gas/expr.c

   1 /* expr.c -operands, expressions-
   2    Copyright (C) 1987, 1990, 1991, 1992, 1993 Free Software Foundation, Inc.
   3
   4    This file is part of GAS, the GNU Assembler.
   5
   6    GAS is free software; you can redistribute it and/or modify
   7    it under the terms of the GNU General Public License as published by
   8    the Free Software Foundation; either version 2, or (at your option)
   9    any later version.
  10
  11    GAS is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14    GNU General Public License for more details.
  15
  16    You should have received a copy of the GNU General Public License
  17    along with GAS; see the file COPYING.  If not, write to
  18    the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
  19
  20 /*
  21  * This is really a branch office of as-read.c. I split it out to clearly
  22  * distinguish the world of expressions from the world of statements.
  23  * (It also gives smaller files to re-compile.)
  24  * Here, "operand"s are of expressions, not instructions.
  25  */
  26
  27 #include <ctype.h>
  28 #include <string.h>
  29
  30 #include "as.h"
  31
  32 #include "obstack.h"
  33
  34 static void floating_constant PARAMS ((expressionS * expressionP));
  35 static void integer_constant PARAMS ((int radix, expressionS * expressionP));
  36 static void clean_up_expression PARAMS ((expressionS * expressionP));
  37 static symbolS *make_expr_symbol PARAMS ((expressionS * expressionP));
  38
  39 extern const char EXP_CHARS[], FLT_CHARS[];
  40 \f
  41 /* Build a dummy symbol to hold a complex expression.  This is how we
  42    build expressions up out of other expressions.  The symbol is put
  43    into the fake section expr_section.  */
  44
  45 static symbolS *
  46 make_expr_symbol (expressionP)
  47      expressionS *expressionP;
  48 {
  49   const char *fake;
  50   symbolS *symbolP;
  51
  52   /* FIXME: This should be something which decode_local_label_name
  53      will handle.  */
  54   fake = FAKE_LABEL_NAME;
  55
  56   /* Putting constant symbols in absolute_section rather than
  57      expr_section is convenient for the old a.out code, for which
  58      S_GET_SEGMENT does not always retrieve the value put in by
  59      S_SET_SEGMENT.  */
  60   symbolP = symbol_new (fake,
  61                         (expressionP->X_op == O_constant
  62                          ? absolute_section
  63                          : expr_section),
  64                         0, &zero_address_frag);
  65   symbolP->sy_value = *expressionP;
  66   return symbolP;
  67 }
  68 \f
  69 /*
  70  * Build any floating-point literal here.
  71  * Also build any bignum literal here.
  72  */
  73
  74 /* Seems atof_machine can backscan through generic_bignum and hit whatever
  75    happens to be loaded before it in memory.  And its way too complicated
  76    for me to fix right.  Thus a hack.  JF:  Just make generic_bignum bigger,
  77    and never write into the early words, thus they'll always be zero.
  78    I hate Dean's floating-point code.  Bleh.  */
  79 LITTLENUM_TYPE generic_bignum[SIZE_OF_LARGE_NUMBER + 6];
  80 FLONUM_TYPE generic_floating_point_number =
  81 {
  82   &generic_bignum[6],           /* low (JF: Was 0) */
  83   &generic_bignum[SIZE_OF_LARGE_NUMBER + 6 - 1],        /* high JF: (added +6) */
  84   0,                            /* leader */
  85   0,                            /* exponent */
  86   0                             /* sign */
  87 };
  88 /* If nonzero, we've been asked to assemble nan, +inf or -inf */
  89 int generic_floating_point_magic;
  90 \f
  91 static void
  92 floating_constant (expressionP)
  93      expressionS *expressionP;
  94 {
  95   /* input_line_pointer->*/
  96   /* floating-point constant. */
  97   int error_code;
  98
  99   error_code = atof_generic
 100     (&input_line_pointer, ".", EXP_CHARS,
 101      &generic_floating_point_number);
 102
 103   if (error_code)
 104     {
 105       if (error_code == ERROR_EXPONENT_OVERFLOW)
 106         {
 107           as_bad ("bad floating-point constant: exponent overflow, probably assembling junk");
 108         }
 109       else
 110         {
 111           as_bad ("bad floating-point constant: unknown error code=%d.", error_code);
 112         }
 113     }
 114   expressionP->X_op = O_big;
 115   /* input_line_pointer->just after constant, */
 116   /* which may point to whitespace. */
 117   expressionP->X_add_number = -1;
 118 }
 119
 120 static void
 121 integer_constant (radix, expressionP)
 122      int radix;
 123      expressionS *expressionP;
 124 {
 125   char *digit_2;        /*->2nd digit of number. */
 126   char c;
 127
 128   valueT number;        /* offset or (absolute) value */
 129   short int digit;      /* value of next digit in current radix */
 130   short int maxdig = 0;/* highest permitted digit value. */
 131   int too_many_digits = 0;      /* if we see >= this number of */
 132   char *name;           /* points to name of symbol */
 133   symbolS *symbolP;     /* points to symbol */
 134
 135   int small;                    /* true if fits in 32 bits. */
 136   extern const char hex_value[]; /* in hex_value.c */
 137
 138   /* May be bignum, or may fit in 32 bits. */
 139   /* Most numbers fit into 32 bits, and we want this case to be fast.
