/* 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
#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
&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)
# endif
/* iconv_carefully_1 is like iconv_carefully, except that it stops after
- converting one character. */
+ 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 = *inbuf;
- const char *inptr_end = inptr + *inbytesleft;
+ const char *inptr_before = *inbuf;
+ const char *inptr = inptr_before;
+ const char *inptr_end = inptr_before + *inbytesleft;
char *outptr = *outbuf;
size_t outsize = *outbytesleft;
- const char *inptr_before = inptr;
size_t res = (size_t)(-1);
size_t insize;
- for (insize = 1; inptr + insize <= inptr_end; 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;
- /* 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;
+ }
}
*inbuf = inptr;
in1ptr++;
in1size--;
}
- utf8buf[utf8len++] = '?';
+ *out1ptr++ = '?';
+ res1 = 0;
}
errno1 = errno;
utf8len = out1ptr - utf8buf;
/* 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);
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;
}
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,
- offsets, 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
char *result = NULL;
size_t length = 0;
int retval = mem_cd_iconveh_internal (src, strlen (src),
- cd, cd1, cd2, handler, 1, NULL,
- &result, &length);
+ cd->cd, cd->cd1, cd->cd2, handler, 1,
+ NULL, &result, &length);
if (retval < 0)
{
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 (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 (cd);
- errno = saved_errno;
- return -1;
- }
- }
-
- if (STRCASEEQ (to_codeset, "UTF-8", 'U','T','F','-','8',0,0,0,0))
- 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);
- if (cd != (iconv_t)(-1))
- iconv_close (cd);
- errno = saved_errno;
- return -1;
- }
- }
+ if (iconveh_open (to_codeset, from_codeset, &cd) < 0)
+ return -1;
result = *resultp;
length = *lengthp;
- retval = mem_cd_iconveh (src, srclen, cd, cd1, cd2, handler, offsets,
+ 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);
- if (cd != (iconv_t)(-1))
- iconv_close (cd);
+ iconveh_close (&cd);
errno = saved_errno;
}
else
{
- if (cd2 != (iconv_t)(-1) && iconv_close (cd2) < 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);
- if (cd != (iconv_t)(-1))
- 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;
- if (cd != (iconv_t)(-1))
- iconv_close (cd);
- if (result != *resultp && result != NULL)
- free (result);
- errno = saved_errno;
- return -1;
- }
- if (cd != (iconv_t)(-1) && iconv_close (cd) < 0)
+ if (iconveh_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);
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
+ if (iconveh_open (to_codeset, from_codeset, &cd) < 0)
+ return NULL;
- 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 (cd);
- errno = saved_errno;
- return NULL;
- }
- }
-
- if (STRCASEEQ (to_codeset, "UTF-8", 'U','T','F','-','8',0,0,0,0))
- 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);
- if (cd != (iconv_t)(-1))
- 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);
- if (cd != (iconv_t)(-1))
- 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);
- if (cd != (iconv_t)(-1))
- 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;
- if (cd != (iconv_t)(-1))
- iconv_close (cd);
- free (result);
- errno = saved_errno;
- return NULL;
- }
- if (cd != (iconv_t)(-1) && 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;
projects / gnulib.git / blobdiff