1 /* Copyright (C) 1995,1996,1998,2000,2001, 2004, 2006 Free Software Foundation, Inc.
3 * This library is free software; you can redistribute it and/or
4 * modify it under the terms of the GNU Lesser General Public
5 * License as published by the Free Software Foundation; either
6 * version 2.1 of the License, or (at your option) any later version.
8 * This library is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * Lesser General Public License for more details.
13 * You should have received a copy of the GNU Lesser General Public
14 * License along with this library; if not, write to the Free Software
15 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
24 #include "libguile/_scm.h"
25 #include "libguile/chars.h"
26 #include "libguile/root.h"
27 #include "libguile/strings.h"
28 #include "libguile/deprecation.h"
29 #include "libguile/validate.h"
30 #include "libguile/dynwind.h"
40 * XXX - keeping an accurate refcount during GC seems to be quite
41 * tricky, so we just keep score of whether a stringbuf might be
42 * shared, not wether it definitely is.
44 * The scheme I (mvo) tried to keep an accurate reference count would
45 * recount all strings that point to a stringbuf during the mark-phase
46 * of the GC. This was done since one cannot access the stringbuf of
47 * a string when that string is freed (in order to decrease the
48 * reference count). The memory of the stringbuf might have been
49 * reused already for something completely different.
51 * This recounted worked for a small number of threads beating on
52 * cow-strings, but it failed randomly with more than 10 threads, say.
53 * I couldn't figure out what went wrong, so I used the conservative
54 * approach implemented below.
56 * A stringbuf needs to know its length, but only so that it can be
57 * reported when the stringbuf is freed.
59 * Stringbufs (and strings) are not stored very compactly: a stringbuf
60 * has room for about 2*sizeof(scm_t_bits)-1 bytes additional
61 * information. As a compensation, the code below is made more
62 * complicated by storing small strings inline in the double cell of a
63 * stringbuf. So we have fixstrings and bigstrings...
66 #define STRINGBUF_F_SHARED 0x100
67 #define STRINGBUF_F_INLINE 0x200
69 #define STRINGBUF_TAG scm_tc7_stringbuf
70 #define STRINGBUF_SHARED(buf) (SCM_CELL_WORD_0(buf) & STRINGBUF_F_SHARED)
71 #define STRINGBUF_INLINE(buf) (SCM_CELL_WORD_0(buf) & STRINGBUF_F_INLINE)
73 #define STRINGBUF_OUTLINE_CHARS(buf) ((char *)SCM_CELL_WORD_1(buf))
74 #define STRINGBUF_OUTLINE_LENGTH(buf) (SCM_CELL_WORD_2(buf))
75 #define STRINGBUF_INLINE_CHARS(buf) ((char *)SCM_CELL_OBJECT_LOC(buf,1))
76 #define STRINGBUF_INLINE_LENGTH(buf) (((size_t)SCM_CELL_WORD_0(buf))>>16)
78 #define STRINGBUF_CHARS(buf) (STRINGBUF_INLINE (buf) \
79 ? STRINGBUF_INLINE_CHARS (buf) \
80 : STRINGBUF_OUTLINE_CHARS (buf))
81 #define STRINGBUF_LENGTH(buf) (STRINGBUF_INLINE (buf) \
82 ? STRINGBUF_INLINE_LENGTH (buf) \
83 : STRINGBUF_OUTLINE_LENGTH (buf))
85 #define STRINGBUF_MAX_INLINE_LEN (3*sizeof(scm_t_bits))
87 #define SET_STRINGBUF_SHARED(buf) \
88 (SCM_SET_CELL_WORD_0 ((buf), SCM_CELL_WORD_0 (buf) | STRINGBUF_F_SHARED))
91 static size_t lenhist
[1001];
95 make_stringbuf (size_t len
)
97 /* XXX - for the benefit of SCM_STRING_CHARS, SCM_SYMBOL_CHARS and
