src/casefiddle.c

   1 /* GNU Emacs case conversion functions.
   2
   3 Copyright (C) 1985, 1994, 1997-1999, 2001-2012 Free Software Foundation, Inc.
   4
   5 This file is part of GNU Emacs.
   6
   7 GNU Emacs is free software: you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License as published by
   9 the Free Software Foundation, either version 3 of the License, or
  10 (at your option) any later version.
  11
  12 GNU Emacs 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
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.  */
  19
  20
  21 #include <config.h>
  22 #include <setjmp.h>
  23 #include "lisp.h"
  24 #include "character.h"
  25 #include "buffer.h"
  26 #include "commands.h"
  27 #include "syntax.h"
  28 #include "composite.h"
  29 #include "keymap.h"
  30
  31 enum case_action {CASE_UP, CASE_DOWN, CASE_CAPITALIZE, CASE_CAPITALIZE_UP};
  32
  33 Lisp_Object Qidentity;
  34 \f
  35 static Lisp_Object
  36 casify_object (enum case_action flag, Lisp_Object obj)
  37 {
  38   register int c, c1;
  39   register int inword = flag == CASE_DOWN;
  40
  41   /* If the case table is flagged as modified, rescan it.  */
  42   if (NILP (XCHAR_TABLE (BVAR (current_buffer, downcase_table))->extras[1]))
  43     Fset_case_table (BVAR (current_buffer, downcase_table));
  44
  45   if (INTEGERP (obj))
  46     {
  47       int flagbits = (CHAR_ALT | CHAR_SUPER | CHAR_HYPER
  48                       | CHAR_SHIFT | CHAR_CTL | CHAR_META);
  49       int flags = XINT (obj) & flagbits;
  50       int multibyte = ! NILP (BVAR (current_buffer, enable_multibyte_characters));
  51
  52       /* If the character has higher bits set
  53          above the flags, return it unchanged.
  54          It is not a real character.  */
  55       if (UNSIGNED_CMP (XFASTINT (obj), >, flagbits))
  56         return obj;
  57
  58       c1 = XFASTINT (obj) & ~flagbits;
  59       /* FIXME: Even if enable-multibyte-characters is nil, we may
  60          manipulate multibyte chars.  This means we have a bug for latin-1
  61          chars since when we receive an int 128-255 we can't tell whether
  62          it's an eight-bit byte or a latin-1 char.  */
  63       if (c1 >= 256)
  64         multibyte = 1;
  65       if (! multibyte)
  66         MAKE_CHAR_MULTIBYTE (c1);
  67       c = downcase (c1);
  68       if (inword)
  69         XSETFASTINT (obj, c | flags);
  70       else if (c == (XFASTINT (obj) & ~flagbits))
  71         {
  72           if (! inword)
  73             c = upcase1 (c1);
  74           if (! multibyte)
  75             MAKE_CHAR_UNIBYTE (c);
  76           XSETFASTINT (obj, c | flags);
  77         }
  78       return obj;
  79     }
  80
  81   if (!STRINGP (obj))
  82     wrong_type_argument (Qchar_or_string_p, obj);
  83   else if (!STRING_MULTIBYTE (obj))
  84     {
  85       ptrdiff_t i;
  86       ptrdiff_t size = SCHARS (obj);
  87
  88       obj = Fcopy_sequence (obj);
  89       for (i = 0; i < size; i++)
  90         {
  91           c = SREF (obj, i);
  92           MAKE_CHAR_MULTIBYTE (c);
  93           c1 = c;
  94           if (inword && flag != CASE_CAPITALIZE_UP)
  95             c = downcase (c);
  96           else if (!uppercasep (c)
  97                    && (!inword || flag != CASE_CAPITALIZE_UP))
  98             c = upcase1 (c1);
  99           if ((int) flag >= (int) CASE_CAPITALIZE)
 100             inword = (SYNTAX (c) == Sword);
