2 Unix SMB/CIFS implementation.
3 Character set conversion Extensions
4 Copyright (C) Igor Vergeichik <iverg@mail.ru> 2001
5 Copyright (C) Andrew Tridgell 2001
6 Copyright (C) Simo Sorce 2001
7 Copyright (C) Martin Pool 2003
9 This program is free software; you can redistribute it and/or modify
10 it under the terms of the GNU General Public License as published by
11 the Free Software Foundation; either version 2 of the License, or
12 (at your option) any later version.
14 This program is distributed in the hope that it will be useful,
15 but WITHOUT ANY WARRANTY; without even the implied warranty of
16 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 GNU General Public License for more details.
19 You should have received a copy of the GNU General Public License
20 along with this program; if not, write to the Free Software
21 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
29 * @brief Character-set conversion routines built on our iconv.
31 * @note Samba's internal character set (at least in the 3.0 series)
32 * is always the same as the one for the Unix filesystem. It is
33 * <b>not</b> necessarily UTF-8 and may be different on machines that
34 * need i18n filenames to be compatible with Unix software. It does
35 * have to be a superset of ASCII. All multibyte sequences must start
36 * with a byte with the high bit set.
42 static smb_iconv_t conv_handles
[NUM_CHARSETS
][NUM_CHARSETS
];
43 static BOOL conv_silent
; /* Should we do a debug if the conversion fails ? */
46 * Return the name of a charset to give to iconv().
48 static const char *charset_name(charset_t ch
)
50 const char *ret
= NULL
;
52 if (ch
== CH_UCS2
) ret
= "UCS-2LE";
53 else if (ch
== CH_UNIX
) ret
= lp_unix_charset();
54 else if (ch
== CH_DOS
) ret
= lp_dos_charset();
55 else if (ch
== CH_DISPLAY
) ret
= lp_display_charset();
56 else if (ch
== CH_UTF8
) ret
= "UTF8";
58 #if defined(HAVE_NL_LANGINFO) && defined(CODESET)
59 if (ret
&& !strcmp(ret
, "LOCALE")) {
60 const char *ln
= NULL
;
63 setlocale(LC_ALL
, "");
65 ln
= nl_langinfo(CODESET
);
67 /* Check whether the charset name is supported
69 smb_iconv_t handle
= smb_iconv_open(ln
,"UCS-2LE");
70 if (handle
== (smb_iconv_t
) -1) {
71 DEBUG(5,("Locale charset '%s' unsupported, using ASCII instead\n", ln
));
74 DEBUG(5,("Substituting charset '%s' for LOCALE\n", ln
));
75 smb_iconv_close(handle
);
82 if (!ret
|| !*ret
) ret
= "ASCII";
86 void lazy_initialize_conv(void)
88 static int initialized
= False
;
98 * Initialize iconv conversion descriptors.
100 * This is called the first time it is needed, and also called again
101 * every time the configuration is reloaded, because the charset or
102 * codepage might have changed.
104 void init_iconv(void)
107 BOOL did_reload
= False
;
109 /* so that charset_name() works we need to get the UNIX<->UCS2 going
111 if (!conv_handles
[CH_UNIX
][CH_UCS2
])
112 conv_handles
[CH_UNIX
][CH_UCS2
] = smb_iconv_open("UCS-2LE", "ASCII");
114 if (!conv_handles
[CH_UCS2
][CH_UNIX
])
115 conv_handles
[CH_UCS2
][CH_UNIX
] = smb_iconv_open("ASCII", "UCS-2LE");
117 for (c1
=0;c1
<NUM_CHARSETS
;c1
++) {
118 for (c2
=0;c2
<NUM_CHARSETS
;c2
++) {
119 const char *n1
= charset_name((charset_t
)c1
);
120 const char *n2
= charset_name((charset_t
)c2
);
121 if (conv_handles
[c1
][c2
] &&
122 strcmp(n1
, conv_handles
[c1
][c2
]->from_name
) == 0 &&
123 strcmp(n2
, conv_handles
[c1
][c2
]->to_name
) == 0)
128 if (conv_handles
[c1
][c2
])
129 smb_iconv_close(conv_handles
[c1
][c2
]);
131 conv_handles
[c1
][c2
] = smb_iconv_open(n2
,n1
);
132 if (conv_handles
[c1
][c2
] == (smb_iconv_t
)-1) {
133 DEBUG(0,("Conversion from %s to %s not supported\n",
134 charset_name((charset_t
)c1
), charset_name((charset_t
)c2
)));
135 conv_handles
[c1
][c2
] = NULL
;
141 /* XXX: Does this really get called every time the dos
142 * codepage changes? */
143 /* XXX: Is the did_reload test too strict? */
145 init_doschar_table();
152 * Convert string from one encoding to another, making error checking etc
153 * Slow path version - uses (slow) iconv.