98 scm_i_symbol_chars, all stringbufs are null-terminated. Once
99 SCM_STRING_CHARS and SCM_SYMBOL_CHARS are removed and the code
100 has been changed for scm_i_symbol_chars, this null-termination
111 if (len
<= STRINGBUF_MAX_INLINE_LEN
-1)
113 return scm_double_cell (STRINGBUF_TAG
| STRINGBUF_F_INLINE
| (len
<< 16),
118 char *mem
= scm_gc_malloc (len
+1, "string");
120 return scm_double_cell (STRINGBUF_TAG
, (scm_t_bits
) mem
,
121 (scm_t_bits
) len
, (scm_t_bits
) 0);
125 /* Return a new stringbuf whose underlying storage consists of the LEN+1
126 octets pointed to by STR (the last octet is zero). */
127 SCM_C_INLINE_KEYWORD SCM
128 scm_i_take_stringbufn (char *str
, size_t len
)
130 scm_gc_register_collectable_memory (str
, len
+ 1, "stringbuf");
132 return scm_double_cell (STRINGBUF_TAG
, (scm_t_bits
) str
,
133 (scm_t_bits
) len
, (scm_t_bits
) 0);
137 scm_i_stringbuf_mark (SCM buf
)
143 scm_i_stringbuf_free (SCM buf
)
145 if (!STRINGBUF_INLINE (buf
))
146 scm_gc_free (STRINGBUF_OUTLINE_CHARS (buf
),
147 STRINGBUF_OUTLINE_LENGTH (buf
) + 1, "string");
150 scm_i_pthread_mutex_t stringbuf_write_mutex
= SCM_I_PTHREAD_MUTEX_INITIALIZER
;
152 /* Copy-on-write strings.
155 #define STRING_TAG scm_tc7_string
157 #define STRING_STRINGBUF(str) (SCM_CELL_OBJECT_1(str))
158 #define STRING_START(str) ((size_t)SCM_CELL_WORD_2(str))
159 #define STRING_LENGTH(str) ((size_t)SCM_CELL_WORD_3(str))
161 #define SET_STRING_STRINGBUF(str,buf) (SCM_SET_CELL_OBJECT_1(str,buf))
162 #define SET_STRING_START(str,start) (SCM_SET_CELL_WORD_2(str,start))
164 #define IS_STRING(str) (SCM_NIMP(str) && SCM_TYP7(str) == STRING_TAG)
166 /* Read-only strings.
169 #define RO_STRING_TAG (scm_tc7_string + 0x200)
170 #define IS_RO_STRING(str) (SCM_CELL_TYPE(str)==RO_STRING_TAG)
172 /* Mutation-sharing substrings
175 #define SH_STRING_TAG (scm_tc7_string + 0x100)
177 #define SH_STRING_STRING(sh) (SCM_CELL_OBJECT_1(sh))
178 /* START and LENGTH as for STRINGs. */
180 #define IS_SH_STRING(str) (SCM_CELL_TYPE(str)==SH_STRING_TAG)
183 scm_i_make_string (size_t len
, char **charsp
)
185 SCM buf
= make_stringbuf (len
);
188 *charsp
= STRINGBUF_CHARS (buf
);
189 res
= scm_double_cell (STRING_TAG
, SCM_UNPACK(buf
),
190 (scm_t_bits
)0, (scm_t_bits
) len
);
195 validate_substring_args (SCM str
, size_t start
, size_t end
)
197 if (!IS_STRING (str
))
198 scm_wrong_type_arg_msg (NULL
, 0, str
, "string");
199 if (start
> STRING_LENGTH (str
))
200 scm_out_of_range (NULL
, scm_from_size_t (start
));
201 if (end
> STRING_LENGTH (str
) || end
< start
)
202 scm_out_of_range (NULL
, scm_from_size_t (end
));
206 get_str_buf_start (SCM
*str
, SCM
*buf
, size_t *start
)
208 *start
= STRING_START (*str
);
209 if (IS_SH_STRING (*str
))
211 *str
= SH_STRING_STRING (*str
);
212 *start
+= STRING_START (*str
);
214 *buf
= STRING_STRINGBUF (*str
);
218 scm_i_substring (SCM str
, size_t start
, size_t end
)
222 get_str_buf_start (&str
, &buf
, &str_start
);
223 scm_i_pthread_mutex_lock (&stringbuf_write_mutex
);
224 SET_STRINGBUF_SHARED (buf
);
225 scm_i_pthread_mutex_unlock (&stringbuf_write_mutex
);
226 return scm_double_cell (STRING_TAG
, SCM_UNPACK(buf
),
227 (scm_t_bits
)str_start
+ start
,
228 (scm_t_bits
) end
- start
);
232 scm_i_substring_read_only (SCM str
, size_t start
, size_t end
)
236 get_str_buf_start (&str
, &buf
