X-Git-Url: http://erislabs.net/gitweb/?a=blobdiff_plain;f=lib%2Fstriconveh.c;h=05dbb73fb9bf3c562bd9e78ca8d20846c5dfa236;hb=d1d75d5540bf75df8bc41c85b5a9fd111ca1f2dd;hp=b02a182bbd3f9e37c09999a6c9681e222a4f1ba8;hpb=36a6f6825953d52a32710a6c38d3ef3a5870d3ac;p=gnulib.git
diff --git a/lib/striconveh.c b/lib/striconveh.c
index b02a182bb..05dbb73fb 100644
--- a/lib/striconveh.c
+++ b/lib/striconveh.c
@@ -1,11 +1,11 @@
/* Character set conversion with error handling.
- Copyright (C) 2001-2007 Free Software Foundation, Inc.
+ Copyright (C) 2001-2009 Free Software Foundation, Inc.
Written by Bruno Haible and Simon Josefsson.
- This program is free software; you can redistribute it and/or modify
+ This program 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 2, or (at your option)
- any later version.
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -13,8 +13,7 @@
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software Foundation,
- Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
+ along with this program. If not, see . */
#include
@@ -28,13 +27,11 @@
#if HAVE_ICONV
# include
-# include "utf8-ucs4-safe.h"
-# include "ucs4-utf8.h"
# include "unistr.h"
#endif
-#include "strdup.h"
#include "c-strcase.h"
+#include "c-strcaseeq.h"
#ifndef SIZE_MAX
# define SIZE_MAX ((size_t) -1)
@@ -43,9 +40,98 @@
#if HAVE_ICONV
-/* The caller must provide CD, CD1, CD2, not just CD, because when a conversion
- error occurs, we may have to determine the Unicode representation of the
- inconvertible character. */
+/* The caller must provide an iconveh_t, not just an iconv_t, because when a
+ conversion error occurs, we may have to determine the Unicode representation
+ of the inconvertible character. */
+
+int
+iconveh_open (const char *to_codeset, const char *from_codeset, iconveh_t *cdp)
+{
+ iconv_t cd;
+ iconv_t cd1;
+ iconv_t cd2;
+
+ /* Avoid glibc-2.1 bug with EUC-KR. */
+# if (__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) && !defined _LIBICONV_VERSION
+ if (c_strcasecmp (from_codeset, "EUC-KR") == 0
+ || c_strcasecmp (to_codeset, "EUC-KR") == 0)
+ {
+ errno = EINVAL;
+ return -1;
+ }
+# endif
+
+ cd = iconv_open (to_codeset, from_codeset);
+
+ if (STRCASEEQ (from_codeset, "UTF-8", 'U','T','F','-','8',0,0,0,0))
+ cd1 = (iconv_t)(-1);
+ else
+ {
+ cd1 = iconv_open ("UTF-8", from_codeset);
+ if (cd1 == (iconv_t)(-1))
+ {
+ int saved_errno = errno;
+ if (cd != (iconv_t)(-1))
+ iconv_close (cdp->cd);
+ errno = saved_errno;
+ return -1;
+ }
+ }
+
+ if (STRCASEEQ (to_codeset, "UTF-8", 'U','T','F','-','8',0,0,0,0)
+# if (__GLIBC__ == 2 && __GLIBC_MINOR__ >= 2) || __GLIBC__ > 2 || _LIBICONV_VERSION >= 0x0105
+ || c_strcasecmp (to_codeset, "UTF-8//TRANSLIT") == 0
+# endif
+ )
+ cd2 = (iconv_t)(-1);
+ else
+ {
+ cd2 = iconv_open (to_codeset, "UTF-8");
+ if (cd2 == (iconv_t)(-1))
+ {
