1 /* GLIB - Library of useful routines for C programming
3 * gconvert.c: Convert between character sets using iconv
4 * Copyright Red Hat Inc., 2000
5 * Authors: Havoc Pennington <hp@redhat.com>, Owen Taylor <otaylor@redhat.com
7 * This library is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2 of the License, or (at your option) any later version.
12 * This library is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with this library; if not, write to the
19 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 * Boston, MA 02111-1307, USA.
38 g_convert_error_quark()
42 quark
= g_quark_from_static_string ("g_convert_error");
47 #if defined(USE_LIBICONV) && !defined (_LIBICONV_H)
48 #error libiconv in use but included iconv.h not from libiconv
50 #if !defined(USE_LIBICONV) && defined (_LIBICONV_H)
51 #error libiconv not in use but included iconv.h is from libiconv
55 g_iconv_open (const gchar
*to_codeset
,
56 const gchar
*from_codeset
)
58 iconv_t cd
= iconv_open (to_codeset
, from_codeset
);
64 g_iconv (GIConv converter
,
68 size_t *outbytes_left
)
70 iconv_t cd
= (iconv_t
)converter
;
72 return iconv (cd
, inbuf
, inbytes_left
, outbuf
, outbytes_left
);
76 g_iconv_close (GIConv converter
)
78 iconv_t cd
= (iconv_t
)converter
;
80 return iconv_close (cd
);
84 open_converter (const gchar
*to_codeset
,
85 const gchar
*from_codeset
,
88 GIConv cd
= g_iconv_open (to_codeset
, from_codeset
);
90 if (cd
== (iconv_t
) -1)
92 /* Something went wrong. */
94 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_NO_CONVERSION
,
95 _("Conversion from character set `%s' to `%s' is not supported"),
96 from_codeset
, to_codeset
);
98 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_FAILED
,
99 _("Could not open converter from `%s' to `%s': %s"),
100 from_codeset
, to_codeset
, strerror (errno
));
109 * @str: the string to convert
110 * @len: the length of the string
111 * @to_codeset: name of character set into which to convert @str
112 * @from_codeset: character set of @str.
113 * @bytes_read: location to store the number of bytes in the
114 * input string that were successfully converted, or %NULL.
115 * Even if the conversion was succesful, this may be
116 * less than len if there were partial characters
117 * at the end of the input. If the error
118 * G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
119 * stored will the byte fofset after the last valid
121 * @bytes_written: the stored in the output buffer (not including the
123 * @error: location to store the error occuring, or %NULL to ignore
124 * errors. Any of the errors in #GConvertError may occur.
126 * Convert a string from one character set to another.
128 * Return value: If the conversion was successful, a newly allocated
129 * NUL-terminated string, which must be freed with
130 * g_free. Otherwise %NULL and @error will be set.
133 g_convert (const gchar
*str
,
135 const gchar
*to_codeset
,
136 const gchar
*from_codeset
,
144 size_t inbytes_remaining
;
145 size_t outbytes_remaining
;
149 gboolean have_error
= FALSE
;
151 g_return_val_if_fail (str
!= NULL
, NULL
);
152 g_return_val_if_fail (to_codeset
!= NULL
, NULL
);
153 g_return_val_if_fail (from_codeset
!= NULL
, NULL
);
155 cd
= open_converter (to_codeset
, from_codeset
, error
);
157 if (cd
== (GIConv
) -1)
172 inbytes_remaining
= len
;
174 /* Due to a GLIBC bug, round outbuf_size up to a multiple of 4 */
175 /* + 1 for nul in case len == 1 */
176 outbuf_size
= ((len
+ 3) & ~3) + 1;
178 outbytes_remaining
= outbuf_size
- 1; /* -1 for nul */
179 outp
= dest
= g_malloc (outbuf_size
);
183 err
= g_iconv (cd
, (char **)&p
, &inbytes_remaining
, &outp
, &outbytes_remaining
);
185 if (err
== (size_t) -1)
190 /* Incomplete text, do not report an error */
194 size_t used
= outp
- dest
;
196 /* glibc's iconv can return E2BIG even if there is space
197 * remaining if an internal buffer is exhausted. The
198 * folllowing is a heuristic to catch this. The 16 is
201 if (used
+ 16 > outbuf_size
)
203 outbuf_size
= (outbuf_size
- 1) * 2 + 1;
204 dest
= g_realloc (dest
, outbuf_size
);
207 outbytes_remaining
= outbuf_size
- used
- 1; /* -1 for nul */
213 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_ILLEGAL_SEQUENCE
,
214 _("Invalid byte sequence in conversion input"));
218 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_FAILED
,
219 _("Error during conversion: %s"),
231 *bytes_read
= p
- str
;
234 if ((p
- str
) != len
)
238 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_PARTIAL_INPUT
,
239 _("Partial character sequence at end of input"));
246 *bytes_written
= outp
- dest
; /* Doesn't include '\0' */
258 * g_convert_with_fallback:
259 * @str: the string to convert
260 * @len: the length of the string
261 * @to_codeset: name of character set into which to convert @str
262 * @from_codeset: character set of @str.
