gdb/c-lang.c

   1 /* C language support routines for GDB, the GNU debugger.
   2
   3    Copyright (C) 1992-2022 Free Software Foundation, Inc.
   4
   5    This file is part of GDB.
   6
   7    This program 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    This program 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 this program.  If not, see <http://www.gnu.org/licenses/>.  */
  19
  20 #include "defs.h"
  21 #include "symtab.h"
  22 #include "gdbtypes.h"
  23 #include "expression.h"
  24 #include "parser-defs.h"
  25 #include "language.h"
  26 #include "varobj.h"
  27 #include "c-lang.h"
  28 #include "c-support.h"
  29 #include "valprint.h"
  30 #include "macroscope.h"
  31 #include "charset.h"
  32 #include "demangle.h"
  33 #include "cp-abi.h"
  34 #include "cp-support.h"
  35 #include "gdbsupport/gdb_obstack.h"
  36 #include <ctype.h>
  37 #include "gdbcore.h"
  38 #include "gdbarch.h"
  39 #include "compile/compile-internal.h"
  40 #include "c-exp.h"
  41
  42 /* Given a C string type, STR_TYPE, return the corresponding target
  43    character set name.  */
  44
  45 static const char *
  46 charset_for_string_type (c_string_type str_type, struct gdbarch *gdbarch)
  47 {
  48   switch (str_type & ~C_CHAR)
  49     {
  50     case C_STRING:
  51       return target_charset (gdbarch);
  52     case C_WIDE_STRING:
  53       return target_wide_charset (gdbarch);
  54     case C_STRING_16:
  55       /* FIXME: UTF-16 is not always correct.  */
  56       if (gdbarch_byte_order (gdbarch) == BFD_ENDIAN_BIG)
  57         return "UTF-16BE";
  58       else
  59         return "UTF-16LE";
  60     case C_STRING_32:
  61       /* FIXME: UTF-32 is not always correct.  */
  62       if (gdbarch_byte_order (gdbarch) == BFD_ENDIAN_BIG)
  63         return "UTF-32BE";
  64       else
  65         return "UTF-32LE";
  66     }
  67   internal_error (__FILE__, __LINE__, _("unhandled c_string_type"));
  68 }
  69
  70 /* Classify ELTTYPE according to what kind of character it is.  Return
  71    the enum constant representing the character type.  Also set
  72    *ENCODING to the name of the character set to use when converting
  73    characters of this type in target BYTE_ORDER to the host character
  74    set.  */
  75
  76 static c_string_type
  77 classify_type (struct type *elttype, struct gdbarch *gdbarch,
  78                const char **encoding)
  79 {
  80   c_string_type result;
  81
  82   /* We loop because ELTTYPE may be a typedef, and we want to
  83      successively peel each typedef until we reach a type we
  84      understand.  We don't use CHECK_TYPEDEF because that will strip
  85      all typedefs at once -- but in C, wchar_t is itself a typedef, so
  86      that would do the wrong thing.  */
  87   while (elttype)
  88     {
  89       const char *name = elttype->name ();
  90
  91       if (name == nullptr)
  92         {
  93           result = C_CHAR;
  94           goto done;
  95         }
  96
  97       if (!strcmp (name, "wchar_t"))
  98         {
  99           result = C_WIDE_CHAR;
 100           goto done;
 101         }
 102
 103       if (!strcmp (name, "char16_t"))
 104         {
 105           result = C_CHAR_16;
 106           goto done;
 107         }
 108
 109       if (!strcmp (name, "char32_t"))
 110         {
 111           result = C_CHAR_32;
 112           goto done;
 113         }
 114
 115       if (elttype->code () != TYPE_CODE_TYPEDEF)
 116         break;
 117
 118       /* Call for side effects.  */
 119       check_typedef (elttype);
 120
 121       if (TYPE_TARGET_TYPE (elttype))
 122         elttype = TYPE_TARGET_TYPE (elttype);
 123       else
 124         {
 125           /* Perhaps check_typedef did not update the target type.  In
 126              this case, force the lookup again and hope it works out.