+ int saved_errno = errno;
+ if (cd1 != (iconv_t)(-1))
+ iconv_close (cd1);
+ if (cd != (iconv_t)(-1))
+ iconv_close (cd);
+ errno = saved_errno;
+ return -1;
+ }
+ }
+
+ cdp->cd = cd;
+ cdp->cd1 = cd1;
+ cdp->cd2 = cd2;
+ return 0;
+}
+
+int
+iconveh_close (const iconveh_t *cd)
+{
+ if (cd->cd2 != (iconv_t)(-1) && iconv_close (cd->cd2) < 0)
+ {
+ /* Return -1, but preserve the errno from iconv_close. */
+ int saved_errno = errno;
+ if (cd->cd1 != (iconv_t)(-1))
+ iconv_close (cd->cd1);
+ if (cd->cd != (iconv_t)(-1))
+ iconv_close (cd->cd);
+ errno = saved_errno;
+ return -1;
+ }
+ if (cd->cd1 != (iconv_t)(-1) && iconv_close (cd->cd1) < 0)
+ {
+ /* Return -1, but preserve the errno from iconv_close. */
+ int saved_errno = errno;
+ if (cd->cd != (iconv_t)(-1))
+ iconv_close (cd->cd);
+ errno = saved_errno;
+ return -1;
+ }
+ if (cd->cd != (iconv_t)(-1) && iconv_close (cd->cd) < 0)
+ return -1;
+ return 0;
+}
/* iconv_carefully is like iconv, except that it stops as soon as it encounters
a conversion error, and it returns in *INCREMENTED a boolean telling whether
@@ -82,9 +168,13 @@ iconv_carefully (iconv_t cd,
&outptr, &outsize);
if (!(res == (size_t)(-1) && errno == EINVAL))
break;
- /* We expect that no input bytes have been consumed so far. */
- if (inptr != inptr_before)
- abort ();
+ /* iconv can eat up a shift sequence but give EINVAL while attempting
+ to convert the first character. E.g. libiconv does this. */
+ if (inptr > inptr_before)
+ {
+ res = 0;
+ break;
+ }
}
if (res == 0)
@@ -119,11 +209,146 @@ iconv_carefully (iconv_t cd,
iconv (cd, (ICONV_CONST char **) (inbuf), inbytesleft, outbuf, outbytesleft))
# endif
+/* iconv_carefully_1 is like iconv_carefully, except that it stops after
+ converting one character or one shift sequence. */
+static size_t
+iconv_carefully_1 (iconv_t cd,
+ const char **inbuf, size_t *inbytesleft,
+ char **outbuf, size_t *outbytesleft,
+ bool *incremented)
+{
+ const char *inptr_before = *inbuf;
+ const char *inptr = inptr_before;
+ const char *inptr_end = inptr_before + *inbytesleft;
+ char *outptr = *outbuf;
+ size_t outsize = *outbytesleft;
+ size_t res = (size_t)(-1);
+ size_t insize;
+
+ for (insize = 1; inptr_before + insize <= inptr_end; insize++)
+ {
+ inptr = inptr_before;
+ res = iconv (cd,
+ (ICONV_CONST char **) &inptr, &insize,
+ &outptr, &outsize);
+ if (!(res == (size_t)(-1) && errno == EINVAL))
+ break;
+ /* iconv can eat up a shift sequence but give EINVAL while attempting
+ to convert the first character. E.g. libiconv does this. */
+ if (inptr > inptr_before)
+ {
+ res = 0;
+ break;
+ }
+ }
+
+ *inbuf = inptr;
+ *inbytesleft = inptr_end - inptr;
+# if !defined _LIBICONV_VERSION && !defined __GLIBC__
+ /* Irix iconv() inserts a NUL byte if it cannot convert.
+ NetBSD iconv() inserts a question mark if it cannot convert.