, &str_start
);
237 scm_i_pthread_mutex_lock (&stringbuf_write_mutex
);
238 SET_STRINGBUF_SHARED (buf
);
239 scm_i_pthread_mutex_unlock (&stringbuf_write_mutex
);
240 return scm_double_cell (RO_STRING_TAG
, SCM_UNPACK(buf
),
241 (scm_t_bits
)str_start
+ start
,
242 (scm_t_bits
) end
- start
);
246 scm_i_substring_copy (SCM str
, size_t start
, size_t end
)
248 size_t len
= end
- start
;
251 get_str_buf_start (&str
, &buf
, &str_start
);
252 my_buf
= make_stringbuf (len
);
253 memcpy (STRINGBUF_CHARS (my_buf
),
254 STRINGBUF_CHARS (buf
) + str_start
+ start
, len
);
255 scm_remember_upto_here_1 (buf
);
256 return scm_double_cell (STRING_TAG
, SCM_UNPACK(my_buf
),
257 (scm_t_bits
)0, (scm_t_bits
) len
);
261 scm_i_substring_shared (SCM str
, size_t start
, size_t end
)
263 if (start
== 0 && end
== STRING_LENGTH (str
))
267 size_t len
= end
- start
;
268 if (IS_SH_STRING (str
))
270 start
+= STRING_START (str
);
271 str
= SH_STRING_STRING (str
);
273 return scm_double_cell (SH_STRING_TAG
, SCM_UNPACK(str
),
274 (scm_t_bits
)start
, (scm_t_bits
) len
);
279 scm_c_substring (SCM str
, size_t start
, size_t end
)
281 validate_substring_args (str
, start
, end
);
282 return scm_i_substring (str
, start
, end
);
286 scm_c_substring_read_only (SCM str
, size_t start
, size_t end
)
288 validate_substring_args (str
, start
, end
);
289 return scm_i_substring_read_only (str
, start
, end
);
293 scm_c_substring_copy (SCM str
, size_t start
, size_t end
)
295 validate_substring_args (str
, start
, end
);
296 return scm_i_substring_copy (str
, start
, end
);
300 scm_c_substring_shared (SCM str
, size_t start
, size_t end
)
302 validate_substring_args (str
, start
, end
);
303 return scm_i_substring_shared (str
, start
, end
);
307 scm_i_string_mark (SCM str
)
309 if (IS_SH_STRING (str
))
310 return SH_STRING_STRING (str
);
312 return STRING_STRINGBUF (str
);
316 scm_i_string_free (SCM str
)
320 /* Internal accessors
324 scm_i_string_length (SCM str
)
326 return STRING_LENGTH (str
);
330 scm_i_string_chars (SCM str
)
334 get_str_buf_start (&str
, &buf
, &start
);
335 return STRINGBUF_CHARS (buf
) + start
;
339 scm_i_string_writable_chars (SCM orig_str
)
341 SCM buf
, str
= orig_str
;
344 get_str_buf_start (&str
, &buf
, &start
);
345 if (IS_RO_STRING (str
))
346 scm_misc_error (NULL
, "string is read-only: ~s", scm_list_1 (orig_str
));
348 scm_i_pthread_mutex_lock (&stringbuf_write_mutex
);
349 if (STRINGBUF_SHARED (buf
))
351 /* Clone stringbuf. For this, we put all threads to sleep.
354 size_t len
= STRING_LENGTH (str
);
357 scm_i_pthread_mutex_unlock (&stringbuf_write_mutex
);
359 new_buf
= make_stringbuf (len
);
360 memcpy (STRINGBUF_CHARS (new_buf
),
361 STRINGBUF_CHARS (buf
) + STRING_START (str
), len
);
363 scm_i_thread_put_to_sleep ();
364 SET_STRING_STRINGBUF (str
, new_buf
);
365 start
-= STRING_START (str
);
366 SET_STRING_START (str
, 0);
367 scm_i_thread_wake_up ();
371 scm_i_pthread_mutex_lock (&stringbuf_write_mutex
);
374 return STRINGBUF_CHARS (buf
) + start
;
378 scm_i_string_stop_writing (void)
380 scm_i_pthread_mutex_unlock (&stringbuf_write_mutex
);
385 Basic symbol creation and accessing is done here, the rest is in
386 symbols.[hc]. This has been done to keep stringbufs and the
387 internals of strings and string-like objects confined to this file.