 127              It never will for C, but it might for C++.  */
 128           elttype = check_typedef (elttype);
 129         }
 130     }
 131
 132   /* Punt.  */
 133   result = C_CHAR;
 134
 135  done:
 136   if (encoding)
 137     *encoding = charset_for_string_type (result, gdbarch);
 138
 139   return result;
 140 }
 141
 142 /* Print the character C on STREAM as part of the contents of a
 143    literal string whose delimiter is QUOTER.  Note that that format
 144    for printing characters and strings is language specific.  */
 145
 146 void
 147 c_emit_char (int c, struct type *type,
 148              struct ui_file *stream, int quoter)
 149 {
 150   const char *encoding;
 151
 152   classify_type (type, type->arch (), &encoding);
 153   generic_emit_char (c, type, stream, quoter, encoding);
 154 }
 155
 156 /* See language.h.  */
 157
 158 void
 159 language_defn::printchar (int c, struct type *type,
 160                           struct ui_file * stream) const
 161 {
 162   c_string_type str_type;
 163
 164   str_type = classify_type (type, type->arch (), NULL);
 165   switch (str_type)
 166     {
 167     case C_CHAR:
 168       break;
 169     case C_WIDE_CHAR:
 170       gdb_putc ('L', stream);
 171       break;
 172     case C_CHAR_16:
 173       gdb_putc ('u', stream);
 174       break;
 175     case C_CHAR_32:
 176       gdb_putc ('U', stream);
 177       break;
 178     }
 179
 180   gdb_putc ('\'', stream);
 181   emitchar (c, type, stream, '\'');
 182   gdb_putc ('\'', stream);
 183 }
 184
 185 /* Print the character string STRING, printing at most LENGTH
 186    characters.  LENGTH is -1 if the string is nul terminated.  Each
 187    character is WIDTH bytes long.  Printing stops early if the number
 188    hits print_max; repeat counts are printed as appropriate.  Print
 189    ellipses at the end if we had to stop before printing LENGTH
 190    characters, or if FORCE_ELLIPSES.  */
 191
 192 void
 193 c_printstr (struct ui_file *stream, struct type *type,
 194             const gdb_byte *string, unsigned int length,
 195             const char *user_encoding, int force_ellipses,
 196             const struct value_print_options *options)
 197 {
 198   c_string_type str_type;
 199   const char *type_encoding;
 200   const char *encoding;
 201
 202   str_type = (classify_type (type, type->arch (), &type_encoding)
 203               & ~C_CHAR);
 204   switch (str_type)
 205     {
 206     case C_STRING:
 207       break;
 208     case C_WIDE_STRING:
 209       gdb_puts ("L", stream);
 210       break;
 211     case C_STRING_16:
 212       gdb_puts ("u", stream);
 213       break;
 214     case C_STRING_32:
 215       gdb_puts ("U", stream);
 216       break;
 217     }
 218
 219   encoding = (user_encoding && *user_encoding) ? user_encoding : type_encoding;
 220
 221   generic_printstr (stream, type, string, length, encoding, force_ellipses,
 222                     '"', 1, options);
 223 }
 224
 225 /* Obtain a C string from the inferior storing it in a newly allocated
 226    buffer in BUFFER, which should be freed by the caller.  If the in-
 227    and out-parameter *LENGTH is specified at -1, the string is read
 228    until a null character of the appropriate width is found, otherwise
 229    the string is read to the length of characters specified.  The size
 230    of a character is determined by the length of the target type of
 231    the pointer or array.
 232
 233    If VALUE is an array with a known length, and *LENGTH is -1,
 234    the function will not read past the end of the array.  However, any
 235    declared size of the array is ignored if *LENGTH > 0.
 236
 237    On completion, *LENGTH will be set to the size of the string read in
 238    characters.  (If a length of -1 is specified, the length returned
 239    will not include the null character).  CHARSET is always set to the
 240    target charset.  */
 241
 242 void
 243 c_get_string (struct value *value, gdb::unique_xmalloc_ptr<gdb_byte> *buffer,
 244               int *length, struct type **char_type,
 245               const char **charset)
 246 {
 247   int err, width;
 248   unsigned int fetchlimit;
 249   struct type *type = check_typedef (value_type (value));
 250   struct type *element_type = TYPE_TARGET_TYPE (type);
 251   int req_length = *length;
 252   enum bfd_endian byte_order
 253     = type_byte_order (type);
 254
 255   if (element_type == NULL)
 256     goto error;
 257
 258   if (type->code () == TYPE_CODE_ARRAY)
 259     {
 260       /* If we know the size of the array, we can use it as a limit on
 261          the number of characters to be fetched.  */
 262       if (type->num_fields () == 1
 263           && type->field (0).type ()->code () == TYPE_CODE_RANGE)
 264         {
 265           LONGEST low_bound, high_bound;
 266
 267           get_discrete_bounds (type->field (0).type (),
 268                                &low_bound, &high_bound);
 269           fetchlimit = high_bound - low_bound + 1;
 270         }
 271       else
 272         fetchlimit = UINT_MAX;
 273     }
 274   else if (type->code () == TYPE_CODE_PTR)
 275     fetchlimit = UINT_MAX;
 276   else
 277     /* We work only with arrays and pointers.  */
 278     goto error;
 279
 280   if (! c_textual_element_type (element_type, 0))
 281     goto error;
 282   classify_type (element_type, element_type->arch (), charset);
 283   width = TYPE_LENGTH (element_type);
 284
 285   /* If the string lives in GDB's memory instead of the inferior's,
 286      then we just need to copy it to BUFFER.  Also, since such strings
 287      are arrays with known size, FETCHLIMIT will hold the size of the
 288      array.
