+2000-04-09 Ulrich Drepper <drepper@redhat.com>
+
+ Implement handling of restartable conversion functions according to
+ ISO C.
+ * iconv/gconv.h (__gconv_fct): Add additional parameter.
+ * iconv/gconv_int.h (__BUILTIN_TRANS): Likewise.
+ * iconv/gconv.c: Pass additional parameter to conversion function.
+ * iconv/gconv_simple.c (internal_ucs4_loop_single): New function.
+ (internal_ucs4le_loop_single): New function.
+ (__gconv_transform_ascii_internal): Define ONE_DIRECTION.
+ (__gconv_transform_internal_ascii): Likewise.
+ (__gconv_transform_internal_utf8): Likewise.
+ (__gconv_transform_utf8_internal): Likewise.
+ (__gconv_transform_ucs2_internal): Likewise.
+ (__gconv_transform_internal_ucs2): Likewise.
+ (__gconv_transform_ucs2reverse_internal): Likewise.
+ (__gconv_transform_internal_ucs2reverse): Likewise.
+ (internal_ucs4le_loop_unaligned): Before return
+ __GCONV_INCOMPLETE_INPUT check that the remaining bytes really form
+ a valid character. Otherwise return __GCONV_ILLEGAL_INPUT.
+ (__gconv_transform_utf8_internal): Define STORE_REST and UNPACK_BYTES.
+ * iconv/loop.c: Fit in definition of function to convert one character
+ for processing of left-over bytes from the state object.
+ * iconv/skeleton.c (gconv): Rename inbuf to inptrp and inbufend to
+ inend to match names in loop functions.
+ (RESET_INPUT_BUFFER): Change apprpriately.
+ (gconv): If needed, call function to process bytes from the state
+ object. Similar at the end: store left over bytes if input is
+ incomplete.
+ Take extra argument and add new argument to all calls of the
+ conversion function.
+ * iconvdata/iso-2022-cn.c: Adjust numeric values used to store
+ information in the state object to not conflict with length count.
+ * iconvdata/iso-2022-jp.c: Likewise.
+ * iconvdata/iso-2022-kr.c: Likewise.
+ * iconvdata/unicode.c: Adjust for change change in parameters of
+ skeleton function.
+ * iconvdata/utf-16.c: Likewise.
+ * libio/iofwide.c: Add new parameter to all calls of conversion
+ function.
+ * wcsmbs/btowc.c: Likewise.
+ * wcsmbs/mbrtowc.c: Likewise.
+ * wcsmbs/mbsnrtowcs.c: Likewise.
+ * wcsmbs/mbsrtowcs.c: Likewise.
+ * wcsmbs/wcrtomb.c: Likewise.
+ * wcsmbs/wcsnrtombs.c: Likewise.
+ * wcsmbs/wcsrtombs.c: Likewise.
+ * wcsmbs/wctob.c: Likewise.
+
+ * iconvdata/gbgbk.c: Always define MAX_NEEDED_OUTPUT and
+ MAX_NEEDED_INPUT.
+
2000-04-07 Andreas Jaeger <aj@suse.de>
* sysdeps/unix/sysv/linux/mips/kernel_sigaction.h: Updated for
/* Type of a conversion function. */
typedef int (*__gconv_fct) (struct __gconv_step *, struct __gconv_step_data *,
__const unsigned char **, __const unsigned char *,
- size_t *, int);
+ size_t *, int, int);
/* Constructor and destructor for local data for conversion step. */
typedef int (*__gconv_init_fct) (struct __gconv_step *);
struct __gconv_step_data *__data, \
const unsigned char **__inbuf, \
const unsigned char *__inbufend, size_t *__written, \
- int __do_flush)
+ int __do_flush, int __consume_incomplete)
__BUILTIN_TRANS (__gconv_transform_ascii_internal);
__BUILTIN_TRANS (__gconv_transform_internal_ascii);
}
#endif
+
+static inline int
+internal_ucs4_loop_single (const unsigned char **inptrp,
+ const unsigned char *inend,
+ unsigned char **outptrp, unsigned char *outend,
+ mbstate_t *state, void *data, size_t *converted)
+{
+ size_t cnt = state->__count & 7;
+
+ while (*inptrp < inend && cnt < 4)
+ state->__value.__wchb[cnt++] = *(*inptrp)++;
+
+ if (cnt < 4)
+ {
+ /* Still not enough bytes. Store the ones in the input buffer. */
+ state->__count &= ~7;
+ state->__count |= cnt;
+
+ return __GCONV_INCOMPLETE_INPUT;
+ }
+
+#if __BYTE_ORDER == __LITTLE_ENDIAN
+ (*outptrp)[0] = state->__value.__wchb[3];
+ (*outptrp)[1] = state->__value.__wchb[2];
+ (*outptrp)[2] = state->__value.__wchb[1];
+ (*outptrp)[3] = state->__value.__wchb[0];
+#elif __BYTE_ORDER == __BIG_ENDIAN
+ /* XXX unaligned */
+ *(*((uint32_t **) outptrp)++) = state->__value.__wch;
+#else
+# error "This endianess is not supported."
