/* YACC parser for C expressions, for GDB.
Copyright (C) 1986, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
- 1998, 1999, 2000, 2003, 2004, 2006, 2007, 2008
+ 1998, 1999, 2000, 2003, 2004, 2006, 2007, 2008, 2009
Free Software Foundation, Inc.
-This file is part of GDB.
+ This file is part of GDB.
-This program is free software; you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 2 of the License, or
-(at your option) any later version.
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
-This program is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-GNU General Public License for more details.
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
-You should have received a copy of the GNU General Public License
-along with this program; if not, write to the Free Software
-Foundation, Inc., 51 Franklin Street, Fifth Floor,
-Boston, MA 02110-1301, USA. */
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
/* Parse a C expression from text in a string,
and return the result as a struct expression pointer.
#include "block.h"
#include "cp-support.h"
#include "dfp.h"
+#include "gdb_assert.h"
+#include "macroscope.h"
+
+#define parse_type builtin_type (parse_gdbarch)
/* Remap normal yacc parser interface names (yyparse, yylex, yyerror, etc),
as well as gratuitiously global symbol names, so we can have multiple
struct symbol *sym;
struct type *tval;
struct stoken sval;
+ struct typed_stoken tsval;
struct ttype tsym;
struct symtoken ssym;
int voidval;
enum exp_opcode opcode;
struct internalvar *ivar;
+ struct stoken_vector svec;
struct type **tvec;
int *ivec;
}
Contexts where this distinction is not important can use the
nonterminal "name", which matches either NAME or TYPENAME. */
-%token <sval> STRING
+%token <tsval> STRING
+%token <tsval> CHAR
%token <ssym> NAME /* BLOCKNAME defined below to give it higher precedence. */
%token <voidval> COMPLETE
%token <tsym> TYPENAME
%type <sval> name
+%type <svec> string_exp
%type <ssym> name_not_typename
%type <tsym> typename
write_exp_elt_opcode (OP_LONG); }
;
+exp : CHAR
+ {
+ struct stoken_vector vec;
+ vec.len = 1;
+ vec.tokens = &$1;
+ write_exp_string_vector ($1.type, &vec);
+ }
+ ;
+
exp : NAME_OR_INT
{ YYSTYPE val;
parse_number ($1.stoken.ptr, $1.stoken.length, 0, &val);
exp : SIZEOF '(' type ')' %prec UNARY
{ write_exp_elt_opcode (OP_LONG);
- write_exp_elt_type (builtin_type (current_gdbarch)->builtin_int);
+ write_exp_elt_type (parse_type->builtin_int);
CHECK_TYPEDEF ($3);
write_exp_elt_longcst ((LONGEST) TYPE_LENGTH ($3));
write_exp_elt_opcode (OP_LONG); }
;
-exp : STRING
- { /* C strings are converted into array constants with
- an explicit null byte added at the end. Thus
- the array upper bound is the string length.
- There is no such thing in C as a completely empty
- string. */
- char *sp = $1.ptr; int count = $1.length;
- while (count-- > 0)
+string_exp:
+ STRING
+ {
+ /* We copy the string here, and not in the
+ lexer, to guarantee that we do not leak a
+ string. Note that we follow the
+ NUL-termination convention of the
+ lexer. */
+ struct typed_stoken *vec = XNEW (struct typed_stoken);
+ $$.len = 1;
+ $$.tokens = vec;
+
+ vec->type = $1.type;
+ vec->length = $1.length;
+ vec->ptr = malloc ($1.length + 1);
+ memcpy (vec->ptr, $1.ptr, $1.length + 1);
+ }
+
+ | string_exp STRING
+ {
+ /* Note that we NUL-terminate here, but just
+ for convenience. */
+ char *p;
+ ++$$.len;
+ $$.tokens = realloc ($$.tokens,
+ $$.len * sizeof (struct typed_stoken));
+
+ p = malloc ($2.length + 1);
+ memcpy (p, $2.ptr, $2.length + 1);
+
+ $$.tokens[$$.len - 1].type = $2.type;
+ $$.tokens[$$.len - 1].length = $2.length;
+ $$.tokens[$$.len - 1].ptr = p;
+ }
+ ;
+
+exp : string_exp
+ {
+ int i;
+ enum c_string_type type = C_STRING;
+
+ for (i = 0; i < $1.len; ++i)