155 * @param src pointer to source string (multibyte or singlebyte)
156 * @param srclen length of the source string in bytes
157 * @param dest pointer to destination string (multibyte or singlebyte)
158 * @param destlen maximal length allowed for string
159 * @param allow_bad_conv determines if a "best effort" conversion is acceptable (never returns errors)
160 * @returns the number of bytes occupied in the destination
162 * Ensure the srclen contains the terminating zero.
166 static size_t convert_string_internal(charset_t from
, charset_t to
,
167 void const *src
, size_t srclen
,
168 void *dest
, size_t destlen
, BOOL allow_bad_conv
)
172 const char* inbuf
= (const char*)src
;
173 char* outbuf
= (char*)dest
;
174 smb_iconv_t descriptor
;
176 lazy_initialize_conv();
178 descriptor
= conv_handles
[from
][to
];
180 if (srclen
== (size_t)-1) {
181 if (from
== CH_UCS2
) {
182 srclen
= (strlen_w((const smb_ucs2_t
*)src
)+1) * 2;
184 srclen
= strlen((const char *)src
)+1;
189 if (descriptor
== (smb_iconv_t
)-1 || descriptor
== (smb_iconv_t
)0) {
191 DEBUG(0,("convert_string_internal: Conversion not supported.\n"));
200 retval
= smb_iconv(descriptor
, (char **)&inbuf
, &i_len
, &outbuf
, &o_len
);
201 if(retval
==(size_t)-1) {
202 const char *reason
="unknown error";
205 reason
="Incomplete multibyte sequence";
207 DEBUG(3,("convert_string_internal: Conversion error: %s(%s)\n",reason
,inbuf
));
212 reason
="No more room";
214 DEBUG(3, ("convert_string_internal: Required %lu, available %lu\n",
215 (unsigned long)srclen
, (unsigned long)destlen
));
216 /* we are not sure we need srclen bytes,
217 may be more, may be less.
218 We only know we need more than destlen
222 reason
="Illegal multibyte sequence";
224 DEBUG(3,("convert_string_internal: Conversion error: %s(%s)\n",reason
,inbuf
));
230 DEBUG(0,("convert_string_internal: Conversion error: %s(%s)\n",reason
,inbuf
));
233 /* smb_panic(reason); */
235 return destlen
-o_len
;
240 * Conversion not supported. This is actually an error, but there are so
241 * many misconfigured iconv systems and smb.conf's out there we can't just
242 * fail. Do a very bad conversion instead.... JRA.
246 if (o_len
== 0 || i_len
== 0)
247 return destlen
- o_len
;
249 if (from
== CH_UCS2
&& to
!= CH_UCS2
) {
250 /* Can't convert from ucs2 to multibyte. Just truncate this char to ascii. */
252 return destlen
- o_len
;
263 if (o_len
== 0 || i_len
== 0)
264 return destlen
- o_len
;
266 /* Keep trying with the next char... */
269 } else if (from
!= CH_UCS2
&& to
== CH_UCS2
) {
270 /* Can't convert to ucs2 - just widen by adding zero. */
272 return destlen
- o_len
;
274 outbuf
[0] = inbuf
[0];
283 if (o_len
== 0 || i_len
== 0)
284 return destlen
- o_len
;
286 /* Keep trying with the next char... */
289 } else if (from
!= CH_UCS2
&& to
!= CH_UCS2
) {
290 /* Failed multibyte to multibyte. Just copy 1 char and
292 outbuf
[0] = inbuf
[0];
300 if (o_len
== 0 || i_len
== 0)
301 return destlen
- o_len
;
303 /* Keep trying with the next char... */
307 /* Keep compiler happy.... */
308 return destlen
- o_len
;
314 * Convert string from one encoding to another, making error checking etc
315 * Fast path version - handles ASCII first.
317 * @param src pointer to source string (multibyte or singlebyte)
318 * @param srclen length of the source string in bytes, or -1 for nul terminated.