263 * @fallback: UTF-8 string to use in place of character not
264 * present in the target encoding. (This must be
265 * in the target encoding), if %NULL, characters
266 * not in the target encoding will be represented
267 * as Unicode escapes \x{XXXX} or \x{XXXXXX}.
268 * @bytes_read: location to store the number of bytes in the
269 * input string that were successfully converted, or %NULL.
270 * Even if the conversion was succesful, this may be
271 * less than len if there were partial characters
272 * at the end of the input. If the error
273 * G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
274 * stored will the byte fofset after the last valid
276 * @bytes_written: the stored in the output buffer (not including the
278 * @error: location to store the error occuring, or %NULL to ignore
279 * errors. Any of the errors in #GConvertError may occur.
281 * Convert a string from one character set to another, possibly
282 * including fallback sequences for characters not representable
283 * in the output. Note that it is not guaranteed that the specification
284 * for the fallback sequences in @fallback will be honored. Some
285 * systems may do a approximate conversion from @from_codeset
286 * to @to_codeset in their iconv() functions, in which case GLib
287 * will simply return that approximate conversion.
289 * Return value: If the conversion was successful, a newly allocated
290 * NUL-terminated string, which must be freed with
291 * g_free. Otherwise %NULL and @error will be set.
294 g_convert_with_fallback (const gchar
*str
,
296 const gchar
*to_codeset
,
297 const gchar
*from_codeset
,
306 const gchar
*insert_str
= NULL
;
308 int inbytes_remaining
;
309 const gchar
*save_p
= NULL
;
310 size_t save_inbytes
= 0;
311 size_t outbytes_remaining
;
315 gboolean have_error
= FALSE
;
316 gboolean done
= FALSE
;
318 GError
*local_error
= NULL
;
320 g_return_val_if_fail (str
!= NULL
, NULL
);
321 g_return_val_if_fail (to_codeset
!= NULL
, NULL
);
322 g_return_val_if_fail (from_codeset
!= NULL
, NULL
);
327 /* Try an exact conversion; we only proceed if this fails
328 * due to an illegal sequence in the input string.
330 dest
= g_convert (str
, len
, to_codeset
, from_codeset
,
331 bytes_read
, bytes_written
, &local_error
);
335 if (!g_error_matches (local_error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_ILLEGAL_SEQUENCE
))
337 g_propagate_error (error
, local_error
);
341 g_error_free (local_error
);
345 /* No go; to proceed, we need a converter from "UTF-8" to
346 * to_codeset, and the string as UTF-8.
348 cd
= open_converter (to_codeset
, "UTF-8", error
);
349 if (cd
== (GIConv
) -1)
360 utf8
= g_convert (str
, len
, "UTF-8", from_codeset
,
361 bytes_read
, &inbytes_remaining
, error
);
365 /* Now the heart of the code. We loop through the UTF-8 string, and
366 * whenever we hit an offending character, we form fallback, convert
367 * the fallback to the target codeset, and then go back to
368 * converting the original string after finishing with the fallback.