 289
 290      An array is assumed to live in GDB's memory, so we take this path
 291      here.
 292
 293      However, it's possible for the caller to request more array
 294      elements than apparently exist -- this can happen when using the
 295      C struct hack.  So, only do this if either no length was
 296      specified, or the length is within the existing bounds.  This
 297      avoids running off the end of the value's contents.  */
 298   if ((VALUE_LVAL (value) == not_lval
 299        || VALUE_LVAL (value) == lval_internalvar
 300        || type->code () == TYPE_CODE_ARRAY)
 301       && fetchlimit != UINT_MAX
 302       && (*length < 0 || *length <= fetchlimit))
 303     {
 304       int i;
 305       const gdb_byte *contents = value_contents (value).data ();
 306
 307       /* If a length is specified, use that.  */
 308       if (*length >= 0)
 309         i  = *length;
 310       else
 311         /* Otherwise, look for a null character.  */
 312         for (i = 0; i < fetchlimit; i++)
 313           if (extract_unsigned_integer (contents + i * width,
 314                                         width, byte_order) == 0)
 315             break;
 316
 317       /* I is now either a user-defined length, the number of non-null
 318          characters, or FETCHLIMIT.  */
 319       *length = i * width;
 320       buffer->reset ((gdb_byte *) xmalloc (*length));
 321       memcpy (buffer->get (), contents, *length);
 322       err = 0;
 323     }
 324   else
 325     {
 326       /* value_as_address does not return an address for an array when
 327          c_style_arrays is false, so we handle that specially
 328          here.  */
 329       CORE_ADDR addr;
 330       if (type->code () == TYPE_CODE_ARRAY)
 331         {
 332           if (VALUE_LVAL (value) != lval_memory)
 333             error (_("Attempt to take address of value "
 334                      "not located in memory."));
 335           addr = value_address (value);
 336         }
 337       else
 338         addr = value_as_address (value);
 339
 340       /* Prior to the fix for PR 16196 read_string would ignore fetchlimit
 341          if length > 0.  The old "broken" behaviour is the behaviour we want:
 342          The caller may want to fetch 100 bytes from a variable length array
 343          implemented using the common idiom of having an array of length 1 at
 344          the end of a struct.  In this case we want to ignore the declared
 345          size of the array.  However, it's counterintuitive to implement that
 346          behaviour in read_string: what does fetchlimit otherwise mean if
 347          length > 0.  Therefore we implement the behaviour we want here:
 348          If *length > 0, don't specify a fetchlimit.  This preserves the
 349          previous behaviour.  We could move this check above where we know
 350          whether the array is declared with a fixed size, but we only want
 351          to apply this behaviour when calling read_string.  PR 16286.  */
 352       if (*length > 0)
 353         fetchlimit = UINT_MAX;
 354
 355       err = target_read_string (addr, *length, width, fetchlimit,
 356                                 buffer, length);
 357       if (err != 0)
 358         memory_error (TARGET_XFER_E_IO, addr);
 359     }
 360
 361   /* If the LENGTH is specified at -1, we want to return the string
 362      length up to the terminating null character.  If an actual length
 363      was specified, we want to return the length of exactly what was
 364      read.  */
 365   if (req_length == -1)
 366     /* If the last character is null, subtract it from LENGTH.  */
 367     if (*length > 0
 368         && extract_unsigned_integer (buffer->get () + *length - width,
 369                                      width, byte_order) == 0)
 370       *length -= width;
 371
 372   /* The read_string function will return the number of bytes read.
 373      If length returned from read_string was > 0, return the number of
 374      characters read by dividing the number of bytes by width.  */
 375   if (*length != 0)
 376      *length = *length / width;
 377
 378   *char_type = element_type;
 379
 380   return;
 381
 382  error:
 383   {
 384     std::string type_str = type_to_string (type);
 385     if (!type_str.empty ())
 386       {
 387         error (_("Trying to read string with inappropriate type `%s'."),
 388                type_str.c_str ());
 389       }
 390     else
 391       error (_("Trying to read string with inappropriate type."));