 140      so we pretend it will fit into 32 bits.  If, after making up a 32
 141      bit number, we realise that we have scanned more digits than
 142      comfortably fit into 32 bits, we re-scan the digits coding them
 143      into a bignum.  For decimal and octal numbers we are
 144      conservative: Some numbers may be assumed bignums when in fact
 145      they do fit into 32 bits.  Numbers of any radix can have excess
 146      leading zeros: We strive to recognise this and cast them back
 147      into 32 bits.  We must check that the bignum really is more than
 148      32 bits, and change it back to a 32-bit number if it fits.  The
 149      number we are looking for is expected to be positive, but if it
 150      fits into 32 bits as an unsigned number, we let it be a 32-bit
 151      number.  The cavalier approach is for speed in ordinary cases. */
 152   /* This has been extended for 64 bits.  We blindly assume that if
 153      you're compiling in 64-bit mode, the target is a 64-bit machine.
 154      This should be cleaned up.  */
 155
 156 #ifdef BFD64
 157 #define valuesize 64
 158 #else /* includes non-bfd case, mostly */
 159 #define valuesize 32
 160 #endif
 161
 162   switch (radix)
 163     {
 164     case 2:
 165       maxdig = 2;
 166       too_many_digits = valuesize + 1;
 167       break;
 168     case 8:
 169       maxdig = radix = 8;
 170       too_many_digits = (valuesize + 2) / 3;
 171       break;
 172     case 16:
 173       maxdig = radix = 16;
 174       too_many_digits = (valuesize + 3) / 4;
 175       break;
 176     case 10:
 177       maxdig = radix = 10;
 178       too_many_digits = (valuesize + 12) / 4; /* very rough */
 179     }
 180 #undef valuesize
 181   c = *input_line_pointer;
 182   input_line_pointer++;
 183   digit_2 = input_line_pointer;
 184   for (number = 0;
 185        (digit = hex_value[(unsigned char) c]) < maxdig;
 186        c = *input_line_pointer++)
 187     {
 188       number = number * radix + digit;
 189     }
 190   /* c contains character after number. */
 191   /* input_line_pointer->char after c. */
 192   small = input_line_pointer - digit_2 < too_many_digits;
 193   if (!small)
 194     {
 195       /*
 196        * we saw a lot of digits. manufacture a bignum the hard way.
 197        */
 198       LITTLENUM_TYPE *leader;   /*->high order littlenum of the bignum. */
 199       LITTLENUM_TYPE *pointer;  /*->littlenum we are frobbing now. */
 200       long carry;
 201
 202       leader = generic_bignum;
 203       generic_bignum[0] = 0;
 204       generic_bignum[1] = 0;
 205       /* we could just use digit_2, but lets be mnemonic. */
 206       input_line_pointer = --digit_2;   /*->1st digit. */
 207       c = *input_line_pointer++;
 208       for (;
 209            (carry = hex_value[(unsigned char) c]) < maxdig;
 210            c = *input_line_pointer++)
 211         {
 212           for (pointer = generic_bignum;
 213                pointer <= leader;
 214                pointer++)
 215             {
 216               long work;
 217
 218               work = carry + radix * *pointer;
 219               *pointer = work & LITTLENUM_MASK;
 220               carry = work >> LITTLENUM_NUMBER_OF_BITS;
 221             }
 222           if (carry)
 223             {
 224               if (leader < generic_bignum + SIZE_OF_LARGE_NUMBER - 1)
 225                 {               /* room to grow a longer bignum. */
 226                   *++leader = carry;
 227                 }
 228             }
 229         }
 230       /* again, c is char after number, */
 231       /* input_line_pointer->after c. */
 232       know (LITTLENUM_NUMBER_OF_BITS == 16);
 233       if (leader < generic_bignum + sizeof (valueT) / 2)
 234         {                       /* will fit into 32 bits. */
 235           number =
 236             ((generic_bignum[1] & LITTLENUM_MASK) << LITTLENUM_NUMBER_OF_BITS)
 237             | (generic_bignum[0] & LITTLENUM_MASK);
 238           small = 1;
 239         }
 240       else
 241         {
 242           number = leader - generic_bignum + 1; /* number of littlenums in the bignum. */
 243         }
 244     }
 245   if (small)
 246     {
 247       /*
 248        * here with number, in correct radix. c is the next char.