319 * @param dest pointer to destination string (multibyte or singlebyte)
320 * @param destlen maximal length allowed for string - *NEVER* -1.
321 * @param allow_bad_conv determines if a "best effort" conversion is acceptable (never returns errors)
322 * @returns the number of bytes occupied in the destination
324 * Ensure the srclen contains the terminating zero.
326 * This function has been hand-tuned to provide a fast path.
327 * Don't change unless you really know what you are doing. JRA.
330 size_t convert_string(charset_t from
, charset_t to
,
331 void const *src
, size_t srclen
,
332 void *dest
, size_t destlen
, BOOL allow_bad_conv
)
335 * NB. We deliberately don't do a strlen here if srclen == -1.
336 * This is very expensive over millions of calls and is taken
337 * care of in the slow path in convert_string_internal. JRA.
341 SMB_ASSERT(destlen
!= (size_t)-1);
347 if (from
!= CH_UCS2
&& to
!= CH_UCS2
) {
348 const unsigned char *p
= (const unsigned char *)src
;
349 unsigned char *q
= (unsigned char *)dest
;
350 size_t slen
= srclen
;
351 size_t dlen
= destlen
;
355 /* If all characters are ascii, fast path here. */
356 while (slen
&& dlen
) {
357 if ((lastp
= *p
) <= 0x7f) {
359 if (slen
!= (size_t)-1) {
367 #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
370 return retval
+ convert_string_internal(from
, to
, p
, slen
, q
, dlen
, allow_bad_conv
);
375 } else if (from
== CH_UCS2
&& to
!= CH_UCS2
) {
376 const unsigned char *p
= (const unsigned char *)src
;
377 unsigned char *q
= (unsigned char *)dest
;
379 size_t slen
= srclen
;
380 size_t dlen
= destlen
;
383 /* If all characters are ascii, fast path here. */
384 while (((slen
== (size_t)-1) || (slen
>= 2)) && dlen
) {
385 if (((lastp
= *p
) <= 0x7f) && (p
[1] == 0)) {
387 if (slen
!= (size_t)-1) {
396 #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
399 return retval
+ convert_string_internal(from
, to
, p
, slen
, q
, dlen
, allow_bad_conv
);
404 } else if (from
!= CH_UCS2
&& to
== CH_UCS2
) {
405 const unsigned char *p
= (const unsigned char *)src
;
406 unsigned char *q
= (unsigned char *)dest
;
408 size_t slen
= srclen
;
409 size_t dlen
= destlen
;
412 /* If all characters are ascii, fast path here. */
413 while (slen
&& (dlen
>= 2)) {
414 if ((lastp
= *p
) <= 0x7F) {
417 if (slen
!= (size_t)-1) {
425 #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
428 return retval
+ convert_string_internal(from
, to
, p
, slen
, q
, dlen
, allow_bad_conv
);
435 #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
438 return convert_string_internal(from
, to
, src
, srclen
, dest
, destlen
, allow_bad_conv
);
442 * Convert between character sets, allocating a new buffer for the result.
444 * @param ctx TALLOC_CTX to use to allocate with. If NULL use malloc.
445 * @param srclen length of source buffer.
446 * @param dest always set at least to NULL
447 * @note -1 is not accepted for srclen.
449 * @returns Size in bytes of the converted string; or -1 in case of error.
451 * Ensure the srclen contains the terminating zero.
453 * I hate the goto's in this function. It's embarressing.....
454 * There has to be a cleaner way to do this. JRA.
