src/editor/format.c

   1 /*
   2    Dynamic paragraph formatting.
   3
   4    Copyright (C) 2011-2016
   5    Free Software Foundation, Inc.
   6
   7    Copyright (C) 1996 Paul Sheer
   8
   9    Writen by:
  10    Paul Sheer, 1996
  11    Andrew Borodin <aborodin@vmail.ru>, 2013, 2014
  12
  13    This file is part of the Midnight Commander.
  14
  15    The Midnight Commander is free software: you can redistribute it
  16    and/or modify it under the terms of the GNU General Public License as
  17    published by the Free Software Foundation, either version 3 of the License,
  18    or (at your option) any later version.
  19
  20    The Midnight Commander is distributed in the hope that it will be useful,
  21    but WITHOUT ANY WARRANTY; without even the implied warranty of
  22    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  23    GNU General Public License for more details.
  24
  25    You should have received a copy of the GNU General Public License
  26    along with this program.  If not, see <http://www.gnu.org/licenses/>.
  27  */
  28
  29 /** \file
  30  *  \brief Source: Dynamic paragraph formatting
  31  *  \author Paul Sheer
  32  *  \date 1996
  33  *  \author Andrew Borodin
  34  *  \date 2013, 2014
  35  */
  36
  37 #include <config.h>
  38
  39 #include <stdio.h>
  40 #include <stdarg.h>
  41 #include <sys/types.h>
  42 #include <unistd.h>
  43 #include <string.h>
  44 #include <ctype.h>
  45 #include <errno.h>
  46 #include <sys/stat.h>
  47
  48 #include <stdlib.h>
  49
  50 #include "lib/global.h"
  51
  52 #include "src/setup.h"          /* option_tab_spacing */
  53
  54 #include "edit-impl.h"
  55 #include "editwidget.h"
  56
  57 /*** global variables ****************************************************************************/
  58
  59 char *option_stop_format_chars = NULL;
  60
  61 /*** file scope macro definitions ****************************************************************/
  62
  63 #define tab_width option_tab_spacing
  64
  65 #define FONT_MEAN_WIDTH 1
  66
  67 /*** file scope type declarations ****************************************************************/
  68
  69 /*** file scope variables ************************************************************************/
  70
  71 /*** file scope functions ************************************************************************/
  72 /* --------------------------------------------------------------------------------------------- */
  73
  74 static off_t
  75 line_start (const edit_buffer_t * buf, long line)
  76 {
  77     off_t p;
  78     long l;
  79
  80     l = buf->curs_line;
  81     p = buf->curs1;
  82
  83     if (line < l)
  84         p = edit_buffer_get_backward_offset (buf, p, l - line);
  85     else if (line > l)
  86         p = edit_buffer_get_forward_offset (buf, p, line - l, 0);
  87
  88     p = edit_buffer_get_bol (buf, p);
  89     while (strchr ("\t ", edit_buffer_get_byte (buf, p)) != NULL)
  90         p++;
  91     return p;
  92 }
  93
  94 /* --------------------------------------------------------------------------------------------- */
  95
  96 static gboolean
  97 bad_line_start (const edit_buffer_t * buf, off_t p)
  98 {
  99     int c;
 100
 101     c = edit_buffer_get_byte (buf, p);
 102     if (c == '.')
 103     {
 104         /* `...' is acceptable */
 105         return !(edit_buffer_get_byte (buf, p + 1) == '.'
 106                  && edit_buffer_get_byte (buf, p + 2) == '.');
 107     }
 108     if (c == '-')
 109     {
 110         /* `---' is acceptable */
 111         return !(edit_buffer_get_byte (buf, p + 1) == '-'
 112                  && edit_buffer_get_byte (buf, p + 2) == '-');
 113     }
 114
 115     return (option_stop_format_chars != NULL && strchr (option_stop_format_chars, c) != NULL);
 116 }
 117
 118 /* --------------------------------------------------------------------------------------------- */
 119 /**
 120  * Find the start of the current paragraph for the purpose of formatting.
 121  * Return position in the file.
 122  */
 123
 124 static off_t
 125 begin_paragraph (WEdit * edit, gboolean force, long *lines)
 126 {
 127     long i;
 128
 129     for (i = edit->buffer.curs_line - 1; i >= 0; i--)
 130         if (edit_line_is_blank (edit, i) ||
 131             (force && bad_line_start (&edit->buffer, line_start (&edit->buffer, i))))
 132         {
 133             i++;
 134             break;
 135         }
 136
 137     *lines = edit->buffer.curs_line - i;
 138
 139     return edit_buffer_get_backward_offset (&edit->buffer,
 140                                             edit_buffer_get_current_bol (&edit->buffer), *lines);
 141 }
 142
 143 /* --------------------------------------------------------------------------------------------- */
 144 /**
 145  * Find the end of the current paragraph for the purpose of formatting.
