libguile/strings.h

   1 /* classes: h_files */
   2
   3 #ifndef SCM_STRINGS_H
   4 #define SCM_STRINGS_H
   5
   6 /* Copyright (C) 1995,1996,1997,1998,2000,2001, 2004, 2005, 2006, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
   7  *
   8  * This library is free software; you can redistribute it and/or
   9  * modify it under the terms of the GNU Lesser General Public License
  10  * as published by the Free Software Foundation; either version 3 of
  11  * the License, or (at your option) any later version.
  12  *
  13  * This library is distributed in the hope that it will be useful, but
  14  * WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  16  * Lesser General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU Lesser General Public
  19  * License along with this library; if not, write to the Free Software
  20  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  21  * 02110-1301 USA
  22  */
  23
  24 \f
  25
  26 #include "libguile/__scm.h"
  27
  28 \f
  29
  30 /* String representation.
  31
  32    A string is a piece of a stringbuf.  A stringbuf can be used by
  33    more than one string.  When a string is written to and the
  34    stringbuf of that string is used by more than one string, a new
  35    stringbuf is created.  That is, strings are copy-on-write.  This
  36    behavior can be used to make the substring operation quite
  37    efficient.
  38
  39    The implementation is tuned so that mutating a string is costly,
  40    but just reading it is cheap and lock-free.
  41
  42    There are also mutation-sharing strings.  They refer to a part of
  43    an ordinary string.  Writing to a mutation-sharing string just
  44    writes to the ordinary string.
  45
  46
  47    Internal, low level interface to the character arrays
  48
  49    - Use scm_is_narrow_string to determine is the string is narrow or
  50      wide.
  51
  52    - Use scm_i_string_chars or scm_i_string_wide_chars to get a
  53      pointer to the byte or scm_t_wchar array of a string for reading.
  54      Use scm_i_string_length to get the number of characters in that
  55      array.  The array is not null-terminated.
  56
  57    - The array is valid as long as the corresponding SCM object is
  58      protected but only until the next SCM_TICK.  During such a 'safe
  59      point', strings might change their representation.
  60
  61    - Use scm_i_string_start_writing to get a version of the string
  62      ready for reading and writing.  This is a potentially costly
  63      operation since it implements the copy-on-write behavior.  When
  64      done with the writing, call scm_i_string_stop_writing.  You must
  65      do this before the next SCM_TICK.  (This means, before calling
  66      almost any other scm_ function and you can't allow throws, of
  67      course.)
  68
  69    - New strings can be created with scm_i_make_string or
  70      scm_i_make_wide_string.  This gives access to a writable pointer
  71      that remains valid as long as nobody else makes a copy-on-write
  72      substring of the string.  Do not call scm_i_string_stop_writing
  73      for this pointer.
  74
  75    - Alternately, scm_i_string_ref and scm_i_string_set_x can be used
  76      to read and write strings without worrying about whether the
  77      string is narrow or wide.  scm_i_string_set_x still needs to be
  78      bracketed by scm_i_string_start_writing and
  79      scm_i_string_stop_writing.
  80
  81    Legacy interface
  82
  83    - SCM_STRINGP is just scm_is_string.
  84
  85    - SCM_STRING_CHARS uses scm_i_string_writable_chars and immediately
  86      calls scm_i_stop_writing, hoping for the best.  SCM_STRING_LENGTH
  87      is the same as scm_i_string_length.  SCM_STRING_CHARS will throw
  88      an error for strings that are not null-terminated.  There is
  89      no wide version of this interface.