+ Only GNU libiconv and GNU libc are known to prefer to fail rather
+ than doing a lossy conversion. */
+ if (res != (size_t)(-1) && res > 0)
+ {
+ /* iconv() has already incremented INPTR. We cannot go back to a
+ previous INPTR, otherwise the state inside CD would become invalid,
+ if FROM_CODESET is a stateful encoding. So, tell the caller that
+ *INBUF has already been incremented. */
+ *incremented = (inptr > inptr_before);
+ errno = EILSEQ;
+ return (size_t)(-1);
+ }
+# endif
+
+ if (res != (size_t)(-1))
+ {
+ *outbuf = outptr;
+ *outbytesleft = outsize;
+ }
+ *incremented = false;
+ return res;
+}
+
+/* utf8conv_carefully is like iconv, except that
+ - it converts from UTF-8 to UTF-8,
+ - it stops as soon as it encounters a conversion error, and it returns
+ in *INCREMENTED a boolean telling whether it has incremented the input
+ pointers past the error location,
+ - if one_character_only is true, it stops after converting one
+ character. */
+static size_t
+utf8conv_carefully (bool one_character_only,
+ const char **inbuf, size_t *inbytesleft,
+ char **outbuf, size_t *outbytesleft,
+ bool *incremented)
+{
+ const char *inptr = *inbuf;
+ size_t insize = *inbytesleft;
+ char *outptr = *outbuf;
+ size_t outsize = *outbytesleft;
+ size_t res;
+
+ res = 0;
+ do
+ {
+ ucs4_t uc;
+ int n;
+ int m;
+
+ n = u8_mbtoucr (&uc, (const uint8_t *) inptr, insize);
+ if (n < 0)
+ {
+ errno = (n == -2 ? EINVAL : EILSEQ);
+ n = u8_mbtouc (&uc, (const uint8_t *) inptr, insize);
+ inptr += n;
+ insize -= n;
+ res = (size_t)(-1);
+ *incremented = true;
+ break;
+ }
+ if (outsize == 0)
+ {
+ errno = E2BIG;
+ res = (size_t)(-1);
+ *incremented = false;
+ break;
+ }
+ m = u8_uctomb ((uint8_t *) outptr, uc, outsize);
+ if (m == -2)
+ {
+ errno = E2BIG;
+ res = (size_t)(-1);
+ *incremented = false;
+ break;
+ }
+ inptr += n;
+ insize -= n;
+ if (m == -1)
+ {
+ errno = EILSEQ;
+ res = (size_t)(-1);
+ *incremented = true;
+ break;
+ }
+ outptr += m;
+ outsize -= m;
+ }
+ while (!one_character_only && insize > 0);
+
+ *inbuf = inptr;
+ *inbytesleft = insize;
+ *outbuf = outptr;
+ *outbytesleft = outsize;
+ return res;
+}
+
static int
mem_cd_iconveh_internal (const char *src, size_t srclen,
iconv_t cd, iconv_t cd1, iconv_t cd2,
enum iconv_ilseq_handler handler,
size_t extra_alloc,
+ size_t *offsets,
char **resultp, size_t *lengthp)
{
/* When a conversion error occurs, we cannot start using CD1 and CD2 at
@@ -141,8 +366,9 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
char *result;
size_t allocated;
size_t length;
+ size_t last_length = (size_t)(-1); /* only needed if offsets != NULL */
- if (*lengthp >= sizeof (tmpbuf))
+ if (*resultp != NULL && *lengthp >= sizeof (tmpbuf))
{
initial_result = *resultp;
allocated = *lengthp;
@@ -153,6 +379,20 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
allocated = sizeof (tmpbuf);
}
result = initial_result;
+
+ /* Test whether a direct conversion is possible at all. */
+ if (cd == (iconv_t)(-1))
+ goto indirectly;
+
+ if (offsets != NULL)
+ {
+ size_t i;
+
+ for (i = 0; i < srclen; i++)
+ offsets[i] = (size_t)(-1);
+
+ last_length = (size_t)(-1);
+ }
length = 0;
/* First, try a direct conversion, and see whether a conversion error
@@ -176,16 +416,29 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
size_t res;
bool grow;
- /* Use iconv_carefully instead of iconv here, because:
- - If TO_CODESET is UTF-8, we can do the error handling in this loop,
- no need for a second loop,
- - With iconv() implementations other than GNU libiconv and GNU libc,
- if we use iconv() in a big swoop, checking for an E2BIG return,
- we lose the number of irreversible conversions. */
- res = iconv_carefully (cd,
- &inptr, &insize,
- &outptr, &outsize,
- &incremented);
+ if (offsets != NULL)
+ {
+ if (length != last_length) /* ensure that offset[] be increasing */
+ {
+ offsets[inptr - src] = length;
+ last_length = length;
+ }
+ res = iconv_carefully_1 (cd,
+ &inptr, &insize,
+ &outptr, &outsize,
+ &incremented);
+ }
+ else
+ /* Use iconv_carefully instead of iconv here, because:
+ - If TO_CODESET is UTF-8, we can do the error handling in this
+ loop, no need for a second loop,
+ - With iconv() implementations other than GNU libiconv and GNU
+ libc, if we use iconv() in a big swoop, checking for an E2BIG
+ return, we lose the number of irreversible conversions. */
+ res = iconv_carefully (cd,
+ &inptr, &insize,
+ &outptr, &outsize,
+ &incremented);
length = outptr - result;
grow = (length + extra_alloc > allocated / 2);
@@ -329,11 +582,20 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
goto done;
indirectly:
- /* The direct conversion failed, handler != iconveh_error,
- and cd2 != (iconv_t)(-1).