457 size_t convert_string_allocate(TALLOC_CTX
*ctx
, charset_t from
, charset_t to
,
458 void const *src
, size_t srclen
, void **dest
, BOOL allow_bad_conv
)
460 size_t i_len
, o_len
, destlen
= MAX(srclen
, 512);
462 const char *inbuf
= (const char *)src
;
463 char *outbuf
= NULL
, *ob
= NULL
;
464 smb_iconv_t descriptor
;
468 if (src
== NULL
|| srclen
== (size_t)-1)
473 lazy_initialize_conv();
475 descriptor
= conv_handles
[from
][to
];
477 if (descriptor
== (smb_iconv_t
)-1 || descriptor
== (smb_iconv_t
)0) {
479 DEBUG(0,("convert_string_allocate: Conversion not supported.\n"));
487 if ((destlen
*2) < destlen
) {
488 /* wrapped ! abort. */
490 DEBUG(0, ("convert_string_allocate: destlen wrapped !\n"));
495 destlen
= destlen
* 2;
499 ob
= (char *)talloc_realloc(ctx
, ob
, destlen
);
501 ob
= (char *)Realloc(ob
, destlen
);
504 DEBUG(0, ("convert_string_allocate: realloc failed!\n"));
516 retval
= smb_iconv(descriptor
,
517 (char **)&inbuf
, &i_len
,
519 if(retval
== (size_t)-1) {
520 const char *reason
="unknown error";
523 reason
="Incomplete multibyte sequence";
525 DEBUG(3,("convert_string_allocate: Conversion error: %s(%s)\n",reason
,inbuf
));
532 reason
="Illegal multibyte sequence";
534 DEBUG(3,("convert_string_allocate: Conversion error: %s(%s)\n",reason
,inbuf
));
540 DEBUG(0,("Conversion error: %s(%s)\n",reason
,inbuf
));
541 /* smb_panic(reason); */
547 destlen
= destlen
- o_len
;
549 *dest
= (char *)talloc_realloc(ctx
,ob
,destlen
);
551 *dest
= (char *)Realloc(ob
,destlen
);
552 if (destlen
&& !*dest
) {
553 DEBUG(0, ("convert_string_allocate: out of memory!\n"));
564 * Conversion not supported. This is actually an error, but there are so
565 * many misconfigured iconv systems and smb.conf's out there we can't just
566 * fail. Do a very bad conversion instead.... JRA.
570 if (o_len
== 0 || i_len
== 0)
573 if (from
== CH_UCS2
&& to
!= CH_UCS2
) {
574 /* Can't convert from ucs2 to multibyte. Just truncate this char to ascii. */
588 if (o_len
== 0 || i_len
== 0)
591 /* Keep trying with the next char... */
594 } else if (from
!= CH_UCS2
&& to
== CH_UCS2
) {
595 /* Can't convert to ucs2 - just widen by adding zero. */
599 outbuf
[0] = inbuf
[0];
608 if (o_len
== 0 || i_len
== 0)
611 /* Keep trying with the next char... */
614 } else if (from
!= CH_UCS2
&& to
!= CH_UCS2
) {
615 /* Failed multibyte to multibyte. Just copy 1 char and
617 outbuf
[0] = inbuf
[0];
625 if (o_len
== 0 || i_len
== 0)
628 /* Keep trying with the next char... */
632 /* Keep compiler happy.... */
639 * Convert between character sets, allocating a new buffer using talloc for the result.
641 * @param srclen length of source buffer.
642 * @param dest always set at least to NULL
643 * @note -1 is not accepted for srclen.
645 * @returns Size in bytes of the converted string; or -1 in case of error.
647 static size_t convert_string_talloc(TALLOC_CTX
*ctx
, charset_t from
, charset_t to
,
648 void const *src
, size_t srclen
, void **dest
, BOOL allow_bad_conv
)
653 dest_len
=convert_string_allocate(ctx
, from
, to
, src
, srclen
, dest
, allow_bad_conv
);
654 if (dest_len
== (size_t)-1)
661 size_t unix_strupper(const char *src
, size_t srclen
, char *dest
, size_t destlen
)
666 size
= push_ucs2_allocate(&buffer
, src
);
667 if (size
== (size_t)-1) {
668 smb_panic("failed to create UCS2 buffer");
670 if (!strupper_w(buffer
) && (dest
== src
)) {
675 size
= convert_string(CH_UCS2
, CH_UNIX
, buffer
, size
, dest
, destlen
, True
);
681 strdup() a unix string to upper case.