 249        * note that unlike un*x, we allow "011f" "0x9f" to
 250        * both mean the same as the (conventional) "9f". this is simply easier
 251        * than checking for strict canonical form. syntax sux!
 252        */
 253
 254       switch (c)
 255         {
 256
 257 #ifdef LOCAL_LABELS_FB
 258         case 'b':
 259           {
 260             /*
 261              * backward ref to local label.
 262              * because it is backward, expect it to be defined.
 263              */
 264             /* Construct a local label.  */
 265             name = fb_label_name ((int) number, 0);
 266
 267             /* seen before, or symbol is defined: ok */
 268             symbolP = symbol_find (name);
 269             if ((symbolP != NULL) && (S_IS_DEFINED (symbolP)))
 270               {
 271
 272                 /* local labels are never absolute. don't waste time
 273                    checking absoluteness. */
 274                 know (SEG_NORMAL (S_GET_SEGMENT (symbolP)));
 275
 276                 expressionP->X_op = O_symbol;
 277                 expressionP->X_add_symbol = symbolP;
 278
 279               }
 280             else
 281               {
 282                 /* either not seen or not defined. */
 283                 /* @@ Should print out the original string instead of
 284                    the parsed number.  */
 285                 as_bad ("backw. ref to unknown label \"%d:\", 0 assumed.",
 286                         (int) number);
 287                 expressionP->X_op = O_constant;
 288               }
 289
 290             expressionP->X_add_number = 0;
 291             break;
 292           }                     /* case 'b' */
 293
 294         case 'f':
 295           {
 296             /*
 297              * forward reference. expect symbol to be undefined or
 298              * unknown. undefined: seen it before. unknown: never seen
 299              * it before.
 300              * construct a local label name, then an undefined symbol.
 301              * don't create a xseg frag for it: caller may do that.
 302              * just return it as never seen before.
 303              */
 304             name = fb_label_name ((int) number, 1);
 305             symbolP = symbol_find_or_make (name);
 306             /* we have no need to check symbol properties. */
 307 #ifndef many_segments
 308             /* since "know" puts its arg into a "string", we
 309                can't have newlines in the argument.  */
 310             know (S_GET_SEGMENT (symbolP) == undefined_section || S_GET_SEGMENT (symbolP) == text_section || S_GET_SEGMENT (symbolP) == data_section);
 311 #endif
 312             expressionP->X_op = O_symbol;
 313             expressionP->X_add_symbol = symbolP;
 314             expressionP->X_add_number = 0;
 315
 316             break;
 317           }                     /* case 'f' */
 318
 319 #endif /* LOCAL_LABELS_FB */
 320
 321 #ifdef LOCAL_LABELS_DOLLAR
 322
 323         case '$':
 324           {
 325
 326             /* If the dollar label is *currently* defined, then this is just
 327                another reference to it.  If it is not *currently* defined,
 328                then this is a fresh instantiation of that number, so create
 329                it.  */
 330
 331             if (dollar_label_defined ((long) number))
 332               {
 333                 name = dollar_label_name ((long) number, 0);
 334                 symbolP = symbol_find (name);
 335                 know (symbolP != NULL);
 336               }
 337             else
 338               {
 339                 name = dollar_label_name ((long) number, 1);
 340                 symbolP = symbol_find_or_make (name);
 341               }
 342
 343             expressionP->X_op = O_symbol;
 344             expressionP->X_add_symbol = symbolP;
 345             expressionP->X_add_number = 0;
 346
 347             break;
 348           }                     /* case '$' */
 349
 350 #endif /* LOCAL_LABELS_DOLLAR */
 351
 352         default:
 353           {
 354             expressionP->X_op = O_constant;
 355             expressionP->X_add_number = number;
 356             input_line_pointer--;       /* restore following character. */
 357             break;
 358           }                     /* really just a number */
 359
 360         }                       /* switch on char following the number */
 361
 362
 363     }
 364   else
 365     {
 366       /* not a small number */
 367       expressionP->X_op = O_big;
 368       expressionP->X_add_number = number;
 369       input_line_pointer--;     /*->char following number. */
 370     }
 371 }
 372
 373
 374 /*
 375  * Summary of operand().