 392   }
 393 }
 394
 395 \f
 396 /* Evaluating C and C++ expressions.  */
 397
 398 /* Convert a UCN.  The digits of the UCN start at P and extend no
 399    farther than LIMIT.  DEST_CHARSET is the name of the character set
 400    into which the UCN should be converted.  The results are written to
 401    OUTPUT.  LENGTH is the maximum length of the UCN, either 4 or 8.
 402    Returns a pointer to just after the final digit of the UCN.  */
 403
 404 static const char *
 405 convert_ucn (const char *p, const char *limit, const char *dest_charset,
 406              struct obstack *output, int length)
 407 {
 408   unsigned long result = 0;
 409   gdb_byte data[4];
 410   int i;
 411
 412   for (i = 0; i < length && p < limit && ISXDIGIT (*p); ++i, ++p)
 413     result = (result << 4) + fromhex (*p);
 414
 415   for (i = 3; i >= 0; --i)
 416     {
 417       data[i] = result & 0xff;
 418       result >>= 8;
 419     }
 420
 421   convert_between_encodings ("UTF-32BE", dest_charset, data,
 422                              4, 4, output, translit_none);
 423
 424   return p;
 425 }
 426
 427 /* Emit a character, VALUE, which was specified numerically, to
 428    OUTPUT.  TYPE is the target character type.  */
 429
 430 static void
 431 emit_numeric_character (struct type *type, unsigned long value,
 432                         struct obstack *output)
 433 {
 434   gdb_byte *buffer;
 435
 436   buffer = (gdb_byte *) alloca (TYPE_LENGTH (type));
 437   pack_long (buffer, type, value);
 438   obstack_grow (output, buffer, TYPE_LENGTH (type));
 439 }
 440
 441 /* Convert an octal escape sequence.  TYPE is the target character
 442    type.  The digits of the escape sequence begin at P and extend no
 443    farther than LIMIT.  The result is written to OUTPUT.  Returns a
 444    pointer to just after the final digit of the escape sequence.  */
 445
 446 static const char *
 447 convert_octal (struct type *type, const char *p,
 448                const char *limit, struct obstack *output)
 449 {
 450   int i;
 451   unsigned long value = 0;
 452
 453   for (i = 0;
 454        i < 3 && p < limit && ISDIGIT (*p) && *p != '8' && *p != '9';
 455        ++i)
 456     {
 457       value = 8 * value + fromhex (*p);
 458       ++p;
 459     }
 460
 461   emit_numeric_character (type, value, output);
 462
 463   return p;
 464 }
 465
 466 /* Convert a hex escape sequence.  TYPE is the target character type.
 467    The digits of the escape sequence begin at P and extend no farther
 468    than LIMIT.  The result is written to OUTPUT.  Returns a pointer to
 469    just after the final digit of the escape sequence.  */
 470
 471 static const char *
 472 convert_hex (struct type *type, const char *p,
 473              const char *limit, struct obstack *output)
 474 {
 475   unsigned long value = 0;
 476
 477   while (p < limit && ISXDIGIT (*p))
 478     {
 479       value = 16 * value + fromhex (*p);
 480       ++p;
 481     }
 482
 483   emit_numeric_character (type, value, output);
 484
 485   return p;
 486 }
 487
 488 #define ADVANCE                                 \
 489   do {                                          \
 490     ++p;                                        \
 491     if (p == limit)                             \
 492       error (_("Malformed escape sequence"));   \
 493   } while (0)
 494
 495 /* Convert an escape sequence to a target format.  TYPE is the target
 496    character type to use, and DEST_CHARSET is the name of the target
 497    character set.  The backslash of the escape sequence is at *P, and
 498    the escape sequence will not extend past LIMIT.  The results are
 499    written to OUTPUT.  Returns a pointer to just past the final
 500    character of the escape sequence.  */
 501
 502 static const char *
 503 convert_escape (struct type *type, const char *dest_charset,
 504                 const char *p, const char *limit, struct obstack *output)
 505 {
 506   /* Skip the backslash.  */
 507   ADVANCE;
 508
 509   switch (*p)
 510     {
 511     case '\\':
 512       obstack_1grow (output, '\\');
 513       ++p;
 514       break;
 515
 516     case 'x':
 517       ADVANCE;
 518       if (!ISXDIGIT (*p))
 519         error (_("\\x used with no following hex digits."));
 520       p = convert_hex (type, p, limit, output);
 521       break;
 522
 523     case '0':
 524     case '1':
 525     case '2':
 526     case '3':
 527     case '4':
 528     case '5':
 529     case '6':
 530     case '7':
 531       p = convert_octal (type, p, limit, output);
 532       break;
 533
 534     case 'u':
 535     case 'U':
 536       {
 537         int length = *p == 'u' ? 4 : 8;
 538
 539         ADVANCE;
 540         if (!ISXDIGIT (*p))
 541           error (_("\\u used with no following hex digits"));
