src/casefiddle.c

   1 /* GNU Emacs case conversion functions.
   2    Copyright (C) 1985,94,97,98,99, 2001, 2002, 2004
   3    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 2, or (at your option)
  10 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; see the file COPYING.  If not, write to
  19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  20 Boston, MA 02111-1307, USA.  */
  21
  22
  23 #include <config.h>
  24 #include "lisp.h"
  25 #include "buffer.h"
  26 #include "charset.h"
  27 #include "commands.h"
  28 #include "syntax.h"
  29 #include "composite.h"
  30 #include "keymap.h"
  31
  32 enum case_action {CASE_UP, CASE_DOWN, CASE_CAPITALIZE, CASE_CAPITALIZE_UP};
  33
  34 Lisp_Object Qidentity;
  35 \f
  36 Lisp_Object
  37 casify_object (flag, obj)
  38      enum case_action flag;
  39      Lisp_Object obj;
  40 {
  41   register int i, c, len;
  42   register int inword = flag == CASE_DOWN;
  43
  44   /* If the case table is flagged as modified, rescan it.  */
  45   if (NILP (XCHAR_TABLE (current_buffer->downcase_table)->extras[1]))
  46     Fset_case_table (current_buffer->downcase_table);
  47
  48   while (1)
  49     {
  50       if (INTEGERP (obj))
  51         {
  52           int flagbits = (CHAR_ALT | CHAR_SUPER | CHAR_HYPER
  53                           | CHAR_SHIFT | CHAR_CTL | CHAR_META);
  54           int flags = XINT (obj) & flagbits;
  55
  56           /* If the character has higher bits set
  57              above the flags, return it unchanged.
  58              It is not a real character.  */
  59           if ((unsigned) XFASTINT (obj) > (unsigned) flagbits)
  60             return obj;
  61
  62           c = DOWNCASE (XFASTINT (obj) & ~flagbits);
  63           if (inword)
  64             XSETFASTINT (obj, c | flags);
  65           else if (c == (XFASTINT (obj) & ~flagbits))
  66             {
  67               c = UPCASE1 ((XFASTINT (obj) & ~flagbits));
  68               XSETFASTINT (obj, c | flags);
  69             }
  70           return obj;
  71         }
  72
  73       if (STRINGP (obj))
  74         {
  75           int multibyte = STRING_MULTIBYTE (obj);
  76
  77           obj = Fcopy_sequence (obj);
  78           len = SBYTES (obj);
  79
  80           /* Scan all single-byte characters from start of string.  */
  81           for (i = 0; i < len;)
  82             {
  83               c = SREF (obj, i);
  84
  85               if (multibyte && c >= 0x80)
  86                 /* A multibyte character can't be handled in this
  87                    simple loop.  */
  88                 break;
  89               if (inword && flag != CASE_CAPITALIZE_UP)
  90                 c = DOWNCASE (c);
  91               else if (!UPPERCASEP (c)
  92                        && (!inword || flag != CASE_CAPITALIZE_UP))
  93                 c = UPCASE1 (c);
  94               /* If this char won't fit in a single-byte string.
  95                  fall out to the multibyte case.  */
  96               if (multibyte ? ! ASCII_BYTE_P (c)
  97                   : ! SINGLE_BYTE_CHAR_P (c))
  98                 break;
  99
 100               SSET (obj, i, c);
 101               if ((int) flag >= (int) CASE_CAPITALIZE)
 102                 inword = SYNTAX (c) == Sword;
 103               i++;
 104             }
 105
 106           /* If we didn't do the whole string as single-byte,
 107              scan the rest in a more complex way.  */
 108           if (i < len)
 109             {
 110               /* The work is not yet finished because of a multibyte
 111                  character just encountered.  */
 112               int fromlen, j_byte = i;
 113               char *buf;
 114               int bufsize;
 115               USE_SAFE_ALLOCA;
 116
 117               bufsize = (len - i) * MAX_MULTIBYTE_LENGTH + i;
 118               SAFE_ALLOCA (buf, char *, bufsize);
 119
 120               /* Copy data already handled.  */
 121               bcopy (SDATA (obj), buf, i);
 122
 123               /* From now on, I counts bytes.  */
 124               while (i < len)
 125                 {
 126                   c = STRING_CHAR_AND_LENGTH (SDATA (obj) + i,
 127                                               len - i, fromlen);
 128                   if (inword && flag != CASE_CAPITALIZE_UP)
 129                     c = DOWNCASE (c);
 130                   else if (!UPPERCASEP (c)
 131                            && (!inword || flag != CASE_CAPITALIZE_UP))
 132                     c = UPCASE1 (c);
 133                   i += fromlen;
 134                   j_byte += CHAR_STRING (c, buf + j_byte);
 135                   if ((int) flag >= (int) CASE_CAPITALIZE)
 136                     inword = SYNTAX (c) == Sword;
 137                 }
 138               obj = make_multibyte_string (buf, SCHARS (obj),
 139                                            j_byte);
 140               SAFE_FREE (bufsize);
 141             }
 142           return obj;
 143         }
 144       obj = wrong_type_argument (Qchar_or_string_p, obj);
 145     }
 146 }
 147
 148 DEFUN ("upcase", Fupcase, Supcase, 1, 1, 0,
 149        doc: /* Convert argument to upper case and return that.
