1 /* Rust expression parsing for GDB, the GNU debugger.
3 Copyright (C) 2016-2021 Free Software Foundation, Inc.
5 This file is part of GDB.
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.
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.
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/>. */
24 #include "cp-support.h"
25 #include "gdb_obstack.h"
26 #include "gdb_regex.h"
27 #include "rust-lang.h"
28 #include "parser-defs.h"
29 #include "gdbsupport/selftest.h"
36 /* A regular expression for matching Rust numbers. This is split up
37 since it is very long and this gives us a way to comment the
40 static const char number_regex_text
[] =
41 /* subexpression 1: allows use of alternation, otherwise uninteresting */
43 /* First comes floating point. */
44 /* Recognize number after the decimal point, with optional
45 exponent and optional type suffix.
46 subexpression 2: allows "?", otherwise uninteresting
47 subexpression 3: if present, type suffix
49 "[0-9][0-9_]*\\.[0-9][0-9_]*([eE][-+]?[0-9][0-9_]*)?(f32|f64)?"
52 /* Recognize exponent without decimal point, with optional type
54 subexpression 4: if present, type suffix
57 "[0-9][0-9_]*[eE][-+]?[0-9][0-9_]*(f32|f64)?"
59 /* "23." is a valid floating point number, but "23.e5" and
60 "23.f32" are not. So, handle the trailing-. case
64 /* Finally come integers.
65 subexpression 5: text of integer
66 subexpression 6: if present, type suffix
67 subexpression 7: allows use of alternation, otherwise uninteresting
71 "(0x[a-fA-F0-9_]+|0o[0-7_]+|0b[01_]+|[0-9][0-9_]*)"
72 "([iu](size|8|16|32|64))?"
74 /* The number of subexpressions to allocate space for, including the
75 "0th" whole match subexpression. */
76 #define NUM_SUBEXPRESSIONS 8
78 /* The compiled number-matching regex. */
80 static regex_t number_regex
;
82 /* The kinds of tokens. Note that single-character tokens are
83 represented by themselves, so for instance '[' is a token. */
86 /* Make sure to start after any ASCII character. */
110 /* Operator tokens. */
125 /* A typed integer constant. */
133 /* A typed floating point constant. */
135 struct typed_val_float
141 /* A struct of this type is used to describe a token. */
147 enum exp_opcode opcode
;
150 /* Identifier tokens. */
152 static const struct token_info identifier_tokens
[] =
154 { "as", KW_AS
, OP_NULL
},
155 { "false", KW_FALSE
, OP_NULL
},
156 { "if", 0, OP_NULL
},
157 { "mut", KW_MUT
, OP_NULL
},
158 { "const", KW_CONST
, OP_NULL
},
159 { "self", KW_SELF
, OP_NULL
},
160 { "super", KW_SUPER
, OP_NULL
},
161 { "true", KW_TRUE
, OP_NULL
},
162 { "extern", KW_EXTERN
, OP_NULL
},
163 { "fn", KW_FN
, OP_NULL
},
164 { "sizeof", KW_SIZEOF
, OP_NULL
},
167 /* Operator tokens, sorted longest first. */
169 static const struct token_info operator_tokens
[] =
171 { ">>=", COMPOUND_ASSIGN
, BINOP_RSH
},
172 { "<<=", COMPOUND_ASSIGN
, BINOP_LSH
},
174 { "<<", LSH
, OP_NULL
},
175 { ">>", RSH
, OP_NULL
},
176 { "&&", ANDAND
, OP_NULL
},
177 { "||", OROR
, OP_NULL
},
178 { "==", EQEQ
, OP_NULL
},
179 { "!=", NOTEQ
, OP_NULL
},
180 { "<=", LTEQ
, OP_NULL
},
181 { ">=", GTEQ
, OP_NULL
},
182 { "+=", COMPOUND_ASSIGN
, BINOP_ADD
},
183 { "-=", COMPOUND_ASSIGN
, BINOP_SUB
},
184 { "*=", COMPOUND_ASSIGN
, BINOP_MUL
},
185 { "/=", COMPOUND_ASSIGN
, BINOP_DIV
},
186 { "%=", COMPOUND_ASSIGN
, BINOP_REM
},
187 { "&=", COMPOUND_ASSIGN
, BINOP_BITWISE_AND
},
188 { "|=", COMPOUND_ASSIGN
, BINOP_BITWISE_IOR
},
189 { "^=", COMPOUND_ASSIGN
, BINOP_BITWISE_XOR
},
190 { "..=", DOTDOTEQ
, OP_NULL
},
192 { "::", COLONCOLON
, OP_NULL
},
193 { "..", DOTDOT
, OP_NULL
},
194 { "->", ARROW
, OP_NULL
}
197 /* An instance of this is created before parsing, and destroyed when
198 parsing is finished. */
202 explicit rust_parser (struct parser_state
*state
)
207 DISABLE_COPY_AND_ASSIGN (rust_parser
);
209 /* Return the parser's language. */
210 const struct language_defn
*language () const
212 return pstate
->language ();
215 /* Return the parser's gdbarch. */
216 struct gdbarch
*arch () const
218 return pstate