{
- write_exp_elt_opcode (OP_LONG);
- write_exp_elt_type (builtin_type (current_gdbarch)->builtin_char);
- write_exp_elt_longcst ((LONGEST)(*sp++));
- write_exp_elt_opcode (OP_LONG);
+ switch ($1.tokens[i].type)
+ {
+ case C_STRING:
+ break;
+ case C_WIDE_STRING:
+ case C_STRING_16:
+ case C_STRING_32:
+ if (type != C_STRING
+ && type != $1.tokens[i].type)
+ error ("Undefined string concatenation.");
+ type = $1.tokens[i].type;
+ break;
+ default:
+ /* internal error */
+ internal_error (__FILE__, __LINE__,
+ "unrecognized type in string concatenation");
+ }
}
- write_exp_elt_opcode (OP_LONG);
- write_exp_elt_type (builtin_type (current_gdbarch)->builtin_char);
- write_exp_elt_longcst ((LONGEST)'\0');
- write_exp_elt_opcode (OP_LONG);
- write_exp_elt_opcode (OP_ARRAY);
- write_exp_elt_longcst ((LONGEST) 0);
- write_exp_elt_longcst ((LONGEST) ($1.length));
- write_exp_elt_opcode (OP_ARRAY); }
+
+ write_exp_string_vector (type, &$1);
+ for (i = 0; i < $1.len; ++i)
+ free ($1.tokens[i].ptr);
+ free ($1.tokens);
+ }
;
/* C++. */
exp : TRUEKEYWORD
{ write_exp_elt_opcode (OP_LONG);
- write_exp_elt_type (builtin_type (current_gdbarch)->builtin_bool);
+ write_exp_elt_type (parse_type->builtin_bool);
write_exp_elt_longcst ((LONGEST) 1);
write_exp_elt_opcode (OP_LONG); }
;
exp : FALSEKEYWORD
{ write_exp_elt_opcode (OP_LONG);
- write_exp_elt_type (builtin_type (current_gdbarch)->builtin_bool);
+ write_exp_elt_type (parse_type->builtin_bool);
write_exp_elt_longcst ((LONGEST) 0);
write_exp_elt_opcode (OP_LONG); }
;
: TYPENAME
{ $$ = $1.type; }
| INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_int; }
+ { $$ = parse_type->builtin_int; }
| LONG
- { $$ = builtin_type (current_gdbarch)->builtin_long; }
+ { $$ = parse_type->builtin_long; }
| SHORT
- { $$ = builtin_type (current_gdbarch)->builtin_short; }
+ { $$ = parse_type->builtin_short; }
| LONG INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long; }
+ { $$ = parse_type->builtin_long; }
| LONG SIGNED_KEYWORD INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long; }
+ { $$ = parse_type->builtin_long; }
| LONG SIGNED_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long; }
+ { $$ = parse_type->builtin_long; }
| SIGNED_KEYWORD LONG INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long; }
+ { $$ = parse_type->builtin_long; }
| UNSIGNED LONG INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long; }
+ { $$ = parse_type->builtin_unsigned_long; }
| LONG UNSIGNED INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long; }
+ { $$ = parse_type->builtin_unsigned_long; }
| LONG UNSIGNED
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long; }
+ { $$ = parse_type->builtin_unsigned_long; }
| LONG LONG
- { $$ = builtin_type (current_gdbarch)->builtin_long_long; }
+ { $$ = parse_type->builtin_long_long; }
| LONG LONG INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long_long; }
+ { $$ = parse_type->builtin_long_long; }
| LONG LONG SIGNED_KEYWORD INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long_long; }
+ { $$ = parse_type->builtin_long_long; }
| LONG LONG SIGNED_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long_long; }
+ { $$ = parse_type->builtin_long_long; }
| SIGNED_KEYWORD LONG LONG
- { $$ = builtin_type (current_gdbarch)->builtin_long_long; }
+ { $$ = parse_type->builtin_long_long; }
| SIGNED_KEYWORD LONG LONG INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long_long; }
+ { $$ = parse_type->builtin_long_long; }
| UNSIGNED LONG LONG
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long_long; }
+ { $$ = parse_type->builtin_unsigned_long_long; }
| UNSIGNED LONG LONG INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long_long; }
+ { $$ = parse_type->builtin_unsigned_long_long; }
| LONG LONG UNSIGNED