  90 */
  91
  92 /* A type indicating what strategy to take when string locale
  93    conversion is unsuccessful.  */
  94 typedef enum
  95 {
  96   SCM_FAILED_CONVERSION_ERROR = SCM_ICONVEH_ERROR,
  97   SCM_FAILED_CONVERSION_QUESTION_MARK = SCM_ICONVEH_QUESTION_MARK,
  98   SCM_FAILED_CONVERSION_ESCAPE_SEQUENCE = SCM_ICONVEH_ESCAPE_SEQUENCE
  99 } scm_t_string_failed_conversion_handler;
 100
 101 SCM_INTERNAL SCM scm_nullstr;
 102
 103 SCM_API SCM scm_string_p (SCM x);
 104 SCM_API SCM scm_string (SCM chrs);
 105 SCM_API SCM scm_make_string (SCM k, SCM chr);
 106 SCM_API SCM scm_string_length (SCM str);
 107 SCM_API SCM scm_string_bytes_per_char (SCM str);
 108 SCM_API SCM scm_string_ref (SCM str, SCM k);
 109 SCM_API SCM scm_string_set_x (SCM str, SCM k, SCM chr);
 110 SCM_API SCM scm_substring (SCM str, SCM start, SCM end);
 111 SCM_API SCM scm_substring_read_only (SCM str, SCM start, SCM end);
 112 SCM_API SCM scm_substring_shared (SCM str, SCM start, SCM end);
 113 SCM_API SCM scm_substring_copy (SCM str, SCM start, SCM end);
 114 SCM_API SCM scm_string_append (SCM args);
 115
 116 SCM_API SCM scm_from_stringn (const char *str, size_t len, const char *encoding,
 117                               scm_t_string_failed_conversion_handler handler);
 118 SCM_API SCM scm_c_make_string (size_t len, SCM chr);
 119 SCM_API size_t scm_c_string_length (SCM str);
 120 SCM_API size_t scm_c_symbol_length (SCM sym);
 121 SCM_API SCM scm_c_string_ref (SCM str, size_t pos);
 122 SCM_API void scm_c_string_set_x (SCM str, size_t pos, SCM chr);
 123 SCM_API SCM scm_c_substring (SCM str, size_t start, size_t end);
 124 SCM_API SCM scm_c_substring_read_only (SCM str, size_t start, size_t end);
 125 SCM_API SCM scm_c_substring_shared (SCM str, size_t start, size_t end);
 126 SCM_API SCM scm_c_substring_copy (SCM str, size_t start, size_t end);
 127
 128 /* Use locale encoding for user input, user output, or interacting with
 129    the C library.  Use latin1 for ASCII, and for literals in source
 130    code.  Use utf8 for interaction with modern libraries which deal in
 131    UTF-8.  Otherwise use scm_to_stringn or scm_from_stringn with a
 132    specific encoding. */
 133
 134 SCM_API SCM scm_from_locale_string (const char *str);
 135 SCM_API SCM scm_from_locale_stringn (const char *str, size_t len);
 136 SCM_API SCM scm_take_locale_string (char *str);
 137 SCM_API SCM scm_take_locale_stringn (char *str, size_t len);
 138 SCM_API char *scm_to_locale_string (SCM str);
 139 SCM_API char *scm_to_locale_stringn (SCM str, size_t *lenp);
 140
 141 SCM_API SCM scm_from_latin1_string (const char *str);
 142 SCM_API SCM scm_from_latin1_stringn (const char *str, size_t len);
 143 SCM_API char *scm_to_latin1_string (SCM str);
 144 SCM_API char *scm_to_latin1_stringn (SCM str, size_t *lenp);
 145
 146 SCM_API char *scm_to_utf8_string (SCM str);
 147 SCM_API char *scm_to_utf8_stringn (SCM str, size_t *lenp);
 148 SCM_API SCM scm_from_utf8_string (const char *str);
 149 SCM_API SCM scm_from_utf8_stringn (const char *str, size_t len);
 150
 151 SCM_API scm_t_wchar *scm_to_utf32_string (SCM str);
 152 SCM_API scm_t_wchar *scm_to_utf32_stringn (SCM str, size_t *lenp);
 153 SCM_API SCM scm_from_utf32_string (const scm_t_wchar *str);
 154 SCM_API SCM scm_from_utf32_stringn (const scm_t_wchar *str, size_t len);
 155
 156 SCM_API char *scm_to_stringn (SCM str, size_t *lenp, const char *encoding,
 157                               scm_t_string_failed_conversion_handler handler);
 158 SCM_API size_t scm_to_locale_stringbuf (SCM str, char *buf, size_t max_len);
 159
 160 SCM_API SCM scm_string_normalize_nfd (SCM str);
 161 SCM_API SCM scm_string_normalize_nfkd (SCM str);
 162 SCM_API SCM scm_string_normalize_nfc (SCM str);
 163 SCM_API SCM scm_string_normalize_nfkc (SCM str);
 164
 165 SCM_API SCM scm_makfromstrs (int argc, char **argv);
 166
 167 \f
 168 /* internal constants */
 169
 170 /* Type tag for read-only strings.  */
 171 #define scm_tc7_ro_string             (scm_tc7_string + 0x200)
 172
 173 /* Flags for shared and wide strings.  */