+ /* The direct conversion failed.
Use a conversion through UTF-8. */
+ if (offsets != NULL)
+ {
+ size_t i;
+
+ for (i = 0; i < srclen; i++)
+ offsets[i] = (size_t)(-1);
+
+ last_length = (size_t)(-1);
+ }
length = 0;
{
+ const bool slowly = (offsets != NULL || handler == iconveh_error);
# define utf8bufsize 4096 /* may also be smaller or larger than tmpbufsize */
char utf8buf[utf8bufsize + 1];
size_t utf8len = 0;
@@ -348,7 +610,8 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
/* Set to the initial state. */
if (cd1 != (iconv_t)(-1))
iconv (cd1, NULL, NULL, NULL, NULL);
- iconv (cd2, NULL, NULL, NULL, NULL);
+ if (cd2 != (iconv_t)(-1))
+ iconv (cd2, NULL, NULL, NULL, NULL);
# endif
while (in1size > 0 || do_final_flush1 || utf8len > 0 || do_final_flush2)
@@ -362,63 +625,32 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
/* Conversion step 1: from FROM_CODESET to UTF-8. */
if (in1size > 0)
{
+ if (offsets != NULL
+ && length != last_length) /* ensure that offset[] be increasing */
+ {
+ offsets[in1ptr - src] = length;
+ last_length = length;
+ }
if (cd1 != (iconv_t)(-1))
- res1 = iconv_carefully (cd1,
- (ICONV_CONST char **) &in1ptr, &in1size,
- &out1ptr, &out1size,
- &incremented1);
+ {
+ if (slowly)
+ res1 = iconv_carefully_1 (cd1,
+ &in1ptr, &in1size,
+ &out1ptr, &out1size,
+ &incremented1);
+ else
+ res1 = iconv_carefully (cd1,
+ &in1ptr, &in1size,
+ &out1ptr, &out1size,
+ &incremented1);
+ }
else
{
/* FROM_CODESET is UTF-8. */
- res1 = 0;
- do
- {
- ucs4_t uc;
- int n;
- int m;
-
- n = u8_mbtouc_safe (&uc, (const uint8_t *) in1ptr, in1size);
- if (uc == 0xfffd
- && !(n >= 3
- && (uint8_t)in1ptr[0] == 0xEF
- && (uint8_t)in1ptr[1] == 0xBF
- && (uint8_t)in1ptr[2] == 0xBD))
- {
- in1ptr += n;
- in1size -= n;
- errno = EILSEQ;
- res1 = (size_t)(-1);
- incremented1 = true;
- break;
- }
- if (out1size == 0)
- {
- errno = E2BIG;
- res1 = (size_t)(-1);
- incremented1 = false;
- break;
- }
- m = u8_uctomb ((uint8_t *) out1ptr, uc, out1size);
- if (m == -2)
- {
- errno = E2BIG;
- res1 = (size_t)(-1);
- incremented1 = false;
- break;
- }
- in1ptr += n;
- in1size -= n;
- if (m == -1)
- {
- errno = EILSEQ;
- res1 = (size_t)(-1);
- incremented1 = true;
- break;
- }
- out1ptr += m;
- out1size -= m;
- }
- while (in1size > 0);
+ res1 = utf8conv_carefully (slowly,
+ &in1ptr, &in1size,
+ &out1ptr, &out1size,
+ &incremented1);
}
}
else if (do_final_flush1)
@@ -464,12 +696,14 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
in1ptr++;
in1size--;
}
- utf8buf[utf8len++] = '?';
+ *out1ptr++ = '?';
+ res1 = 0;
}
errno1 = errno;
utf8len = out1ptr - utf8buf;
- if (in1size == 0
+ if (offsets != NULL
+ || in1size == 0
|| utf8len > utf8bufsize / 2
|| (res1 == (size_t)(-1) && errno1 == E2BIG))
{
@@ -487,10 +721,19 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
bool grow;
if (in2size > 0)
- res2 = iconv_carefully (cd2,
- &in2ptr, &in2size,
- &out2ptr, &out2size,
- &incremented2);
+ {
+ if (cd2 != (iconv_t)(-1))
+ res2 = iconv_carefully (cd2,
+ &in2ptr, &in2size,
+ &out2ptr, &out2size,
+ &incremented2);
+ else
+ /* TO_CODESET is UTF-8. */
+ res2 = utf8conv_carefully (false,
+ &in2ptr, &in2size,