 150 The argument may be a character or string.  The result has the same type.
 151 The argument object is not altered--the value is a copy.
 152 See also `capitalize', `downcase' and `upcase-initials'.  */)
 153      (obj)
 154      Lisp_Object obj;
 155 {
 156   return casify_object (CASE_UP, obj);
 157 }
 158
 159 DEFUN ("downcase", Fdowncase, Sdowncase, 1, 1, 0,
 160        doc: /* Convert argument to lower case and return that.
 161 The argument may be a character or string.  The result has the same type.
 162 The argument object is not altered--the value is a copy.  */)
 163      (obj)
 164      Lisp_Object obj;
 165 {
 166   return casify_object (CASE_DOWN, obj);
 167 }
 168
 169 DEFUN ("capitalize", Fcapitalize, Scapitalize, 1, 1, 0,
 170        doc: /* Convert argument to capitalized form and return that.
 171 This means that each word's first character is upper case
 172 and the rest is lower case.
 173 The argument may be a character or string.  The result has the same type.
 174 The argument object is not altered--the value is a copy.  */)
 175      (obj)
 176      Lisp_Object obj;
 177 {
 178   return casify_object (CASE_CAPITALIZE, obj);
 179 }
 180
 181 /* Like Fcapitalize but change only the initials.  */
 182
 183 DEFUN ("upcase-initials", Fupcase_initials, Supcase_initials, 1, 1, 0,
 184        doc: /* Convert the initial of each word in the argument to upper case.
 185 Do not change the other letters of each word.
 186 The argument may be a character or string.  The result has the same type.
 187 The argument object is not altered--the value is a copy.  */)
 188      (obj)
 189      Lisp_Object obj;
 190 {
 191   return casify_object (CASE_CAPITALIZE_UP, obj);
 192 }
 193 \f
 194 /* flag is CASE_UP, CASE_DOWN or CASE_CAPITALIZE or CASE_CAPITALIZE_UP.
 195    b and e specify range of buffer to operate on. */
 196
 197 void
 198 casify_region (flag, b, e)
 199      enum case_action flag;
 200      Lisp_Object b, e;
 201 {
 202   register int i;
 203   register int c;
 204   register int inword = flag == CASE_DOWN;
 205   register int multibyte = !NILP (current_buffer->enable_multibyte_characters);
 206   int start, end;
 207   int start_byte, end_byte;
 208   int changed = 0;
 209
 210   if (EQ (b, e))
 211     /* Not modifying because nothing marked */
 212     return;
 213
 214   /* If the case table is flagged as modified, rescan it.  */
 215   if (NILP (XCHAR_TABLE (current_buffer->downcase_table)->extras[1]))
 216     Fset_case_table (current_buffer->downcase_table);
 217
 218   validate_region (&b, &e);
 219   start = XFASTINT (b);
 220   end = XFASTINT (e);
 221   modify_region (current_buffer, start, end);
 222   record_change (start, end - start);
 223   start_byte = CHAR_TO_BYTE (start);
 224   end_byte = CHAR_TO_BYTE (end);
 225
 226   for (i = start_byte; i < end_byte; i++, start++)
 227     {
 228       int c2;
 229       c = c2 = FETCH_BYTE (i);
 230       if (multibyte && c >= 0x80)
 231         /* A multibyte character can't be handled in this simple loop.  */
 232         break;
 233       if (inword && flag != CASE_CAPITALIZE_UP)
 234         c = DOWNCASE (c);
 235       else if (!UPPERCASEP (c)
 236                && (!inword || flag != CASE_CAPITALIZE_UP))
 237         c = UPCASE1 (c);
 238       FETCH_BYTE (i) = c;
 239       if (c != c2)
 240         changed = 1;
 241       if ((int) flag >= (int) CASE_CAPITALIZE)
 242         inword = SYNTAX (c) == Sword && (inword || !SYNTAX_PREFIX (c));
 243     }
 244   if (i < end_byte)
 245     {
 246       /* The work is not yet finished because of a multibyte character
 247          just encountered.  */
 248       int opoint = PT;
 249       int opoint_byte = PT_BYTE;
 250       int c2;
 251
 252       while (i < end_byte)
 253         {