370 * The variables save_p and save_inbytes store the input state
371 * for the original string while we are converting the fallback
374 /* Due to a GLIBC bug, round outbuf_size up to a multiple of 4 */
375 /* + 1 for nul in case len == 1 */
376 outbuf_size
= ((len
+ 3) & ~3) + 1;
377 outbytes_remaining
= outbuf_size
- 1; /* -1 for nul */
378 outp
= dest
= g_malloc (outbuf_size
);
380 while (!done
&& !have_error
)
382 size_t inbytes_tmp
= inbytes_remaining
;
383 err
= g_iconv (cd
, (char **)&p
, &inbytes_tmp
, &outp
, &outbytes_remaining
);
384 inbytes_remaining
= inbytes_tmp
;
386 if (err
== (size_t) -1)
391 g_assert_not_reached();
395 size_t used
= outp
- dest
;
397 /* glibc's iconv can return E2BIG even if there is space
398 * remaining if an internal buffer is exhausted. The
399 * folllowing is a heuristic to catch this. The 16 is
402 if (used
+ 16 > outbuf_size
)
404 outbuf_size
= (outbuf_size
- 1) * 2 + 1;
405 dest
= g_realloc (dest
, outbuf_size
);
408 outbytes_remaining
= outbuf_size
- used
- 1; /* -1 for nul */
416 /* Error converting fallback string - fatal
418 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_ILLEGAL_SEQUENCE
,
419 _("Cannot convert fallback '%s' to codeset '%s'"),
420 insert_str
, to_codeset
);
428 gunichar ch
= g_utf8_get_char (p
);
429 insert_str
= g_strdup_printf ("\\x{%0*X}",
430 (ch
< 0x10000) ? 4 : 6,
434 insert_str
= fallback
;
436 save_p
= g_utf8_next_char (p
);
437 save_inbytes
= inbytes_remaining
- (save_p
- p
);
439 inbytes_remaining
= strlen (p
);
443 g_set_error (error
, G_CONVERT_ERROR
, G_CONVERT_ERROR_FAILED
,
444 _("Error during conversion: %s"),
455 g_free ((gchar
*)insert_str
);
457 inbytes_remaining
= save_inbytes
;
472 *bytes_written
= outp
- str
; /* Doesn't include '\0' */
478 if (save_p
&& !fallback
)
479 g_free ((gchar
*)insert_str
);
495 * @opsysstring: a string in the encoding of the current locale
496 * @len: the length of the string, or -1 if the string is
498 * @bytes_read: location to store the number of bytes in the
499 * input string that were successfully converted, or %NULL.
500 * Even if the conversion was succesful, this may be
501 * less than len if there were partial characters
502 * at the end of the input. If the error
503 * G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
504 * stored will the byte fofset after the last valid
506 * @bytes_written: the stored in the output buffer (not including the
508 * @error: location to store the error occuring, or %NULL to ignore
509 * errors. Any of the errors in #GConvertError may occur.
511 * Converts a string which is in the encoding used for strings by
512 * the C runtime (usually the same as that used by the operating
513 * system) in the current locale into a UTF-8 string.
515 * Return value: The converted string, or %NULL on an error.
518 g_locale_to_utf8 (const gchar
*opsysstring
,
526 gint i
, clen
, total_len
, wclen
, first
;
532 len
= strlen (opsysstring
);
534 wcs
= g_new (wchar_t, len
);
535 wclen
= MultiByteToWideChar (CP_ACP
, 0, opsysstring
, len
, wcs
, len
);
539 for (i
= 0; i
< wclen
; i
++)
547 else if (wc
< 0x10000)
549 else if (wc
< 0x200000)
551 else if (wc
< 0x4000000)
557 result
= g_malloc (total_len
+ 1);
561 for (i
= 0; i
< wclen
; i
++)
575 else if (wc
< 0x10000)
580 else if (wc
< 0x200000)
585 else if (wc
< 0x4000000)
599 case 6: bp
[5] = (wc
& 0x3f) | 0x80; wc
>>= 6; /* Fall through */
600 case 5: bp
[4] = (wc
& 0x3f) | 0x80; wc
>>= 6; /* Fall through */
601 case 4: bp
[3] = (wc
& 0x3f) | 0x80; wc
>>= 6; /* Fall through */
602 case 3: bp
[2] = (wc
& 0x3f) | 0x80; wc
>>= 6; /* Fall through */
603 case 2: bp
[1] = (wc
& 0x3f) | 0x80; wc
>>= 6; /* Fall through */
604 case 1: bp
[0] = wc
| first
;
616 *bytes_written
= total_len
;
624 if (g_get_charset (&charset
))
625 return g_strdup (opsysstring
);
627 str
= g_convert (opsysstring
, len
,
628 "UTF-8", charset
, bytes_read
, bytes_written
, error
);
635 * g_locale_from_utf8:
636 * @utf8string: a UTF-8 encoded string
637 * @len: the length of the string, or -1 if the string is
639 * @bytes_read: location to store the number of bytes in the
640 * input string that were successfully converted, or %NULL.
641 * Even if the conversion was succesful, this may be
642 * less than len if there were partial characters
643 * at the end of the input. If the error
644 * G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
645 * stored will the byte fofset after the last valid
647 * @bytes_written: the stored in the output buffer (not including the
649 * @error: location to store the error occuring, or %NULL to ignore
650 * errors. Any of the errors in #GConvertError may occur.