+#endif
+
+ /* Clear the state buffer. */
+ state->__count &= ~7;
+
+ return __GCONV_OK;
+}
+
#include <iconv/skeleton.c>
}
#endif
+
+static inline int
+internal_ucs4le_loop_single (const unsigned char **inptrp,
+ const unsigned char *inend,
+ unsigned char **outptrp, unsigned char *outend,
+ mbstate_t *state, void *data, size_t *converted)
+{
+ size_t cnt = state->__count & 7;
+
+ while (*inptrp < inend && cnt < 4)
+ state->__value.__wchb[cnt++] = *(*inptrp)++;
+
+ if (cnt < 4)
+ {
+ /* Still not enough bytes. Store the ones in the input buffer. */
+ state->__count &= ~7;
+ state->__count |= cnt;
+
+ return __GCONV_INCOMPLETE_INPUT;
+ }
+
+#if __BYTE_ORDER == __BIG_ENDIAN
+ (*outptrp)[0] = state->__value.__wchb[3];
+ (*outptrp)[1] = state->__value.__wchb[2];
+ (*outptrp)[2] = state->__value.__wchb[1];
+ (*outptrp)[3] = state->__value.__wchb[0];
+#else
+ /* XXX unaligned */
+ *(*((uint32_t **) outptrp)++) = state->__value.__wch;
+#endif
+
+ /* Clear the state buffer. */
+ state->__count &= ~7;
+
+ return __GCONV_OK;
+}
+
#include <iconv/skeleton.c>
#define FROM_LOOP ascii_internal_loop
#define TO_LOOP ascii_internal_loop /* This is not used. */
#define FUNCTION_NAME __gconv_transform_ascii_internal
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
} \
\
/* It's an one byte sequence. */ \
+ /* XXX unaligned. */ \
*((uint32_t *) outptr)++ = *inptr++; \
}
#include <iconv/loop.c>
#define FROM_LOOP internal_ascii_loop
#define TO_LOOP internal_ascii_loop /* This is not used. */
#define FUNCTION_NAME __gconv_transform_internal_ascii
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
#define FROM_LOOP internal_utf8_loop
#define TO_LOOP internal_utf8_loop /* This is not used. */
#define FUNCTION_NAME __gconv_transform_internal_utf8
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
#define FROM_LOOP utf8_internal_loop
#define TO_LOOP utf8_internal_loop /* This is not used. */
#define FUNCTION_NAME __gconv_transform_utf8_internal
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MAX_NEEDED_INPUT MAX_NEEDED_FROM
\
if (NEED_LENGTH_TEST && inptr + cnt > inend) \
{ \
- /* We don't have enough input. */ \
- result = __GCONV_INCOMPLETE_INPUT; \
+ /* We don't have enough input. But before we report that check \
+ that all the bytes are correct. */ \
+ for (i = 1; inptr + i < inend; ++i) \
+ if ((inptr[i] & 0xc0) != 0x80) \
+ break; \
+ result = (inptr + i == inend \
+ ? __GCONV_INCOMPLETE_INPUT : __GCONV_ILLEGAL_INPUT); \
break; \
} \
\
/* Now adjust the pointers and store the result. */ \
*((uint32_t *) outptr)++ = ch; \
}
+
+#define STORE_REST \
+ { \
+ /* We store the remaining bytes while converting them into the UCS4 \
+ format. We can assume that the first byte in the buffer is \
+ correct and that it requires a larger number of bytes than there \
+ are in the input buffer. */ \
+ wint_t ch = **inptrp; \
+ size_t cnt; \
+ \
+ state->__count = inend - *inptrp; \
+ \
+ if (ch >= 0xc2 && ch < 0xe0) \
+ { \
+ /* We expect two bytes. The first byte cannot be 0xc0 or \
+ 0xc1, otherwise the wide character could have been \
+ represented using a single byte. */ \
+ cnt = 2; \
+ ch &= 0x1f; \
+ } \
+ else if ((ch & 0xf0) == 0xe0) \
+ { \
+ /* We expect three bytes. */ \
+ cnt = 3; \
+ ch &= 0x0f; \
+ } \
+ else if ((ch & 0xf8) == 0xf0) \
+ { \
+ /* We expect four bytes. */ \
+ cnt = 4; \
+ ch &= 0x07; \
+ } \
+ else if ((ch & 0xfc) == 0xf8) \
+ { \
+ /* We expect five bytes. */ \
+ cnt = 5; \
+ ch &= 0x03; \
+ } \
+ else \
+ { \