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long_long; }
+ { $$ = parse_type->builtin_unsigned_long_long; }
| LONG LONG UNSIGNED INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_long_long; }
+ { $$ = parse_type->builtin_unsigned_long_long; }
| SHORT INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_short; }
+ { $$ = parse_type->builtin_short; }
| SHORT SIGNED_KEYWORD INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_short; }
+ { $$ = parse_type->builtin_short; }
| SHORT SIGNED_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_short; }
+ { $$ = parse_type->builtin_short; }
| UNSIGNED SHORT INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_short; }
+ { $$ = parse_type->builtin_unsigned_short; }
| SHORT UNSIGNED
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_short; }
+ { $$ = parse_type->builtin_unsigned_short; }
| SHORT UNSIGNED INT_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_short; }
+ { $$ = parse_type->builtin_unsigned_short; }
| DOUBLE_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_double; }
+ { $$ = parse_type->builtin_double; }
| LONG DOUBLE_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_long_double; }
+ { $$ = parse_type->builtin_long_double; }
| STRUCT name
{ $$ = lookup_struct (copy_name ($2),
expression_context_block); }
| UNSIGNED typename
{ $$ = lookup_unsigned_typename (TYPE_NAME($2.type)); }
| UNSIGNED
- { $$ = builtin_type (current_gdbarch)->builtin_unsigned_int; }
+ { $$ = parse_type->builtin_unsigned_int; }
| SIGNED_KEYWORD typename
{ $$ = lookup_signed_typename (TYPE_NAME($2.type)); }
| SIGNED_KEYWORD
- { $$ = builtin_type (current_gdbarch)->builtin_int; }
+ { $$ = parse_type->builtin_int; }
/* It appears that this rule for templates is never
reduced; template recognition happens by lookahead
in the token processing code in yylex. */
{
$$.stoken.ptr = "int";
$$.stoken.length = 3;
- $$.type = builtin_type (current_gdbarch)->builtin_int;
+ $$.type = parse_type->builtin_int;
}
| LONG
{
$$.stoken.ptr = "long";
$$.stoken.length = 4;
- $$.type = builtin_type (current_gdbarch)->builtin_long;
+ $$.type = parse_type->builtin_long;
}
| SHORT
{
$$.stoken.ptr = "short";
$$.stoken.length = 5;
- $$.type = builtin_type (current_gdbarch)->builtin_short;
+ $$.type = parse_type->builtin_short;
}
;
{
p[len - 2] = '\0';
putithere->typed_val_decfloat.type
- = builtin_type (current_gdbarch)->builtin_decfloat;
+ = parse_type->builtin_decfloat;
decimal_from_string (putithere->typed_val_decfloat.val, 4, p);
p[len - 2] = 'd';
return DECFLOAT;
{
p[len - 2] = '\0';
putithere->typed_val_decfloat.type
- = builtin_type (current_gdbarch)->builtin_decdouble;
+ = parse_type->builtin_decdouble;
decimal_from_string (putithere->typed_val_decfloat.val, 8, p);
p[len - 2] = 'd';
return DECFLOAT;
{
p[len - 2] = '\0';
putithere->typed_val_decfloat.type
- = builtin_type (current_gdbarch)->builtin_declong;
+ = parse_type->builtin_declong;
decimal_from_string (putithere->typed_val_decfloat.val, 16, p);
p[len - 2] = 'd';
return DECFLOAT;
if (num == 1)
putithere->typed_val_float.type =
- builtin_type (current_gdbarch)->builtin_double;
+ parse_type->builtin_double;
if (num == 2 )
{
double. */
if (!strcasecmp (s, "f"))
putithere->typed_val_float.type =
- builtin_type (current_gdbarch)->builtin_float;
+ parse_type->builtin_float;
else if (!strcasecmp (s, "l"))
putithere->typed_val_float.type =
- builtin_type (current_gdbarch)->builtin_long_double;
+ parse_type->builtin_long_double;
else
{
free (s);
un = (ULONGEST)n >> 2;
if (long_p == 0
- && (un >> (gdbarch_int_bit (current_gdbarch) - 2)) == 0)
+ && (un >> (gdbarch_int_bit (parse_gdbarch) - 2)) == 0)
{
- high_bit = ((ULONGEST)1) << (gdbarch_int_bit (current_gdbarch) - 1);
+ high_bit = ((ULONGEST)1) << (gdbarch_int_bit (parse_gdbarch) - 1);
