gas/input-scrub.c

   1 /* input_scrub.c - Break up input buffers into whole numbers of lines.
   2    Copyright 1987, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
   3    2000, 2001, 2003, 2006, 2007
   4    Free Software Foundation, Inc.
   5
   6    This file is part of GAS, the GNU Assembler.
   7
   8    GAS is free software; you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation; either version 3, or (at your option)
  11    any later version.
  12
  13    GAS is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with GAS; see the file COPYING.  If not, write to the Free
  20    Software Foundation, 51 Franklin Street - Fifth Floor, Boston, MA
  21    02110-1301, USA.  */
  22
  23 #include "as.h"
  24 #include "input-file.h"
  25 #include "sb.h"
  26 #include "listing.h"
  27
  28 /*
  29  * O/S independent module to supply buffers of sanitised source code
  30  * to rest of assembler.  We get sanitised input data of arbitrary length.
  31  * We break these buffers on line boundaries, recombine pieces that
  32  * were broken across buffers, and return a buffer of full lines to
  33  * the caller.
  34  * The last partial line begins the next buffer we build and return to caller.
  35  * The buffer returned to caller is preceded by BEFORE_STRING and followed
  36  * by AFTER_STRING, as sentinels. The last character before AFTER_STRING
  37  * is a newline.
  38  * Also looks after line numbers, for e.g. error messages.
  39  */
  40
  41 /*
  42  * We don't care how filthy our buffers are, but our callers assume
  43  * that the following sanitation has already been done.
  44  *
  45  * No comments, reduce a comment to a space.
  46  * Reduce a tab to a space unless it is 1st char of line.
  47  * All multiple tabs and spaces collapsed into 1 char. Tab only
  48  *   legal if 1st char of line.
  49  * # line file statements converted to .line x;.file y; statements.
  50  * Escaped newlines at end of line: remove them but add as many newlines
  51  *   to end of statement as you removed in the middle, to synch line numbers.
  52  */
  53 \f
  54 #define BEFORE_STRING ("\n")
  55 #define AFTER_STRING ("\0")     /* memcpy of 0 chars might choke.  */
  56 #define BEFORE_SIZE (1)
  57 #define AFTER_SIZE  (1)
  58
  59 #ifndef TC_EOL_IN_INSN
  60 #define TC_EOL_IN_INSN(P) 0
  61 #endif
  62
  63 static char *buffer_start;      /*->1st char of full buffer area.  */
  64 static char *partial_where;     /*->after last full line in buffer.  */
  65 static int partial_size;        /* >=0. Number of chars in partial line in buffer.  */
  66
  67 /* Because we need AFTER_STRING just after last full line, it clobbers
  68    1st part of partial line. So we preserve 1st part of partial line
  69    here.  */
  70 static char save_source[AFTER_SIZE];
  71
  72 /* What is the largest size buffer that input_file_give_next_buffer()
  73    could return to us?  */
  74 static unsigned int buffer_length;
  75
  76 /* The index into an sb structure we are reading from.  -1 if none.  */
  77 static int sb_index = -1;
  78
  79 /* If we are reading from an sb structure, this is it.  */
  80 static sb from_sb;
  81
  82 /* Should we do a conditional check on from_sb? */
  83 static int from_sb_is_expansion = 1;
  84
  85 /* The number of nested sb structures we have included.  */
  86 int macro_nest;
  87
  88 /* We can have more than one source file open at once, though the info for all
  89    but the latest one are saved off in a struct input_save.  These files remain
  90    open, so we are limited by the number of open files allowed by the
  91    underlying OS. We may also sequentially read more than one source file in an
  92    assembly.  */
  93
  94 /* We must track the physical file and line number for error messages. We also
  95    track a "logical" file and line number corresponding to (C?)  compiler
  96    source line numbers.  Whenever we open a file we must fill in
  97    physical_input_file. So if it is NULL we have not opened any files yet.  */
  98
  99 static char *physical_input_file;
 100 static char *logical_input_file;
 101
 102 typedef unsigned int line_numberT;      /* 1-origin line number in a source file.  */
 103 /* A line ends in '\n' or eof.  */
 104
 105 static line_numberT physical_input_line;
 106 static int logical_input_line;
 107
 108 /* Struct used to save the state of the input handler during include files */
 109 struct input_save {
 110   char *              buffer_start;
 111   char *              partial_where;
 112   int                 partial_size;
 113   char                save_source[AFTER_SIZE];
 114   unsigned int        buffer_length;
 115   char *              physical_input_file;
 116   char *              logical_input_file;
 117   line_numberT        physical_input_line;
 118   int                 logical_input_line;
 119   int                 sb_index;
 120   sb                  from_sb;
 121   int                 from_sb_is_expansion; /* Should we do a conditional check?  */
 122   struct input_save * next_saved_file;  /* Chain of input_saves.  */
 123   char *              input_file_save;  /* Saved state of input routines.  */
 124   char *              saved_position;   /* Caller's saved position in buf.  */
 125 };
 126
 127 static struct input_save *input_scrub_push (char *saved_position);
 128 static char *input_scrub_pop (struct input_save *arg);
 129
 130 /* Saved information about the file that .include'd this one.  When we hit EOF,
 131    we automatically pop to that file.  */
 132
 133 static struct input_save *next_saved_file;
 134
 135 /* Push the state of input reading and scrubbing so that we can #include.