 542         p = convert_ucn (p, limit, dest_charset, output, length);
 543       }
 544     }
 545
 546   return p;
 547 }
 548
 549 /* Given a single string from a (C-specific) OP_STRING list, convert
 550    it to a target string, handling escape sequences specially.  The
 551    output is written to OUTPUT.  DATA is the input string, which has
 552    length LEN.  DEST_CHARSET is the name of the target character set,
 553    and TYPE is the type of target character to use.  */
 554
 555 static void
 556 parse_one_string (struct obstack *output, const char *data, int len,
 557                   const char *dest_charset, struct type *type)
 558 {
 559   const char *limit;
 560
 561   limit = data + len;
 562
 563   while (data < limit)
 564     {
 565       const char *p = data;
 566
 567       /* Look for next escape, or the end of the input.  */
 568       while (p < limit && *p != '\\')
 569         ++p;
 570       /* If we saw a run of characters, convert them all.  */
 571       if (p > data)
 572         convert_between_encodings (host_charset (), dest_charset,
 573                                    (const gdb_byte *) data, p - data, 1,
 574                                    output, translit_none);
 575       /* If we saw an escape, convert it.  */
 576       if (p < limit)
 577         p = convert_escape (type, dest_charset, p, limit, output);
 578       data = p;
 579     }
 580 }
 581
 582 namespace expr
 583 {
 584
 585 value *
 586 c_string_operation::evaluate (struct type *expect_type,
 587                               struct expression *exp,
 588                               enum noside noside)
 589 {
 590   struct type *type;
 591   struct value *result;
 592   c_string_type dest_type;
 593   const char *dest_charset;
 594   int satisfy_expected = 0;
 595
 596   auto_obstack output;
 597
 598   dest_type = std::get<0> (m_storage);
 599
 600   switch (dest_type & ~C_CHAR)
 601     {
 602     case C_STRING:
 603       type = language_string_char_type (exp->language_defn,
 604                                         exp->gdbarch);
 605       break;
 606     case C_WIDE_STRING:
 607       type = lookup_typename (exp->language_defn, "wchar_t", NULL, 0);
 608       break;
 609     case C_STRING_16:
 610       type = lookup_typename (exp->language_defn, "char16_t", NULL, 0);
 611       break;
 612     case C_STRING_32:
 613       type = lookup_typename (exp->language_defn, "char32_t", NULL, 0);
 614       break;
 615     default:
 616       internal_error (__FILE__, __LINE__, _("unhandled c_string_type"));
 617     }
 618
 619   /* Ensure TYPE_LENGTH is valid for TYPE.  */
 620   check_typedef (type);
 621
 622   /* If the caller expects an array of some integral type,
 623      satisfy them.  If something odder is expected, rely on the
 624      caller to cast.  */
 625   if (expect_type && expect_type->code () == TYPE_CODE_ARRAY)
 626     {
 627       struct type *element_type
 628         = check_typedef (TYPE_TARGET_TYPE (expect_type));
 629
 630       if (element_type->code () == TYPE_CODE_INT
 631           || element_type->code () == TYPE_CODE_CHAR)
 632         {
 633           type = element_type;
 634           satisfy_expected = 1;
 635         }
 636     }
 637
 638   dest_charset = charset_for_string_type (dest_type, exp->gdbarch);
 639
 640   for (const std::string &item : std::get<1> (m_storage))
 641     parse_one_string (&output, item.c_str (), item.size (),
 642                       dest_charset, type);
 643
 644   if ((dest_type & C_CHAR) != 0)
 645     {
 646       LONGEST value;
 647
 648       if (obstack_object_size (&output) != TYPE_LENGTH (type))
 649         error (_("Could not convert character "
 650                  "constant to target character set"));
 651       value = unpack_long (type, (gdb_byte *) obstack_base (&output));
 652       result = value_from_longest (type, value);
 653     }
 654   else
 655     {
 656       int i;
 657
 658       /* Write the terminating character.  */
 659       for (i = 0; i < TYPE_LENGTH (type); ++i)
 660         obstack_1grow (&output, 0);
 661
 662       if (satisfy_expected)
 663         {
 664           LONGEST low_bound, high_bound;
 665           int element_size = TYPE_LENGTH (type);
 666
 667           if (!get_discrete_bounds (expect_type->index_type (),
 668                                     &low_bound, &high_bound))
 669             {
 670               low_bound = 0;
 671               high_bound = (TYPE_LENGTH (expect_type) / element_size) - 1;
 672             }
 673           if (obstack_object_size (&output) / element_size
 674               > (high_bound - low_bound + 1))
 675             error (_("Too many array elements"));
 676
 677           result = allocate_value (expect_type);
 678           memcpy (value_contents_raw (result).data (), obstack_base (&output),