/* A large decimal (not hex or octal) constant (between INT_MAX
and UINT_MAX) is a long or unsigned long, according to ANSI,
int. This probably should be fixed. GCC gives a warning on
such constants. */
- unsigned_type = builtin_type (current_gdbarch)->builtin_unsigned_int;
- signed_type = builtin_type (current_gdbarch)->builtin_int;
+ unsigned_type = parse_type->builtin_unsigned_int;
+ signed_type = parse_type->builtin_int;
}
else if (long_p <= 1
- && (un >> (gdbarch_long_bit (current_gdbarch) - 2)) == 0)
+ && (un >> (gdbarch_long_bit (parse_gdbarch) - 2)) == 0)
{
- high_bit = ((ULONGEST)1) << (gdbarch_long_bit (current_gdbarch) - 1);
- unsigned_type = builtin_type (current_gdbarch)->builtin_unsigned_long;
- signed_type = builtin_type (current_gdbarch)->builtin_long;
+ high_bit = ((ULONGEST)1) << (gdbarch_long_bit (parse_gdbarch) - 1);
+ unsigned_type = parse_type->builtin_unsigned_long;
+ signed_type = parse_type->builtin_long;
}
else
{
int shift;
if (sizeof (ULONGEST) * HOST_CHAR_BIT
- < gdbarch_long_long_bit (current_gdbarch))
+ < gdbarch_long_long_bit (parse_gdbarch))
/* A long long does not fit in a LONGEST. */
shift = (sizeof (ULONGEST) * HOST_CHAR_BIT - 1);
else
- shift = (gdbarch_long_long_bit (current_gdbarch) - 1);
+ shift = (gdbarch_long_long_bit (parse_gdbarch) - 1);
high_bit = (ULONGEST) 1 << shift;
- unsigned_type = builtin_type (current_gdbarch)->builtin_unsigned_long_long;
- signed_type = builtin_type (current_gdbarch)->builtin_long_long;
+ unsigned_type = parse_type->builtin_unsigned_long_long;
+ signed_type = parse_type->builtin_long_long;
}
putithere->typed_val_int.val = n;
return INT;
}
+/* Temporary obstack used for holding strings. */
+static struct obstack tempbuf;
+static int tempbuf_init;
+
+/* Parse a C escape sequence. The initial backslash of the sequence
+ is at (*PTR)[-1]. *PTR will be updated to point to just after the
+ last character of the sequence. If OUTPUT is not NULL, the
+ translated form of the escape sequence will be written there. If
+ OUTPUT is NULL, no output is written and the call will only affect
+ *PTR. If an escape sequence is expressed in target bytes, then the
+ entire sequence will simply be copied to OUTPUT. Return 1 if any
+ character was emitted, 0 otherwise. */
+
+int
+c_parse_escape (char **ptr, struct obstack *output)
+{
+ char *tokptr = *ptr;
+ int result = 1;
+
+ /* Some escape sequences undergo character set conversion. Those we
+ translate here. */
+ switch (*tokptr)
+ {
+ /* Hex escapes do not undergo character set conversion, so keep
+ the escape sequence for later. */
+ case 'x':
+ if (output)
+ obstack_grow_str (output, "\\x");
+ ++tokptr;
+ if (!isxdigit (*tokptr))
+ error (_("\\x escape without a following hex digit"));
+ while (isxdigit (*tokptr))
+ {
+ if (output)
+ obstack_1grow (output, *tokptr);
+ ++tokptr;
+ }
+ break;
+
+ /* Octal escapes do not undergo character set conversion, so
+ keep the escape sequence for later. */
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ if (output)
+ obstack_grow_str (output, "\\");
+ while (isdigit (*tokptr) && *tokptr != '8' && *tokptr != '9')
+ {
+ if (output)
+ obstack_1grow (output, *tokptr);
+ ++tokptr;
+ }
+ break;
+
+ /* We handle UCNs later. We could handle them here, but that
+ would mean a spurious error in the case where the UCN could
+ be converted to the target charset but not the host
+ charset. */
+ case 'u':
+ case 'U':
+ {
+ char c = *tokptr;
+ int i, len = c == 'U' ? 8 : 4;
+ if (output)
+ {
+ obstack_1grow (output, '\\');
+ obstack_1grow (output, *tokptr);
+ }
+ ++tokptr;
+ if (!isxdigit (*tokptr))
+ error (_("\\%c escape without a following hex digit"), c);
+ for (i = 0; i < len && isxdigit (*tokptr); ++i)
+ {
+ if (output)
