}, {0x0183, 0x0182}, {0x0185, 0x0184},
{0x0188, 0x0187}, {0x018c, 0x018b}, {0x0192, 0x0191}, {0x0195, 0x01f6},
{0x0199, 0x0198}, {0x019a, 0x023d}, {0x019e, 0x0220}, {0x01a8, 0x01a7},
{0x026b, 0x2c62}, {0x026f, 0x019c}, {0x0272, 0x019d}, {0x0275, 0x019f},
{0x027d, 0x2c64}, {0x0280, 0x01a6}, {0x0283, 0x01a9}, {0x0288, 0x01ae},
{0x0289, 0x0244}, {0x028c, 0x0245}, {0x0292, 0x01b7}, {0x03f2, 0x03f9},
- {0x04cf, 0x04c0}, {0x1d7d, 0x2c63}, {0x214e, 0x2132}, {0}
+ {0x04cf, 0x04c0}, {0x1d7d, 0x2c63}, {0x214e, 0x2132},
};
- int i, r;
- int k, off;
-
- memset((char*)uc, 0, uc_len);
- uc_len >>= 1;
- if (uc_len > 65536)
- uc_len = 65536;
- for (i = 0; (u32)i < uc_len; i++)
- uc[i] = cpu_to_le16(i);
- for (r = 0; uc_run_table[r][0]; r++) {
- off = uc_run_table[r][2];
- for (i = uc_run_table[r][0]; i < uc_run_table[r][1]; i++)
- uc[i] = cpu_to_le16(i + off);
- }
- for (r = 0; uc_dup_table[r][0]; r++)
- for (i = uc_dup_table[r][0]; i < uc_dup_table[r][1]; i += 2)
- uc[i + 1] = cpu_to_le16(i);
- for (r = 0; uc_byte_table[r][0]; r++) {
- k = uc_byte_table[r][1];
- uc[uc_byte_table[r][0]] = cpu_to_le16(k);
- }
-}
-
-static utf16lechar upcase[65536];
-void
-init_upcase(void)
-{
- ntfs_upcase_table_build(upcase, ARRAY_LEN(upcase));
+ for (u32 i = 0; i < 65536; i++)
+ upcase[i] = i;
+
+ for (u32 r = 0; r < ARRAY_LEN(uc_run_table); r++)
+ for (u32 i = uc_run_table[r][0]; i < uc_run_table[r][1]; i++)
+ upcase[i] = i + uc_run_table[r][2];
+
+ for (u32 r = 0; r < ARRAY_LEN(uc_dup_table); r++)
+ for (u32 i = uc_dup_table[r][0]; i < uc_dup_table[r][1]; i += 2)
+ upcase[i + 1] = i;
+
+ for (u32 r = 0; r < ARRAY_LEN(uc_byte_table); r++)
+ upcase[uc_byte_table[r][0]] = uc_byte_table[r][1];
+
+#if 0
+ /* Sanity checks */
+ wimlib_assert(upcase['a'] == 'A');
+ wimlib_assert(upcase['A'] == 'A');
+ wimlib_assert(upcase['z'] == 'Z');
+ wimlib_assert(upcase['Z'] == 'Z');
+ wimlib_assert(upcase['1'] == '1');
+ wimlib_assert(upcase[0x00e9] == 0x00c9); /* Latin letter e, with acute accent */
+ wimlib_assert(upcase[0x00c9] == 0x00c9);
+ wimlib_assert(upcase[0x03c1] == 0x03a1); /* Greek letter rho */
+ wimlib_assert(upcase[0x03a1] == 0x03a1);
+ wimlib_assert(upcase[0x0436] == 0x0416); /* Cyrillic letter zhe */
+ wimlib_assert(upcase[0x0416] == 0x0416);
+ wimlib_assert(upcase[0x0567] == 0x0537); /* Armenian letter eh */
+ wimlib_assert(upcase[0x0537] == 0x0537);
+ wimlib_assert(upcase[0x24d0] == 0x24b6); /* Circled Latin letter A
+ (is that a real character???) */
+ wimlib_assert(upcase[0x24b6] == 0x24b6);
+ wimlib_assert(upcase[0x2603] == 0x2603); /* Note to self: Upper case
+ snowman symbol does not
+ exist. */
+#endif
}
/* Compare UTF-16LE strings case-sensitively (%ignore_case == false) or
if (ignore_case) {
for (size_t i = 0; i < n; i++) {
- u16 c1 = le16_to_cpu(upcase[le16_to_cpu(s1[i])]);
- u16 c2 = le16_to_cpu(upcase[le16_to_cpu(s2[i])]);
+ u16 c1 = upcase[le16_to_cpu(s1[i])];
+ u16 c2 = upcase[le16_to_cpu(s2[i])];
if (c1 != c2)
return (c1 < c2) ? -1 : 1;
}
return 0;
return (n1 < n2) ? -1 : 1;
}
+
+/* Duplicate a UTF16-LE string. The input string might not be null terminated
+ * and might be misaligned, but the returned string is guaranteed to be null
+ * terminated and properly aligned. */
+utf16lechar *
+utf16le_dupz(const void *ustr, size_t usize)
+{
+ utf16lechar *dup = MALLOC(usize + sizeof(utf16lechar));
+ if (dup) {
+ memcpy(dup, ustr, usize);
+ dup[usize / sizeof(utf16lechar)] = 0;
+ }
+ return dup;
+}
git://wimlib.net / wimlib / blobdiff