+ /* We expect six bytes. */ \
+ cnt = 6; \
+ ch &= 0x01; \
+ } \
+ \
+ /* The first byte is already consumed. */ \
+ --cnt; \
+ while (++(*inptrp) < inend) \
+ { \
+ ch <<= 6; \
+ ch |= **inptrp & 0x3f; \
+ --cnt; \
+ } \
+ \
+ /* Shift for the so far missing bytes. */ \
+ ch <<= cnt * 6; \
+ \
+ /* Store the value. */ \
+ state->__value.__wch = ch; \
+ }
+
+#define UNPACK_BYTES \
+ { \
+ wint_t wch = state->__value.__wch; \
+ inlen = state->__count; \
+ \
+ if (state->__value.__wch <= 0x7ff) \
+ bytebuf[0] = 0xc0; \
+ else if (state->__value.__wch <= 0xffff) \
+ bytebuf[0] = 0xe0; \
+ else if (state->__value.__wch <= 0x1fffff) \
+ bytebuf[0] = 0xf0; \
+ else if (state->__value.__wch <= 0x3ffffff) \
+ bytebuf[0] = 0xf8; \
+ else \
+ bytebuf[0] = 0xfc; \
+ \
+ while (inlen-- > 1) \
+ bytebuf[inlen] = 0x80 | (wch & 0x3f); \
+ \
+ bytebuf[0] |= wch; \
+ }
+
#include <iconv/loop.c>
#include <iconv/skeleton.c>
#define FROM_LOOP ucs2_internal_loop
#define TO_LOOP ucs2_internal_loop /* This is not used. */
#define FUNCTION_NAME __gconv_transform_ucs2_internal
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
#define FROM_LOOP internal_ucs2_loop
#define TO_LOOP internal_ucs2_loop /* This is not used. */
#define FUNCTION_NAME __gconv_transform_internal_ucs2
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
#define FROM_LOOP ucs2reverse_internal_loop
#define TO_LOOP ucs2reverse_internal_loop/* This is not used.*/
#define FUNCTION_NAME __gconv_transform_ucs2reverse_internal
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
#define FROM_LOOP internal_ucs2reverse_loop
#define TO_LOOP internal_ucs2reverse_loop/* This is not used.*/
#define FUNCTION_NAME __gconv_transform_internal_ucs2reverse
+#define ONE_DIRECTION 1
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
UPDATE_PARAMS code to store result in params.
*/
+#include <assert.h>
#include <endian.h>
#include <gconv.h>
#include <stdint.h>
}
-#undef get16
-#undef get32
-#undef put16
-#undef put32
-#undef unaligned
-
/* Include the file a second time to define the function to define the
function to handle unaligned access. */
#if !defined DEFINE_UNALIGNED && !defined _STRING_ARCH_unaligned \
&& MIN_NEEDED_FROM != 1 && MAX_NEEDED_FROM % MIN_NEEDED_FROM == 0 \
&& MIN_NEEDED_TO != 1 && MAX_NEEDED_TO % MIN_NEEDED_TO == 0
+# undef get16
+# undef get32
+# undef put16
+# undef put32
+# undef unaligned
+
# define DEFINE_UNALIGNED
# include "loop.c"
# undef DEFINE_UNALIGNED
#endif
+#if MAX_NEEDED_INPUT > 1
+# define SINGLE(fct) SINGLE2 (fct)
+# define SINGLE2(fct) fct##_single
+static inline int
+SINGLE(LOOPFCT) (const unsigned char **inptrp, const unsigned char *inend,
+ unsigned char **outptrp, unsigned char *outend,
+ mbstate_t *state, void *data, size_t *converted
+ EXTRA_LOOP_DECLS)
+{
+ int result = __GCONV_OK;
+ unsigned char bytebuf[MAX_NEEDED_INPUT];
+ const unsigned char *inptr = *inptrp;
+ unsigned char *outptr = *outptrp;
+ size_t inlen;
+
+#ifdef INIT_PARAMS
+ INIT_PARAMS;
+#endif
+
+#ifdef UNPACK_BYTES
+ UNPACK_BYTES
+#else
+ /* Add the bytes from the state to the input buffer. */
+ for (inlen = 0; inlen < (state->__count & 7); ++ inlen)
+ bytebuf[inlen] = state->__value.__wchb[inlen];
+#endif
+
+ /* Are there enough bytes in the input buffer? */
+ if (inptr + (MAX_NEEDED_INPUT - inlen) > inend)
+ {
+#ifdef STORE_REST
+ *inptrp = inend;
+ inptr = bytebuf;
+ inptrp = &inptr;
+ inend = &bytebuf[inlen];
+
+ STORE_REST
+#else
+ /* We don't have enough input for another complete input