->gdbarch ();
221 /* A helper to look up a Rust type, or fail. This only works for
222 types defined by rust_language_arch_info. */
224 struct type
*get_type (const char *name
)
228 type
= language_lookup_primitive_type (language (), arch (), name
);
230 error (_("Could not find Rust type %s"), name
);
234 std::string
crate_name (const std::string
&name
);
235 std::string
super_name (const std::string
&ident
, unsigned int n_supers
);
237 int lex_character ();
240 int lex_identifier ();
241 uint32_t lex_hex (int min
, int max
);
242 uint32_t lex_escape (int is_byte
);
244 int lex_one_token ();
245 void push_back (char c
);
247 /* The main interface to lexing. Lexes one token and updates the
251 current_token
= lex_one_token ();
254 /* Assuming the current token is TYPE, lex the next token. */
255 void assume (int type
)
257 gdb_assert (current_token
== type
);
261 /* Require the single-character token C, and lex the next token; or
262 throw an exception. */
263 void require (char type
)
265 if (current_token
!= type
)
266 error (_("'%c' expected"), type
);
270 /* Entry point for all parsing. */
271 operation_up
parse_entry_point ()
274 return parse_expr ();
277 operation_up
parse_tuple ();
278 operation_up
parse_array ();
279 operation_up
name_to_operation (const std::string
&name
);
280 operation_up
parse_struct_expr (struct type
*type
);
281 operation_up
parse_binop (bool required
);
282 operation_up
parse_range ();
283 operation_up
parse_expr ();
284 operation_up
parse_sizeof ();
285 operation_up
parse_addr ();
286 operation_up
parse_field (operation_up
&&);
287 operation_up
parse_index (operation_up
&&);
288 std::vector
<operation_up
> parse_paren_args ();
289 operation_up
parse_call (operation_up
&&);
290 std::vector
<struct type
*> parse_type_list ();
291 std::vector
<struct type
*> parse_maybe_type_list ();
292 struct type
*parse_array_type ();
293 struct type
*parse_slice_type ();
294 struct type
*parse_pointer_type ();
295 struct type
*parse_function_type ();
296 struct type
*parse_tuple_type ();
297 struct type
*parse_type ();
298 std::string
parse_path (bool for_expr
);
299 operation_up
parse_string ();
300 operation_up
parse_tuple_struct (struct type
*type
);
301 operation_up
parse_path_expr ();
302 operation_up
parse_atom (bool required
);
304 void update_innermost_block (struct block_symbol sym
);
305 struct block_symbol
lookup_symbol (const char *name
,
306 const struct block
*block
,
307 const domain_enum domain
);
308 struct type
*rust_lookup_type (const char *name
);
310 /* Clear some state. This is only used for testing. */
312 void reset (const char *input
)
314 pstate
->prev_lexptr
= nullptr;
315 pstate
->lexptr
= input
;
318 current_int_val
= {};
319 current_float_val
= {};
320 current_string_val
= {};
321 current_opcode
= OP_NULL
;
323 #endif /* GDB_SELF_TEST */
325 /* Return the token's string value as a string. */
326 std::string
get_string () const
328 return std::string (current_string_val
.ptr
, current_string_val
.length
);
331 /* A pointer to this is installed globally. */
332 auto_obstack obstack
;
334 /* The parser state gdb gave us. */
335 struct parser_state
*pstate
;
337 /* Depth of parentheses. */
340 /* The current token's type. */
341 int current_token
= 0;
342 /* The current token's payload, if any. */
343 typed_val_int current_int_val
{};
344 typed_val_float current_float_val
{};
345 struct stoken current_string_val
{};
346 enum exp_opcode current_opcode
= OP_NULL
;
348 /* When completing, this may be set to the field operation to
350 operation_up completion_op
;
353 /* Return an string referring to NAME, but relative to the crate's
357 rust_parser::crate_name (const std::string
&name
)
359 std::string crate
= rust_crate_for_block (pstate
->expression_context_block
);
362 error (_("Could not find crate for current location"));
363 return "::" + crate
+ "::" + name
;
366 /* Return a string referring to a "super::" qualified name. IDENT is
367 the base name and N_SUPERS is how many "super::"s were provided.