 101           if (c != c1)
 102             {
 103                   MAKE_CHAR_UNIBYTE (c);
 104               /* If the char can't be converted to a valid byte, just don't
 105                  change it.  */
 106               if (c >= 0 && c < 256)
 107                 SSET (obj, i, c);
 108             }
 109         }
 110       return obj;
 111     }
 112   else
 113     {
 114       ptrdiff_t i, i_byte, size = SCHARS (obj);
 115       int len;
 116       USE_SAFE_ALLOCA;
 117       unsigned char *dst, *o;
 118       ptrdiff_t o_size = (size < STRING_BYTES_BOUND / MAX_MULTIBYTE_LENGTH
 119                           ? size * MAX_MULTIBYTE_LENGTH
 120                           : STRING_BYTES_BOUND);
 121       SAFE_ALLOCA (dst, void *, o_size);
 122       o = dst;
 123
 124       for (i = i_byte = 0; i < size; i++, i_byte += len)
 125         {
 126           if (o_size - (o - dst) < MAX_MULTIBYTE_LENGTH)
 127             string_overflow ();
 128           c = STRING_CHAR_AND_LENGTH (SDATA (obj) + i_byte, len);
 129           if (inword && flag != CASE_CAPITALIZE_UP)
 130             c = downcase (c);
 131           else if (!uppercasep (c)
 132                    && (!inword || flag != CASE_CAPITALIZE_UP))
 133             c = upcase1 (c);
 134           if ((int) flag >= (int) CASE_CAPITALIZE)
 135             inword = (SYNTAX (c) == Sword);
 136           o += CHAR_STRING (c, o);
 137         }
 138       eassert (o - dst <= o_size);
 139       obj = make_multibyte_string ((char *) dst, size, o - dst);
 140       SAFE_FREE ();
 141       return obj;
 142     }
 143 }
 144
 145 DEFUN ("upcase", Fupcase, Supcase, 1, 1, 0,
 146        doc: /* Convert argument to upper case and return that.
 147 The argument may be a character or string.  The result has the same type.
 148 The argument object is not altered--the value is a copy.
 149 See also `capitalize', `downcase' and `upcase-initials'.  */)
 150   (Lisp_Object obj)
 151 {
 152   return casify_object (CASE_UP, obj);
 153 }
 154
 155 DEFUN ("downcase", Fdowncase, Sdowncase, 1, 1, 0,
 156        doc: /* Convert argument to lower case and return that.
 157 The argument may be a character or string.  The result has the same type.
 158 The argument object is not altered--the value is a copy.  */)
 159   (Lisp_Object obj)
 160 {
 161   return casify_object (CASE_DOWN, obj);
 162 }
 163
 164 DEFUN ("capitalize", Fcapitalize, Scapitalize, 1, 1, 0,
 165        doc: /* Convert argument to capitalized form and return that.
 166 This means that each word's first character is upper case
 167 and the rest is lower case.
 168 The argument may be a character or string.  The result has the same type.
 169 The argument object is not altered--the value is a copy.  */)
 170   (Lisp_Object obj)
 171 {
 172   return casify_object (CASE_CAPITALIZE, obj);
 173 }
 174
 175 /* Like Fcapitalize but change only the initials.  */
 176
 177 DEFUN ("upcase-initials", Fupcase_initials, Supcase_initials, 1, 1, 0,
 178        doc: /* Convert the initial of each word in the argument to upper case.
 179 Do not change the other letters of each word.
 180 The argument may be a character or string.  The result has the same type.
 181 The argument object is not altered--the value is a copy.  */)
 182   (Lisp_Object obj)
 183 {
 184   return casify_object (CASE_CAPITALIZE_UP, obj);
 185 }
 186 \f
 187 /* flag is CASE_UP, CASE_DOWN or CASE_CAPITALIZE or CASE_CAPITALIZE_UP.