 254           if ((c = FETCH_BYTE (i)) >= 0x80)
 255             c = FETCH_MULTIBYTE_CHAR (i);
 256           c2 = c;
 257           if (inword && flag != CASE_CAPITALIZE_UP)
 258             c2 = DOWNCASE (c);
 259           else if (!UPPERCASEP (c)
 260                    && (!inword || flag != CASE_CAPITALIZE_UP))
 261             c2 = UPCASE1 (c);
 262           if (c != c2)
 263             {
 264               int fromlen, tolen, j;
 265               unsigned char str[MAX_MULTIBYTE_LENGTH];
 266
 267               changed = 1;
 268               /* Handle the most likely case */
 269               if (c < 0400 && c2 < 0400)
 270                 FETCH_BYTE (i) = c2;
 271               else if (fromlen = CHAR_STRING (c, str),
 272                        tolen = CHAR_STRING (c2, str),
 273                        fromlen == tolen)
 274                 {
 275                   for (j = 0; j < tolen; ++j)
 276                     FETCH_BYTE (i + j) = str[j];
 277                 }
 278               else
 279                 {
 280                   error ("Can't casify letters that change length");
 281 #if 0 /* This is approximately what we'd like to be able to do here */
 282                   if (tolen < fromlen)
 283                     del_range_1 (i + tolen, i + fromlen, 0, 0);
 284                   else if (tolen > fromlen)
 285                     {
 286                       TEMP_SET_PT (i + fromlen);
 287                       insert_1 (str + fromlen, tolen - fromlen, 1, 0, 0);
 288                     }
 289 #endif
 290                 }
 291             }
 292           if ((int) flag >= (int) CASE_CAPITALIZE)
 293             inword = SYNTAX (c2) == Sword;
 294           INC_BOTH (start, i);
 295         }
 296       TEMP_SET_PT_BOTH (opoint, opoint_byte);
 297     }
 298
 299   start = XFASTINT (b);
 300   if (changed)
 301     {
 302       signal_after_change (start, end - start, end - start);
 303       update_compositions (start, end, CHECK_ALL);
 304     }
 305 }
 306
 307 DEFUN ("upcase-region", Fupcase_region, Supcase_region, 2, 2, "r",
 308        doc: /* Convert the region to upper case.  In programs, wants two arguments.
 309 These arguments specify the starting and ending character numbers of
 310 the region to operate on.  When used as a command, the text between
 311 point and the mark is operated on.
 312 See also `capitalize-region'.  */)
 313      (beg, end)
 314      Lisp_Object beg, end;
 315 {
 316   casify_region (CASE_UP, beg, end);
 317   return Qnil;
 318 }
 319
 320 DEFUN ("downcase-region", Fdowncase_region, Sdowncase_region, 2, 2, "r",
 321        doc: /* Convert the region to lower case.  In programs, wants two arguments.
 322 These arguments specify the starting and ending character numbers of
 323 the region to operate on.  When used as a command, the text between
 324 point and the mark is operated on.  */)
 325      (beg, end)
 326      Lisp_Object beg, end;
 327 {
 328   casify_region (CASE_DOWN, beg, end);
 329   return Qnil;
 330 }
 331
 332 DEFUN ("capitalize-region", Fcapitalize_region, Scapitalize_region, 2, 2, "r",
 333        doc: /* Convert the region to capitalized form.
 334 Capitalized form means each word's first character is upper case
 335 and the rest of it is lower case.
 336 In programs, give two arguments, the starting and ending
 337 character positions to operate on.  */)
 338      (beg, end)
 339      Lisp_Object beg, end;
 340 {
 341   casify_region (CASE_CAPITALIZE, beg, end);
 342   return Qnil;
 343 }
 344
 345 /* Like Fcapitalize_region but change only the initials.  */
 346
 347 DEFUN ("upcase-initials-region", Fupcase_initials_region,
 348        Supcase_initials_region, 2, 2, "r",
 349        doc: /* Upcase the initial of each word in the region.