+ obstack_1grow (output, *tokptr);
+ ++tokptr;
+ }
+ }
+ break;
+
+ /* We must pass backslash through so that it does not
+ cause quoting during the second expansion. */
+ case '\\':
+ if (output)
+ obstack_grow_str (output, "\\\\");
+ ++tokptr;
+ break;
+
+ /* Escapes which undergo conversion. */
+ case 'a':
+ if (output)
+ obstack_1grow (output, '\a');
+ ++tokptr;
+ break;
+ case 'b':
+ if (output)
+ obstack_1grow (output, '\b');
+ ++tokptr;
+ break;
+ case 'f':
+ if (output)
+ obstack_1grow (output, '\f');
+ ++tokptr;
+ break;
+ case 'n':
+ if (output)
+ obstack_1grow (output, '\n');
+ ++tokptr;
+ break;
+ case 'r':
+ if (output)
+ obstack_1grow (output, '\r');
+ ++tokptr;
+ break;
+ case 't':
+ if (output)
+ obstack_1grow (output, '\t');
+ ++tokptr;
+ break;
+ case 'v':
+ if (output)
+ obstack_1grow (output, '\v');
+ ++tokptr;
+ break;
+
+ /* GCC extension. */
+ case 'e':
+ if (output)
+ obstack_1grow (output, HOST_ESCAPE_CHAR);
+ ++tokptr;
+ break;
+
+ /* Backslash-newline expands to nothing at all. */
+ case '\n':
+ ++tokptr;
+ result = 0;
+ break;
+
+ /* A few escapes just expand to the character itself. */
+ case '\'':
+ case '\"':
+ case '?':
+ /* GCC extensions. */
+ case '(':
+ case '{':
+ case '[':
+ case '%':
+ /* Unrecognized escapes turn into the character itself. */
+ default:
+ if (output)
+ obstack_1grow (output, *tokptr);
+ ++tokptr;
+ break;
+ }
+ *ptr = tokptr;
+ return result;
+}
+
+/* Parse a string or character literal from TOKPTR. The string or
+ character may be wide or unicode. *OUTPTR is set to just after the
+ end of the literal in the input string. The resulting token is
+ stored in VALUE. This returns a token value, either STRING or
+ CHAR, depending on what was parsed. *HOST_CHARS is set to the
+ number of host characters in the literal. */
+static int
+parse_string_or_char (char *tokptr, char **outptr, struct typed_stoken *value,
+ int *host_chars)
+{
+ int quote, i;
+ enum c_string_type type;
+
+ /* Build the gdb internal form of the input string in tempbuf. Note
+ that the buffer is null byte terminated *only* for the
+ convenience of debugging gdb itself and printing the buffer
+ contents when the buffer contains no embedded nulls. Gdb does
+ not depend upon the buffer being null byte terminated, it uses
+ the length string instead. This allows gdb to handle C strings
+ (as well as strings in other languages) with embedded null
+ bytes */
+
+ if (!tempbuf_init)
+ tempbuf_init = 1;
+ else
+ obstack_free (&tempbuf, NULL);
+ obstack_init (&tempbuf);
+
+ /* Record the string type. */
+ if (*tokptr == 'L')
+ {
+ type = C_WIDE_STRING;
+ ++tokptr;
+ }
+ else if (*tokptr == 'u')
+ {
+ type = C_STRING_16;
+ ++tokptr;
+ }
+ else if (*tokptr == 'U')
+ {
+ type = C_STRING_32;
+ ++tokptr;
+ }
+ else
+ type = C_STRING;
+
+ /* Skip the quote. */
+ quote = *tokptr;
+ if (quote == '\'')
+ type |= C_CHAR;
+ ++tokptr;
+
+ *host_chars = 0;
+
+ while (*tokptr)
+ {
+ char c = *tokptr;
+ if (c == '\\')
+ {
+ ++tokptr;
+ *host_chars += c_parse_escape (&tokptr, &tempbuf);
+ }
+ else if (c == quote)
+ break;
+ else
+ {
+ obstack_1grow (&tempbuf, c);
+ ++tokptr;
+ /* FIXME: this does the wrong thing with multi-byte host
+ characters. We could use mbrlen here, but that would
+ make "set host-charset" a bit less useful. */
+ ++*host_chars;
+ }
+ }
+
+ if (*tokptr != quote)
+ {
+ if (quote == '"')
+ error ("Unterminated string in expression.");
+ else
+ error ("Unmatched single quote.");
+ }
+ ++tokptr;
+
+ value->type = type;
+ value->ptr = obstack_base (&tempbuf);
+ value->length = obstack_object_size (&tempbuf);
+
+ *outptr = tokptr;
+
+ return quote == '"' ? STRING : CHAR;
+}
+
struct token
{
char *operator;
int token;
enum exp_opcode opcode;
+ int cxx_only;
};
static const struct token tokentab3[] =
{