390 #define SYMBOL_STRINGBUF SCM_CELL_OBJECT_1
393 scm_i_make_symbol (SCM name
, scm_t_bits flags
,
394 unsigned long hash
, SCM props
)
397 size_t start
= STRING_START (name
);
398 size_t length
= STRING_LENGTH (name
);
400 if (IS_SH_STRING (name
))
402 name
= SH_STRING_STRING (name
);
403 start
+= STRING_START (name
);
405 buf
= SYMBOL_STRINGBUF (name
);
407 if (start
== 0 && length
== STRINGBUF_LENGTH (buf
))
410 scm_i_pthread_mutex_lock (&stringbuf_write_mutex
);
411 SET_STRINGBUF_SHARED (buf
);
412 scm_i_pthread_mutex_unlock (&stringbuf_write_mutex
);
417 SCM new_buf
= make_stringbuf (length
);
418 memcpy (STRINGBUF_CHARS (new_buf
),
419 STRINGBUF_CHARS (buf
) + start
, length
);
422 return scm_double_cell (scm_tc7_symbol
| flags
, SCM_UNPACK (buf
),
423 (scm_t_bits
) hash
, SCM_UNPACK (props
));
427 scm_i_c_make_symbol (const char *name
, size_t len
,
428 scm_t_bits flags
, unsigned long hash
, SCM props
)
430 SCM buf
= make_stringbuf (len
);
431 memcpy (STRINGBUF_CHARS (buf
), name
, len
);
433 return scm_double_cell (scm_tc7_symbol
| flags
, SCM_UNPACK (buf
),
434 (scm_t_bits
) hash
, SCM_UNPACK (props
));
437 /* Return a new symbol that uses the LEN bytes pointed to by NAME as its
438 underlying storage. */
440 scm_i_c_take_symbol (char *name
, size_t len
,
441 scm_t_bits flags
, unsigned long hash
, SCM props
)
443 SCM buf
= scm_i_take_stringbufn (name
, len
);
445 return scm_double_cell (scm_tc7_symbol
| flags
, SCM_UNPACK (buf
),
446 (scm_t_bits
) hash
, SCM_UNPACK (props
));
450 scm_i_symbol_length (SCM sym
)
452 return STRINGBUF_LENGTH (SYMBOL_STRINGBUF (sym
));
456 scm_i_symbol_chars (SCM sym
)
458 SCM buf
= SYMBOL_STRINGBUF (sym
);
459 return STRINGBUF_CHARS (buf
);
463 scm_i_symbol_mark (SCM sym
)
465 scm_gc_mark (SYMBOL_STRINGBUF (sym
));
466 return SCM_CELL_OBJECT_3 (sym
);
470 scm_i_symbol_free (SCM sym
)
475 scm_i_symbol_substring (SCM sym
, size_t start
, size_t end
)
477 SCM buf
= SYMBOL_STRINGBUF (sym
);
478 scm_i_pthread_mutex_lock (&stringbuf_write_mutex
);
479 SET_STRINGBUF_SHARED (buf
);
480 scm_i_pthread_mutex_unlock (&stringbuf_write_mutex
);
481 return scm_double_cell (STRING_TAG
, SCM_UNPACK(buf
),
482 (scm_t_bits
)start
, (scm_t_bits
) end
- start
);
490 SCM
scm_sys_string_dump (SCM
);
491 SCM
scm_sys_symbol_dump (SCM
);
492 SCM
scm_sys_stringbuf_hist (void);
494 SCM_DEFINE (scm_sys_string_dump
, "%string-dump", 1, 0, 0,
497 #define FUNC_NAME s_scm_sys_string_dump
499 SCM_VALIDATE_STRING (1, str
);
500 fprintf (stderr
, "%p:\n", str
);
501 fprintf (stderr
, " start: %u\n", STRING_START (str
));
502 fprintf (stderr
, " len: %u\n", STRING_LENGTH (str
));
503 if (IS_SH_STRING (str
))
505 fprintf (stderr
, " string: %p\n", SH_STRING_STRING (str
));
506 fprintf (stderr
, "\n");
507 scm_sys_string_dump (SH_STRING_STRING (str
));
511 SCM buf