368 N_SUPERS can be zero. */
371 rust_parser::super_name (const std::string
&ident
, unsigned int n_supers
)
373 const char *scope
= block_scope (pstate
->expression_context_block
);
376 if (scope
[0] == '\0')
377 error (_("Couldn't find namespace scope for self::"));
382 std::vector
<int> offsets
;
383 unsigned int current_len
;
385 current_len
= cp_find_first_component (scope
);
386 while (scope
[current_len
] != '\0')
388 offsets
.push_back (current_len
);
389 gdb_assert (scope
[current_len
] == ':');
392 current_len
+= cp_find_first_component (scope
396 len
= offsets
.size ();
398 error (_("Too many super:: uses from '%s'"), scope
);
400 offset
= offsets
[len
- n_supers
];
403 offset
= strlen (scope
);
405 return "::" + std::string (scope
, offset
) + "::" + ident
;
408 /* A helper to appropriately munge NAME and BLOCK depending on the
409 presence of a leading "::". */
412 munge_name_and_block (const char **name
, const struct block
**block
)
414 /* If it is a global reference, skip the current block in favor of
416 if (startswith (*name
, "::"))
419 *block
= block_static_block (*block
);
423 /* Like lookup_symbol, but handles Rust namespace conventions, and
424 doesn't require field_of_this_result. */
427 rust_parser::lookup_symbol (const char *name
, const struct block
*block
,
428 const domain_enum domain
)
430 struct block_symbol result
;
432 munge_name_and_block (&name
, &block
);
434 result
= ::lookup_symbol (name
, block
, domain
, NULL
);
435 if (result
.symbol
!= NULL
)
436 update_innermost_block (result
);
440 /* Look up a type, following Rust namespace conventions. */
443 rust_parser::rust_lookup_type (const char *name
)
445 struct block_symbol result
;
448 const struct block
*block
= pstate
->expression_context_block
;
449 munge_name_and_block (&name
, &block
);
451 result
= ::lookup_symbol (name
, block
, STRUCT_DOMAIN
, NULL
);
452 if (result
.symbol
!= NULL
)
454 update_innermost_block (result
);
455 return SYMBOL_TYPE (result
.symbol
);
458 type
= lookup_typename (language (), name
, NULL
, 1);
462 /* Last chance, try a built-in type. */
463 return language_lookup_primitive_type (language (), arch (), name
);
466 /* A helper that updates the innermost block as appropriate. */
469 rust_parser::update_innermost_block (struct block_symbol sym
)
471 if (symbol_read_needs_frame (sym
.symbol
))
472 pstate
->block_tracker
->update (sym
);
475 /* Lex a hex number with at least MIN digits and at most MAX
479 rust_parser::lex_hex (int min
, int max
)
483 /* We only want to stop at MAX if we're lexing a byte escape. */
484 int check_max
= min
== max
;
486 while ((check_max
? len
<= max
: 1)
487 && ((pstate
->lexptr
[0] >= 'a' && pstate
->lexptr
[0] <= 'f')
488 || (pstate
->lexptr
[0] >= 'A' && pstate
->lexptr
[0] <= 'F')
489 || (pstate
->lexptr
[0] >= '0' && pstate
->lexptr
[0] <= '9')))
492 if (pstate
->lexptr
[0] >= 'a' && pstate
->lexptr
[0] <= 'f')
493 result
= result
+ 10 + pstate
->lexptr
[0] - 'a';
494 else if (pstate
->lexptr
[0] >= 'A' && pstate
->lexptr
[0] <= 'F')
495 result
= result
+ 10 + pstate
->lexptr
[0] - 'A';
497 result
= result
+ pstate
->lexptr
[0] - '0';
503 error (_("Not enough hex digits seen"));
506 gdb_assert (min
!= max
);
507 error (_("Overlong hex escape"));
513 /* Lex an escape. IS_BYTE is true if we're lexing a byte escape;
514 otherwise we're lexing a character escape. */
517 rust_parser::lex_escape (int is_byte
)
521 gdb_assert (pstate
->lexptr
[0] == '\\');
523 switch (pstate
->lexptr
[0])
527 result
= lex_hex (2, 2);
532 error (_("Unicode escape in byte literal"));
534 if (pstate
->lexptr
[0] != '{')
535 error (_("Missing '{' in Unicode escape"));
537 result
= lex_hex (1, 6);
538 /* Could do range checks here. */
539 if (pstate
->lexptr
[0] != '}')
540 error (_("Missing '}' in Unicode escape"));
574 error (_("Invalid escape \\%c in literal"), pstate
->lexptr
[0]);
580 /* Lex a character constant. */
583 rust_parser::lex_character ()
588 if (pstate
->lexptr
[0] == 'b')