- {">>=", ASSIGN_MODIFY, BINOP_RSH},
- {"<<=", ASSIGN_MODIFY, BINOP_LSH}
+ {">>=", ASSIGN_MODIFY, BINOP_RSH, 0},
+ {"<<=", ASSIGN_MODIFY, BINOP_LSH, 0}
};
static const struct token tokentab2[] =
{
- {"+=", ASSIGN_MODIFY, BINOP_ADD},
- {"-=", ASSIGN_MODIFY, BINOP_SUB},
- {"*=", ASSIGN_MODIFY, BINOP_MUL},
- {"/=", ASSIGN_MODIFY, BINOP_DIV},
- {"%=", ASSIGN_MODIFY, BINOP_REM},
- {"|=", ASSIGN_MODIFY, BINOP_BITWISE_IOR},
- {"&=", ASSIGN_MODIFY, BINOP_BITWISE_AND},
- {"^=", ASSIGN_MODIFY, BINOP_BITWISE_XOR},
- {"++", INCREMENT, BINOP_END},
- {"--", DECREMENT, BINOP_END},
- {"->", ARROW, BINOP_END},
- {"&&", ANDAND, BINOP_END},
- {"||", OROR, BINOP_END},
- {"::", COLONCOLON, BINOP_END},
- {"<<", LSH, BINOP_END},
- {">>", RSH, BINOP_END},
- {"==", EQUAL, BINOP_END},
- {"!=", NOTEQUAL, BINOP_END},
- {"<=", LEQ, BINOP_END},
- {">=", GEQ, BINOP_END}
+ {"+=", ASSIGN_MODIFY, BINOP_ADD, 0},
+ {"-=", ASSIGN_MODIFY, BINOP_SUB, 0},
+ {"*=", ASSIGN_MODIFY, BINOP_MUL, 0},
+ {"/=", ASSIGN_MODIFY, BINOP_DIV, 0},
+ {"%=", ASSIGN_MODIFY, BINOP_REM, 0},
+ {"|=", ASSIGN_MODIFY, BINOP_BITWISE_IOR, 0},
+ {"&=", ASSIGN_MODIFY, BINOP_BITWISE_AND, 0},
+ {"^=", ASSIGN_MODIFY, BINOP_BITWISE_XOR, 0},
+ {"++", INCREMENT, BINOP_END, 0},
+ {"--", DECREMENT, BINOP_END, 0},
+ {"->", ARROW, BINOP_END, 0},
+ {"&&", ANDAND, BINOP_END, 0},
+ {"||", OROR, BINOP_END, 0},
+ {"::", COLONCOLON, BINOP_END, 0},
+ {"<<", LSH, BINOP_END, 0},
+ {">>", RSH, BINOP_END, 0},
+ {"==", EQUAL, BINOP_END, 0},
+ {"!=", NOTEQUAL, BINOP_END, 0},
+ {"<=", LEQ, BINOP_END, 0},
+ {">=", GEQ, BINOP_END, 0}
};
+/* Identifier-like tokens. */
+static const struct token ident_tokens[] =
+ {
+ {"unsigned", UNSIGNED, OP_NULL, 0},
+ {"template", TEMPLATE, OP_NULL, 1},
+ {"volatile", VOLATILE_KEYWORD, OP_NULL, 0},
+ {"struct", STRUCT, OP_NULL, 0},
+ {"signed", SIGNED_KEYWORD, OP_NULL, 0},
+ {"sizeof", SIZEOF, OP_NULL, 0},
+ {"double", DOUBLE_KEYWORD, OP_NULL, 0},
+ {"false", FALSEKEYWORD, OP_NULL, 1},
+ {"class", CLASS, OP_NULL, 1},
+ {"union", UNION, OP_NULL, 0},
+ {"short", SHORT, OP_NULL, 0},
+ {"const", CONST_KEYWORD, OP_NULL, 0},
+ {"enum", ENUM, OP_NULL, 0},
+ {"long", LONG, OP_NULL, 0},
+ {"true", TRUEKEYWORD, OP_NULL, 1},
+ {"int", INT_KEYWORD, OP_NULL, 0},
+
+ {"and", ANDAND, BINOP_END, 1},
+ {"and_eq", ASSIGN_MODIFY, BINOP_BITWISE_AND, 1},
+ {"bitand", '&', OP_NULL, 1},
+ {"bitor", '|', OP_NULL, 1},
+ {"compl", '~', OP_NULL, 1},
+ {"not", '!', OP_NULL, 1},
+ {"not_eq", NOTEQUAL, BINOP_END, 1},
+ {"or", OROR, BINOP_END, 1},
+ {"or_eq", ASSIGN_MODIFY, BINOP_BITWISE_IOR, 1},
+ {"xor", '^', OP_NULL, 1},
+ {"xor_eq", ASSIGN_MODIFY, BINOP_BITWISE_XOR, 1}
+ };
+
+/* When we find that lexptr (the global var defined in parse.c) is
+ pointing at a macro invocation, we expand the invocation, and call
+ scan_macro_expansion to save the old lexptr here and point lexptr
+ into the expanded text. When we reach the end of that, we call
+ end_macro_expansion to pop back to the value we saved here. The
+ macro expansion code promises to return only fully-expanded text,
+ so we don't need to "push" more than one level.
+
+ This is disgusting, of course. It would be cleaner to do all macro
+ expansion beforehand, and then hand that to lexptr. But we don't
+ really know where the expression ends. Remember, in a command like
+
+ (gdb) break *ADDRESS if CONDITION
+
+ we evaluate ADDRESS in the scope of the current frame, but we
+ evaluate CONDITION in the scope of the breakpoint's location. So
+ it's simply wrong to try to macro-expand the whole thing at once. */
+static char *macro_original_text;
+
+/* We save all intermediate macro expansions on this obstack for the
+ duration of a single parse. The expansion text may sometimes have
+ to live past the end of the expansion, due to yacc lookahead.