 376  *
 377  * in:  Input_line_pointer points to 1st char of operand, which may
 378  *      be a space.
 379  *
 380  * out: A expressionS.
 381  *      The operand may have been empty: in this case X_op == O_absent.
 382  *      Input_line_pointer->(next non-blank) char after operand.
 383  */
 384
 385 static segT
 386 operand (expressionP)
 387      expressionS *expressionP;
 388 {
 389   char c;
 390   symbolS *symbolP;     /* points to symbol */
 391   char *name;           /* points to name of symbol */
 392   segT segment;
 393
 394   /* All integers are regarded as unsigned unless they are negated.
 395      This is because the only thing which cares whether a number is
 396      unsigned is the code in emit_expr which extends constants into
 397      bignums.  It should only sign extend negative numbers, so that
 398      something like ``.quad 0x80000000'' is not sign extended even
 399      though it appears negative if valueT is 32 bits.  */
 400   expressionP->X_unsigned = 1;
 401
 402   /* digits, assume it is a bignum. */
 403
 404   SKIP_WHITESPACE ();           /* leading whitespace is part of operand. */
 405   c = *input_line_pointer++;    /* input_line_pointer->past char in c. */
 406
 407   switch (c)
 408     {
 409 #ifdef MRI
 410     case '%':
 411       integer_constant (2, expressionP);
 412       break;
 413     case '@':
 414       integer_constant (8, expressionP);
 415       break;
 416     case '$':
 417       integer_constant (16, expressionP);
 418       break;
 419 #endif
 420     case '1':
 421     case '2':
 422     case '3':
 423     case '4':
 424     case '5':
 425     case '6':
 426     case '7':
 427     case '8':
 428     case '9':
 429       input_line_pointer--;
 430
 431       integer_constant (10, expressionP);
 432       break;
 433
 434     case '0':
 435       /* non-decimal radix */
 436
 437       c = *input_line_pointer;
 438       switch (c)
 439         {
 440
 441         default:
 442           if (c && strchr (FLT_CHARS, c))
 443             {
 444               input_line_pointer++;
 445               floating_constant (expressionP);
 446             }
 447           else
 448             {
 449               /* The string was only zero */
 450               expressionP->X_op = O_constant;
 451               expressionP->X_add_number = 0;
 452             }
 453
 454           break;
 455
 456         case 'x':
 457         case 'X':
 458           input_line_pointer++;
 459           integer_constant (16, expressionP);
 460           break;
 461
 462         case 'b':
 463 #ifdef LOCAL_LABELS_FB
 464           /* FIXME: This seems to be nonsense.  At this point we know
 465              for sure that *input_line_pointer is 'b'.  So why are we
 466              checking it?  What is this code supposed to do?  */
 467           if (!*input_line_pointer
 468               || (!strchr ("+-.0123456789", *input_line_pointer)
 469                   && !strchr (EXP_CHARS, *input_line_pointer)))
 470             {
 471               input_line_pointer--;
 472               integer_constant (10, expressionP);
 473               break;
 474             }
 475 #endif
 476         case 'B':
 477           input_line_pointer++;
 478           integer_constant (2, expressionP);
 479           break;
 480
 481         case '0':
 482         case '1':
 483         case '2':
 484         case '3':
 485         case '4':
 486         case '5':
 487         case '6':
 488         case '7':
 489           integer_constant (8, expressionP);
 490           break;
 491
 492         case 'f':
 493 #ifdef LOCAL_LABELS_FB
 494           /* if it says '0f' and the line ends or it doesn't look like
 495              a floating point #, its a local label ref.  dtrt */
 496           /* likewise for the b's.  xoxorich. */
 497           /* FIXME: As in the 'b' case, we know that the
 498              *input_line_pointer is 'f'.  What is this code really
 499              trying to do?  */
 500           if (c == 'f'
 501               && (!*input_line_pointer ||
 502                   (!strchr ("+-.0123456789", *input_line_pointer) &&
 503                    !strchr (EXP_CHARS, *input_line_pointer))))
 504             {
 505               input_line_pointer -= 1;
 506               integer_constant (10, expressionP);
 507               break;
 508             }
 509 #endif
 510
 511         case 'd':
 512         case 'D':
 513         case 'F':
 514         case 'r':
 515         case 'e':
 516         case 'E':
 517         case 'g':
 518         case 'G':
 519
 520           input_line_pointer++;
 521           floating_constant (expressionP);
 522           expressionP->X_add_number = -(isupper (c) ? tolower (c) : c);
 523           break;
 524
 525 #ifdef LOCAL_LABELS_DOLLAR
 526         case '$':
 527           integer_constant (10, expressionP);