 136    The return value is a 'void *' (fudged for old compilers) to a save
 137    area, which can be restored by passing it to input_scrub_pop().  */
 138
 139 static struct input_save *
 140 input_scrub_push (char *saved_position)
 141 {
 142   register struct input_save *saved;
 143
 144   saved = (struct input_save *) xmalloc (sizeof *saved);
 145
 146   saved->saved_position = saved_position;
 147   saved->buffer_start = buffer_start;
 148   saved->partial_where = partial_where;
 149   saved->partial_size = partial_size;
 150   saved->buffer_length = buffer_length;
 151   saved->physical_input_file = physical_input_file;
 152   saved->logical_input_file = logical_input_file;
 153   saved->physical_input_line = physical_input_line;
 154   saved->logical_input_line = logical_input_line;
 155   saved->sb_index = sb_index;
 156   saved->from_sb = from_sb;
 157   saved->from_sb_is_expansion = from_sb_is_expansion;
 158   memcpy (saved->save_source, save_source, sizeof (save_source));
 159   saved->next_saved_file = next_saved_file;
 160   saved->input_file_save = input_file_push ();
 161
 162   input_file_begin ();          /* Reinitialize! */
 163   logical_input_line = -1;
 164   logical_input_file = (char *) NULL;
 165   buffer_length = input_file_buffer_size ();
 166   sb_index = -1;
 167
 168   buffer_start = xmalloc ((BEFORE_SIZE + buffer_length + buffer_length + AFTER_SIZE));
 169   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 170
 171   return saved;
 172 }
 173
 174 static char *
 175 input_scrub_pop (struct input_save *saved)
 176 {
 177   char *saved_position;
 178
 179   input_scrub_end ();           /* Finish off old buffer */
 180
 181   input_file_pop (saved->input_file_save);
 182   saved_position = saved->saved_position;
 183   buffer_start = saved->buffer_start;
 184   buffer_length = saved->buffer_length;
 185   physical_input_file = saved->physical_input_file;
 186   logical_input_file = saved->logical_input_file;
 187   physical_input_line = saved->physical_input_line;
 188   logical_input_line = saved->logical_input_line;
 189   sb_index = saved->sb_index;
 190   from_sb = saved->from_sb;
 191   from_sb_is_expansion = saved->from_sb_is_expansion;
 192   partial_where = saved->partial_where;
 193   partial_size = saved->partial_size;
 194   next_saved_file = saved->next_saved_file;
 195   memcpy (save_source, saved->save_source, sizeof (save_source));
 196
 197   free (saved);
 198   return saved_position;
 199 }
 200 \f
 201 void
 202 input_scrub_begin (void)
 203 {
 204   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 205   know (strlen (AFTER_STRING) == AFTER_SIZE
 206         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 207
 208   input_file_begin ();
 209
 210   buffer_length = input_file_buffer_size ();
 211
 212   buffer_start = xmalloc ((BEFORE_SIZE + buffer_length + buffer_length + AFTER_SIZE));
 213   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 214
 215   /* Line number things.  */
 216   logical_input_line = -1;
 217   logical_input_file = (char *) NULL;
 218   physical_input_file = NULL;   /* No file read yet.  */
 219   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 220   do_scrub_begin (flag_m68k_mri);
 221 }
 222
 223 void
 224 input_scrub_end (void)
 225 {
 226   if (buffer_start)
 227     {
 228       free (buffer_start);
 229       buffer_start = 0;
 230       input_file_end ();
 231     }
 232 }
 233
 234 /* Start reading input from a new file.