593 gdb_assert (pstate
->lexptr
[0] == '\'');
595 /* This should handle UTF-8 here. */
596 if (pstate
->lexptr
[0] == '\\')
597 value
= lex_escape (is_byte
);
600 value
= pstate
->lexptr
[0] & 0xff;
604 if (pstate
->lexptr
[0] != '\'')
605 error (_("Unterminated character literal"));
608 current_int_val
.val
= value
;
609 current_int_val
.type
= get_type (is_byte
? "u8" : "char");
614 /* Return the offset of the double quote if STR looks like the start
615 of a raw string, or 0 if STR does not start a raw string. */
618 starts_raw_string (const char *str
)
620 const char *save
= str
;
625 while (str
[0] == '#')
632 /* Return true if STR looks like the end of a raw string that had N
633 hashes at the start. */
636 ends_raw_string (const char *str
, int n
)
640 gdb_assert (str
[0] == '"');
641 for (i
= 0; i
< n
; ++i
)
642 if (str
[i
+ 1] != '#')
647 /* Lex a string constant. */
650 rust_parser::lex_string ()
652 int is_byte
= pstate
->lexptr
[0] == 'b';
657 raw_length
= starts_raw_string (pstate
->lexptr
);
658 pstate
->lexptr
+= raw_length
;
659 gdb_assert (pstate
->lexptr
[0] == '"');
668 if (pstate
->lexptr
[0] == '"' && ends_raw_string (pstate
->lexptr
,
671 /* Exit with lexptr pointing after the final "#". */
672 pstate
->lexptr
+= raw_length
;
675 else if (pstate
->lexptr
[0] == '\0')
676 error (_("Unexpected EOF in string"));
678 value
= pstate
->lexptr
[0] & 0xff;
679 if (is_byte
&& value
> 127)
680 error (_("Non-ASCII value in raw byte string"));
681 obstack_1grow (&obstack
, value
);
685 else if (pstate
->lexptr
[0] == '"')
687 /* Make sure to skip the quote. */
691 else if (pstate
->lexptr
[0] == '\\')
693 value
= lex_escape (is_byte
);
696 obstack_1grow (&obstack
, value
);
698 convert_between_encodings ("UTF-32", "UTF-8", (gdb_byte
*) &value
,
699 sizeof (value
), sizeof (value
),
700 &obstack
, translit_none
);
702 else if (pstate
->lexptr
[0] == '\0')
703 error (_("Unexpected EOF in string"));
706 value
= pstate
->lexptr
[0] & 0xff;
707 if (is_byte
&& value
> 127)
708 error (_("Non-ASCII value in byte string"));
709 obstack_1grow (&obstack
, value
);
714 current_string_val
.length
= obstack_object_size (&obstack
);
715 current_string_val
.ptr
= (const char *) obstack_finish (&obstack
);
716 return is_byte
? BYTESTRING
: STRING
;
719 /* Return true if STRING starts with whitespace followed by a digit. */
722 space_then_number (const char *string
)
724 const char *p
= string
;
726 while (p
[0] == ' ' || p
[0] == '\t')
731 return *p
>= '0' && *p
<= '9';
734 /* Return true if C can start an identifier. */
737 rust_identifier_start_p (char c
)
739 return ((c
>= 'a' && c
<= 'z')
740 || (c
>= 'A' && c
<= 'Z')
745 /* Lex an identifier. */
748 rust_parser::lex_identifier ()
750 const char *start
= pstate
->lexptr
;
752 const struct token_info
*token
;
754 int is_gdb_var
= pstate
->lexptr
[0] == '$';
756 gdb_assert (rust_identifier_start_p (pstate
->lexptr
[0]));
760 /* For the time being this doesn't handle Unicode rules. Non-ASCII
761 identifiers are gated anyway. */
762 while ((pstate
->lexptr
[0] >= 'a' && pstate
->lexptr
[0] <= 'z')
763 || (pstate
->lexptr
[0] >= 'A' && pstate
->lexptr
[0] <= 'Z')
764 || pstate
->lexptr
[0] == '_'
765 || (is_gdb_var
&& pstate
->lexptr
[0] == '$')
766 || (pstate
->lexptr
[0] >= '0' && pstate
->lexptr
[0] <= '9'))
770 length
= pstate
->lexptr
- start
;
772 for (i
= 0; i
< ARRAY_SIZE (identifier_tokens
); ++i
)
774 if (length
== strlen (identifier_tokens
[i
].name
)
775 && strncmp (identifier_tokens
[i
].name
, start
, length
) == 0)
777 token
= &identifier_tokens
[i
];
784 if (token
->value
== 0)
786 /* Leave the terminating token alone. */
787 pstate
->lexptr
= start
;
791 else if (token
== NULL
792 && (strncmp (start
, "thread", length
) == 0
793 || strncmp (start
, "task", length
) == 0)
794 && space_then_number (pstate
->lexptr
))
796 /* "task" or "thread" followed by a number terminates the
797 parse, per gdb rules. */
798 pstate
->lexptr
= start
;
802 if (token