685 char *strdup_upper(const char *s
)
688 const unsigned char *p
= (const unsigned char *)s
;
689 unsigned char *q
= (unsigned char *)out_buffer
;
691 /* this is quite a common operation, so we want it to be
692 fast. We optimise for the ascii case, knowing that all our
693 supported multi-byte character sets are ascii-compatible
694 (ie. they match for the first 128 chars) */
703 if (p
- ( const unsigned char *)s
>= sizeof(pstring
))
711 size
= convert_string(CH_UNIX
, CH_UCS2
, s
, -1, buffer
, sizeof(buffer
), True
);
712 if (size
== (size_t)-1) {
718 size
= convert_string(CH_UCS2
, CH_UNIX
, buffer
, -1, out_buffer
, sizeof(out_buffer
), True
);
719 if (size
== (size_t)-1) {
724 return strdup(out_buffer
);
727 size_t unix_strlower(const char *src
, size_t srclen
, char *dest
, size_t destlen
)
730 smb_ucs2_t
*buffer
= NULL
;
732 size
= convert_string_allocate(NULL
, CH_UNIX
, CH_UCS2
, src
, srclen
,
733 (void **) &buffer
, True
);
734 if (size
== (size_t)-1 || !buffer
) {
735 smb_panic("failed to create UCS2 buffer");
737 if (!strlower_w(buffer
) && (dest
== src
)) {
741 size
= convert_string(CH_UCS2
, CH_UNIX
, buffer
, size
, dest
, destlen
, True
);
747 strdup() a unix string to lower case.
750 char *strdup_lower(const char *s
)
753 smb_ucs2_t
*buffer
= NULL
;
756 size
= push_ucs2_allocate(&buffer
, s
);
757 if (size
== -1 || !buffer
) {
763 size
= pull_ucs2_allocate(&out_buffer
, buffer
);
766 if (size
== (size_t)-1) {
773 static size_t ucs2_align(const void *base_ptr
, const void *p
, int flags
)
775 if (flags
& (STR_NOALIGN
|STR_ASCII
))
777 return PTR_DIFF(p
, base_ptr
) & 1;
782 * Copy a string from a char* unix src to a dos codepage string destination.
784 * @return the number of bytes occupied by the string in the destination.
786 * @param flags can include
788 * <dt>STR_TERMINATE</dt> <dd>means include the null termination</dd>
789 * <dt>STR_UPPER</dt> <dd>means uppercase in the destination</dd>
792 * @param dest_len the maximum length in bytes allowed in the
793 * destination. If @p dest_len is -1 then no maximum is used.
795 size_t push_ascii(void *dest
, const char *src
, size_t dest_len
, int flags
)
797 size_t src_len
= strlen(src
);
800 /* treat a pstring as "unlimited" length */
801 if (dest_len
== (size_t)-1)
802 dest_len
= sizeof(pstring
);
804 if (flags
& STR_UPPER
) {
805 pstrcpy(tmpbuf
, src
);
810 if (flags
& (STR_TERMINATE
| STR_TERMINATE_ASCII
))
813 return convert_string(CH_UNIX
, CH_DOS
, src
, src_len
, dest
, dest_len
, True
);
816 size_t push_ascii_fstring(void *dest
, const char *src
)
818 return push_ascii(dest
, src
, sizeof(fstring
), STR_TERMINATE
);
821 size_t push_ascii_pstring(void *dest
, const char *src
)
823 return push_ascii(dest
, src
, sizeof(pstring
), STR_TERMINATE
);
826 /********************************************************************
827 Push an nstring - ensure null terminated. Written by
828 moriyama@miraclelinux.com (MORIYAMA Masayuki).
829 ********************************************************************/
831 size_t push_ascii_nstring(void *dest
, const char *src
)
833 size_t i
, buffer_len
, dest_len
;
836 buffer_len
= push_ucs2_allocate(&buffer
, src
);
837 if (buffer_len
== (size_t)-1) {
838 smb_panic("failed to create UCS2 buffer");
842 for (i
= 0; buffer
[i
] != 0 && (i
< buffer_len
); i
++) {
843 unsigned char mb
[10];
844 /* Convert one smb_ucs2_t character at a time. */
845 size_t mb_len
= convert_string(CH_UCS2
, CH_DOS
, buffer
+i
, sizeof(smb_ucs2_t
), mb
, sizeof(mb
), False
);
846 if ((mb_len
!= (size_t)-1) && (dest_len
+ mb_len
<= MAX_NETBIOSNAME_LEN
- 1)) {
847 memcpy((char *)dest
+ dest_len
, mb
, mb_len
);
854 ((char *)dest
)[dest_len
] = '\0';
861 * Copy a string from a dos codepage source to a unix char* destination.
863 * The resulting string in "dest" is always null terminated.
865 * @param flags can have:
867 * <dt>STR_TERMINATE</dt>
868 * <dd>STR_TERMINATE means the string in @p src
869 * is null terminated, and src_len is ignored.</dd>
872 * @param src_len is the length of the source area in bytes.
873 * @returns the number of bytes occupied by the string in @p src.