 679                   obstack_object_size (&output));
 680         }
 681       else
 682         result = value_cstring ((const char *) obstack_base (&output),
 683                                 obstack_object_size (&output),
 684                                 type);
 685     }
 686   return result;
 687 }
 688
 689 } /* namespace expr */
 690
 691 \f
 692 /* See c-lang.h.  */
 693
 694 bool
 695 c_is_string_type_p (struct type *type)
 696 {
 697   type = check_typedef (type);
 698   while (type->code () == TYPE_CODE_REF)
 699     {
 700       type = TYPE_TARGET_TYPE (type);
 701       type = check_typedef (type);
 702     }
 703
 704   switch (type->code ())
 705     {
 706     case TYPE_CODE_ARRAY:
 707       {
 708         /* See if target type looks like a string.  */
 709         struct type *array_target_type = TYPE_TARGET_TYPE (type);
 710         return (TYPE_LENGTH (type) > 0
 711                 && TYPE_LENGTH (array_target_type) > 0
 712                 && c_textual_element_type (array_target_type, 0));
 713       }
 714     case TYPE_CODE_STRING:
 715       return true;
 716     case TYPE_CODE_PTR:
 717       {
 718         struct type *element_type = TYPE_TARGET_TYPE (type);
 719         return c_textual_element_type (element_type, 0);
 720       }
 721     default:
 722       break;
 723     }
 724
 725   return false;
 726 }
 727
 728 \f
 729
 730 void
 731 c_language_arch_info (struct gdbarch *gdbarch,
 732                       struct language_arch_info *lai)
 733 {
 734   const struct builtin_type *builtin = builtin_type (gdbarch);
 735
 736   /* Helper function to allow shorter lines below.  */
 737   auto add  = [&] (struct type * t)
 738   {
 739     lai->add_primitive_type (t);
 740   };
 741
 742   add (builtin->builtin_int);
 743   add (builtin->builtin_long);
 744   add (builtin->builtin_short);
 745   add (builtin->builtin_char);
 746   add (builtin->builtin_float);
 747   add (builtin->builtin_double);
 748   add (builtin->builtin_void);
 749   add (builtin->builtin_long_long);
 750   add (builtin->builtin_signed_char);
 751   add (builtin->builtin_unsigned_char);
 752   add (builtin->builtin_unsigned_short);
 753   add (builtin->builtin_unsigned_int);
 754   add (builtin->builtin_unsigned_long);
 755   add (builtin->builtin_unsigned_long_long);
 756   add (builtin->builtin_long_double);
 757   add (builtin->builtin_complex);
 758   add (builtin->builtin_double_complex);
 759   add (builtin->builtin_decfloat);
 760   add (builtin->builtin_decdouble);
 761   add (builtin->builtin_declong);
 762
 763   lai->set_string_char_type (builtin->builtin_char);
 764   lai->set_bool_type (builtin->builtin_int);
 765 }
 766
 767 /* Class representing the C language.  */
 768
 769 class c_language : public language_defn
 770 {
 771 public:
 772   c_language ()
 773     : language_defn (language_c)
 774   { /* Nothing.  */ }
 775
 776   /* See language.h.  */
 777
 778   const char *name () const override
 779   { return "c"; }
 780
 781   /* See language.h.  */
 782
 783   const char *natural_name () const override
 784   { return "C"; }
 785
 786   /* See language.h.  */
 787
 788   const std::vector<const char *> &filename_extensions () const override
 789   {
 790     static const std::vector<const char *> extensions = { ".c" };
 791     return extensions;
 792   }
 793
 794   /* See language.h.  */
 795   void language_arch_info (struct gdbarch *gdbarch,
 796                            struct language_arch_info *lai) const override
 797   {
 798     c_language_arch_info (gdbarch, lai);
 799   }
 800
 801   /* See language.h.  */
 802   std::unique_ptr<compile_instance> get_compile_instance () const override
 803   {
 804     return c_get_compile_context ();
 805   }
 806
 807   /* See language.h.  */
 808   std::string compute_program (compile_instance *inst,
 809                                const char *input,
 810                                struct gdbarch *gdbarch,
 811                                const struct block *expr_block,
 812                                CORE_ADDR expr_pc) const override
 813   {
 814     return c_compute_program (inst, input, gdbarch, expr_block, expr_pc);
 815   }
 816
 817   /* See language.h.  */
 818
 819   void print_type (struct type *type, const char *varstring,
 820                    struct ui_file *stream, int show, int level,
 821                    const struct type_print_options *flags) const override
 822   {
 823     c_print_type (type, varstring, stream, show, level, la_language, flags);
 824   }
 825
 826   /* See language.h.  */
 827
 828   bool store_sym_names_in_linkage_form_p () const override
 829   { return true; }
 830
 831   /* See language.h.  */
 832
 833   enum macro_expansion macro_expansion () const override