== NULL
|| (pstate
->parse_completion
&& pstate
->lexptr
[0] == '\0'))
804 current_string_val
.length
= length
;
805 current_string_val
.ptr
= start
;
808 if (pstate
->parse_completion
&& pstate
->lexptr
[0] == '\0')
810 /* Prevent rustyylex from returning two COMPLETE tokens. */
811 pstate
->prev_lexptr
= pstate
->lexptr
;
822 /* Lex an operator. */
825 rust_parser::lex_operator ()
827 const struct token_info
*token
= NULL
;
830 for (i
= 0; i
< ARRAY_SIZE (operator_tokens
); ++i
)
832 if (strncmp (operator_tokens
[i
].name
, pstate
->lexptr
,
833 strlen (operator_tokens
[i
].name
)) == 0)
835 pstate
->lexptr
+= strlen (operator_tokens
[i
].name
);
836 token
= &operator_tokens
[i
];
843 current_opcode
= token
->opcode
;
847 return *pstate
->lexptr
++;
853 rust_parser::lex_number ()
855 regmatch_t subexps
[NUM_SUBEXPRESSIONS
];
858 int could_be_decimal
= 1;
859 int implicit_i32
= 0;
860 const char *type_name
= NULL
;
866 match
= regexec (&number_regex
, pstate
->lexptr
, ARRAY_SIZE (subexps
),
868 /* Failure means the regexp is broken. */
869 gdb_assert (match
== 0);
871 if (subexps
[INT_TEXT
].rm_so
!= -1)
873 /* Integer part matched. */
875 end_index
= subexps
[INT_TEXT
].rm_eo
;
876 if (subexps
[INT_TYPE
].rm_so
== -1)
883 type_index
= INT_TYPE
;
884 could_be_decimal
= 0;
887 else if (subexps
[FLOAT_TYPE1
].rm_so
!= -1)
889 /* Found floating point type suffix. */
890 end_index
= subexps
[FLOAT_TYPE1
].rm_so
;
891 type_index
= FLOAT_TYPE1
;
893 else if (subexps
[FLOAT_TYPE2
].rm_so
!= -1)
895 /* Found floating point type suffix. */
896 end_index
= subexps
[FLOAT_TYPE2
].rm_so
;
897 type_index
= FLOAT_TYPE2
;
901 /* Any other floating point match. */
902 end_index
= subexps
[0].rm_eo
;
906 /* We need a special case if the final character is ".". In this
907 case we might need to parse an integer. For example, "23.f()" is
908 a request for a trait method call, not a syntax error involving
909 the floating point number "23.". */
910 gdb_assert (subexps
[0].rm_eo
> 0);
911 if (pstate
->lexptr
[subexps
[0].rm_eo
- 1] == '.')
913 const char *next
= skip_spaces (&pstate
->lexptr
[subexps
[0].rm_eo
]);
915 if (rust_identifier_start_p (*next
) || *next
== '.')
919 end_index
= subexps
[0].rm_eo
;
921 could_be_decimal
= 1;
926 /* Compute the type name if we haven't already. */
927 std::string type_name_holder
;
928 if (type_name
== NULL
)
930 gdb_assert (type_index
!= -1);
931 type_name_holder
= std::string ((pstate
->lexptr
932 + subexps
[type_index
].rm_so
),
933 (subexps
[type_index
].rm_eo
934 - subexps
[type_index
].rm_so
));
935 type_name
= type_name_holder
.c_str ();
938 /* Look up the type. */
939 type
= get_type (type_name
);
941 /* Copy the text of the number and remove the "_"s. */
943 for (i
= 0; i
< end_index
&& pstate
->lexptr
[i
]; ++i
)
945 if (pstate
->lexptr
[i
] == '_')
946 could_be_decimal
= 0;
948 number
.push_back (pstate
->lexptr
[i
]);
951 /* Advance past the match. */
952 pstate
->lexptr
+= subexps
[0].rm_eo
;
954 /* Parse the number. */
961 if (number
[0] == '0')
963 if (number
[1] == 'x')
965 else if (number
[1] == 'o')
967 else if (number
[1] == 'b')
972 could_be_decimal
= 0;
976 value
= strtoulst (number
.c_str () + offset
, NULL
, radix
);
977 if (implicit_i32
&& value
>= ((uint64_t) 1) << 31)
978 type
= get_type ("i64");
980 current_int_val
.val
= value
;
981 current_int_val
.type
= type
;
985 current_float_val
.type
= type
;
986 bool parsed
= parse_float (number
.c_str (), number
.length (),
987 current_float_val
.type
,
988 current_float_val
.val
.data ());
992 return is_integer
? (could_be_decimal
? DECIMAL_INTEGER
: INTEGER
) : FLOAT
;
998 rust_parser::lex_one_token ()
1000 /* Skip all leading whitespace. */
1001 while (pstate
->lexptr
[0] == ' '
1002 || pstate
->lexptr
[0] == '\t'
1003 || pstate
->lexptr
[0] == '\r'
1004 || pstate
->lexptr
[0] == '\n')
1007 /* If we hit EOF and we're completing, then return COMPLETE -- maybe
1008 we're completing an empty string at the end of a field_expr.