875 size_t pull_ascii(char *dest
, const void *src
, size_t dest_len
, size_t src_len
, int flags
)
879 if (dest_len
== (size_t)-1)
880 dest_len
= sizeof(pstring
);
882 if (flags
& STR_TERMINATE
) {
883 if (src_len
== (size_t)-1) {
884 src_len
= strlen(src
) + 1;
886 size_t len
= strnlen(src
, src_len
);
893 ret
= convert_string(CH_DOS
, CH_UNIX
, src
, src_len
, dest
, dest_len
, True
);
894 if (ret
== (size_t)-1) {
899 dest
[MIN(ret
, dest_len
-1)] = 0;
906 size_t pull_ascii_pstring(char *dest
, const void *src
)
908 return pull_ascii(dest
, src
, sizeof(pstring
), -1, STR_TERMINATE
);
911 size_t pull_ascii_fstring(char *dest
, const void *src
)
913 return pull_ascii(dest
, src
, sizeof(fstring
), -1, STR_TERMINATE
);
916 /* When pulling an nstring it can expand into a larger size (dos cp -> utf8). Cope with this. */
918 size_t pull_ascii_nstring(char *dest
, size_t dest_len
, const void *src
)
920 return pull_ascii(dest
, src
, dest_len
, sizeof(nstring
), STR_TERMINATE
);
924 * Copy a string from a char* src to a unicode destination.
926 * @returns the number of bytes occupied by the string in the destination.
928 * @param flags can have:
931 * <dt>STR_TERMINATE <dd>means include the null termination.
932 * <dt>STR_UPPER <dd>means uppercase in the destination.
933 * <dt>STR_NOALIGN <dd>means don't do alignment.
936 * @param dest_len is the maximum length allowed in the
937 * destination. If dest_len is -1 then no maxiumum is used.
940 size_t push_ucs2(const void *base_ptr
, void *dest
, const char *src
, size_t dest_len
, int flags
)
946 /* treat a pstring as "unlimited" length */
947 if (dest_len
== (size_t)-1)
948 dest_len
= sizeof(pstring
);
950 if (flags
& STR_TERMINATE
)
951 src_len
= (size_t)-1;
953 src_len
= strlen(src
);
955 if (ucs2_align(base_ptr
, dest
, flags
)) {
957 dest
= (void *)((char *)dest
+ 1);
963 /* ucs2 is always a multiple of 2 bytes */
966 ret
= convert_string(CH_UNIX
, CH_UCS2
, src
, src_len
, dest
, dest_len
, True
);
967 if (ret
== (size_t)-1) {
973 if (flags
& STR_UPPER
) {
974 smb_ucs2_t
*dest_ucs2
= dest
;
976 for (i
= 0; i
< (dest_len
/ 2) && dest_ucs2
[i
]; i
++) {
977 smb_ucs2_t v
= toupper_w(dest_ucs2
[i
]);
978 if (v
!= dest_ucs2
[i
]) {
989 * Copy a string from a unix char* src to a UCS2 destination,
990 * allocating a buffer using talloc().
992 * @param dest always set at least to NULL
994 * @returns The number of bytes occupied by the string in the destination
995 * or -1 in case of error.
997 size_t push_ucs2_talloc(TALLOC_CTX
*ctx
, smb_ucs2_t
**dest
, const char *src
)
999 size_t src_len
= strlen(src
)+1;
1002 return convert_string_talloc(ctx
, CH_UNIX
, CH_UCS2
, src
, src_len
, (void **)dest
, True
);
1007 * Copy a string from a unix char* src to a UCS2 destination, allocating a buffer
1009 * @param dest always set at least to NULL
1011 * @returns The number of bytes occupied by the string in the destination
1012 * or -1 in case of error.
1015 size_t push_ucs2_allocate(smb_ucs2_t
**dest
, const char *src
)
1017 size_t src_len
= strlen(src
)+1;
1020 return convert_string_allocate(NULL
, CH_UNIX
, CH_UCS2
, src
, src_len
, (void **)dest
, True
);
1024 Copy a string from a char* src to a UTF-8 destination.
1025 Return the number of bytes occupied by the string in the destination
1027 STR_TERMINATE means include the null termination
1028 STR_UPPER means uppercase in the destination
1029 dest_len is the maximum length allowed in the destination. If dest_len
1030 is -1 then no maxiumum is used.