652 * Converts a string from UTF-8 to the encoding used for strings by
653 * the C runtime (usually the same as that used by the operating
654 * system) in the current locale.
656 * Return value: The converted string, or %NULL on an error.
659 g_locale_from_utf8 (const gchar
*utf8string
,
667 gint i
, mask
, clen
, mblen
;
674 len
= strlen (utf8string
);
676 /* First convert to wide chars */
677 cp
= (guchar
*) utf8string
;
680 wcs
= g_new (wchar_t, len
+ 1);
692 else if ((c
& 0xe0) == 0xc0)
697 else if ((c
& 0xf0) == 0xe0)
702 else if ((c
& 0xf8) == 0xf0)
707 else if ((c
& 0xfc) == 0xf8)
712 else if ((c
& 0xfc) == 0xfc)
729 *wcp
= (cp
[0] & mask
);
730 for (i
= 1; i
< clen
; i
++)
732 if ((cp
[i
] & 0xc0) != 0x80)
738 *wcp
|= (cp
[i
] & 0x3f);
751 /* n is the number of wide chars constructed */
753 /* Convert to a string in the current ANSI codepage */
755 result
= g_new (gchar
, 3 * n
+ 1);
756 mblen
= WideCharToMultiByte (CP_ACP
, 0, wcs
, n
, result
, 3*n
, NULL
, NULL
);
763 *bytes_written
= mblen
;
769 gchar
*charset
, *str
;
771 if (g_get_charset (&charset
))
772 return g_strdup (utf8string
);
774 str
= g_convert (utf8string
, strlen (utf8string
),
775 charset
, "UTF-8", bytes_read
, bytes_written
, error
);
783 * g_filename_to_utf8:
784 * @opsysstring: a string in the encoding for filenames
785 * @len: the length of the string, or -1 if the string is
787 * @bytes_read: location to store the number of bytes in the
788 * input string that were successfully converted, or %NULL.
789 * Even if the conversion was succesful, this may be
790 * less than len if there were partial characters
791 * at the end of the input. If the error
792 * G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
793 * stored will the byte fofset after the last valid
795 * @bytes_written: the stored in the output buffer (not including the
797 * @error: location to store the error occuring, or %NULL to ignore
798 * errors. Any of the errors in #GConvertError may occur.
800 * Converts a string which is in the encoding used for filenames
801 * into a UTF-8 string.
803 * Return value: The converted string, or %NULL on an error.
806 g_filename_to_utf8 (const gchar
*opsysstring
,
813 return g_locale_to_utf8 (opsysstring
, len
,
814 bytes_read
, bytes_written
,
817 if (getenv ("G_BROKEN_FILENAMES"))
818 return g_locale_to_utf8 (opsysstring
, len
,
819 bytes_read
, bytes_written
,
822 if (bytes_read
|| bytes_written
)
824 gint len
= strlen (opsysstring
);
829 *bytes_written
= len
;
833 return g_strdup (opsysstring
);
835 return g_strndup (opsysstring
, len
);
840 * g_filename_from_utf8:
841 * @utf8string: a UTF-8 encoded string
842 * @len: the length of the string, or -1 if the string is
844 * @bytes_read: location to store the number of bytes in the
845 * input string that were successfully converted, or %NULL.
846 * Even if the conversion was succesful, this may be
847 * less than len if there were partial characters
848 * at the end of the input. If the error
849 * G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
850 * stored will the byte fofset after the last valid
852 * @bytes_written: the stored in the output buffer (not including the
854 * @error: location to store the error occuring, or %NULL to ignore
855 * errors. Any of the errors in #GConvertError may occur.
857 * Converts a string from UTF-8 to the encoding used for filenames.
859 * Return value: The converted string, or %NULL on an error.
862 g_filename_from_utf8 (const gchar
*utf8string
,
869 return g_locale_from_utf8 (utf8string
, len
,
870 bytes_read
, bytes_written
,
873 if (getenv ("G_BROKEN_FILENAMES"))
874 return g_locale_from_utf8 (utf8string
, len
,
875 bytes_read
, bytes_written
,
878 if (bytes_read
|| bytes_written
)
880 gint len
= strlen (utf8string
);
885 *bytes_written
= len
;
889 return g_strdup (utf8string
);
891 return g_strndup (utf8string
, len
);