+ character. */
+ while (inptr < inend)
+ state->__value.__wchb[inlen++] = *inptr++;
+#endif
+
+ return __GCONV_INCOMPLETE_INPUT;
+ }
+
+ /* Enough space in output buffer. */
+ if ((MIN_NEEDED_OUTPUT != 1 && outptr + MIN_NEEDED_OUTPUT > outend)
+ || (MIN_NEEDED_OUTPUT == 1 && outptr >= outend))
+ /* Overflow in the output buffer. */
+ return __GCONV_FULL_OUTPUT;
+
+ /* Now add characters from the normal input buffer. */
+ do
+ bytebuf[inlen++] = *inptr++;
+ while (inlen < MAX_NEEDED_INPUT);
+
+ inptr = bytebuf;
+ inend = &inptr[MAX_NEEDED_INPUT];
+ do
+ {
+ BODY
+ }
+ while (0);
+
+ if (result == __GCONV_OK)
+ {
+ /* We successfully converted the character (maybe even more).
+ Update the pointers passed in. */
+ assert (inptr - bytebuf > (state->__count & 7));
+
+ *inptrp += inptr - bytebuf - (state->__count & 7);
+ *outptrp = outptr;
+
+ /* Clear the state buffer. */
+ state->__count &= ~7;
+ }
+
+ return result;
+}
+# undef SINGLE
+# undef SINGLE2
+#endif
+
+
/* We remove the macro definitions so that we can include this file again
for the definition of another function. */
#undef MIN_NEEDED_INPUT
#undef EXTRA_LOOP_DECLS
#undef INIT_PARAMS
#undef UPDATE_PARAMS
+#undef get16
+#undef get32
+#undef put16
+#undef put32
+#undef unaligned
+#undef UNPACK_BYTES
(outbuf - outerr) is always divisible by MIN_NEEDED_TO. */
# define RESET_INPUT_BUFFER \
if (MIN_NEEDED_FROM % MIN_NEEDED_TO == 0) \
- *inbuf -= (outbuf - outerr) * (MIN_NEEDED_FROM / MIN_NEEDED_TO); \
+ *inptrp -= (outbuf - outerr) * (MIN_NEEDED_FROM / MIN_NEEDED_TO); \
else if (MIN_NEEDED_TO % MIN_NEEDED_FROM == 0) \
- *inbuf -= (outbuf - outerr) / (MIN_NEEDED_TO / MIN_NEEDED_FROM); \
+ *inptrp -= (outbuf - outerr) / (MIN_NEEDED_TO / MIN_NEEDED_FROM); \
else \
- *inbuf -= ((outbuf - outerr) / MIN_NEEDED_TO) * MIN_NEEDED_FROM
+ *inptrp -= ((outbuf - outerr) / MIN_NEEDED_TO) * MIN_NEEDED_FROM
# endif
#endif
# define FUNCTION_NAME gconv
#endif
+/* The macros are used to access the function to convert single characters. */
+#define SINGLE(fct) SINGLE2 (fct)
+#define SINGLE2(fct) fct##_single
+
+
int
FUNCTION_NAME (struct __gconv_step *step, struct __gconv_step_data *data,
- const unsigned char **inbuf, const unsigned char *inbufend,
- size_t *written, int do_flush)
+ const unsigned char **inptrp, const unsigned char *inend,
+ size_t *written, int do_flush, int consume_incomplete)
{
struct __gconv_step *next_step = step + 1;
struct __gconv_step_data *next_data = data + 1;
successfully emitted the escape sequence. */
if (status == __GCONV_OK && ! data->__is_last)
status = DL_CALL_FCT (fct, (next_step, next_data, NULL, NULL,
- written, 1));
+ written, 1, consume_incomplete));
}
else
{
/* We preserve the initial values of the pointer variables. */
- const unsigned char *inptr = *inbuf;
+ const unsigned char *inptr = *inptrp;
unsigned char *outbuf = data->__outbuf;
unsigned char *outend = data->__outbufend;
unsigned char *outstart;
PREPARE_LOOP
#endif
+#if MAX_NEEDED_FROM > 1 || MAX_NEEDED_TO > 1
+ /* If the function is used to implement the mb*towc*() or wc*tomb*()
+ functions we must test whether any bytes from the last call are
+ stored in the `state' object. */
+ if (((MAX_NEEDED_FROM > 1 && FROM_DIRECTION)
+ || (MAX_NEEDED_TO > 1 && !FROM_DIRECTION))
+ && consume_incomplete && (data->__statep->__count & 7) != 0)
+ {
+ /* Yep, we have some bytes left over. Process them now. */
+
+# if MAX_NEEDED_FROM > 1
+ if (MAX_NEEDED_TO == 1 || FROM_DIRECTION)