 146  * Return position in the file.
 147  */
 148
 149 static off_t
 150 end_paragraph (WEdit * edit, gboolean force)
 151 {
 152     long i;
 153
 154     for (i = edit->buffer.curs_line + 1; i <= edit->buffer.lines; i++)
 155         if (edit_line_is_blank (edit, i) ||
 156             (force && bad_line_start (&edit->buffer, line_start (&edit->buffer, i))))
 157         {
 158             i--;
 159             break;
 160         }
 161
 162     return edit_buffer_get_eol (&edit->buffer,
 163                                 edit_buffer_get_forward_offset (&edit->buffer,
 164                                                                 edit_buffer_get_current_bol
 165                                                                 (&edit->buffer),
 166                                                                 i - edit->buffer.curs_line, 0));
 167 }
 168
 169 /* --------------------------------------------------------------------------------------------- */
 170
 171 static GString *
 172 get_paragraph (const edit_buffer_t * buf, off_t p, off_t q, gboolean indent)
 173 {
 174     GString *t;
 175
 176     t = g_string_sized_new (128);
 177
 178     for (; p < q; p++)
 179     {
 180         if (indent && edit_buffer_get_byte (buf, p - 1) == '\n')
 181             while (strchr ("\t ", edit_buffer_get_byte (buf, p)) != NULL)
 182                 p++;
 183
 184         g_string_append_c (t, edit_buffer_get_byte (buf, p));
 185     }
 186
 187     g_string_append_c (t, '\n');
 188
 189     return t;
 190 }
 191
 192 /* --------------------------------------------------------------------------------------------- */
 193
 194 static inline void
 195 strip_newlines (unsigned char *t, off_t size)
 196 {
 197     unsigned char *p;
 198
 199     for (p = t; size-- != 0; p++)
 200         if (*p == '\n')
 201             *p = ' ';
 202 }
 203
 204 /* --------------------------------------------------------------------------------------------- */
 205 /**
 206    This function calculates the number of chars in a line specified to length l in pixels
 207  */
 208
 209 static inline off_t
 210 next_tab_pos (off_t x)
 211 {
 212     x += tab_width - x % tab_width;
 213     return x;
 214 }
 215
 216 /* --------------------------------------------------------------------------------------------- */
 217
 218 static inline off_t
 219 line_pixel_length (unsigned char *t, off_t b, off_t l, gboolean utf8)
 220 {
 221     off_t xn, x;                /* position conters */
 222     off_t char_length;          /* character length in bytes */
 223
 224 #ifndef HAVE_CHARSET
 225     (void) utf8;
 226 #endif
 227
 228     for (xn = 0, x = 0; xn <= l; x = xn, b += char_length)
 229     {
 230         char *tb;
 231
 232         tb = (char *) t + b;
 233         char_length = 1;
 234
 235         switch (tb[0])
 236         {
 237         case '\n':
 238             return b;
 239         case '\t':
 240             xn = next_tab_pos (x);
 241             break;
 242         default:
 243 #ifdef HAVE_CHARSET
 244             if (utf8)
 245             {
 246                 gunichar ch;
 247
 248                 ch = g_utf8_get_char_validated (tb, -1);
 249                 if (ch != (gunichar) (-2) && ch != (gunichar) (-1))
 250                 {
 251                     char *next_ch;
 252
 253                     /* Calculate UTF-8 char length */
 254                     next_ch = g_utf8_next_char (tb);
 255                     char_length = next_ch - tb;
 256
 257                     if (g_unichar_iswide (ch))
 258                         x++;
 259                 }
 260             }
 261 #endif
 262
 263             xn = x + 1;
 264             break;
 265         }
 266     }
 267
 268     return b;
 269 }
 270
 271 /* --------------------------------------------------------------------------------------------- */
 272
 273 static off_t
 274 next_word_start (unsigned char *t, off_t q, off_t size)
 275 {
 276     off_t i;
 277     gboolean saw_ws = FALSE;
 278
 279     for (i = q; i < size; i++)
 280     {
 281         switch (t[i])
 282         {
 283         case '\n':
 284             return -1;
 285         case '\t':
 286         case ' ':
 287             saw_ws = TRUE;
 288             break;
 289         default:
 290             if (saw_ws)
 291                 return i;
 292             break;
 293         }
 294     }
 295     return (-1);