+ &out2ptr, &out2size,
+ &incremented2);
+ }
else /* in1size == 0 && !do_final_flush1
&& in2size == 0 && do_final_flush2 */
{
@@ -498,10 +741,11 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
state. But avoid glibc-2.1 bug and Solaris 2.7 bug. */
# if defined _LIBICONV_VERSION \
|| !((__GLIBC__ == 2 && __GLIBC_MINOR__ <= 1) || defined __sun)
- res2 = iconv (cd2, NULL, NULL, &out2ptr, &out2size);
-# else
- res2 = 0;
+ if (cd2 != (iconv_t)(-1))
+ res2 = iconv (cd2, NULL, NULL, &out2ptr, &out2size);
+ else
# endif
+ res2 = 0;
do_final_flush2 = false;
incremented2 = true;
}
@@ -538,8 +782,8 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
int n;
if (in2size == 0)
abort ();
- n = u8_mbtouc (&uc, (const uint8_t *) in2ptr,
- in2size);
+ n = u8_mbtouc_unsafe (&uc, (const uint8_t *) in2ptr,
+ in2size);
in2ptr += n;
in2size -= n;
}
@@ -572,9 +816,28 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
inptr = scratchbuf;
insize = scratchlen;
- res = iconv (cd2,
- (ICONV_CONST char **) &inptr, &insize,
- &out2ptr, &out2size);
+ if (cd2 != (iconv_t)(-1))
+ res = iconv (cd2,
+ (ICONV_CONST char **) &inptr, &insize,
+ &out2ptr, &out2size);
+ else
+ {
+ /* TO_CODESET is UTF-8. */
+ if (out2size >= insize)
+ {
+ memcpy (out2ptr, inptr, insize);
+ out2ptr += insize;
+ out2size -= insize;
+ inptr += insize;
+ insize = 0;
+ res = 0;
+ }
+ else
+ {
+ errno = E2BIG;
+ res = (size_t)(-1);
+ }
+ }
length = out2ptr - result;
if (res == (size_t)(-1) && errno == E2BIG)
{
@@ -601,9 +864,23 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
out2ptr = result + length;
out2size = allocated - extra_alloc - length;
- res = iconv (cd2,
- (ICONV_CONST char **) &inptr, &insize,
- &out2ptr, &out2size);
+ if (cd2 != (iconv_t)(-1))
+ res = iconv (cd2,
+ (ICONV_CONST char **) &inptr,
+ &insize,
+ &out2ptr, &out2size);
+ else
+ {
+ /* TO_CODESET is UTF-8. */
+ if (!(out2size >= insize))
+ abort ();
+ memcpy (out2ptr, inptr, insize);
+ out2ptr += insize;
+ out2size -= insize;
+ inptr += insize;
+ insize = 0;
+ res = 0;
+ }
length = out2ptr - result;
}
# if !defined _LIBICONV_VERSION && !defined __GLIBC__
@@ -692,9 +969,10 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
/* Now the final memory allocation. */
if (result == tmpbuf)
{
+ size_t memsize = length + extra_alloc;
char *memory;
- memory = (char *) malloc (length + extra_alloc);
+ memory = (char *) malloc (memsize > 0 ? memsize : 1);
if (memory != NULL)
{
memcpy (memory, tmpbuf, length);
@@ -709,9 +987,10 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
else if (result != *resultp && length + extra_alloc < allocated)
{
/* Shrink the allocated memory if possible. */
+ size_t memsize = length + extra_alloc;
char *memory;
- memory = (char *) realloc (result, length + extra_alloc);
+ memory = (char *) realloc (result, memsize > 0 ? memsize : 1);
if (memory != NULL)
result = memory;
}
@@ -724,17 +1003,18 @@ mem_cd_iconveh_internal (const char *src, size_t srclen,
int
mem_cd_iconveh (const char *src, size_t srclen,
- iconv_t cd, iconv_t cd1, iconv_t cd2,
+ const iconveh_t *cd,
enum iconv_ilseq_handler handler,
+ size_t *offsets,
char **resultp, size_t *lengthp)
{