+ status = SINGLE(FROM_LOOP) (inptrp, inend, &outbuf, outend,
+ data->__statep, step->__data,
+ &converted EXTRA_LOOP_ARGS);
+# endif
+# if MAX_NEEDED_FROM > 1 && MAX_NEEDED_TO > 1 && !ONE_DIRECTION
+ else
+# endif
+# if MAX_NEEDED_TO > 1 && !ONE_DIRECTION
+ status = SINGLE(TO_LOOP) (inptrp, inend, &outbuf, outend,
+ data->__statep, step->__data,
+ &converted EXTRA_LOOP_ARGS);
+# endif
+
+ if (status != __GCONV_OK)
+ return status;
+ }
+#endif
+
#if !defined _STRING_ARCH_unaligned \
&& MIN_NEEDED_FROM != 1 && MAX_NEEDED_FROM % MIN_NEEDED_FROM == 0 \
&& MIN_NEEDED_TO != 1 && MAX_NEEDED_TO % MIN_NEEDED_TO == 0
do
{
/* Remember the start value for this round. */
- inptr = *inbuf;
+ inptr = *inptrp;
/* The outbuf buffer is empty. */
outstart = outbuf;
{
if (FROM_DIRECTION)
/* Run the conversion loop. */
- status = FROM_LOOP (inbuf, inbufend, &outbuf, outend,
+ status = FROM_LOOP (inptrp, inend, &outbuf, outend,
data->__statep, step->__data, &converted
EXTRA_LOOP_ARGS);
else
/* Run the conversion loop. */
- status = TO_LOOP (inbuf, inbufend, &outbuf, outend,
+ status = TO_LOOP (inptrp, inend, &outbuf, outend,
data->__statep, step->__data, &converted
EXTRA_LOOP_ARGS);
}
{
if (FROM_DIRECTION)
/* Run the conversion loop. */
- status = GEN_unaligned (FROM_LOOP) (inbuf, inbufend, &outbuf,
+ status = GEN_unaligned (FROM_LOOP) (inptrp, inend, &outbuf,
outend, data->__statep,
step->__data, &converted
EXTRA_LOOP_ARGS);
else
/* Run the conversion loop. */
- status = GEN_unaligned (TO_LOOP) (inbuf, inbufend, &outbuf,
+ status = GEN_unaligned (TO_LOOP) (inptrp, inend, &outbuf,
outend, data->__statep,
step->__data, &converted
EXTRA_LOOP_ARGS);
int result;
result = DL_CALL_FCT (fct, (next_step, next_data, &outerr,
- outbuf, written, 0));
+ outbuf, written, 0,
+ consume_incomplete));
if (result != __GCONV_EMPTY_INPUT)
{
size_t nstatus;
/* Reload the pointers. */
- *inbuf = inptr;
+ *inptrp = inptr;
outbuf = outstart;
/* Reset the state. */
if (FROM_DIRECTION)
/* Run the conversion loop. */
- nstatus = FROM_LOOP ((const unsigned char **) inbuf,
- (const unsigned char *) inbufend,
+ nstatus = FROM_LOOP ((const unsigned char **) inptrp,
+ (const unsigned char *) inend,
(unsigned char **) &outbuf,
(unsigned char *) outerr,
data->__statep, step->__data,
&converted EXTRA_LOOP_ARGS);
else
/* Run the conversion loop. */
- nstatus = TO_LOOP ((const unsigned char **) inbuf,
- (const unsigned char *) inbufend,
+ nstatus = TO_LOOP ((const unsigned char **) inptrp,
+ (const unsigned char *) inend,
(unsigned char **) &outbuf,
(unsigned char *) outerr,
data->__statep, step->__data,
#ifdef END_LOOP
END_LOOP
#endif
+
+ /* If we are supposed to consume all character store now all of the
+ remaining characters in the `state' object. */
+#if MAX_NEEDED_FROM > 1 || MAX_NEEDED_TO > 1
+ if (((MAX_NEEDED_FROM > 1 && FROM_DIRECTION)
+ || (MAX_NEEDED_TO > 1 && !FROM_DIRECTION))
+ && consume_incomplete && status == __GCONV_INCOMPLETE_INPUT)
+ {
+# ifdef STORE_REST
+ mbstate_t *state = data->__statep;
+
+ STORE_REST
+# else
+ size_t cnt;
+
+ /* Make sure the remaining bytes fit into the state objects
+ buffer. */
+ assert (inend - *inptrp < 4);
+
+ for (cnt = 0; *inptrp < inend; ++cnt)
+ data->__statep->__value.__wchb[cnt] = *(*inptrp)++;
+ data->__statep->__count &= ~7;
+ data->__statep->__count |= cnt;
+# endif
+ }
+#endif
}
return status;
#undef FUNCTION_NAME
#undef PREPARE_LOOP
#undef END_LOOP
+#undef ONE_DIRECTION
+#undef STORE_REST
/* Mapping tables from GBK to GB2312 and vice versa.
