+ /* 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':
+ {
+ int i;
+ if (output)
+ obstack_grow_str (output, "\\");
+ for (i = 0;
+ i < 3 && isdigit (*tokptr) && *tokptr != '8' && *tokptr != '9';
+ ++i)
+ {
+ 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 (const char *tokptr, const char **outptr,
+ struct typed_stoken *value, int *host_chars)
+{
+ int quote;
+ enum c_string_type type;
+ int is_objc = 0;
+
+ /* 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 if (*tokptr == '@')
+ {
+ /* An Objective C string. */
+ is_objc = 1;
+ type = C_STRING;
+ ++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 == '"' ? (is_objc ? NSSTRING : STRING) : CHAR;
+}
+
+/* This is used to associate some attributes with a token. */
+
+enum token_flags
+{
+ /* If this bit is set, the token is C++-only. */
+
+ FLAG_CXX = 1,
+
+ /* If this bit is set, the token is conditional: if there is a
+ symbol of the same name, then the token is a symbol; otherwise,
+ the token is a keyword. */
+
+ FLAG_SHADOW = 2
+};
+
+struct token
+{
+ char *oper;
+ int token;
+ enum exp_opcode opcode;
+ enum token_flags flags;
+};
+
+static const struct token tokentab3[] =
+ {
+ {">>=", ASSIGN_MODIFY, BINOP_RSH, 0},
+ {"<<=", ASSIGN_MODIFY, BINOP_LSH, 0},
+ {"->*", ARROW_STAR, BINOP_END, FLAG_CXX},
+ {"...", DOTDOTDOT, BINOP_END, 0}
+ };
+
+static const struct token tokentab2[] =
+ {
+ {"+=", 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},
+ /* "::" is *not* only C++: gdb overrides its meaning in several
+ different ways, e.g., 'filename'::func, function::variable. */
+ {"::", 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},
+ {".*", DOT_STAR, BINOP_END, FLAG_CXX}
+ };
+
+/* Identifier-like tokens. */
+static const struct token ident_tokens[] =
+ {
+ {"unsigned", UNSIGNED, OP_NULL, 0},
+ {"template", TEMPLATE, OP_NULL, FLAG_CXX},
+ {"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, FLAG_CXX},
+ {"class", CLASS, OP_NULL, FLAG_CXX},
+ {"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, FLAG_CXX},
+ {"int", INT_KEYWORD, OP_NULL, 0},
+ {"new", NEW, OP_NULL, FLAG_CXX},
+ {"delete", DELETE, OP_NULL, FLAG_CXX},
+ {"operator", OPERATOR, OP_NULL, FLAG_CXX},
+
+ {"and", ANDAND, BINOP_END, FLAG_CXX},
+ {"and_eq", ASSIGN_MODIFY, BINOP_BITWISE_AND, FLAG_CXX},
+ {"bitand", '&', OP_NULL, FLAG_CXX},
+ {"bitor", '|', OP_NULL, FLAG_CXX},
+ {"compl", '~', OP_NULL, FLAG_CXX},
+ {"not", '!', OP_NULL, FLAG_CXX},
+ {"not_eq", NOTEQUAL, BINOP_END, FLAG_CXX},
+ {"or", OROR, BINOP_END, FLAG_CXX},
+ {"or_eq", ASSIGN_MODIFY, BINOP_BITWISE_IOR, FLAG_CXX},
+ {"xor", '^', OP_NULL, FLAG_CXX},
+ {"xor_eq", ASSIGN_MODIFY, BINOP_BITWISE_XOR, FLAG_CXX},
+
+ {"const_cast", CONST_CAST, OP_NULL, FLAG_CXX },
+ {"dynamic_cast", DYNAMIC_CAST, OP_NULL, FLAG_CXX },
+ {"static_cast", STATIC_CAST, OP_NULL, FLAG_CXX },
+ {"reinterpret_cast", REINTERPRET_CAST, OP_NULL, FLAG_CXX },
+
+ {"__typeof__", TYPEOF, OP_TYPEOF, 0 },
+ {"__typeof", TYPEOF, OP_TYPEOF, 0 },
+ {"typeof", TYPEOF, OP_TYPEOF, FLAG_SHADOW },
+ {"__decltype", DECLTYPE, OP_DECLTYPE, FLAG_CXX },
+ {"decltype", DECLTYPE, OP_DECLTYPE, FLAG_CXX | FLAG_SHADOW },
+
+ {"typeid", TYPEID, OP_TYPEID, FLAG_CXX}
+ };
+
+/* 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