1033 static size_t push_utf8(void *dest
, const char *src
, size_t dest_len
, int flags
)
1035 size_t src_len
= strlen(src
);
1038 /* treat a pstring as "unlimited" length */
1039 if (dest_len
== (size_t)-1)
1040 dest_len
= sizeof(pstring
);
1042 if (flags
& STR_UPPER
) {
1043 pstrcpy(tmpbuf
, src
);
1048 if (flags
& STR_TERMINATE
)
1051 return convert_string(CH_UNIX
, CH_UTF8
, src
, src_len
, dest
, dest_len
, True
);
1054 size_t push_utf8_fstring(void *dest
, const char *src
)
1056 return push_utf8(dest
, src
, sizeof(fstring
), STR_TERMINATE
);
1060 * Copy a string from a unix char* src to a UTF-8 destination, allocating a buffer using talloc
1062 * @param dest always set at least to NULL
1064 * @returns The number of bytes occupied by the string in the destination
1067 size_t push_utf8_talloc(TALLOC_CTX
*ctx
, char **dest
, const char *src
)
1069 size_t src_len
= strlen(src
)+1;
1072 return convert_string_talloc(ctx
, CH_UNIX
, CH_UTF8
, src
, src_len
, (void**)dest
, True
);
1076 * Copy a string from a unix char* src to a UTF-8 destination, allocating a buffer
1078 * @param dest always set at least to NULL
1080 * @returns The number of bytes occupied by the string in the destination
1083 size_t push_utf8_allocate(char **dest
, const char *src
)
1085 size_t src_len
= strlen(src
)+1;
1088 return convert_string_allocate(NULL
, CH_UNIX
, CH_UTF8
, src
, src_len
, (void **)dest
, True
);
1092 Copy a string from a ucs2 source to a unix char* destination.
1094 STR_TERMINATE means the string in src is null terminated.
1095 STR_NOALIGN means don't try to align.
1096 if STR_TERMINATE is set then src_len is ignored if it is -1.
1097 src_len is the length of the source area in bytes
1098 Return the number of bytes occupied by the string in src.
1099 The resulting string in "dest" is always null terminated.
1102 size_t pull_ucs2(const void *base_ptr
, char *dest
, const void *src
, size_t dest_len
, size_t src_len
, int flags
)
1106 if (dest_len
== (size_t)-1)
1107 dest_len
= sizeof(pstring
);
1109 if (ucs2_align(base_ptr
, src
, flags
)) {
1110 src
= (const void *)((const char *)src
+ 1);
1111 if (src_len
!= (size_t)-1)
1115 if (flags
& STR_TERMINATE
) {
1116 /* src_len -1 is the default for null terminated strings. */
1117 if (src_len
!= (size_t)-1) {
1118 size_t len
= strnlen_w(src
, src_len
/2);
1119 if (len
< src_len
/2)
1125 /* ucs2 is always a multiple of 2 bytes */
1126 if (src_len
!= (size_t)-1)
1129 ret
= convert_string(CH_UCS2
, CH_UNIX
, src
, src_len
, dest
, dest_len
, True
);
1130 if (ret
== (size_t)-1) {
1134 if (src_len
== (size_t)-1)
1138 dest
[MIN(ret
, dest_len
-1)] = 0;
1145 size_t pull_ucs2_pstring(char *dest
, const void *src
)
1147 return pull_ucs2(NULL
, dest
, src
, sizeof(pstring
), -1, STR_TERMINATE
);
1150 size_t pull_ucs2_fstring(char *dest
, const void *src
)
1152 return pull_ucs2(NULL
, dest
, src
, sizeof(fstring
), -1, STR_TERMINATE
);
1156 * Copy a string from a UCS2 src to a unix char * destination, allocating a buffer using talloc
1158 * @param dest always set at least to NULL
1160 * @returns The number of bytes occupied by the string in the destination
1163 size_t pull_ucs2_talloc(TALLOC_CTX
*ctx
, char **dest
, const smb_ucs2_t
*src
)
1165 size_t src_len
= (strlen_w(src
)+1) * sizeof(smb_ucs2_t
);
1167 return convert_string_talloc(ctx
, CH_UCS2
, CH_UNIX
, src
, src_len
, (void **)dest
, True
);
1171 * Copy a string from a UCS2 src to a unix char * destination, allocating a buffer
1173 * @param dest always set at least to NULL
1175 * @returns The number of bytes occupied by the string in the destination
1178 size_t pull_ucs2_allocate(char **dest
, const smb_ucs2_t
*src
)