 350 Subsequent letters of each word are not changed.
 351 In programs, give two arguments, the starting and ending
 352 character positions to operate on.  */)
 353      (beg, end)
 354      Lisp_Object beg, end;
 355 {
 356   casify_region (CASE_CAPITALIZE_UP, beg, end);
 357   return Qnil;
 358 }
 359 \f
 360 Lisp_Object
 361 operate_on_word (arg, newpoint)
 362      Lisp_Object arg;
 363      int *newpoint;
 364 {
 365   Lisp_Object val;
 366   int farend;
 367   int iarg;
 368
 369   CHECK_NUMBER (arg);
 370   iarg = XINT (arg);
 371   farend = scan_words (PT, iarg);
 372   if (!farend)
 373     farend = iarg > 0 ? ZV : BEGV;
 374
 375   *newpoint = PT > farend ? PT : farend;
 376   XSETFASTINT (val, farend);
 377
 378   return val;
 379 }
 380
 381 DEFUN ("upcase-word", Fupcase_word, Supcase_word, 1, 1, "p",
 382        doc: /* Convert following word (or ARG words) to upper case, moving over.
 383 With negative argument, convert previous words but do not move.
 384 See also `capitalize-word'.  */)
 385      (arg)
 386      Lisp_Object arg;
 387 {
 388   Lisp_Object beg, end;
 389   int newpoint;
 390   XSETFASTINT (beg, PT);
 391   end = operate_on_word (arg, &newpoint);
 392   casify_region (CASE_UP, beg, end);
 393   SET_PT (newpoint);
 394   return Qnil;
 395 }
 396
 397 DEFUN ("downcase-word", Fdowncase_word, Sdowncase_word, 1, 1, "p",
 398        doc: /* Convert following word (or ARG words) to lower case, moving over.
 399 With negative argument, convert previous words but do not move.  */)
 400      (arg)
 401      Lisp_Object arg;
 402 {
 403   Lisp_Object beg, end;
 404   int newpoint;
 405   XSETFASTINT (beg, PT);
 406   end = operate_on_word (arg, &newpoint);
 407   casify_region (CASE_DOWN, beg, end);
 408   SET_PT (newpoint);
 409   return Qnil;
 410 }
 411
 412 DEFUN ("capitalize-word", Fcapitalize_word, Scapitalize_word, 1, 1, "p",
 413        doc: /* Capitalize the following word (or ARG words), moving over.
 414 This gives the word(s) a first character in upper case
 415 and the rest lower case.
 416 With negative argument, capitalize previous words but do not move.  */)
 417      (arg)
 418      Lisp_Object arg;
 419 {
 420   Lisp_Object beg, end;
 421   int newpoint;
 422   XSETFASTINT (beg, PT);
 423   end = operate_on_word (arg, &newpoint);
 424   casify_region (CASE_CAPITALIZE, beg, end);
 425   SET_PT (newpoint);
 426   return Qnil;
 427 }
 428 \f
 429 void
 430 syms_of_casefiddle ()
 431 {
 432   Qidentity = intern ("identity");
 433   staticpro (&Qidentity);
 434   defsubr (&Supcase);
 435   defsubr (&Sdowncase);
 436   defsubr (&Scapitalize);
 437   defsubr (&Supcase_initials);
 438   defsubr (&Supcase_region);
 439   defsubr (&Sdowncase_region);
 440   defsubr (&Scapitalize_region);
 441   defsubr (&Supcase_initials_region);
 442   defsubr (&Supcase_word);
 443   defsubr (&Sdowncase_word);
 444   defsubr (&Scapitalize_word);
 445 }
 446
 447 void
 448 keys_of_casefiddle ()
 449 {
 450   initial_define_key (control_x_map, Ctl('U'), "upcase-region");
 451   Fput (intern ("upcase-region"), Qdisabled, Qt);
 452   initial_define_key (control_x_map, Ctl('L'), "downcase-region");
 453   Fput (intern ("downcase-region"), Qdisabled, Qt);
 454
 455   initial_define_key (meta_map, 'u', "upcase-word");
 456   initial_define_key (meta_map, 'l', "downcase-word");
 457   initial_define_key (meta_map, 'c', "capitalize-word");
 458 }
 459
 460 /* arch-tag: 60a73c66-5489-47e7-a81f-cead4057c526
 461    (do not change this comment) */