 834   { return macro_expansion_c; }
 835 };
 836
 837 /* Single instance of the C language class.  */
 838
 839 static c_language c_language_defn;
 840
 841 /* A class for the C++ language.  */
 842
 843 class cplus_language : public language_defn
 844 {
 845 public:
 846   cplus_language ()
 847     : language_defn (language_cplus)
 848   { /* Nothing.  */ }
 849
 850   /* See language.h.  */
 851
 852   const char *name () const override
 853   { return "c++"; }
 854
 855   /* See language.h.  */
 856
 857   const char *natural_name () const override
 858   { return "C++"; }
 859
 860   /* See language.h  */
 861   const char *get_digit_separator () const override
 862   { return "\'"; }
 863
 864   /* See language.h.  */
 865
 866   const std::vector<const char *> &filename_extensions () const override
 867   {
 868     static const std::vector<const char *> extensions
 869       = { ".C", ".cc", ".cp", ".cpp", ".cxx", ".c++" };
 870     return extensions;
 871   }
 872
 873   /* See language.h.  */
 874
 875   struct language_pass_by_ref_info pass_by_reference_info
 876         (struct type *type) const override
 877   {
 878     return cp_pass_by_reference (type);
 879   }
 880
 881   /* See language.h.  */
 882   void language_arch_info (struct gdbarch *gdbarch,
 883                            struct language_arch_info *lai) const override
 884   {
 885     const struct builtin_type *builtin = builtin_type (gdbarch);
 886
 887     /* Helper function to allow shorter lines below.  */
 888     auto add  = [&] (struct type * t)
 889     {
 890       lai->add_primitive_type (t);
 891     };
 892
 893     add (builtin->builtin_int);
 894     add (builtin->builtin_long);
 895     add (builtin->builtin_short);
 896     add (builtin->builtin_char);
 897     add (builtin->builtin_float);
 898     add (builtin->builtin_double);
 899     add (builtin->builtin_void);
 900     add (builtin->builtin_long_long);
 901     add (builtin->builtin_signed_char);
 902     add (builtin->builtin_unsigned_char);
 903     add (builtin->builtin_unsigned_short);
 904     add (builtin->builtin_unsigned_int);
 905     add (builtin->builtin_unsigned_long);
 906     add (builtin->builtin_unsigned_long_long);
 907     add (builtin->builtin_long_double);
 908     add (builtin->builtin_complex);
 909     add (builtin->builtin_double_complex);
 910     add (builtin->builtin_bool);
 911     add (builtin->builtin_decfloat);
 912     add (builtin->builtin_decdouble);
 913     add (builtin->builtin_declong);
 914     add (builtin->builtin_char16);
 915     add (builtin->builtin_char32);
 916     add (builtin->builtin_wchar);
 917
 918     lai->set_string_char_type (builtin->builtin_char);
 919     lai->set_bool_type (builtin->builtin_bool, "bool");
 920   }
 921
 922   /* See language.h.  */
 923   struct type *lookup_transparent_type (const char *name) const override
 924   {
 925     return cp_lookup_transparent_type (name);
 926   }
 927
 928   /* See language.h.  */
 929   std::unique_ptr<compile_instance> get_compile_instance () const override
 930   {
 931     return cplus_get_compile_context ();
 932   }
 933
 934   /* See language.h.  */
 935   std::string compute_program (compile_instance *inst,
 936                                const char *input,
 937                                struct gdbarch *gdbarch,
 938                                const struct block *expr_block,
 939                                CORE_ADDR expr_pc) const override
 940   {
 941     return cplus_compute_program (inst, input, gdbarch, expr_block, expr_pc);
 942   }
 943
 944   /* See language.h.  */
 945   unsigned int search_name_hash (const char *name) const override
 946   {
 947     return cp_search_name_hash (name);
 948   }
 949
 950   /* See language.h.  */
 951   bool sniff_from_mangled_name
 952        (const char *mangled,
 953         gdb::unique_xmalloc_ptr<char> *demangled) const override
 954   {
 955     *demangled = gdb_demangle (mangled, DMGL_PARAMS | DMGL_ANSI);
 956     return *demangled != NULL;
 957   }
 958
 959   /* See language.h.  */
 960
 961   gdb::unique_xmalloc_ptr<char> demangle_symbol (const char *mangled,
 962                                                  int options) const override
 963   {
 964     return gdb_demangle (mangled, options);
 965   }
 966
 967   /* See language.h.  */
 968
 969   void print_type (struct type *type, const char *varstring,
 970                    struct ui_file *stream, int show, int level,
 971                    const struct type_print_options *flags) const override
 972   {
 973     c_print_type (type, varstring, stream, show, level, la_language, flags);
 974   }
 975
 976   /* See language.h.  */
 977
 978   CORE_ADDR skip_trampoline (struct frame_info *fi,