+ Rather than try to be clever about saving the data for a single
+ token, we simply keep it all and delete it after parsing has
+ completed. */
+static struct obstack expansion_obstack;
+
+static void
+scan_macro_expansion (char *expansion)
+{
+ char *copy;
+
+ /* We'd better not be trying to push the stack twice. */
+ gdb_assert (! macro_original_text);
+
+ /* Copy to the obstack, and then free the intermediate
+ expansion. */
+ copy = obstack_copy0 (&expansion_obstack, expansion, strlen (expansion));
+ xfree (expansion);
+
+ /* Save the old lexptr value, so we can return to it when we're done
+ parsing the expanded text. */
+ macro_original_text = lexptr;
+ lexptr = copy;
+}
+
+
+static int
+scanning_macro_expansion (void)
+{
+ return macro_original_text != 0;
+}
+
+
+static void
+finished_macro_expansion (void)
+{
+ /* There'd better be something to pop back to. */
+ gdb_assert (macro_original_text);
+
+ /* Pop back to the original text. */
+ lexptr = macro_original_text;
+ macro_original_text = 0;
+}
+
+
+static void
+scan_macro_cleanup (void *dummy)
+{
+ if (macro_original_text)
+ finished_macro_expansion ();
+
+ obstack_free (&expansion_obstack, NULL);
+}
+
+
+/* The scope used for macro expansion. */
+static struct macro_scope *expression_macro_scope;
+
/* This is set if a NAME token appeared at the very end of the input
string, with no whitespace separating the name from the EOF. This
is used only when parsing to do field name completion. */
int namelen;
unsigned int i;
char *tokstart;
- char *tokptr;
- int tempbufindex;
- static char *tempbuf;
- static int tempbufsize;
- char * token_string = NULL;
- int class_prefix = 0;
int saw_structop = last_was_structop;
+ char *copy;
last_was_structop = 0;
if (! scanning_macro_expansion ())
{
char *expanded = macro_expand_next (&lexptr,
- expression_macro_lookup_func,
- expression_macro_lookup_baton);
+ standard_macro_lookup,
+ expression_macro_scope);
if (expanded)
scan_macro_expansion (expanded);
lexptr++;
goto retry;
- case '\'':
- /* We either have a character constant ('0' or '\177' for example)
- or we have a quoted symbol reference ('foo(int,int)' in C++
- for example). */
- lexptr++;
- c = *lexptr++;
- if (c == '\\')
- c = parse_escape (&lexptr);
- else if (c == '\'')
- error ("Empty character constant.");
- else if (! host_char_to_target (c, &c))
- {
- int toklen = lexptr - tokstart + 1;
- char *tok = alloca (toklen + 1);
- memcpy (tok, tokstart, toklen);
- tok[toklen] = '\0';
- error ("There is no character corresponding to %s in the target "
- "character set `%s'.", tok, target_charset ());
- }
-
- yylval.typed_val_int.val = c;
- yylval.typed_val_int.type = builtin_type (current_gdbarch)->builtin_char;
-
- c = *lexptr++;
- if (c != '\'')
- {
- namelen = skip_quoted (tokstart) - tokstart;
- if (namelen > 2)
- {
- lexptr = tokstart + namelen;
- if (lexptr[-1] != '\'')
- error ("Unmatched single quote.");
- namelen -= 2;
- tokstart++;
- goto tryname;
- }
- error ("Invalid character constant.");
- }
- return INT;
-
case '(':
paren_depth++;
lexptr++;
lexptr++;
return c;
+ case 'L':
+ case 'u':
+ case 'U':
+ if (tokstart[1] != '"' && tokstart[1] != '\'')
+ break;
+ /* Fall through. */
+ case '\'':
case '"':
-
- /* Build the gdb internal form of the input string in tempbuf,
- translating any standard C escape forms seen. Note that the
- buffer is null byte terminated *only* for the convenience of
- debugging gdb itself and printing the buffer contents when
- the buffer contains no embedded nulls. Gdb does not depend
- upon the buffer being null byte terminated, it uses the length
- string instead. This allows gdb to handle C strings (as well
- as strings in other languages) with embedded null bytes */
-
- tokptr = ++tokstart;
- tempbufindex = 0;
-
- do {
- char *char_start_pos = tokptr;
-
- /* Grow the static temp buffer if necessary, including allocating
- the first one on demand. */
- if (tempbufindex + 1 >= tempbufsize)
- {
- tempbuf = (char *) realloc (tempbuf, tempbufsize += 64);
- }
- switch (*tokptr)
+ {
+ int host_len;
+ int result = parse_string_or_char (tokstart, &lexptr, &yylval.tsval,
+ &host_len);
+ if (result == CHAR)
{
- case '\0':
- case '"':
- /* Do nothing, loop will terminate. */
- break;
- case '\\':
- tokptr++;
- c = parse_escape (&tokptr);
- if (c == -1)
+ if (host_len == 0)
+ error ("Empty character constant.");
+ else if (host_len > 2 && c == '\'')
{
- continue;
+ ++tokstart;
+ namelen = lexptr - tokstart - 1;
+ goto tryname;
}
- tempbuf[tempbufindex++] = c;
- break;
- default:
- c = *tokptr++;
- if (! host_char_to_target (c, &c))