1009 But, we don't want to return two COMPLETE tokens in a row. */
1010 if (pstate
->lexptr
[0] == '\0' && pstate
->lexptr
== pstate
->prev_lexptr
)
1012 pstate
->prev_lexptr
= pstate
->lexptr
;
1013 if (pstate
->lexptr
[0] == '\0')
1015 if (pstate
->parse_completion
)
1017 current_string_val
.length
=0;
1018 current_string_val
.ptr
= "";
1024 if (pstate
->lexptr
[0] >= '0' && pstate
->lexptr
[0] <= '9')
1025 return lex_number ();
1026 else if (pstate
->lexptr
[0] == 'b' && pstate
->lexptr
[1] == '\'')
1027 return lex_character ();
1028 else if (pstate
->lexptr
[0] == 'b' && pstate
->lexptr
[1] == '"')
1029 return lex_string ();
1030 else if (pstate
->lexptr
[0] == 'b' && starts_raw_string (pstate
->lexptr
+ 1))
1031 return lex_string ();
1032 else if (starts_raw_string (pstate
->lexptr
))
1033 return lex_string ();
1034 else if (rust_identifier_start_p (pstate
->lexptr
[0]))
1035 return lex_identifier ();
1036 else if (pstate
->lexptr
[0] == '"')
1037 return lex_string ();
1038 else if (pstate
->lexptr
[0] == '\'')
1039 return lex_character ();
1040 else if (pstate
->lexptr
[0] == '}' || pstate
->lexptr
[0] == ']')
1042 /* Falls through to lex_operator. */
1045 else if (pstate
->lexptr
[0] == '(' || pstate
->lexptr
[0] == '{')
1047 /* Falls through to lex_operator. */
1050 else if (pstate
->lexptr
[0] == ',' && pstate
->comma_terminates
1051 && paren_depth
== 0)
1054 return lex_operator ();
1057 /* Push back a single character to be re-lexed. */
1060 rust_parser::push_back (char c
)
1062 /* Can't be called before any lexing. */
1063 gdb_assert (pstate
->prev_lexptr
!= NULL
);
1066 gdb_assert (*pstate
->lexptr
== c
);
1071 /* Parse a tuple or paren expression. */
1074 rust_parser::parse_tuple ()
1078 if (current_token
== ')')
1081 struct type
*unit
= get_type ("()");
1082 return make_operation
<long_const_operation
> (unit
, 0);
1085 operation_up expr
= parse_expr ();
1086 if (current_token
== ')')
1088 /* Parenthesized expression. */
1093 std::vector
<operation_up
> ops
;
1094 ops
.push_back (std::move (expr
));
1095 while (current_token
!= ')')
1097 if (current_token
!= ',')
1098 error (_("',' or ')' expected"));
1101 /* A trailing "," is ok. */
1102 if (current_token
!= ')')
1103 ops
.push_back (parse_expr ());
1108 error (_("Tuple expressions not supported yet"));
1111 /* Parse an array expression. */
1114 rust_parser::parse_array ()
1118 if (current_token
== KW_MUT
)
1121 operation_up result
;
1122 operation_up expr
= parse_expr ();
1123 if (current_token
== ';')
1126 operation_up rhs
= parse_expr ();
1127 result
= make_operation
<rust_array_operation
> (std::move (expr
),
1130 else if (current_token
== ',')
1132 std::vector
<operation_up
> ops
;
1133 ops
.push_back (std::move (expr
));
1134 while (current_token
!= ']')
1136 if (current_token
!= ',')
1137 error (_("',' or ']' expected"));
1139 ops
.push_back (parse_expr ());
1141 ops
.shrink_to_fit ();
1142 int len
= ops
.size () - 1;
1143 result
= make_operation
<array_operation
> (0, len
, std::move (ops
));
1145 else if (current_token
!= ']')
1146 error (_("',', ';', or ']' expected"));
1153 /* Turn a name into an operation. */
1156 rust_parser::name_to_operation (const std::string
&name
)
1158 struct block_symbol sym
= lookup_symbol (name
.c_str (),
1159 pstate
->expression_context_block
,
1161 if (sym
.symbol
!= nullptr && SYMBOL_CLASS (sym
.symbol
) != LOC_TYPEDEF
)
1162 return make_operation
<var_value_operation
> (sym
);
1164 struct type
*type
= nullptr;
1166 if (sym
.symbol
!= nullptr)
1168 gdb_assert (SYMBOL_CLASS (sym
.symbol
) == LOC_TYPEDEF
);
1169 type
= SYMBOL_TYPE (sym
.symbol
);
1171 if (type
== nullptr)
1172 type
= rust_lookup_type (name
.c_str ());
1173 if (type
== nullptr)
1174 error (_("No symbol '%s' in current context"), name
.c_str ());
1176 if (type
->code () == TYPE_CODE_STRUCT
&& type
->num_fields () == 0)
1178 /* A unit-like struct. */
1179 operation_up
result (new rust_aggregate_operation (type
, {}, {}));
1183 return make_operation
<type_operation
> (type
);
1186 /* Parse a struct expression. */
1189 rust_parser::parse_struct_expr (struct type
*type
)
1193 if (type
->code () != TYPE_CODE_STRUCT
1194 || rust_tuple_type_p (type
)
1195 || rust_tuple_struct_type_p (type
))
1196 error (_("Struct expression applied to non-struct type"));
1198 std::vector
<std::pair
<std::string
, operation_up
>> field_v
;
1199 while (current_token
!= '}' && current_token
!= DOTDOT
)
1201 if (current_token
!= IDENT
)
1202 error (_("'}', '..', or identifier expected"));
1204 std::string name
= get_string ();
1208 if (current_token
== ',' || current_token
== '}'
1209 || current_token
== DOTDOT
)
1210 expr
= name_to_operation (name
);
1214 expr
= parse_expr ();
1216 field_v
.emplace_back (std::move (name
), std::move (expr
));
1218 /* A trailing "," is ok. */
1219 if (current_token
== ',')
1223 operation_up others
;
1224 if (current_token
== DOTDOT
)
1227 others
= parse_expr ();
1232 return make_operation
<rust_aggregate_operation
> (type
,
1234 std::move (field_v
));
1237 /* Used by the operator precedence parser. */
1240 rustop_item (int token_
, int precedence_
, enum exp_opcode opcode_
,
1243 precedence (precedence_
),
1245 op (std::move (op_
))
1249 /* The token value. */
1251 /* Precedence of this operator. */
1253 /* This is used only for assign-modify. */
1254 enum exp_opcode opcode
;
1255 /* The right hand side of this operation. */
1259 /* An operator precedence parser for binary operations, including
1263 rust_parser::parse_binop (bool required
)