- return mem_cd_iconveh_internal (src, srclen, cd, cd1, cd2, handler, 0,
- resultp, lengthp);
+ return mem_cd_iconveh_internal (src, srclen, cd->cd, cd->cd1, cd->cd2,
+ handler, 0, offsets, resultp, lengthp);
}
char *
str_cd_iconveh (const char *src,
- iconv_t cd, iconv_t cd1, iconv_t cd2,
+ const iconveh_t *cd,
enum iconv_ilseq_handler handler)
{
/* For most encodings, a trailing NUL byte in the input will be converted
@@ -744,8 +1024,8 @@ str_cd_iconveh (const char *src,
char *result = NULL;
size_t length = 0;
int retval = mem_cd_iconveh_internal (src, strlen (src),
- cd, cd1, cd2, handler, 1,
- &result, &length);
+ cd->cd, cd->cd1, cd->cd2, handler, 1,
+ NULL, &result, &length);
if (retval < 0)
{
@@ -770,6 +1050,7 @@ int
mem_iconveh (const char *src, size_t srclen,
const char *from_codeset, const char *to_codeset,
enum iconv_ilseq_handler handler,
+ size_t *offsets,
char **resultp, size_t *lengthp)
{
if (srclen == 0)
@@ -778,7 +1059,7 @@ mem_iconveh (const char *src, size_t srclen,
*lengthp = 0;
return 0;
}
- else if (c_strcasecmp (from_codeset, to_codeset) == 0)
+ else if (offsets == NULL && c_strcasecmp (from_codeset, to_codeset) == 0)
{
char *result;
@@ -801,103 +1082,32 @@ mem_iconveh (const char *src, size_t srclen,
else
{
#if HAVE_ICONV
- iconv_t cd;
- iconv_t cd1;
- iconv_t cd2;
+ iconveh_t cd;
char *result;
size_t length;
int retval;
- /* Avoid glibc-2.1 bug with EUC-KR. */
-# if (__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) && !defined _LIBICONV_VERSION
- if (c_strcasecmp (from_codeset, "EUC-KR") == 0
- || c_strcasecmp (to_codeset, "EUC-KR") == 0)
- {
- errno = EINVAL;
- return -1;
- }
-# endif
-
- cd = iconv_open (to_codeset, from_codeset);
- if (cd == (iconv_t)(-1))
+ if (iconveh_open (to_codeset, from_codeset, &cd) < 0)
return -1;
- if (c_strcasecmp (from_codeset, "UTF-8") == 0)
- cd1 = (iconv_t)(-1);
- else
- {
- cd1 = iconv_open ("UTF-8", from_codeset);
- if (cd1 == (iconv_t)(-1))
- {
- int saved_errno = errno;
- iconv_close (cd);
- errno = saved_errno;
- return -1;
- }
- }
-
- if (c_strcasecmp (to_codeset, "UTF-8") == 0)
- cd2 = (iconv_t)(-1);
- else
- {
- cd2 = iconv_open (to_codeset, "UTF-8");
- if (cd2 == (iconv_t)(-1))
- {
- int saved_errno = errno;
- if (cd1 != (iconv_t)(-1))
- iconv_close (cd1);
- iconv_close (cd);
- errno = saved_errno;
- return -1;
- }
- }
-
result = *resultp;
length = *lengthp;
- retval =
- mem_cd_iconveh (src, srclen, cd, cd1, cd2, handler, &result, &length);
+ retval = mem_cd_iconveh (src, srclen, &cd, handler, offsets,
+ &result, &length);
if (retval < 0)
{
- /* Close cd, cd1, cd2, but preserve the errno from str_cd_iconv. */
+ /* Close cd, but preserve the errno from str_cd_iconv. */
int saved_errno = errno;
- if (cd2 != (iconv_t)(-1))
- iconv_close (cd2);
- if (cd1 != (iconv_t)(-1))
- iconv_close (cd1);
- iconv_close (cd);
+ iconveh_close (&cd);
errno = saved_errno;
}
else
{
- if (cd2 != (iconv_t)(-1) && iconv_close (cd2) < 0)
+ if (iconveh_close (&cd) < 0)
{
/* Return -1, but free the allocated memory, and while doing
- that, preserve the errno from iconv_close. */
- int saved_errno = errno;
- if (cd1 != (iconv_t)(-1))
- iconv_close (cd1);