- {
- int len = tokptr - char_start_pos;
- char *copy = alloca (len + 1);
- memcpy (copy, char_start_pos, len);
- copy[len] = '\0';
-
- error ("There is no character corresponding to `%s' "
- "in the target character set `%s'.",
- copy, target_charset ());
- }
- tempbuf[tempbufindex++] = c;
- break;
+ else if (host_len > 1)
+ error ("Invalid character constant.");
}
- } while ((*tokptr != '"') && (*tokptr != '\0'));
- if (*tokptr++ != '"')
- {
- error ("Unterminated string in expression.");
- }
- tempbuf[tempbufindex] = '\0'; /* See note above */
- yylval.sval.ptr = tempbuf;
- yylval.sval.length = tempbufindex;
- lexptr = tokptr;
- return (STRING);
+ return result;
+ }
}
if (!(c == '_' || c == '$'
tryname:
- /* Catch specific keywords. Should be done with a data structure. */
- switch (namelen)
- {
- case 8:
- if (strncmp (tokstart, "unsigned", 8) == 0)
- return UNSIGNED;
- if (current_language->la_language == language_cplus
- && strncmp (tokstart, "template", 8) == 0)
- return TEMPLATE;
- if (strncmp (tokstart, "volatile", 8) == 0)
- return VOLATILE_KEYWORD;
- break;
- case 6:
- if (strncmp (tokstart, "struct", 6) == 0)
- return STRUCT;
- if (strncmp (tokstart, "signed", 6) == 0)
- return SIGNED_KEYWORD;
- if (strncmp (tokstart, "sizeof", 6) == 0)
- return SIZEOF;
- if (strncmp (tokstart, "double", 6) == 0)
- return DOUBLE_KEYWORD;
- break;
- case 5:
- if (current_language->la_language == language_cplus)
- {
- if (strncmp (tokstart, "false", 5) == 0)
- return FALSEKEYWORD;
- if (strncmp (tokstart, "class", 5) == 0)
- return CLASS;
- }
- if (strncmp (tokstart, "union", 5) == 0)
- return UNION;
- if (strncmp (tokstart, "short", 5) == 0)
- return SHORT;
- if (strncmp (tokstart, "const", 5) == 0)
- return CONST_KEYWORD;
- break;
- case 4:
- if (strncmp (tokstart, "enum", 4) == 0)
- return ENUM;
- if (strncmp (tokstart, "long", 4) == 0)
- return LONG;
- if (current_language->la_language == language_cplus)
- {
- if (strncmp (tokstart, "true", 4) == 0)
- return TRUEKEYWORD;
- }
- break;
- case 3:
- if (strncmp (tokstart, "int", 3) == 0)
- return INT_KEYWORD;
- break;
- default:
- break;
- }
-
yylval.sval.ptr = tokstart;
yylval.sval.length = namelen;
+ /* Catch specific keywords. */
+ copy = copy_name (yylval.sval);
+ for (i = 0; i < sizeof ident_tokens / sizeof ident_tokens[0]; i++)
+ if (strcmp (copy, ident_tokens[i].operator) == 0)
+ {
+ if (ident_tokens[i].cxx_only
+ && parse_language->la_language != language_cplus)
+ break;
+
+ /* It is ok to always set this, even though we don't always
+ strictly need to. */
+ yylval.opcode = ident_tokens[i].opcode;
+ return ident_tokens[i].token;
+ }
+
if (*tokstart == '$')
{
write_dollar_variable (yylval.sval);
currently as names of types; NAME for other symbols.
The caller is not constrained to care about the distinction. */
{
- char *tmp = copy_name (yylval.sval);
struct symbol *sym;
int is_a_field_of_this = 0;
int hextype;
- sym = lookup_symbol (tmp, expression_context_block,
+ sym = lookup_symbol (copy, expression_context_block,
VAR_DOMAIN,
- current_language->la_language == language_cplus
+ parse_language->la_language == language_cplus
? &is_a_field_of_this : (int *) NULL);
/* Call lookup_symtab, not lookup_partial_symtab, in case there are
no psymtabs (coff, xcoff, or some future change to blow away the
{ /* See if it's a file name. */
struct symtab *symtab;
- symtab = lookup_symtab (tmp);
+ symtab = lookup_symtab (copy);
if (symtab)
{
return TYPENAME;
}
yylval.tsym.type
- = language_lookup_primitive_type_by_name (current_language,
- current_gdbarch, tmp);
+ = language_lookup_primitive_type_by_name (parse_language,
+ parse_gdbarch, copy);
if (yylval.tsym.type != NULL)
return TYPENAME;
int
c_parse (void)
{
+ int result;
+ struct cleanup *back_to = make_cleanup (free_current_contents,
+ &expression_macro_scope);
+
+ /* Set up the scope for macro expansion. */
+ expression_macro_scope = NULL;
+
+ if (expression_context_block)
+ expression_macro_scope
+ = sal_macro_scope (find_pc_line (expression_context_pc, 0));
+ else
+ expression_macro_scope = default_macro_scope ();
+ if (! expression_macro_scope)
+ expression_macro_scope = user_macro_scope ();
+
+ /* Initialize macro expansion code. */
+ obstack_init (&expansion_obstack);
+ gdb_assert (! macro_original_text);
+ make_cleanup (scan_macro_cleanup, 0);
+
+ /* Initialize some state used by the lexer. */
last_was_structop = 0;
saw_name_at_eof = 0;
- return yyparse ();
+
+ result = yyparse ();
+ do_cleanups (back_to);
+ return result;
}
+
void
yyerror (msg)
char *msg;
projects / deliverable / binutils-gdb.git / blobdiff