 296 }
 297
 298 /* --------------------------------------------------------------------------------------------- */
 299 /** find the start of a word */
 300
 301 static inline int
 302 word_start (unsigned char *t, off_t q, off_t size)
 303 {
 304     off_t i;
 305
 306     if (t[q] == ' ' || t[q] == '\t')
 307         return next_word_start (t, q, size);
 308
 309     for (i = q;; i--)
 310     {
 311         unsigned char c;
 312
 313         if (i == 0)
 314             return (-1);
 315         c = t[i - 1];
 316         if (c == '\n')
 317             return (-1);
 318         if (c == ' ' || c == '\t')
 319             return i;
 320     }
 321 }
 322
 323 /* --------------------------------------------------------------------------------------------- */
 324 /** replaces ' ' with '\n' to properly format a paragraph */
 325
 326 static inline void
 327 format_this (unsigned char *t, off_t size, long indent, gboolean utf8)
 328 {
 329     off_t q = 0, ww;
 330
 331     strip_newlines (t, size);
 332     ww = option_word_wrap_line_length * FONT_MEAN_WIDTH - indent;
 333     if (ww < FONT_MEAN_WIDTH * 2)
 334         ww = FONT_MEAN_WIDTH * 2;
 335
 336     while (TRUE)
 337     {
 338         off_t p;
 339
 340         q = line_pixel_length (t, q, ww, utf8);
 341         if (q > size)
 342             break;
 343         if (t[q] == '\n')
 344             break;
 345         p = word_start (t, q, size);
 346         if (p == -1)
 347             q = next_word_start (t, q, size);   /* Return the end of the word if the beginning
 348                                                    of the word is at the beginning of a line
 349                                                    (i.e. a very long word) */
 350         else
 351             q = p;
 352         if (q == -1)            /* end of paragraph */
 353             break;
 354         if (q != 0)
 355             t[q - 1] = '\n';
 356     }
 357 }
 358
 359 /* --------------------------------------------------------------------------------------------- */
 360
 361 static inline void
 362 replace_at (WEdit * edit, off_t q, int c)
 363 {
 364     edit_cursor_move (edit, q - edit->buffer.curs1);
 365     edit_delete (edit, TRUE);
 366     edit_insert_ahead (edit, c);
 367 }
 368
 369 /* --------------------------------------------------------------------------------------------- */
 370
 371 static long
 372 edit_indent_width (const WEdit * edit, off_t p)
 373 {
 374     off_t q = p;
 375
 376     /* move to the end of the leading whitespace of the line */
 377     while (strchr ("\t ", edit_buffer_get_byte (&edit->buffer, q)) != NULL
 378            && q < edit->buffer.size - 1)
 379         q++;
 380     /* count the number of columns of indentation */
 381     return (long) edit_move_forward3 (edit, p, 0, q);
 382 }
 383
 384 /* --------------------------------------------------------------------------------------------- */
 385
 386 static void
 387 edit_insert_indent (WEdit * edit, long indent)
 388 {
 389     if (!option_fill_tabs_with_spaces)
 390         while (indent >= TAB_SIZE)
 391         {
 392             edit_insert (edit, '\t');
 393             indent -= TAB_SIZE;
 394         }
 395
 396     while (indent-- > 0)
 397         edit_insert (edit, ' ');
 398 }
 399
 400 /* --------------------------------------------------------------------------------------------- */
 401 /** replaces a block of text */
 402
 403 static inline void
 404 put_paragraph (WEdit * edit, unsigned char *t, off_t p, long indent, off_t size)
 405 {
 406     off_t cursor;
 407     off_t i;
 408     int c = '\0';
 409
 410     cursor = edit->buffer.curs1;
 411     if (indent != 0)
 412         while (strchr ("\t ", edit_buffer_get_byte (&edit->buffer, p)) != NULL)
 413             p++;
 414     for (i = 0; i < size; i++, p++)
 415     {
 416         if (i != 0 && indent != 0)
 417         {
 418             if (t[i - 1] == '\n' && c == '\n')
 419             {
 420                 while (strchr ("\t ", edit_buffer_get_byte (&edit->buffer, p)) != NULL)
 421                     p++;
 422             }
 423             else if (t[i - 1] == '\n')
 424             {
 425                 off_t curs;
 426
 427                 edit_cursor_move (edit, p - edit->buffer.curs1);
 428                 curs = edit->buffer.curs1;
 429                 edit_insert_indent (edit, indent);
 430                 if (cursor >= curs)