1180 size_t src_len
= (strlen_w(src
)+1) * sizeof(smb_ucs2_t
);
1182 return convert_string_allocate(NULL
, CH_UCS2
, CH_UNIX
, src
, src_len
, (void **)dest
, True
);
1186 * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer using talloc
1188 * @param dest always set at least to NULL
1190 * @returns The number of bytes occupied by the string in the destination
1193 size_t pull_utf8_talloc(TALLOC_CTX
*ctx
, char **dest
, const char *src
)
1195 size_t src_len
= strlen(src
)+1;
1197 return convert_string_talloc(ctx
, CH_UTF8
, CH_UNIX
, src
, src_len
, (void **)dest
, True
);
1201 * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer
1203 * @param dest always set at least to NULL
1205 * @returns The number of bytes occupied by the string in the destination
1208 size_t pull_utf8_allocate(char **dest
, const char *src
)
1210 size_t src_len
= strlen(src
)+1;
1212 return convert_string_allocate(NULL
, CH_UTF8
, CH_UNIX
, src
, src_len
, (void **)dest
, True
);
1216 Copy a string from a char* src to a unicode or ascii
1217 dos codepage destination choosing unicode or ascii based on the
1218 flags in the SMB buffer starting at base_ptr.
1219 Return the number of bytes occupied by the string in the destination.
1221 STR_TERMINATE means include the null termination.
1222 STR_UPPER means uppercase in the destination.
1223 STR_ASCII use ascii even with unicode packet.
1224 STR_NOALIGN means don't do alignment.
1225 dest_len is the maximum length allowed in the destination. If dest_len
1226 is -1 then no maxiumum is used.
1229 size_t push_string_fn(const char *function
, unsigned int line
, const void *base_ptr
, void *dest
, const char *src
, size_t dest_len
, int flags
)
1232 /* We really need to zero fill here, not clobber
1233 * region, as we want to ensure that valgrind thinks
1234 * all of the outgoing buffer has been written to
1235 * so a send() or write() won't trap an error.
1239 if (dest_len
!= (size_t)-1)
1240 clobber_region(function
, line
, dest
, dest_len
);
1242 if (dest_len
!= (size_t)-1)
1243 memset(dest
, '\0', dest_len
);
1247 if (!(flags
& STR_ASCII
) && \
1248 ((flags
& STR_UNICODE
|| \
1249 (SVAL(base_ptr
, smb_flg2
) & FLAGS2_UNICODE_STRINGS
)))) {
1250 return push_ucs2(base_ptr
, dest
, src
, dest_len
, flags
);
1252 return push_ascii(dest
, src
, dest_len
, flags
);
1257 Copy a string from a unicode or ascii source (depending on
1258 the packet flags) to a char* destination.
1260 STR_TERMINATE means the string in src is null terminated.
1261 STR_UNICODE means to force as unicode.
1262 STR_ASCII use ascii even with unicode packet.
1263 STR_NOALIGN means don't do alignment.
1264 if STR_TERMINATE is set then src_len is ignored is it is -1
1265 src_len is the length of the source area in bytes.
1266 Return the number of bytes occupied by the string in src.
1267 The resulting string in "dest" is always null terminated.
1270 size_t pull_string_fn(const char *function
, unsigned int line
, const void *base_ptr
, char *dest
, const void *src
, size_t dest_len
, size_t src_len
, int flags
)
1273 if (dest_len
!= (size_t)-1)
1274 clobber_region(function
, line
, dest
, dest_len
);
1277 if (!(flags
& STR_ASCII
) && \
1278 ((flags
& STR_UNICODE
|| \
1279 (SVAL(base_ptr
, smb_flg2
) & FLAGS2_UNICODE_STRINGS
)))) {
1280 return pull_ucs2(base_ptr
, dest
, src
, dest_len
, src_len
, flags
);
1282 return pull_ascii(dest
, src
, dest_len
, src_len
, flags
);
1285 size_t align_string(const void *base_ptr
, const char *p
, int flags
)
1287 if (!(flags
& STR_ASCII
) && \
1288 ((flags
& STR_UNICODE
|| \
1289 (SVAL(base_ptr
, smb_flg2
) & FLAGS2_UNICODE_STRINGS
)))) {
1290 return ucs2_align(base_ptr
, p
, flags
);