 188    b and e specify range of buffer to operate on. */
 189
 190 static void
 191 casify_region (enum case_action flag, Lisp_Object b, Lisp_Object e)
 192 {
 193   register int c;
 194   register int inword = flag == CASE_DOWN;
 195   register int multibyte = !NILP (BVAR (current_buffer, enable_multibyte_characters));
 196   ptrdiff_t start, end;
 197   ptrdiff_t start_byte;
 198
 199   /* Position of first and last changes.  */
 200   ptrdiff_t first = -1, last IF_LINT (= 0);
 201
 202   ptrdiff_t opoint = PT;
 203   ptrdiff_t opoint_byte = PT_BYTE;
 204
 205   if (EQ (b, e))
 206     /* Not modifying because nothing marked */
 207     return;
 208
 209   /* If the case table is flagged as modified, rescan it.  */
 210   if (NILP (XCHAR_TABLE (BVAR (current_buffer, downcase_table))->extras[1]))
 211     Fset_case_table (BVAR (current_buffer, downcase_table));
 212
 213   validate_region (&b, &e);
 214   start = XFASTINT (b);
 215   end = XFASTINT (e);
 216   modify_region (current_buffer, start, end, 0);
 217   record_change (start, end - start);
 218   start_byte = CHAR_TO_BYTE (start);
 219
 220   SETUP_BUFFER_SYNTAX_TABLE (); /* For syntax_prefix_flag_p.  */
 221
 222   while (start < end)
 223     {
 224       int c2, len;
 225
 226       if (multibyte)
 227         {
 228           c = FETCH_MULTIBYTE_CHAR (start_byte);
 229           len = CHAR_BYTES (c);
 230         }
 231       else
 232         {
 233           c = FETCH_BYTE (start_byte);
 234           MAKE_CHAR_MULTIBYTE (c);
 235           len = 1;
 236         }
 237       c2 = c;
 238       if (inword && flag != CASE_CAPITALIZE_UP)
 239         c = downcase (c);
 240       else if (!uppercasep (c)
 241                && (!inword || flag != CASE_CAPITALIZE_UP))
 242         c = upcase1 (c);
 243       if ((int) flag >= (int) CASE_CAPITALIZE)
 244         inword = ((SYNTAX (c) == Sword)
 245                   && (inword || !syntax_prefix_flag_p (c)));
 246       if (c != c2)
 247         {
 248           last = start;
 249           if (first < 0)
 250             first = start;
 251
 252           if (! multibyte)
 253             {
 254               MAKE_CHAR_UNIBYTE (c);
 255               FETCH_BYTE (start_byte) = c;
 256             }
 257           else if (ASCII_CHAR_P (c2) && ASCII_CHAR_P (c))
 258             FETCH_BYTE (start_byte) = c;
 259           else
 260             {
 261               int tolen = CHAR_BYTES (c);
 262               int j;
 263               unsigned char str[MAX_MULTIBYTE_LENGTH];
 264
 265               CHAR_STRING (c, str);
 266               if (len == tolen)
 267                 {
 268                   /* Length is unchanged.  */
 269                   for (j = 0; j < len; ++j)
 270                     FETCH_BYTE (start_byte + j) = str[j];
 271                 }
 272               else
 273                 {
 274                   /* Replace one character with the other,
 275                      keeping text properties the same.  */
 276                   replace_range_2 (start, start_byte,
 277                                    start + 1, start_byte + len,
 278                                    (char *) str, 1, tolen,
 279                                    0);
 280                   len = tolen;
 281                 }
 282             }
 283         }
 284       start++;
 285       start_byte += len;
 286     }
 287
 288   if (PT != opoint)
 289     TEMP_SET_PT_BOTH (opoint, opoint_byte);
 290
 291   if (first >= 0)
 292     {
 293       signal_after_change (first, last + 1 - first, last + 1 - first);
 294       update_compositions (first, last + 1, CHECK_ALL);
 295     }
 296 }
 297
 298 DEFUN ("upcase-region", Fupcase_region, Supcase_region, 2, 2, "r",
 299        doc: /* Convert the region to upper case.  In programs, wants two arguments.
 300 These arguments specify the starting and ending character numbers of
 301 the region to operate on.  When used as a command, the text between
 302 point and the mark is operated on.
 303 See also `capitalize-region'.  */)
 304   (Lisp_Object beg, Lisp_Object end)
 305 {
 306   casify_region (CASE_UP, beg, end);
 307   return Qnil;
 308 }
 309
 310 DEFUN ("downcase-region", Fdowncase_region, Sdowncase_region, 2, 2, "r",
 311        doc: /* Convert the region to lower case.  In programs, wants two arguments.
 312 These arguments specify the starting and ending character numbers of
 313 the region to operate on.  When used as a command, the text between
 314 point and the mark is operated on.  */)
 315   (Lisp_Object beg, Lisp_Object end)
 316 {
 317   casify_region (CASE_DOWN, beg, end);
 318   return Qnil;
 319 }
 320
 321 DEFUN ("capitalize-region", Fcapitalize_region, Scapitalize_region, 2, 2, "r",
 322        doc: /* Convert the region to capitalized form.
 323 Capitalized form means each word's first character is upper case
 324 and the rest of it is lower case.
 325 In programs, give two arguments, the starting and ending
 326 character positions to operate on.  */)
 327   (Lisp_Object beg, Lisp_Object end)
 328 {
 329   casify_region (CASE_CAPITALIZE, beg, end);
 330   return Qnil;
 331 }
 332
 333 /* Like Fcapitalize_region but change only the initials.  */
 334
 335 DEFUN ("upcase-initials-region", Fupcase_initials_region,
 336        Supcase_initials_region, 2, 2, "r",
 337        doc: /* Upcase the initial of each word in the region.