= STRING_STRINGBUF (str
);
512 fprintf (stderr
, " buf: %p\n", buf
);
513 fprintf (stderr
, " chars: %p\n", STRINGBUF_CHARS (buf
));
514 fprintf (stderr
, " length: %u\n", STRINGBUF_LENGTH (buf
));
515 fprintf (stderr
, " flags: %x\n", (SCM_CELL_WORD_0 (buf
) & 0x300));
517 return SCM_UNSPECIFIED
;
521 SCM_DEFINE (scm_sys_symbol_dump
, "%symbol-dump", 1, 0, 0,
524 #define FUNC_NAME s_scm_sys_symbol_dump
526 SCM_VALIDATE_SYMBOL (1, sym
);
527 fprintf (stderr
, "%p:\n", sym
);
528 fprintf (stderr
, " hash: %lu\n", scm_i_symbol_hash (sym
));
530 SCM buf
= SYMBOL_STRINGBUF (sym
);
531 fprintf (stderr
, " buf: %p\n", buf
);
532 fprintf (stderr
, " chars: %p\n", STRINGBUF_CHARS (buf
));
533 fprintf (stderr
, " length: %u\n", STRINGBUF_LENGTH (buf
));
534 fprintf (stderr
, " shared: %u\n", STRINGBUF_SHARED (buf
));
536 return SCM_UNSPECIFIED
;
540 SCM_DEFINE (scm_sys_stringbuf_hist
, "%stringbuf-hist", 0, 0, 0,
543 #define FUNC_NAME s_scm_sys_stringbuf_hist
546 for (i
= 0; i
< 1000; i
++)
548 fprintf (stderr
, " %3d: %u\n", i
, lenhist
[i
]);
549 fprintf (stderr
, ">999: %u\n", lenhist
[1000]);
550 return SCM_UNSPECIFIED
;
558 SCM_DEFINE (scm_string_p
, "string?", 1, 0, 0,
560 "Return @code{#t} if @var{obj} is a string, else @code{#f}.")
561 #define FUNC_NAME s_scm_string_p
563 return scm_from_bool (IS_STRING (obj
));
568 SCM_REGISTER_PROC (s_scm_list_to_string
, "list->string", 1, 0, 0, scm_string
);
570 SCM_DEFINE (scm_string
, "string", 0, 0, 1,
572 "@deffnx {Scheme Procedure} list->string chrs\n"
573 "Return a newly allocated string composed of the arguments,\n"
575 #define FUNC_NAME s_scm_string
582 long i
= scm_ilength (chrs
);
584 SCM_ASSERT (i
>= 0, chrs
, SCM_ARG1
, FUNC_NAME
);
588 result
= scm_i_make_string (len
, &data
);
589 while (len
> 0 && scm_is_pair (chrs
))
591 SCM elt
= SCM_CAR (chrs
);
593 SCM_VALIDATE_CHAR (SCM_ARGn
, elt
);
594 *data
++ = SCM_CHAR (elt
);
595 chrs
= SCM_CDR (chrs
);
599 scm_misc_error (NULL
, "list changed while constructing string", SCM_EOL
);
600 if (!scm_is_null (chrs
))
601 scm_wrong_type_arg_msg (NULL
, 0, chrs
, "proper list");
607 SCM_DEFINE (scm_make_string
, "make-string", 1, 1, 0,
609 "Return a newly allocated string of\n"
610 "length @var{k}. If @var{chr} is given, then all elements of\n"
611 "the string are initialized to @var{chr}, otherwise the contents\n"
612 "of the @var{string} are unspecified.")
613 #define FUNC_NAME s_scm_make_string
615 return scm_c_make_string (scm_to_size_t (k
), chr
);
620 scm_c_make_string (size_t len
, SCM chr
)
621 #define FUNC_NAME NULL
624 SCM res
= scm_i_make_string (len
, &dst
);
626 if (!SCM_UNBNDP (chr
))
628 SCM_VALIDATE_CHAR (0, chr
);
629 memset (dst
, SCM_CHAR (chr
), len
);
636 SCM_DEFINE (scm_string_length
, "string-length", 1, 0, 0,
638 "Return the number of characters in @var{string}.")