 235    Return start of caller's part of buffer.  */
 236
 237 char *
 238 input_scrub_new_file (char *filename)
 239 {
 240   input_file_open (filename, !flag_no_comments);
 241   physical_input_file = filename[0] ? filename : _("{standard input}");
 242   physical_input_line = 0;
 243
 244   partial_size = 0;
 245   return (buffer_start + BEFORE_SIZE);
 246 }
 247
 248 /* Include a file from the current file.  Save our state, cause it to
 249    be restored on EOF, and begin handling a new file.  Same result as
 250    input_scrub_new_file.  */
 251
 252 char *
 253 input_scrub_include_file (char *filename, char *position)
 254 {
 255   next_saved_file = input_scrub_push (position);
 256   return input_scrub_new_file (filename);
 257 }
 258
 259 /* Start getting input from an sb structure.  This is used when
 260    expanding a macro.  */
 261
 262 void
 263 input_scrub_include_sb (sb *from, char *position, int is_expansion)
 264 {
 265   if (macro_nest > max_macro_nest)
 266     as_fatal (_("macros nested too deeply"));
 267   ++macro_nest;
 268
 269 #ifdef md_macro_start
 270   if (is_expansion)
 271     {
 272       md_macro_start ();
 273     }
 274 #endif
 275
 276   next_saved_file = input_scrub_push (position);
 277
 278   sb_new (&from_sb);
 279   from_sb_is_expansion = is_expansion;
 280   if (from->len >= 1 && from->ptr[0] != '\n')
 281     {
 282       /* Add the sentinel required by read.c.  */
 283       sb_add_char (&from_sb, '\n');
 284     }
 285   sb_scrub_and_add_sb (&from_sb, from);
 286
 287   /* Make sure the parser looks at defined contents when it scans for
 288      e.g. end-of-line at the end of a macro.  */
 289   sb_add_char (&from_sb, 0);
 290   from_sb.len--;
 291
 292   sb_index = 1;
 293
 294   /* These variables are reset by input_scrub_push.  Restore them
 295      since we are, after all, still at the same point in the file.  */
 296   logical_input_line = next_saved_file->logical_input_line;
 297   logical_input_file = next_saved_file->logical_input_file;
 298 }
 299
 300 void
 301 input_scrub_close (void)
 302 {
 303   input_file_close ();
 304 }
 305
 306 char *
 307 input_scrub_next_buffer (char **bufp)
 308 {
 309   register char *limit;         /*->just after last char of buffer.  */
 310
 311   if (sb_index >= 0)
 312     {
 313       if (sb_index >= from_sb.len)
 314         {
 315           sb_kill (&from_sb);
 316           if (from_sb_is_expansion
 317               )
 318             {
 319               cond_finish_check (macro_nest);
 320 #ifdef md_macro_end
 321               /* Allow the target to clean up per-macro expansion
 322                  data.  */
 323               md_macro_end ();
 324 #endif
 325             }
 326           --macro_nest;
 327           partial_where = NULL;
 328           if (next_saved_file != NULL)
 329             *bufp = input_scrub_pop (next_saved_file);
 330           return partial_where;
 331         }
 332
 333       partial_where = from_sb.ptr + from_sb.len;
 334       partial_size = 0;
 335       *bufp = from_sb.ptr + sb_index;
 336       sb_index = from_sb.len;
 337       return partial_where;
 338     }
 339
 340   *bufp = buffer_start + BEFORE_SIZE;
 341
 342   if (partial_size)
 343     {
 344       memcpy (buffer_start + BEFORE_SIZE, partial_where,
 345               (unsigned int) partial_size);
 346       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 347     }
 348   limit = input_file_give_next_buffer (buffer_start
 349                                        + BEFORE_SIZE
 350                                        + partial_size);
 351   if (limit)
 352     {
 353       register char *p;         /* Find last newline.  */
 354       /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 355       *limit = '\0';
 356       for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 357         ;
 358       ++p;
 359
 360       while (p <= buffer_start + BEFORE_SIZE)
 361         {
 362           int limoff;
 363
 364           limoff = limit - buffer_start;
 365           buffer_length += input_file_buffer_size ();
 366           buffer_start = xrealloc (buffer_start,
 367                                    (BEFORE_SIZE
 368                                     + 2 * buffer_length
 369                                     + AFTER_SIZE));
 370           *bufp = buffer_start + BEFORE_SIZE;