 528           break;
 529 #endif
 530         }
 531
 532       break;
 533
 534     case '(':
 535       /* didn't begin with digit & not a name */
 536       segment = expression (expressionP);
 537       /* Expression() will pass trailing whitespace */
 538       if (*input_line_pointer++ != ')')
 539         {
 540           as_bad ("Missing ')' assumed");
 541           input_line_pointer--;
 542         }
 543       /* here with input_line_pointer->char after "(...)" */
 544       return segment;
 545
 546     case '\'':
 547       /* Warning: to conform to other people's assemblers NO ESCAPEMENT is
 548          permitted for a single quote. The next character, parity errors and
 549          all, is taken as the value of the operand. VERY KINKY.  */
 550       expressionP->X_op = O_constant;
 551       expressionP->X_add_number = *input_line_pointer++;
 552       break;
 553
 554     case '+':
 555       (void) operand (expressionP);
 556       break;
 557
 558     case '~':
 559     case '-':
 560       {
 561         operand (expressionP);
 562         if (expressionP->X_op == O_constant)
 563           {
 564             /* input_line_pointer -> char after operand */
 565             if (c == '-')
 566               {
 567                 expressionP->X_add_number = - expressionP->X_add_number;
 568                 /* Notice: '-' may overflow: no warning is given. This is
 569                    compatible with other people's assemblers. Sigh.  */
 570                 expressionP->X_unsigned = 0;
 571               }
 572             else
 573               expressionP->X_add_number = ~ expressionP->X_add_number;
 574           }
 575         else if (expressionP->X_op != O_illegal
 576                  && expressionP->X_op != O_absent)
 577           {
 578             expressionP->X_add_symbol = make_expr_symbol (expressionP);
 579             if (c == '-')
 580               expressionP->X_op = O_uminus;
 581             else
 582               expressionP->X_op = O_bit_not;
 583             expressionP->X_add_number = 0;
 584           }
 585         else
 586           as_warn ("Unary operator %c ignored because bad operand follows",
 587                    c);
 588       }
 589       break;
 590
 591     case '.':
 592       if (!is_part_of_name (*input_line_pointer))
 593         {
 594           const char *fake;
 595
 596           /* JF: '.' is pseudo symbol with value of current location
 597              in current segment.  */
 598           fake = FAKE_LABEL_NAME;
 599           symbolP = symbol_new (fake,
 600                                 now_seg,
 601                                 (valueT) frag_now_fix (),
 602                                 frag_now);
 603
 604           expressionP->X_op = O_symbol;
 605           expressionP->X_add_symbol = symbolP;
 606           expressionP->X_add_number = 0;
 607           break;
 608         }
 609       else
 610         {
 611           goto isname;
 612         }
 613     case ',':
 614     case '\n':
 615     case '\0':
 616     eol:
 617       /* can't imagine any other kind of operand */
 618       expressionP->X_op = O_absent;
 619       input_line_pointer--;
 620       md_operand (expressionP);
 621       break;
 622
 623     default:
 624       if (is_end_of_line[(unsigned char) c])
 625         goto eol;
 626       if (is_name_beginner (c)) /* here if did not begin with a digit */
 627         {
 628           /*
 629            * Identifier begins here.
 630            * This is kludged for speed, so code is repeated.
 631            */
 632         isname:
 633           name = --input_line_pointer;
 634           c = get_symbol_end ();
 635           symbolP = symbol_find_or_make (name);
 636
 637           /* If we have an absolute symbol or a reg, then we know its
 638              value now.  */
 639           segment = S_GET_SEGMENT (symbolP);
 640           if (segment == absolute_section)
 641             {
 642               expressionP->X_op = O_constant;
 643               expressionP->X_add_number = S_GET_VALUE (symbolP);
 644             }
 645           else if (segment == reg_section)
 646             {
 647               expressionP->X_op = O_register;
 648               expressionP->X_add_number = S_GET_VALUE (symbolP);
 649             }
 650           else
 651             {
 652               expressionP->X_op = O_symbol;
 653               expressionP->X_add_symbol = symbolP;
 654               expressionP->X_add_number = 0;
 655             }
 656           *input_line_pointer = c;
 657         }
 658       else
 659         {
 660           as_bad ("Bad expression");
 661           expressionP->X_op = O_constant;
 662           expressionP->X_add_number = 0;
 663         }
 664     }
 665
 666   /*
 667    * It is more 'efficient' to clean up the expressionS when they are created.