639 #define FUNC_NAME s_scm_string_length
641 SCM_VALIDATE_STRING (1, string
);
642 return scm_from_size_t (STRING_LENGTH (string
));
647 scm_c_string_length (SCM string
)
649 if (!IS_STRING (string
))
650 scm_wrong_type_arg_msg (NULL
, 0, string
, "string");
651 return STRING_LENGTH (string
);
654 SCM_DEFINE (scm_string_ref
, "string-ref", 2, 0, 0,
656 "Return character @var{k} of @var{str} using zero-origin\n"
657 "indexing. @var{k} must be a valid index of @var{str}.")
658 #define FUNC_NAME s_scm_string_ref
662 SCM_VALIDATE_STRING (1, str
);
663 idx
= scm_to_unsigned_integer (k
, 0, scm_i_string_length (str
)-1);
664 return SCM_MAKE_CHAR (scm_i_string_chars (str
)[idx
]);
669 scm_c_string_ref (SCM str
, size_t p
)
671 if (p
>= scm_i_string_length (str
))
672 scm_out_of_range (NULL
, scm_from_size_t (p
));
673 return SCM_MAKE_CHAR (scm_i_string_chars (str
)[p
]);
676 SCM_DEFINE (scm_string_set_x
, "string-set!", 3, 0, 0,
677 (SCM str
, SCM k
, SCM chr
),
678 "Store @var{chr} in element @var{k} of @var{str} and return\n"
679 "an unspecified value. @var{k} must be a valid index of\n"
681 #define FUNC_NAME s_scm_string_set_x
685 SCM_VALIDATE_STRING (1, str
);
686 idx
= scm_to_unsigned_integer (k
, 0, scm_i_string_length(str
)-1);
687 SCM_VALIDATE_CHAR (3, chr
);
689 char *dst
= scm_i_string_writable_chars (str
);
690 dst
[idx
] = SCM_CHAR (chr
);
691 scm_i_string_stop_writing ();
693 return SCM_UNSPECIFIED
;
698 scm_c_string_set_x (SCM str
, size_t p
, SCM chr
)
700 if (p
>= scm_i_string_length (str
))
701 scm_out_of_range (NULL
, scm_from_size_t (p
));
703 char *dst
= scm_i_string_writable_chars (str
);
704 dst
[p
] = SCM_CHAR (chr
);
705 scm_i_string_stop_writing ();
709 SCM_DEFINE (scm_substring
, "substring", 2, 1, 0,
710 (SCM str
, SCM start
, SCM end
),
711 "Return a newly allocated string formed from the characters\n"
712 "of @var{str} beginning with index @var{start} (inclusive) and\n"
713 "ending with index @var{end} (exclusive).\n"
714 "@var{str} must be a string, @var{start} and @var{end} must be\n"
715 "exact integers satisfying:\n\n"
716 "0 <= @var{start} <= @var{end} <= (string-length @var{str}).")
717 #define FUNC_NAME s_scm_substring
719 size_t len
, from
, to
;
721 SCM_VALIDATE_STRING (1, str
);
722 len
= scm_i_string_length (str
);
723 from
= scm_to_unsigned_integer (start
, 0, len
);
724 if (SCM_UNBNDP (end
))
727 to
= scm_to_unsigned_integer (end
, from
, len
);
728 return scm_i_substring (str
, from
, to
);
732 SCM_DEFINE (scm_substring_read_only
, "substring/read-only", 2, 1, 0,
733 (SCM str
, SCM start
, SCM end
),
734 "Return a newly allocated string formed from the characters\n"
735 "of @var{str} beginning with index @var{start} (inclusive) and\n"
736 "ending with index @var{end} (exclusive).\n"
737 "@var{str} must be a string, @var{start} and @var{end} must be\n"
738 "exact integers satisfying:\n"
740 "0 <= @var{start} <= @var{end} <= (string-length @var{str}).\n"
742 "The returned string is read-only.\n")
743 #define FUNC_NAME s_scm_substring_read_only
745 size_t len
, from
, to
;
747 SCM_VALIDATE_STRING (1, str
);
748 len
= scm_i_string_length (str
);
749 from
= scm_to_unsigned_integer (start
, 0, len
);
750 if (SCM_UNBNDP (end
))
753 to
= scm_to_unsigned_integer (end
, from
, len
);
754 return scm_i_substring_read_only (str
, from
, to
);
758 SCM_DEFINE (scm_substring_copy
, "substring/copy", 2, 1, 0,
759 (SCM str
, SCM start
, SCM end
),
760 "Return a newly allocated string formed from the characters\n"
761 "of @var{str} beginning with index @var{start} (inclusive) and\n"
762 "ending with index @var{end} (exclusive).\n"
763 "@var{str} must be a string, @var{start} and @var{end} must be\n"
764 "exact integers satisfying:\n\n"
765 "0 <= @var{start} <= @var{end} <= (string-length @var{str}).")