 371           limit = input_file_give_next_buffer (buffer_start + limoff);
 372
 373           if (limit == NULL)
 374             {
 375               as_warn (_("partial line at end of file ignored"));
 376               partial_where = NULL;
 377               if (next_saved_file)
 378                 *bufp = input_scrub_pop (next_saved_file);
 379               return NULL;
 380             }
 381
 382           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 383           *limit = '\0';
 384           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 385             ;
 386           ++p;
 387         }
 388
 389       partial_where = p;
 390       partial_size = limit - p;
 391       memcpy (save_source, partial_where, (int) AFTER_SIZE);
 392       memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 393     }
 394   else
 395     {
 396       partial_where = 0;
 397       if (partial_size > 0)
 398         {
 399           as_warn (_("partial line at end of file ignored"));
 400         }
 401
 402       /* Tell the listing we've finished the file.  */
 403       LISTING_EOF ();
 404
 405       /* If we should pop to another file at EOF, do it.  */
 406       if (next_saved_file)
 407         {
 408           *bufp = input_scrub_pop (next_saved_file);    /* Pop state */
 409           /* partial_where is now correct to return, since we popped it.  */
 410         }
 411     }
 412   return (partial_where);
 413 }
 414 \f
 415 /* The remaining part of this file deals with line numbers, error
 416    messages and so on.  Return TRUE if we opened any file.  */
 417
 418 int
 419 seen_at_least_1_file (void)
 420 {
 421   return (physical_input_file != NULL);
 422 }
 423
 424 void
 425 bump_line_counters (void)
 426 {
 427   if (sb_index < 0)
 428     {
 429       ++physical_input_line;
 430       if (logical_input_line >= 0)
 431         ++logical_input_line;
 432     }
 433 }
 434 \f
 435 /* Tells us what the new logical line number and file are.
 436    If the line_number is -1, we don't change the current logical line
 437    number.  If it is -2, we decrement the logical line number (this is
 438    to support the .appfile pseudo-op inserted into the stream by
 439    do_scrub_chars).
 440    If the fname is NULL, we don't change the current logical file name.
 441    Returns nonzero if the filename actually changes.  */
 442
 443 int
 444 new_logical_line_flags (char *fname, /* DON'T destroy it!  We point to it!  */
 445                         int line_number,
 446                         int flags)
 447 {
 448   switch (flags)
 449     {
 450     case 0:
 451       break;
 452     case 1:
 453       if (line_number != -1)
 454         abort ();
 455       break;
 456     case 1 << 1:
 457     case 1 << 2:
 458       /* FIXME: we could check that include nesting is correct.  */
 459       break;
 460     default:
 461       abort ();
 462     }
 463
 464   if (line_number >= 0)
 465     logical_input_line = line_number;
 466   else if (line_number == -1 && fname && !*fname && (flags & (1 << 2)))
 467     {
 468       logical_input_file = physical_input_file;
 469       logical_input_line = physical_input_line;
 470       fname = NULL;
 471     }
 472
 473   if (fname
 474       && (logical_input_file == NULL
 475           || strcmp (logical_input_file, fname)))
 476     {
 477       logical_input_file = fname;
 478       return 1;
 479     }
 480   else
 481     return 0;
 482 }
 483
 484 int
 485 new_logical_line (char *fname, int line_number)
 486 {
 487   return new_logical_line_flags (fname, line_number, 0);
 488 }
 489
 490 \f
 491 /* Return the current file name and line number.
 492    namep should be char * const *, but there are compilers which screw
 493    up declarations like that, and it's easier to avoid it.  */
 494
 495 void
 496 as_where (char **namep, unsigned int *linep)
 497 {
 498   if (logical_input_file != NULL
 499       && (linep == NULL || logical_input_line >= 0))
 500     {
 501       *namep = logical_input_file;
 502       if (linep != NULL)
 503         *linep = logical_input_line;
 504     }
 505   else if (physical_input_file != NULL)
 506     {
 507       *namep = physical_input_file;
 508       if (linep != NULL)
 509         *linep = physical_input_line;
 510     }
 511   else
 512     {
 513       *namep = 0;
 514       if (linep != NULL)
 515         *linep = 0;
 516     }
 517 }