 431                     cursor += edit->buffer.curs1 - p;
 432                 p = edit->buffer.curs1;
 433             }
 434             else if (c == '\n')
 435             {
 436                 edit_cursor_move (edit, p - edit->buffer.curs1);
 437                 while (strchr ("\t ", edit_buffer_get_byte (&edit->buffer, p)) != NULL)
 438                 {
 439                     edit_delete (edit, TRUE);
 440                     if (cursor > edit->buffer.curs1)
 441                         cursor--;
 442                 }
 443                 p = edit->buffer.curs1;
 444             }
 445         }
 446
 447         c = edit_buffer_get_byte (&edit->buffer, p);
 448         if (c != t[i])
 449             replace_at (edit, p, t[i]);
 450     }
 451 }
 452
 453 /* --------------------------------------------------------------------------------------------- */
 454
 455 static inline long
 456 test_indent (const WEdit * edit, off_t p, off_t q)
 457 {
 458     long indent;
 459
 460     indent = edit_indent_width (edit, p++);
 461     if (indent == 0)
 462         return 0;
 463
 464     for (; p < q; p++)
 465         if (edit_buffer_get_byte (&edit->buffer, p - 1) == '\n'
 466             && indent != edit_indent_width (edit, p))
 467             return 0;
 468     return indent;
 469 }
 470
 471 /* --------------------------------------------------------------------------------------------- */
 472 /*** public functions ****************************************************************************/
 473 /* --------------------------------------------------------------------------------------------- */
 474
 475 void
 476 format_paragraph (WEdit * edit, gboolean force)
 477 {
 478     off_t p, q;
 479     long lines;
 480     off_t size;
 481     GString *t;
 482     long indent;
 483     unsigned char *t2;
 484     gboolean utf8 = FALSE;
 485
 486     if (option_word_wrap_line_length < 2)
 487         return;
 488     if (edit_line_is_blank (edit, edit->buffer.curs_line))
 489         return;
 490
 491     p = begin_paragraph (edit, force, &lines);
 492     q = end_paragraph (edit, force);
 493     indent = test_indent (edit, p, q);
 494
 495     t = get_paragraph (&edit->buffer, p, q, indent != 0);
 496     size = t->len - 1;
 497
 498     if (!force)
 499     {
 500         off_t i;
 501         char *stop_format_chars;
 502
 503         if (option_stop_format_chars != NULL
 504             && strchr (option_stop_format_chars, t->str[0]) != NULL)
 505         {
 506             g_string_free (t, TRUE);
 507             return;
 508         }
 509
 510         if (option_stop_format_chars == NULL || *option_stop_format_chars == '\0')
 511             stop_format_chars = g_strdup ("\t");
 512         else
 513             stop_format_chars = g_strconcat (option_stop_format_chars, "\t", (char *) NULL);
 514
 515         for (i = 0; i < size - 1; i++)
 516             if (t->str[i] == '\n' && strchr (stop_format_chars, t->str[i + 1]) != NULL)
 517             {
 518                 g_free (stop_format_chars);
 519                 g_string_free (t, TRUE);
 520                 return;
 521             }
 522
 523         g_free (stop_format_chars);
 524     }
 525
 526     t2 = (unsigned char *) g_string_free (t, FALSE);
 527 #ifdef HAVE_CHARSET
 528     utf8 = edit->utf8;
 529 #endif
 530     /* scroll up to show 1st line of paragraph */
 531     edit_move_up (edit, lines, TRUE);
 532     /* scroll left as much as possible to show the formatted paragraph */
 533     edit_scroll_left (edit, -edit->start_col);
 534
 535     format_this (t2, q - p, indent, utf8);
 536     put_paragraph (edit, t2, p, indent, size);
 537     g_free ((char *) t2);
 538
 539     /* move to the end of paragraph */
 540     q = end_paragraph (edit, force);
 541     edit_cursor_move (edit, q - edit->buffer.curs1);
 542
 543     /* try move to the start of next paragraph */
 544     if (edit->buffer.curs_line < edit->buffer.lines)
 545     {
 546         edit_execute_cmd (edit, CK_Home, -1);
 547
 548         do
 549         {
 550             edit_execute_cmd (edit, CK_Down, -1);
 551         }
 552         while (edit->buffer.curs_line < edit->buffer.lines
 553                && edit_line_is_blank (edit, edit->buffer.curs_line));
 554     }
 555 }
 556
 557 /* --------------------------------------------------------------------------------------------- */