766 #define FUNC_NAME s_scm_substring_copy
768 /* For the Scheme version, START is mandatory, but for the C
769 version, it is optional. See scm_string_copy in srfi-13.c for a
775 SCM_VALIDATE_STRING (1, str
);
776 scm_i_get_substring_spec (scm_i_string_length (str
),
777 start
, &from
, end
, &to
);
778 return scm_i_substring_copy (str
, from
, to
);
782 SCM_DEFINE (scm_substring_shared
, "substring/shared", 2, 1, 0,
783 (SCM str
, SCM start
, SCM end
),
784 "Return string that indirectly refers to the characters\n"
785 "of @var{str} beginning with index @var{start} (inclusive) and\n"
786 "ending with index @var{end} (exclusive).\n"
787 "@var{str} must be a string, @var{start} and @var{end} must be\n"
788 "exact integers satisfying:\n\n"
789 "0 <= @var{start} <= @var{end} <= (string-length @var{str}).")
790 #define FUNC_NAME s_scm_substring_shared
792 size_t len
, from
, to
;
794 SCM_VALIDATE_STRING (1, str
);
795 len
= scm_i_string_length (str
);
796 from
= scm_to_unsigned_integer (start
, 0, len
);
797 if (SCM_UNBNDP (end
))
800 to
= scm_to_unsigned_integer (end
, from
, len
);
801 return scm_i_substring_shared (str
, from
, to
);
805 SCM_DEFINE (scm_string_append
, "string-append", 0, 0, 1,
807 "Return a newly allocated string whose characters form the\n"
808 "concatenation of the given strings, @var{args}.")
809 #define FUNC_NAME s_scm_string_append
816 SCM_VALIDATE_REST_ARGUMENT (args
);
817 for (l
= args
; !scm_is_null (l
); l
= SCM_CDR (l
))
820 SCM_VALIDATE_STRING (SCM_ARGn
, s
);
821 i
+= scm_i_string_length (s
);
823 res
= scm_i_make_string (i
, &data
);
824 for (l
= args
; !scm_is_null (l
); l
= SCM_CDR (l
))
828 SCM_VALIDATE_STRING (SCM_ARGn
, s
);
829 len
= scm_i_string_length (s
);
830 memcpy (data
, scm_i_string_chars (s
), len
);
832 scm_remember_upto_here_1 (s
);
839 scm_is_string (SCM obj
)
841 return IS_STRING (obj
);
845 scm_from_locale_stringn (const char *str
, size_t len
)
850 if (len
== (size_t)-1)
852 res
= scm_i_make_string (len
, &dst
);
853 memcpy (dst
, str
, len
);
858 scm_from_locale_string (const char *str
)
860 return scm_from_locale_stringn (str
, -1);
864 scm_take_locale_stringn (char *str
, size_t len
)
868 if (len
== (size_t)-1)
872 /* Ensure STR is null terminated. A realloc for 1 extra byte should
873 often be satisfied from the alignment padding after the block, with
874 no actual data movement. */
875 str
= scm_realloc (str
, len
+1);
879 buf
= scm_i_take_stringbufn (str
, len
);
880 res
= scm_double_cell (STRING_TAG
,
882 (scm_t_bits
) 0, (scm_t_bits
) len
);
887 scm_take_locale_string (char *str
)
889 return scm_take_locale_stringn (str
, -1);
893 scm_to_locale_stringn (SCM str
, size_t *lenp
)
898 if (!scm_is_string (str
))
899 scm_wrong_type_arg_msg (NULL
, 0, str
, "string");
900 len
= scm_i_string_length (str
);
901 res
= scm_malloc (len
+ ((lenp
==NULL
)? 1 : 0));
902 memcpy (res
, scm_i_string_chars (str
), len
);
906 if (strlen (res
) != len
)
909 scm_misc_error (NULL
,
910 "string contains #\\nul character: ~S",
917 scm_remember_upto_here_1 (str
);
922 scm_to_locale_string (SCM str