 338 Subsequent letters of each word are not changed.
 339 In programs, give two arguments, the starting and ending
 340 character positions to operate on.  */)
 341   (Lisp_Object beg, Lisp_Object end)
 342 {
 343   casify_region (CASE_CAPITALIZE_UP, beg, end);
 344   return Qnil;
 345 }
 346 \f
 347 static Lisp_Object
 348 operate_on_word (Lisp_Object arg, ptrdiff_t *newpoint)
 349 {
 350   Lisp_Object val;
 351   ptrdiff_t farend;
 352   EMACS_INT iarg;
 353
 354   CHECK_NUMBER (arg);
 355   iarg = XINT (arg);
 356   farend = scan_words (PT, iarg);
 357   if (!farend)
 358     farend = iarg > 0 ? ZV : BEGV;
 359
 360   *newpoint = PT > farend ? PT : farend;
 361   XSETFASTINT (val, farend);
 362
 363   return val;
 364 }
 365
 366 DEFUN ("upcase-word", Fupcase_word, Supcase_word, 1, 1, "p",
 367        doc: /* Convert following word (or ARG words) to upper case, moving over.
 368 With negative argument, convert previous words but do not move.
 369 See also `capitalize-word'.  */)
 370   (Lisp_Object arg)
 371 {
 372   Lisp_Object beg, end;
 373   ptrdiff_t newpoint;
 374   XSETFASTINT (beg, PT);
 375   end = operate_on_word (arg, &newpoint);
 376   casify_region (CASE_UP, beg, end);
 377   SET_PT (newpoint);
 378   return Qnil;
 379 }
 380
 381 DEFUN ("downcase-word", Fdowncase_word, Sdowncase_word, 1, 1, "p",
 382        doc: /* Convert following word (or ARG words) to lower case, moving over.
 383 With negative argument, convert previous words but do not move.  */)
 384   (Lisp_Object arg)
 385 {
 386   Lisp_Object beg, end;
 387   ptrdiff_t newpoint;
 388   XSETFASTINT (beg, PT);
 389   end = operate_on_word (arg, &newpoint);
 390   casify_region (CASE_DOWN, beg, end);
 391   SET_PT (newpoint);
 392   return Qnil;
 393 }
 394
 395 DEFUN ("capitalize-word", Fcapitalize_word, Scapitalize_word, 1, 1, "p",
 396        doc: /* Capitalize the following word (or ARG words), moving over.
 397 This gives the word(s) a first character in upper case
 398 and the rest lower case.
 399 With negative argument, capitalize previous words but do not move.  */)
 400   (Lisp_Object arg)
 401 {
 402   Lisp_Object beg, end;
 403   ptrdiff_t newpoint;
 404   XSETFASTINT (beg, PT);
 405   end = operate_on_word (arg, &newpoint);
 406   casify_region (CASE_CAPITALIZE, beg, end);
 407   SET_PT (newpoint);
 408   return Qnil;
 409 }
 410 \f
 411 void
 412 syms_of_casefiddle (void)
 413 {
 414   DEFSYM (Qidentity, "identity");
 415   defsubr (&Supcase);
 416   defsubr (&Sdowncase);
 417   defsubr (&Scapitalize);
 418   defsubr (&Supcase_initials);
 419   defsubr (&Supcase_region);
 420   defsubr (&Sdowncase_region);
 421   defsubr (&Scapitalize_region);
 422   defsubr (&Supcase_initials_region);
 423   defsubr (&Supcase_word);
 424   defsubr (&Sdowncase_word);
 425   defsubr (&Scapitalize_word);
 426 }
 427
 428 void
 429 keys_of_casefiddle (void)
 430 {
 431   initial_define_key (control_x_map, Ctl ('U'), "upcase-region");
 432   Fput (intern ("upcase-region"), Qdisabled, Qt);
 433   initial_define_key (control_x_map, Ctl ('L'), "downcase-region");
 434   Fput (intern ("downcase-region"), Qdisabled, Qt);
 435
 436   initial_define_key (meta_map, 'u', "upcase-word");
 437   initial_define_key (meta_map, 'l', "downcase-word");
 438   initial_define_key (meta_map, 'c', "capitalize-word");
 439 }