 979                              CORE_ADDR pc) const override
 980   {
 981     return cplus_skip_trampoline (fi, pc);
 982   }
 983
 984   /* See language.h.  */
 985
 986   char *class_name_from_physname (const char *physname) const override
 987   {
 988     return cp_class_name_from_physname (physname);
 989   }
 990
 991   /* See language.h.  */
 992
 993   struct block_symbol lookup_symbol_nonlocal
 994         (const char *name, const struct block *block,
 995          const domain_enum domain) const override
 996   {
 997     return cp_lookup_symbol_nonlocal (this, name, block, domain);
 998   }
 999
1000   /* See language.h.  */
1001
1002   const char *name_of_this () const override
1003   { return "this"; }
1004
1005   /* See language.h.  */
1006
1007   enum macro_expansion macro_expansion () const override
1008   { return macro_expansion_c; }
1009
1010   /* See language.h.  */
1011
1012   const struct lang_varobj_ops *varobj_ops () const override
1013   { return &cplus_varobj_ops; }
1014
1015 protected:
1016
1017   /* See language.h.  */
1018
1019   symbol_name_matcher_ftype *get_symbol_name_matcher_inner
1020         (const lookup_name_info &lookup_name) const override
1021   {
1022     return cp_get_symbol_name_matcher (lookup_name);
1023   }
1024 };
1025
1026 /* The single instance of the C++ language class.  */
1027
1028 static cplus_language cplus_language_defn;
1029
1030 /* A class for the ASM language.  */
1031
1032 class asm_language : public language_defn
1033 {
1034 public:
1035   asm_language ()
1036     : language_defn (language_asm)
1037   { /* Nothing.  */ }
1038
1039   /* See language.h.  */
1040
1041   const char *name () const override
1042   { return "asm"; }
1043
1044   /* See language.h.  */
1045
1046   const char *natural_name () const override
1047   { return "Assembly"; }
1048
1049   /* See language.h.  */
1050
1051   const std::vector<const char *> &filename_extensions () const override
1052   {
1053     static const std::vector<const char *> extensions
1054       = { ".s", ".sx", ".S" };
1055     return extensions;
1056   }
1057
1058   /* See language.h.
1059
1060      FIXME: Should this have its own arch info method?  */
1061   void language_arch_info (struct gdbarch *gdbarch,
1062                            struct language_arch_info *lai) const override
1063   {
1064     c_language_arch_info (gdbarch, lai);
1065   }
1066
1067   /* See language.h.  */
1068
1069   void print_type (struct type *type, const char *varstring,
1070                    struct ui_file *stream, int show, int level,
1071                    const struct type_print_options *flags) const override
1072   {
1073     c_print_type (type, varstring, stream, show, level, la_language, flags);
1074   }
1075
1076   /* See language.h.  */
1077
1078   bool store_sym_names_in_linkage_form_p () const override
1079   { return true; }
1080
1081   /* See language.h.  */
1082
1083   enum macro_expansion macro_expansion () const override
1084   { return macro_expansion_c; }
1085 };
1086
1087 /* The single instance of the ASM language class.  */
1088 static asm_language asm_language_defn;
1089
1090 /* A class for the minimal language.  This does not represent a real
1091    language.  It just provides a minimal support a-la-C that should allow
1092    users to do some simple operations when debugging applications that use
1093    a language currently not supported by GDB.  */
1094
1095 class minimal_language : public language_defn
1096 {
1097 public:
1098   minimal_language ()
1099     : language_defn (language_minimal)
1100   { /* Nothing.  */ }
1101
1102   /* See language.h.  */
1103
1104   const char *name () const override
1105   { return "minimal"; }
1106
1107   /* See language.h.  */
1108
1109   const char *natural_name () const override
1110   { return "Minimal"; }
1111
1112   /* See language.h.  */
1113   void language_arch_info (struct gdbarch *gdbarch,
1114                            struct language_arch_info *lai) const override
1115   {
1116     c_language_arch_info (gdbarch, lai);
1117   }
1118
1119   /* See language.h.  */
1120
1121   void print_type (struct type *type, const char *varstring,
1122                    struct ui_file *stream, int show, int level,
1123                    const struct type_print_options *flags) const override
1124   {
1125     c_print_type (type, varstring, stream, show, level, la_language, flags);
1126   }
1127
1128   /* See language.h.  */
1129
1130   bool store_sym_names_in_linkage_form_p () const override
1131   { return true; }
1132
1133   /* See language.h.  */
1134
1135   enum macro_expansion macro_expansion () const override
1136   { return macro_expansion_c; }
1137 };
1138
1139 /* The single instance of the minimal language class.  */
1140 static minimal_language minimal_language_defn;