 174 #define SCM_I_STRINGBUF_F_SHARED      0x100
 175 #define SCM_I_STRINGBUF_F_WIDE        0x400
 176
 177
 178 /* internal accessor functions.  Arguments must be valid. */
 179
 180 SCM_INTERNAL SCM scm_i_make_string (size_t len, char **datap,
 181                                     int read_only_p);
 182 SCM_INTERNAL SCM scm_i_make_wide_string (size_t len, scm_t_wchar **datap,
 183                                          int read_only_p);
 184 SCM_INTERNAL SCM scm_i_set_string_read_only_x (SCM str);
 185 SCM_INTERNAL SCM scm_i_substring (SCM str, size_t start, size_t end);
 186 SCM_INTERNAL SCM scm_i_substring_read_only (SCM str, size_t start, size_t end);
 187 SCM_INTERNAL SCM scm_i_substring_shared (SCM str, size_t start, size_t end);
 188 SCM_INTERNAL SCM scm_i_substring_copy (SCM str, size_t start, size_t end);
 189 SCM_INTERNAL size_t scm_i_string_length (SCM str);
 190 SCM_API /* FIXME: not internal */ const char *scm_i_string_chars (SCM str);
 191 SCM_API /* FIXME: not internal */ char *scm_i_string_writable_chars (SCM str);
 192 SCM_INTERNAL const scm_t_wchar *scm_i_string_wide_chars (SCM str);
 193 SCM_INTERNAL const void *scm_i_string_data (SCM str);
 194
 195 SCM_INTERNAL SCM scm_i_string_start_writing (SCM str);
 196 SCM_INTERNAL void scm_i_string_stop_writing (void);
 197 SCM_INTERNAL int scm_i_is_narrow_string (SCM str);
 198 SCM_INTERNAL scm_t_wchar scm_i_string_ref (SCM str, size_t x);
 199 SCM_INTERNAL int scm_i_string_contains_char (SCM str, char c);
 200 SCM_INTERNAL int scm_i_string_strcmp (SCM sstr, size_t start_x, const char *cstr);
 201 SCM_INTERNAL void scm_i_string_set_x (SCM str, size_t p, scm_t_wchar chr);
 202 /* internal functions related to symbols. */
 203
 204 SCM_INTERNAL SCM scm_i_make_symbol (SCM name, scm_t_bits flags,
 205                                     unsigned long hash, SCM props);
 206 SCM_INTERNAL SCM
 207 scm_i_c_make_symbol (const char *name, size_t len,
 208                      scm_t_bits flags, unsigned long hash, SCM props);
 209 SCM_INTERNAL const char *scm_i_symbol_chars (SCM sym);
 210 SCM_INTERNAL const scm_t_wchar *scm_i_symbol_wide_chars (SCM sym);
 211 SCM_INTERNAL size_t scm_i_symbol_length (SCM sym);
 212 SCM_INTERNAL int scm_i_is_narrow_symbol (SCM str);
 213 SCM_INTERNAL int scm_i_try_narrow_string (SCM str);
 214 SCM_INTERNAL SCM scm_i_symbol_substring (SCM sym, size_t start, size_t end);
 215 SCM_INTERNAL scm_t_wchar scm_i_symbol_ref (SCM sym, size_t x);
 216 SCM_INTERNAL void scm_encoding_error (const char *subr, int err,
 217                                       const char *message, SCM port, SCM chr);
 218 SCM_INTERNAL void scm_decoding_error (const char *subr, int err,
 219                                       const char *message, SCM port);
 220
 221 /* internal utility functions. */
 222
 223 SCM_INTERNAL char **scm_i_allocate_string_pointers (SCM list);
 224 SCM_INTERNAL void scm_i_get_substring_spec (size_t len,
 225                                             SCM start, size_t *cstart,
 226                                             SCM end, size_t *cend);
 227
 228 /* Debugging functions */
 229
 230 SCM_API SCM scm_sys_string_dump (SCM);
 231 SCM_API SCM scm_sys_symbol_dump (SCM);
 232 #ifdef SCM_STRING_LENGTH_HISTOGRAM
 233 SCM_API SCM scm_sys_stringbuf_hist (void);
 234 #endif
 235
 236
 237
 238 /* deprecated stuff */
 239
 240 #if SCM_ENABLE_DEPRECATED
 241
 242 SCM_DEPRECATED int scm_i_deprecated_stringp (SCM obj);
 243 SCM_DEPRECATED char *scm_i_deprecated_string_chars (SCM str);
 244 SCM_DEPRECATED size_t scm_i_deprecated_string_length (SCM str);
 245
 246 #define SCM_STRINGP(x)       scm_i_deprecated_stringp(x)
 247 #define SCM_STRING_CHARS(x)  scm_i_deprecated_string_chars(x)
 248 #define SCM_STRING_LENGTH(x) scm_i_deprecated_string_length(x)
 249 #define SCM_STRING_UCHARS(str) ((unsigned char *)SCM_STRING_CHARS (str))
 250
 251 #endif
 252
 253 SCM_INTERNAL void scm_init_strings (void);
 254
 255 #endif  /* SCM_STRINGS_H */
 256
 257 /*
 258   Local Variables:
 259   c-file-style: "gnu"
 260   End:
 261 */