- Copyright (C) 1999 Free Software Foundation, Inc.
+ Copyright (C) 1999, 2000 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1999.
#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
#define MAX_NEEDED_INPUT MAX_NEEDED_FROM
#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
+#define MAX_NEEDED_OUTPUT MAX_NEEDED_TO
#define LOOPFCT FROM_LOOP
#define BODY \
{ \
/* Next, define the other direction. */
#define MIN_NEEDED_INPUT MIN_NEEDED_TO
+#define MAX_NEEDED_INPUT MAX_NEEDED_TO
#define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM
#define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM
#define LOOPFCT TO_LOOP
{ \
/* The second character is not available. Store \
the intermediate result. */ \
- result = __GCONV_INCOMPLETE_INPUT; \
- break; \
+ result = __GCONV_INCOMPLETE_INPUT; \
+ break; \
} \
\
if (NEED_LENGTH_TEST && outend - outptr < 2) \
enum
{
ASCII_set = 0,
- GB2312_set,
- CNS11643_1_set,
- CNS11643_2_set,
- CURRENT_MASK = 3,
- GB2312_ann = 4,
- CNS11643_1_ann = 8,
- CNS11643_2_ann = 16
+ GB2312_set = 8,
+ CNS11643_1_set = 16,
+ CNS11643_2_set = 24,
+ CURRENT_SEL_MASK = 24,
+ GB2312_ann = 32,
+ CNS11643_1_ann = 64,
+ CNS11643_2_ann = 128,
+ CURRENT_ANN_MASK = 224
};
outptr += 4; \
}
#define EXTRA_LOOP_DECLS , int *setp
-#define INIT_PARAMS int set = *setp & CURRENT_MASK; \
- int ann = *setp & ~CURRENT_MASK
+#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \
+ int ann = *setp & CURRENT_ANN_MASK
#define UPDATE_PARAMS *setp = set | ann
#include <iconv/loop.c>
inptr += 4; \
}
#define EXTRA_LOOP_DECLS , int *setp
-#define INIT_PARAMS int set = *setp & CURRENT_MASK; \
- int ann = *setp & ~CURRENT_MASK
+#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \
+ int ann = *setp & CURRENT_ANN_MASK
#define UPDATE_PARAMS *setp = set | ann
#include <iconv/loop.c>
enum
{
ASCII_set = 0,
- JISX0208_1978_set,
- JISX0208_1983_set,
- JISX0201_Roman_set,
- JISX0201_Kana_set,
- GB2312_set,
- KSC5601_set,
- JISX0212_set
+ JISX0208_1978_set = 8,
+ JISX0208_1983_set = 16,
+ JISX0201_Roman_set = 24,
+ JISX0201_Kana_set = 32,
+ GB2312_set = 40,
+ KSC5601_set = 48,
+ JISX0212_set = 56,
+ CURRENT_SEL_MASK = 56
};
/* The second value stored is the designation of the G2 set. The following
enum
{
UNSPECIFIED_set = 0,
- ISO88591_set,
- ISO88597_set
+ ISO88591_set = 64,
+ ISO88597_set = 128,
+ CURRENT_ASSIGN_MASK = 192
};
the output state to the initial state. This has to be done during the
flushing. */
#define EMIT_SHIFT_TO_INIT \
- if (data->__statep->__count != ASCII_set) \
+ if ((data->__statep->__count & ~7) != ASCII_set) \
{ \
enum direction dir = ((struct iso2022jp_data *) step->__data)->dir; \
\
if (dir == from_iso2022jp) \
- /* It's easy, we don't have to emit anything, we just reset the \
- state for the input. Note that this also clears the G2 \
- designation. */ \
- data->__statep->__count = ASCII_set; \
+ { \
+ /* It's easy, we don't have to emit anything, we just reset the \
+ state for the input. Note that this also clears the G2 \
+ designation. */ \
+ data->__statep->__count &= 7; \
+ data->__statep->__count |= ASCII_set; \
+ } \
else \
{ \