 668    * Doing it here saves lines of code.
 669    */
 670   clean_up_expression (expressionP);
 671   SKIP_WHITESPACE ();           /*->1st char after operand. */
 672   know (*input_line_pointer != ' ');
 673
 674   /* The PA port needs this information.  */
 675   if (expressionP->X_add_symbol)
 676     expressionP->X_add_symbol->sy_used = 1;
 677
 678   switch (expressionP->X_op)
 679     {
 680     default:
 681       return absolute_section;
 682     case O_symbol:
 683       return S_GET_SEGMENT (expressionP->X_add_symbol);
 684     case O_register:
 685       return reg_section;
 686     }
 687 }                               /* operand() */
 688 \f
 689 /* Internal. Simplify a struct expression for use by expr() */
 690
 691 /*
 692  * In:  address of a expressionS.
 693  *      The X_op field of the expressionS may only take certain values.
 694  *      Elsewise we waste time special-case testing. Sigh. Ditto SEG_ABSENT.
 695  * Out: expressionS may have been modified:
 696  *      'foo-foo' symbol references cancelled to 0,
 697  *              which changes X_op from O_subtract to O_constant.
 698  *      Unused fields zeroed to help expr().
 699  */
 700
 701 static void
 702 clean_up_expression (expressionP)
 703      expressionS *expressionP;
 704 {
 705   switch (expressionP->X_op)
 706     {
 707     case O_illegal:
 708     case O_absent:
 709       expressionP->X_add_number = 0;
 710       /* Fall through.  */
 711     case O_big:
 712     case O_constant:
 713     case O_register:
 714       expressionP->X_add_symbol = NULL;
 715       /* Fall through.  */
 716     case O_symbol:
 717     case O_uminus:
 718     case O_bit_not:
 719       expressionP->X_op_symbol = NULL;
 720       break;
 721     case O_subtract:
 722       if (expressionP->X_op_symbol == expressionP->X_add_symbol
 723           || ((expressionP->X_op_symbol->sy_frag
 724                == expressionP->X_add_symbol->sy_frag)
 725               && SEG_NORMAL (S_GET_SEGMENT (expressionP->X_add_symbol))
 726               && (S_GET_VALUE (expressionP->X_op_symbol)
 727                   == S_GET_VALUE (expressionP->X_add_symbol))))
 728         {
 729           expressionP->X_op = O_constant;
 730           expressionP->X_add_symbol = NULL;
 731           expressionP->X_op_symbol = NULL;
 732         }
 733       break;
 734     default:
 735       break;
 736     }
 737 }
 738 \f
 739 /* Expression parser. */
 740
 741 /*
 742  * We allow an empty expression, and just assume (absolute,0) silently.
 743  * Unary operators and parenthetical expressions are treated as operands.
 744  * As usual, Q==quantity==operand, O==operator, X==expression mnemonics.
 745  *
 746  * We used to do a aho/ullman shift-reduce parser, but the logic got so
 747  * warped that I flushed it and wrote a recursive-descent parser instead.
 748  * Now things are stable, would anybody like to write a fast parser?
 749  * Most expressions are either register (which does not even reach here)
 750  * or 1 symbol. Then "symbol+constant" and "symbol-symbol" are common.
 751  * So I guess it doesn't really matter how inefficient more complex expressions
 752  * are parsed.
 753  *
 754  * After expr(RANK,resultP) input_line_pointer->operator of rank <= RANK.
 755  * Also, we have consumed any leading or trailing spaces (operand does that)
 756  * and done all intervening operators.
 757  *
 758  * This returns the segment of the result, which will be
 759  * absolute_section or the segment of a symbol.