1265 /* All the binary operators. Each one is of the form
1266 OPERATION(TOKEN, PRECEDENCE, TYPE)
1267 TOKEN is the corresponding operator token.
1268 PRECEDENCE is a value indicating relative precedence.
1269 TYPE is the operation type corresponding to the operator.
1270 Assignment operations are handled specially, not via this
1271 table; they have precedence 0. */
1273 OPERATION ('*', 10, mul_operation) \
1274 OPERATION ('/', 10, div_operation) \
1275 OPERATION ('%', 10, rem_operation) \
1276 OPERATION ('@', 9, repeat_operation) \
1277 OPERATION ('+', 8, add_operation) \
1278 OPERATION ('-', 8, sub_operation) \
1279 OPERATION (LSH, 7, lsh_operation) \
1280 OPERATION (RSH, 7, rsh_operation) \
1281 OPERATION ('&', 6, bitwise_and_operation) \
1282 OPERATION ('^', 5, bitwise_xor_operation) \
1283 OPERATION ('|', 4, bitwise_ior_operation) \
1284 OPERATION (EQEQ, 3, equal_operation) \
1285 OPERATION (NOTEQ, 3, notequal_operation) \
1286 OPERATION ('<', 3, less_operation) \
1287 OPERATION (LTEQ, 3, leq_operation) \
1288 OPERATION ('>', 3, gtr_operation) \
1289 OPERATION (GTEQ, 3, geq_operation) \
1290 OPERATION (ANDAND, 2, logical_and_operation) \
1291 OPERATION (OROR, 1, logical_or_operation)
1293 operation_up start
= parse_atom (required
);
1294 if (start
== nullptr)
1296 gdb_assert (!required
);
1300 std::vector
<rustop_item
> operator_stack
;
1301 operator_stack
.emplace_back (0, -1, OP_NULL
, std::move (start
));
1305 int this_token
= current_token
;
1306 enum exp_opcode compound_assign_op
= OP_NULL
;
1307 int precedence
= -2;
1311 #define OPERATION(TOKEN, PRECEDENCE, TYPE) \
1313 precedence = PRECEDENCE; \
1321 case COMPOUND_ASSIGN
:
1322 compound_assign_op
= current_opcode
;
1329 /* "as" must be handled specially. */
1333 rustop_item
&lhs
= operator_stack
.back ();
1334 struct type
*type
= parse_type ();
1335 lhs
.op
= make_operation
<unop_cast_operation
> (std::move (lhs
.op
),
1338 /* Bypass the rest of the loop. */
1342 /* Arrange to pop the entire stack. */
1347 while (precedence
< operator_stack
.back ().precedence
1348 && operator_stack
.size () > 1)
1350 rustop_item rhs
= std::move (operator_stack
.back ());
1351 operator_stack
.pop_back ();
1353 rustop_item
&lhs
= operator_stack
.back ();
1357 #define OPERATION(TOKEN, PRECEDENCE, TYPE) \
1359 lhs.op = make_operation<TYPE> (std::move (lhs.op), \
1360 std::move (rhs.op)); \
1368 case COMPOUND_ASSIGN
:
1370 if (rhs
.token
== '=')
1371 lhs
.op
= (make_operation
<assign_operation
>
1372 (std::move (lhs
.op
), std::move (rhs
.op
)));
1374 lhs
.op
= (make_operation
<assign_modify_operation
>
1375 (rhs
.opcode
, std::move (lhs
.op
),
1376 std::move (rhs
.op
)));
1378 struct type
*unit_type
= get_type ("()");
1380 operation_up
nil (new long_const_operation (unit_type
, 0));
1381 lhs
.op
= (make_operation
<comma_operation
>
1382 (std::move (lhs
.op
), std::move (nil
)));
1387 gdb_assert_not_reached ("bad binary operator");
1391 if (precedence
== -2)
1394 operator_stack
.emplace_back (this_token
, precedence
, compound_assign_op
,
1398 gdb_assert (operator_stack
.size () == 1);
1399 return std::move (operator_stack
[0].op
);
1403 /* Parse a range expression. */
1406 rust_parser::parse_range ()
1408 enum range_flag kind
= (RANGE_HIGH_BOUND_DEFAULT
1409 | RANGE_LOW_BOUND_DEFAULT
);
1412 if (current_token
!= DOTDOT
&& current_token
!= DOTDOTEQ
)
1414 lhs
= parse_binop (true);
1415 kind
&= ~RANGE_LOW_BOUND_DEFAULT
;
1418 if (current_token
== DOTDOT
)
1419 kind
|= RANGE_HIGH_BOUND_EXCLUSIVE
;
1420 else if (current_token
!= DOTDOTEQ
)
1424 /* A "..=" range requires a high bound, but otherwise it is
1426 operation_up rhs
= parse_binop ((kind
& RANGE_HIGH_BOUND_EXCLUSIVE
) == 0);
1428 kind
&= ~RANGE_HIGH_BOUND_DEFAULT
;
1430 return make_operation
<rust_range_operation
> (kind
,
1435 /* Parse an expression. */