)
924 return scm_to_locale_stringn (str
, NULL
);
928 scm_to_locale_stringbuf (SCM str
, char *buf
, size_t max_len
)
932 if (!scm_is_string (str
))
933 scm_wrong_type_arg_msg (NULL
, 0, str
, "string");
934 len
= scm_i_string_length (str
);
935 memcpy (buf
, scm_i_string_chars (str
), (len
> max_len
)? max_len
: len
);
936 scm_remember_upto_here_1 (str
);
940 /* converts C scm_array of strings to SCM scm_list of strings. */
941 /* If argc < 0, a null terminated scm_array is assumed. */
943 scm_makfromstrs (int argc
, char **argv
)
948 for (i
= 0; argv
[i
]; i
++);
950 lst
= scm_cons (scm_from_locale_string (argv
[i
]), lst
);
954 /* Return a newly allocated array of char pointers to each of the strings
955 in args, with a terminating NULL pointer. */
958 scm_i_allocate_string_pointers (SCM list
)
961 int len
= scm_ilength (list
);
965 scm_wrong_type_arg_msg (NULL
, 0, list
, "proper list");
967 scm_dynwind_begin (0);
969 result
= (char **) scm_malloc ((len
+ 1) * sizeof (char *));
971 scm_dynwind_unwind_handler (free
, result
, 0);
973 /* The list might be have been modified in another thread, so
974 we check LIST before each access.
976 for (i
= 0; i
< len
&& scm_is_pair (list
); i
++)
978 result
[i
] = scm_to_locale_string (SCM_CAR (list
));
979 list
= SCM_CDR (list
);
987 scm_i_free_string_pointers (char **pointers
)
991 for (i
= 0; pointers
[i
]; i
++)
997 scm_i_get_substring_spec (size_t len
,
998 SCM start
, size_t *cstart
,
999 SCM end
, size_t *cend
)
1001 if (SCM_UNBNDP (start
))
1004 *cstart
= scm_to_unsigned_integer (start
, 0, len
);
1006 if (SCM_UNBNDP (end
))
1009 *cend
= scm_to_unsigned_integer (end
, *cstart
, len
);
1012 #if SCM_ENABLE_DEPRECATED
1014 /* When these definitions are removed, it becomes reasonable to use
1015 read-only strings for string literals. For that, change the reader
1016 to create string literals with scm_c_substring_read_only instead of
1017 with scm_c_substring_copy.
1021 scm_i_deprecated_stringp (SCM str
)
1023 scm_c_issue_deprecation_warning
1024 ("SCM_STRINGP is deprecated. Use scm_is_string instead.");
1026 return scm_is_string (str
);
1030 scm_i_deprecated_string_chars (SCM str
)
1034 scm_c_issue_deprecation_warning
1035 ("SCM_STRING_CHARS is deprecated. See the manual for alternatives.");
1037 /* We don't accept shared substrings here since they are not
1040 if (IS_SH_STRING (str
))
1041 scm_misc_error (NULL
,
1042 "SCM_STRING_CHARS does not work with shared substrings.",
1045 /* We explicitely test for read-only strings to produce a better
1049 if (IS_RO_STRING (str
))
1050 scm_misc_error (NULL
,
1051 "SCM_STRING_CHARS does not work with read-only strings.",
1054 /* The following is still wrong, of course...
1056 chars
= scm_i_string_writable_chars (str
);
1057 scm_i_string_stop_writing ();
1062 scm_i_deprecated_string_length (SCM str
)
1064 scm_c_issue_deprecation_warning
1065 ("SCM_STRING_LENGTH is deprecated. Use scm_c_string_length instead.");
1066 return scm_c_string_length (str
);
1074 scm_nullstr
= scm_i_make_string (0, NULL
);
1076 #include "libguile/strings.x"