 760  */
 761
 762 #undef __
 763 #define __ O_illegal
 764
 765 static const operatorT op_encoding[256] =
 766 {                               /* maps ASCII->operators */
 767
 768   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 769   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 770
 771   __, O_bit_or_not, __, __, __, O_modulus, O_bit_and, __,
 772   __, __, O_multiply, O_add, __, O_subtract, __, O_divide,
 773   __, __, __, __, __, __, __, __,
 774   __, __, __, __, O_left_shift, __, O_right_shift, __,
 775   __, __, __, __, __, __, __, __,
 776   __, __, __, __, __, __, __, __,
 777   __, __, __, __, __, __, __, __,
 778   __, __, __, __, __, __, O_bit_exclusive_or, __,
 779   __, __, __, __, __, __, __, __,
 780   __, __, __, __, __, __, __, __,
 781   __, __, __, __, __, __, __, __,
 782   __, __, __, __, O_bit_inclusive_or, __, __, __,
 783
 784   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 785   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 786   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 787   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 788   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 789   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 790   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __,
 791   __, __, __, __, __, __, __, __, __, __, __, __, __, __, __, __
 792 };
 793
 794
 795 /*
 796  *      Rank    Examples
 797  *      0       operand, (expression)
 798  *      1       + -
 799  *      2       & ^ ! |
 800  *      3       * / % << >>
 801  *      4       unary - unary ~
 802  */
 803 static const operator_rankT op_rank[] =
 804 {
 805   0,    /* O_illegal */
 806   0,    /* O_absent */
 807   0,    /* O_constant */
 808   0,    /* O_symbol */
 809   0,    /* O_register */
 810   0,    /* O_bit */
 811   4,    /* O_uminus */
 812   4,    /* O_bit_now */
 813   3,    /* O_multiply */
 814   3,    /* O_divide */
 815   3,    /* O_modulus */
 816   3,    /* O_left_shift */
 817   3,    /* O_right_shift */
 818   2,    /* O_bit_inclusive_or */
 819   2,    /* O_bit_or_not */
 820   2,    /* O_bit_exclusive_or */
 821   2,    /* O_bit_and */
 822   1,    /* O_add */
 823   1,    /* O_subtract */
 824 };
 825 \f
 826 segT
 827 expr (rank, resultP)
 828      operator_rankT rank;       /* Larger # is higher rank. */
 829      expressionS *resultP;      /* Deliver result here. */
 830 {
 831   segT retval;
 832   expressionS right;
 833   operatorT op_left;
 834   char c_left;          /* 1st operator character. */
 835   operatorT op_right;
 836   char c_right;
 837
 838   know (rank >= 0);
 839
 840   retval = operand (resultP);
 841
 842   know (*input_line_pointer != ' ');    /* Operand() gobbles spaces. */
 843
 844   c_left = *input_line_pointer; /* Potential operator character. */
 845   op_left = op_encoding[(unsigned char) c_left];
 846   while (op_left != O_illegal && op_rank[(int) op_left] > rank)
 847     {
 848       segT rightseg;
 849
 850       input_line_pointer++;     /*->after 1st character of operator. */
 851       /* Operators "<<" and ">>" have 2 characters. */
 852       if (*input_line_pointer == c_left && (c_left == '<' || c_left == '>'))
 853         ++input_line_pointer;
 854
 855       rightseg = expr (op_rank[(int) op_left], &right);
 856       if (right.X_op == O_absent)
 857         {
 858           as_warn ("missing operand; zero assumed");
 859           right.X_op = O_constant;
 860           right.X_add_number = 0;
 861           right.X_add_symbol = NULL;
 862           right.X_op_symbol = NULL;
 863         }
 864
 865       know (*input_line_pointer != ' ');
 866
 867       if (retval == undefined_section)
 868         {
 869           if (SEG_NORMAL (rightseg))
 870             retval = rightseg;
 871         }
 872       else if (! SEG_NORMAL (retval))
 873         retval = rightseg;
 874       else if (SEG_NORMAL (rightseg)
 875                && retval != rightseg
 876 #ifdef DIFF_EXPR_OK
 877                && op_left != O_subtract
 878 #endif
 879                )
 880         as_bad ("operation combines symbols in different segments");
 881
 882       c_right = *input_line_pointer;
 883       op_right = op_encoding[(unsigned char) c_right];
 884       if (*input_line_pointer == c_right && (c_right == '<' || c_right == '>'))
 885         ++input_line_pointer;
 886
 887       know (op_right == O_illegal || op_rank[(int) op_right] <= op_rank[(int) op_left]);
 888       know ((int) op_left >= (int) O_multiply && (int) op_left <= (int) O_subtract);
 889
 890       /* input_line_pointer->after right-hand quantity. */
 891       /* left-hand quantity in resultP */
 892       /* right-hand quantity in right. */
 893       /* operator in op_left. */
 894
 895       if (resultP->X_op == O_big)
 896         {
 897           as_warn ("left operand of %c is a %s; integer 0 assumed",
 898                    c_left, resultP->X_add_number > 0 ? "bignum" : "float");
 899           resultP->X_op = O_constant;
 900           resultP->X_add_number = 0;
