/* Find matching transformation algorithms and initialize steps.
- Copyright (C) 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
+ Copyright (C) 1997,1998,1999,2000,2001,2004,2005
+ Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
The GNU C Library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Library General Public License as
- published by the Free Software Foundation; either version 2 of the
- License, or (at your option) any later version.
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
The GNU C Library 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
- Library General Public License for more details.
+ Lesser General Public License for more details.
- You should have received a copy of the GNU Library General Public
- License along with the GNU C Library; see the file COPYING.LIB. If not,
- write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
- Boston, MA 02111-1307, USA. */
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
#include <errno.h>
+#include <locale.h>
+#include "../locale/localeinfo.h"
#include <stdlib.h>
#include <string.h>
size_t cnt = 0;
int res;
int conv_flags = 0;
- const char *runp;
+ const char *errhand;
+ const char *ignore;
+ struct trans_struct *trans = NULL;
- /* Find out whether "IGNORE" is part of the options in the `toset'
- name. If yes, remove the string and remember this in the flag. */
- runp = __strchrnul (__strchrnul (toset, '/'), '/');
- if (strcmp (runp, "IGNORE") == 0)
+ /* Find out whether any error handling method is specified. */
+ errhand = strchr (toset, '/');
+ if (errhand != NULL)
+ errhand = strchr (errhand + 1, '/');
+ if (__builtin_expect (errhand != NULL, 1))
{
- /* Found it. This means we should ignore conversion errors. */
- char *newtoset = (char *) alloca (runp - toset + 1);
+ if (*++errhand == '\0')
+ errhand = NULL;
+ else
+ {
+ /* Make copy without the error handling description. */
+ char *newtoset = (char *) alloca (errhand - toset + 1);
+ char *tok;
+ char *ptr;
+
+ newtoset[errhand - toset] = '\0';
+ toset = memcpy (newtoset, toset, errhand - toset);
+
+ /* Find the appropriate transliteration handlers. */
+ tok = strdupa (errhand);
+
+ tok = __strtok_r (tok, ",", &ptr);
+ while (tok != NULL)
+ {
+ if (__strcasecmp_l (tok, "TRANSLIT", _nl_C_locobj_ptr) == 0)
+ {
+ /* It's the builtin transliteration handling. We only
+ support it for working on the internal encoding. */
+ static const char *const internal_trans_names[1]
+ = { "INTERNAL" };
+ struct trans_struct *lastp = NULL;
+ struct trans_struct *runp;
+
+ for (runp = trans; runp != NULL; runp = runp->next)
+ if (runp->trans_fct == __gconv_transliterate)
+ break;
+ else
+ lastp = runp;
+
+ if (runp == NULL)
+ {
+ struct trans_struct *newp;
+
+ newp = (struct trans_struct *) alloca (sizeof (*newp));
+ memset (newp, '\0', sizeof (*newp));
+
+ /* We leave the `name' field zero to signal that
+ this is an internal transliteration step. */
+ newp->csnames = (const char **) internal_trans_names;
+ newp->ncsnames = 1;
+ newp->trans_fct = __gconv_transliterate;
+
+ if (lastp == NULL)
+ trans = newp;
+ else
+ lastp->next = newp;
+ }
+ }
+ else if (__strcasecmp_l (tok, "IGNORE", _nl_C_locobj_ptr) == 0)
+ /* Set the flag to ignore all errors. */
+ conv_flags |= __GCONV_IGNORE_ERRORS;
+ else
+ {
+ /* `tok' is possibly a module name. We'll see later
+ whether we can find it. But first see that we do
+ not already a module of this name. */
+ struct trans_struct *lastp = NULL;
+ struct trans_struct *runp;
+
+ for (runp = trans; runp != NULL; runp = runp->next)
+ if (runp->name != NULL
+ && __strcasecmp_l (tok, runp->name,
+ _nl_C_locobj_ptr) == 0)
+ break;
+ else
+ lastp = runp;
+
+ if (runp == NULL)
+ {
+ struct trans_struct *newp;
- newtoset[runp - toset] = '\0';
- toset = memcpy (newtoset, toset, runp - toset);
+ newp = (struct trans_struct *) alloca (sizeof (*newp));
+ memset (newp, '\0', sizeof (*newp));
+ newp->name = tok;
- flags = __GCONV_IGNORE_ERRORS;
+ if (lastp == NULL)
+ trans = newp;
+ else
+ lastp->next = newp;
+ }
+ }
+
+ tok = __strtok_r (NULL, ",", &ptr);
+ }
+ }
+ }
+
+ /* For the source character set we ignore the error handler specification.