unsigned char *outbuf = data->__outbuf; \
*written += 3; \
data->__outbuf = outbuf; \
/* Note that this also clears the G2 designation. */ \
- data->__statep->__count = ASCII_set; \
+ data->__statep->__count &= ~7; \
+ data->__statep->__count |= ASCII_set; \
} \
} \
}
outptr += 4; \
}
#define EXTRA_LOOP_DECLS , enum variant var, int *setp
-#define INIT_PARAMS int set = *setp % 0x100, set2 = *setp / 0x100
-#define UPDATE_PARAMS *setp = (set2 << 8) + set
+#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \
+ int set2 = *setp & CURRENT_ASSIGN_MASK
+#define UPDATE_PARAMS *setp = set | set2
#include <iconv/loop.c>
inptr += 4; \
}
#define EXTRA_LOOP_DECLS , enum variant var, int *setp
-#define INIT_PARAMS int set = *setp % 0x100, set2 = *setp / 0x100
-#define UPDATE_PARAMS *setp = (set2 << 8) + set
+#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \
+ int set2 = *setp & CURRENT_ASSIGN_MASK
+#define UPDATE_PARAMS *setp = set | set2
#include <iconv/loop.c>
enum
{
ASCII_set = 0,
- KSC5601_set
+ KSC5601_set = 8
};
if (data->__statep->__count != ASCII_set) \
{ \
if (FROM_DIRECTION) \
- /* It's easy, we don't have to emit anything, we just reset the \
- state for the input. */ \
- data->__statep->__count = ASCII_set; \
+ { \
+ /* It's easy, we don't have to emit anything, we just reset the \
+ state for the input. */ \
+ data->__statep->__count &= 7; \
+ data->__statep->__count |= ASCII_set; \
+ } \
else \
{ \
unsigned char *outbuf = data->__outbuf; \
if (data->__invocation_counter == 0) \
{ \
/* We have to find out which byte order the file is encoded in. */ \
- if (inptr + 2 > inbufend) \
+ if (inptr + 2 > inend) \
return __GCONV_EMPTY_INPUT; \
\
if (get16u (inptr) == BOM) \
if (data->__invocation_counter == 0) \
{ \
/* We have to find out which byte order the file is encoded in. */ \
- if (inptr + 2 > inbufend) \
+ if (inptr + 2 > inend) \
return __GCONV_EMPTY_INPUT; \
\
if (get16u (inptr) == BOM) \
codecvt->__cd_out.__cd.__data[0].__statep = statep;
status = (*gs->__fct) (gs, codecvt->__cd_out.__cd.__data, &from_start_copy,
- (const unsigned char *) from_end, &written, 0);
+ (const unsigned char *) from_end, &written, 0, 0);
*from_stop = (wchar_t *) from_start_copy;
*to_stop = codecvt->__cd_out.__cd.__data[0].__outbuf;
codecvt->__cd_out.__cd.__data[0].__statep = statep;
status = (*gs->__fct) (gs, codecvt->__cd_out.__cd.__data, NULL, NULL,
- &written, 1);
+ &written, 1, 0);
*to_stop = codecvt->__cd_out.__cd.__data[0].__outbuf;
codecvt->__cd_in.__cd.__data[0].__statep = statep;
status = (*gs->__fct) (gs, codecvt->__cd_in.__cd.__data, &from_start_copy,
- from_end, &written, 0);
+ from_end, &written, 0, 0);
*from_stop = from_start_copy;
*to_stop = (wchar_t *) codecvt->__cd_in.__cd.__data[0].__outbuf;
codecvt->__cd_in.__cd.__data[0].__statep = statep;
status = (*gs->__fct) (gs, codecvt->__cd_in.__cd.__data, &cp, from_end,
- &written, 0);
+ &written, 0, 0);
result = cp - (const unsigned char *) from_start;
#else
-/* Copyright (C) 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
+/* Copyright (C) 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@gnu.org>, 1996.