- iconv_close (cd);
- if (result != *resultp && result != NULL)
- free (result);
- errno = saved_errno;
- return -1;
- }
- if (cd1 != (iconv_t)(-1) && iconv_close (cd1) < 0)
- {
- /* Return -1, but free the allocated memory, and while doing
- that, preserve the errno from iconv_close. */
- int saved_errno = errno;
- iconv_close (cd);
- if (result != *resultp && result != NULL)
- free (result);
- errno = saved_errno;
- return -1;
- }
- if (iconv_close (cd) < 0)
- {
- /* Return -1, but free the allocated memory, and while doing
- that, preserve the errno from iconv_close. */
+ that, preserve the errno from iconveh_close. */
int saved_errno = errno;
if (result != *resultp && result != NULL)
free (result);
@@ -936,96 +1146,27 @@ str_iconveh (const char *src,
else
{
#if HAVE_ICONV
- iconv_t cd;
- iconv_t cd1;
- iconv_t cd2;
+ iconveh_t cd;
char *result;
- /* Avoid glibc-2.1 bug with EUC-KR. */
-# if (__GLIBC__ - 0 == 2 && __GLIBC_MINOR__ - 0 <= 1) && !defined _LIBICONV_VERSION
- if (c_strcasecmp (from_codeset, "EUC-KR") == 0
- || c_strcasecmp (to_codeset, "EUC-KR") == 0)
- {
- errno = EINVAL;
- return NULL;
- }
-# endif
-
- cd = iconv_open (to_codeset, from_codeset);
- if (cd == (iconv_t)(-1))
+ if (iconveh_open (to_codeset, from_codeset, &cd) < 0)
return NULL;
- if (c_strcasecmp (from_codeset, "UTF-8") == 0)
- cd1 = (iconv_t)(-1);
- else
- {
- cd1 = iconv_open ("UTF-8", from_codeset);
- if (cd1 == (iconv_t)(-1))
- {
- int saved_errno = errno;
- iconv_close (cd);
- errno = saved_errno;
- return NULL;
- }
- }
-
- if (c_strcasecmp (to_codeset, "UTF-8") == 0)
- cd2 = (iconv_t)(-1);
- else
- {
- cd2 = iconv_open (to_codeset, "UTF-8");
- if (cd2 == (iconv_t)(-1))
- {
- int saved_errno = errno;
- if (cd1 != (iconv_t)(-1))
- iconv_close (cd1);
- iconv_close (cd);
- errno = saved_errno;
- return NULL;
- }
- }
-
- result = str_cd_iconveh (src, cd, cd1, cd2, handler);
+ result = str_cd_iconveh (src, &cd, handler);
if (result == NULL)
{
- /* Close cd, cd1, cd2, but preserve the errno from str_cd_iconv. */
+ /* Close cd, but preserve the errno from str_cd_iconv. */
int saved_errno = errno;
- if (cd2 != (iconv_t)(-1))
- iconv_close (cd2);
- if (cd1 != (iconv_t)(-1))
- iconv_close (cd1);
- iconv_close (cd);
+ iconveh_close (&cd);
errno = saved_errno;
}
else
{
- if (cd2 != (iconv_t)(-1) && iconv_close (cd2) < 0)
- {
- /* Return NULL, but free the allocated memory, and while doing
- that, preserve the errno from iconv_close. */
- int saved_errno = errno;
- if (cd1 != (iconv_t)(-1))
- iconv_close (cd1);
- iconv_close (cd);
- free (result);
- errno = saved_errno;
- return NULL;
- }
- if (cd1 != (iconv_t)(-1) && iconv_close (cd1) < 0)
- {
- /* Return NULL, but free the allocated memory, and while doing
- that, preserve the errno from iconv_close. */
- int saved_errno = errno;
- iconv_close (cd);
- free (result);
- errno = saved_errno;
- return NULL;
- }
- if (iconv_close (cd) < 0)
+ if (iconveh_close (&cd) < 0)
{
/* Return NULL, but free the allocated memory, and while doing
- that, preserve the errno from iconv_close. */
+ that, preserve the errno from iconveh_close. */
int saved_errno = errno;
free (result);
errno = saved_errno;