+ XXX Is this really always the best? */
+ ignore = strchr (fromset, '/');
+ if (ignore != NULL && (ignore = strchr (ignore + 1, '/')) != NULL
+ && *++ignore != '\0')
+ {
+ char *newfromset = (char *) alloca (ignore - fromset + 1);
+
+ newfromset[ignore - fromset] = '\0';
+ fromset = memcpy (newfromset, fromset, ignore - fromset);
+ }
+
+ /* If the string is empty define this to mean the charset of the
+ currently selected locale. */
+ if (strcmp (toset, "//") == 0)
+ {
+ const char *codeset = _NL_CURRENT (LC_CTYPE, CODESET);
+ size_t len = strlen (codeset);
+ char *dest;
+ toset = dest = (char *) alloca (len + 3);
+ memcpy (__mempcpy (dest, codeset, len), "//", 3);
+ }
+ if (strcmp (fromset, "//") == 0)
+ {
+ const char *codeset = _NL_CURRENT (LC_CTYPE, CODESET);
+ size_t len = strlen (codeset);
+ char *dest;
+ fromset = dest = (char *) alloca (len + 3);
+ memcpy (__mempcpy (dest, codeset, len), "//", 3);
}
res = __gconv_find_transform (toset, fromset, &steps, &nsteps, flags);
if (res == __GCONV_OK)
{
+ /* Find the modules. */
+ struct trans_struct *lastp = NULL;
+ struct trans_struct *runp;
+
+ for (runp = trans; runp != NULL; runp = runp->next)
+ {
+ if (runp->name == NULL
+ || __builtin_expect (__gconv_translit_find (runp), 0) == 0)
+ lastp = runp;
+ else
+ {
+ /* This means we haven't found the module. Remove it. */
+ if (lastp == NULL)
+ trans = runp->next;
+ else
+ lastp->next = runp->next;
+ }
+ }
+
/* Allocate room for handle. */
result = (__gconv_t) malloc (sizeof (struct __gconv_info)
+ (nsteps
res = __GCONV_NOMEM;
else
{
+ size_t n;
+
/* Remember the list of steps. */
result->__steps = steps;
result->__nsteps = nsteps;
/* Call all initialization functions for the transformation
step implementations. */
- for (cnt = 0; cnt < nsteps - 1; ++cnt)
+ for (cnt = 0; cnt < nsteps; ++cnt)
{
size_t size;
/* Would have to be done if we would not clear the whole
array above. */
- /* If this is the last step we must not allocate an
- output buffer. */
- result->__data[cnt].__flags = conv_flags;
-
#if 0
/* Reset the counter. */
result->__data[cnt].__invocation_counter = 0;
/* We use the `mbstate_t' member in DATA. */
result->__data[cnt].__statep = &result->__data[cnt].__state;
- /* Allocate the buffer. */
- size = (GCONV_NCHAR_GOAL * steps[cnt].__max_needed_to);
+ /* Now see whether we can use any of the transliteration
+ modules for this step. */
+ for (runp = trans; runp != NULL; runp = runp->next)
+ for (n = 0; n < runp->ncsnames; ++n)
+ if (__strcasecmp_l (steps[cnt].__from_name,
+ runp->csnames[n], _nl_C_locobj_ptr) == 0)
+ {
+ void *data = NULL;
+
+ /* Match! Now try the initializer. */
+ if (runp->trans_init_fct == NULL
+ || (runp->trans_init_fct (&data,
+ steps[cnt].__to_name)
+ == __GCONV_OK))
+ {
+ /* Append at the end of the list. */
+ struct __gconv_trans_data *newp;
+ struct __gconv_trans_data **lastp;
+
+ newp = (struct __gconv_trans_data *)
+ malloc (sizeof (struct __gconv_trans_data));
+ if (newp == NULL)
+ {
+ res = __GCONV_NOMEM;
+ goto bail;
+ }
+
+ newp->__trans_fct = runp->trans_fct;
+ newp->__trans_context_fct = runp->trans_context_fct;
+ newp->__trans_end_fct = runp->trans_end_fct;
+ newp->__data = data;
+ newp->__next = NULL;
+
+ lastp = &result->__data[cnt].__trans;
+ while (*lastp != NULL)
+ lastp = &(*lastp)->__next;
+
+ *lastp = newp;
+ }
+ break;
+ }
+
+ /* If this is the last step we must not allocate an
+ output buffer. */
+ if (cnt < nsteps - 1)
+ {
+ result->__data[cnt].__flags = conv_flags;
+
+ /* Allocate the buffer. */
+ size = (GCONV_NCHAR_GOAL * steps[cnt].__max_needed_to);
+
+ result->__data[cnt].__outbuf = malloc (size);
+ if (result->__data[cnt].__outbuf == NULL)
+ {
+ res = __GCONV_NOMEM;
+ goto bail;
+ }
- result->__data[cnt].__outbuf = (char *) malloc (size);
- if (result->__data[cnt].__outbuf == NULL)
+ result->__data[cnt].__outbufend =
+ result->__data[cnt].__outbuf + size;
+ }
+ else
{
- res = __GCONV_NOMEM;
+ /* Handle the last entry. */
+ result->__data[cnt].__flags = conv_flags | __GCONV_IS_LAST;
+
break;
}
- result->__data[cnt].__outbufend =
- result->__data[cnt].__outbuf + size;
}
-
- /* Now handle the last entry. */
- result->__data[cnt].__flags = conv_flags | __GCONV_IS_LAST;
- /* Would have to be done if we would not clear the whole
- array above. */
-#if 0
- result->__data[cnt].__invocation_counter = 0;
- result->__data[cnt].__internal_use = 0;
-#endif
- result->__data[cnt].__statep = &result->__data[cnt].__state;
}
if (res != __GCONV_OK)
{
/* Something went wrong. Free all the resources. */
- int serrno = errno;
+ int serrno;
+ bail:
+ serrno = errno;
if (result != NULL)
{
while (cnt-- > 0)
- free (result->__data[cnt].__outbuf);
+ {
+ struct __gconv_trans_data *transp;
+
+ transp = result->__data[cnt].__trans;
+ while (transp != NULL)
+ {
+ struct __gconv_trans_data *curp = transp;
+ transp = transp->__next;
+
+ if (__builtin_expect (curp->__trans_end_fct != NULL, 0))
+ curp->__trans_end_fct (curp->__data);
+
+ free (curp);
+ }
+
+ free (result->__data[cnt].__outbuf);
+ }
free (result);
result = NULL;