inbuf[0] = c;
status = (*__wcsmbs_gconv_fcts.towc->__fct) (__wcsmbs_gconv_fcts.towc, &data,
- &inptr, inptr + 1, &dummy, 0);
+ &inptr, inptr + 1, &dummy,
+ 0, 1);
/* The conversion failed. */
if (status != __GCONV_OK && status != __GCONV_FULL_OUTPUT
&& status != __GCONV_EMPTY_INPUT)
inbuf = (const unsigned char *) s;
status = (*__wcsmbs_gconv_fcts.towc->__fct) (__wcsmbs_gconv_fcts.towc,
&data, &inbuf, inbuf + n,
- &dummy, 0);
+ &dummy, 0, 1);
/* There must not be any problems with the conversion but illegal input
characters. The output buffer must be large enough, otherwise the
-/* Copyright (C) 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
+/* Copyright (C) 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@gnu.org>, 1996.
data.__outbuf = (char *) buf;
status = (*towc->__fct) (__wcsmbs_gconv_fcts.towc, &data, &inbuf,
- srcend, &result, 0);
+ srcend, &result, 0, 1);
}
while (status == __GCONV_FULL_OUTPUT);
status = (*towc->__fct) (__wcsmbs_gconv_fcts.towc, &data,
(const unsigned char **) src, srcend,
- &result, 0);
+ &result, 0, 1);
/* We have to determine whether the last character converted
is the NUL character. */
data.__outbuf = (char *) buf;
status = (*towc->__fct) (__wcsmbs_gconv_fcts.towc, &data, &inbuf,
- srcend, &non_reversible, 0);
+ srcend, &non_reversible, 0, 1);
result += (wchar_t *) data.__outbuf - buf;
}
status = (*towc->__fct) (__wcsmbs_gconv_fcts.towc, &data,
(const unsigned char **) src, srcend,
- &non_reversible, 0);
+ &non_reversible, 0, 1);
result = (wchar_t *) data.__outbuf - dst;
{
status = (*__wcsmbs_gconv_fcts.tomb->__fct) (__wcsmbs_gconv_fcts.tomb,
&data, NULL, NULL,
- &dummy, 1);
+ &dummy, 1, 1);
if (status == __GCONV_OK || status == __GCONV_EMPTY_INPUT)
*data.__outbuf++ = '\0';
status = (*__wcsmbs_gconv_fcts.tomb->__fct) (__wcsmbs_gconv_fcts.tomb,
&data, &inbuf,
inbuf + sizeof (wchar_t),
- &dummy, 0);
+ &dummy, 0, 1);
}
/* There must not be any problems with the conversion but illegal input
-/* Copyright (C) 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
+/* Copyright (C) 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@gnu.org>, 1996.
status = (*tomb->__fct) (__wcsmbs_gconv_fcts.tomb, &data,
(const unsigned char **) &inbuf,
- (const unsigned char *) srcend, &dummy, 0);
+ (const unsigned char *) srcend, &dummy,
+ 0, 1);
/* Count the number of bytes. */
result += data.__outbuf - buf;
status = (*tomb->__fct) (__wcsmbs_gconv_fcts.tomb, &data,
(const unsigned char **) src,
- (const unsigned char *) srcend, &dummy, 0);
+ (const unsigned char *) srcend, &dummy, 0, 1);
/* Count the number of bytes. */
result = data.__outbuf - (unsigned char *) dst;
status = (*tomb->__fct) (__wcsmbs_gconv_fcts.tomb, &data,
(const unsigned char **) &inbuf,
- (const unsigned char *) srcend, &dummy, 0);
+ (const unsigned char *) srcend, &dummy,
+ 0, 1);
/* Count the number of bytes. */
result += data.__outbuf - buf;
status = (*tomb->__fct) (__wcsmbs_gconv_fcts.tomb, &data,
(const unsigned char **) src,
- (const unsigned char *) srcend, &dummy, 0);
+ (const unsigned char *) srcend, &dummy, 0, 1);
/* Count the number of bytes. */
result = data.__outbuf - (unsigned char *) dst;
-/* Copyright (C) 1996, 1997, 1998, 1999 Free Software Foundation, Inc.
+/* Copyright (C) 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1996.
status = (*__wcsmbs_gconv_fcts.tomb->__fct) (__wcsmbs_gconv_fcts.tomb, &data,
(const unsigned char **) &inptr,
(const unsigned char *) &inbuf[1],
- &dummy, 0);
+ &dummy, 0, 1);
/* The conversion failed or the output is too long. */
if ((status != __GCONV_OK && status != __GCONV_FULL_OUTPUT
&& status != __GCONV_EMPTY_INPUT)
projects / external / glibc.git / commitdiff