 901           resultP->X_add_symbol = NULL;
 902           resultP->X_op_symbol = NULL;
 903         }
 904       if (right.X_op == O_big)
 905         {
 906           as_warn ("right operand of %c is a %s; integer 0 assumed",
 907                    c_left, right.X_add_number > 0 ? "bignum" : "float");
 908           right.X_op = O_constant;
 909           right.X_add_number = 0;
 910           right.X_add_symbol = NULL;
 911           right.X_op_symbol = NULL;
 912         }
 913
 914       /* Optimize common cases.  */
 915       if (op_left == O_add && right.X_op == O_constant)
 916         {
 917           /* X + constant.  */
 918           resultP->X_add_number += right.X_add_number;
 919         }
 920       else if (op_left == O_subtract && right.X_op == O_constant)
 921         {
 922           /* X - constant.  */
 923           resultP->X_add_number -= right.X_add_number;
 924         }
 925       else if (op_left == O_add && resultP->X_op == O_constant)
 926         {
 927           /* Constant + X.  */
 928           resultP->X_op = right.X_op;
 929           resultP->X_add_symbol = right.X_add_symbol;
 930           resultP->X_op_symbol = right.X_op_symbol;
 931           resultP->X_add_number += right.X_add_number;
 932           retval = rightseg;
 933         }
 934       else if (resultP->X_op == O_constant && right.X_op == O_constant)
 935         {
 936           /* Constant OP constant.  */
 937           offsetT v = right.X_add_number;
 938           if (v == 0 && (op_left == O_divide || op_left == O_modulus))
 939             {
 940               as_warn ("division by zero");
 941               v = 1;
 942             }
 943           switch (op_left)
 944             {
 945             case O_multiply:            resultP->X_add_number *= v; break;
 946             case O_divide:              resultP->X_add_number /= v; break;
 947             case O_modulus:             resultP->X_add_number %= v; break;
 948             case O_left_shift:          resultP->X_add_number <<= v; break;
 949             case O_right_shift:         resultP->X_add_number >>= v; break;
 950             case O_bit_inclusive_or:    resultP->X_add_number |= v; break;
 951             case O_bit_or_not:          resultP->X_add_number |= ~v; break;
 952             case O_bit_exclusive_or:    resultP->X_add_number ^= v; break;
 953             case O_bit_and:             resultP->X_add_number &= v; break;
 954             case O_add:                 resultP->X_add_number += v; break;
 955             case O_subtract:            resultP->X_add_number -= v; break;
 956             default:                    abort ();
 957             }
 958         }
 959       else if (resultP->X_op == O_symbol
 960                && right.X_op == O_symbol
 961                && (op_left == O_add
 962                    || op_left == O_subtract
 963                    || (resultP->X_add_number == 0
 964                        && right.X_add_number == 0)))
 965         {
 966           /* Symbol OP symbol.  */
 967           resultP->X_op = op_left;
 968           resultP->X_op_symbol = right.X_add_symbol;
 969           if (op_left == O_add)
 970             resultP->X_add_number += right.X_add_number;
 971           else if (op_left == O_subtract)
 972             resultP->X_add_number -= right.X_add_number;
 973         }
 974       else
 975         {
 976           /* The general case.  */
 977           resultP->X_add_symbol = make_expr_symbol (resultP);
 978           resultP->X_op_symbol = make_expr_symbol (&right);
 979           resultP->X_op = op_left;
 980           resultP->X_add_number = 0;
 981           resultP->X_unsigned = 1;
 982         }
 983
 984       op_left = op_right;
 985     }                           /* While next operator is >= this rank. */
 986
 987   /* The PA port needs this information.  */
 988   if (resultP->X_add_symbol)
 989     resultP->X_add_symbol->sy_used = 1;
 990
 991   return resultP->X_op == O_constant ? absolute_section : retval;
 992 }
 993 \f
 994 /*
 995  *                      get_symbol_end()
 996  *
 997  * This lives here because it belongs equally in expr.c & read.c.
 998  * Expr.c is just a branch office read.c anyway, and putting it
 999  * here lessens the crowd at read.c.
1000  *
1001  * Assume input_line_pointer is at start of symbol name.
1002  * Advance input_line_pointer past symbol name.
1003  * Turn that character into a '\0', returning its former value.
1004  * This allows a string compare (RMS wants symbol names to be strings)
1005  * of the symbol name.
1006  * There will always be a char following symbol name, because all good
1007  * lines end in end-of-line.
1008  */
1009 char
1010 get_symbol_end ()
1011 {
1012   char c;
1013
1014   while (is_part_of_name (c = *input_line_pointer++))
1015     ;
1016   *--input_line_pointer = 0;
1017   return (c);
1018 }
1019
1020
1021 unsigned int
1022 get_single_number ()
1023 {
1024   expressionS exp;
1025   operand (&exp);
1026   return exp.X_add_number;
1027
1028 }
1029
1030 /* end of expr.c */