1438 rust_parser::parse_expr ()
1440 return parse_range ();
1443 /* Parse a sizeof expression. */
1446 rust_parser::parse_sizeof ()
1451 operation_up result
= make_operation
<unop_sizeof_operation
> (parse_expr ());
1456 /* Parse an address-of operation. */
1459 rust_parser::parse_addr ()
1463 if (current_token
== KW_MUT
)
1466 return make_operation
<rust_unop_addr_operation
> (parse_atom (true));
1469 /* Parse a field expression. */
1472 rust_parser::parse_field (operation_up
&&lhs
)
1476 operation_up result
;
1477 switch (current_token
)
1482 bool is_complete
= current_token
== COMPLETE
;
1483 auto struct_op
= new rust_structop (std::move (lhs
), get_string ());
1487 completion_op
.reset (struct_op
);
1488 pstate
->mark_struct_expression (struct_op
);
1489 /* Throw to the outermost level of the parser. */
1490 error (_("not really an error"));
1492 result
.reset (struct_op
);
1496 case DECIMAL_INTEGER
:
1497 result
= make_operation
<rust_struct_anon
> (current_int_val
.val
,
1503 error (_("'_' not allowed in integers in anonymous field references"));
1506 error (_("field name expected"));
1512 /* Parse an index expression. */
1515 rust_parser::parse_index (operation_up
&&lhs
)
1518 operation_up rhs
= parse_expr ();
1521 return make_operation
<rust_subscript_operation
> (std::move (lhs
),
1525 /* Parse a sequence of comma-separated expressions in parens. */
1527 std::vector
<operation_up
>
1528 rust_parser::parse_paren_args ()
1532 std::vector
<operation_up
> args
;
1533 while (current_token
!= ')')
1537 if (current_token
!= ',')
1538 error (_("',' or ')' expected"));
1542 args
.push_back (parse_expr ());
1550 /* Parse the parenthesized part of a function call. */
1553 rust_parser::parse_call (operation_up
&&lhs
)
1555 std::vector
<operation_up
> args
= parse_paren_args ();
1557 return make_operation
<funcall_operation
> (std::move (lhs
),
1561 /* Parse a list of types. */
1563 std::vector
<struct type
*>
1564 rust_parser::parse_type_list ()
1566 std::vector
<struct type
*> result
;
1567 result
.push_back (parse_type ());
1568 while (current_token
== ',')
1571 result
.push_back (parse_type ());
1576 /* Parse a possibly-empty list of types, surrounded in parens. */
1578 std::vector
<struct type
*>
1579 rust_parser::parse_maybe_type_list ()
1582 std::vector
<struct type
*> types
;
1583 if (current_token
!= ')')
1584 types
= parse_type_list ();
1589 /* Parse an array type. */
1592 rust_parser::parse_array_type ()
1595 struct type
*elt_type
= parse_type ();
1598 if (current_token
!= INTEGER
&& current_token
!= DECIMAL_INTEGER
)
1599 error (_("integer expected"));
1600 ULONGEST val
= current_int_val
.val
;
1604 return lookup_array_range_type (elt_type
, 0, val
- 1);
1607 /* Parse a slice type. */
1610 rust_parser::parse_slice_type ()
1614 bool is_slice
= current_token
== '[';
1618 struct type
*target
= parse_type ();
1623 return rust_slice_type ("&[*gdb*]", target
, get_type ("usize"));
1626 /* For now we treat &x and *x identically. */
1627 return lookup_pointer_type (target
);
1630 /* Parse a pointer type. */
1633 rust_parser::parse_pointer_type ()
1637 if (current_token
== KW_MUT
|| current_token
== KW_CONST
)
1640 struct type
*target
= parse_type ();
1641 /* For the time being we ignore mut/const. */
1642 return lookup_pointer_type (target
);
1645 /* Parse a function type. */
1648 rust_parser::parse_function_type ()
1652 if (current_token
!= '(')
1653 error (_("'(' expected"));
1655 std::vector
<struct type
*> types
= parse_maybe_type_list ();
1657 if (current_token
!= ARROW
)
1658 error (_("'->' expected"));
1661 struct type
*result_type
= parse_type ();
1663 struct type
**argtypes
= nullptr;
1664 if (!types
.empty ())
1665 argtypes
= types
.data ();
1667 result_type
= lookup_function_type_with_arguments (result_type
,
1670 return lookup_pointer_type (result_type
);
1673 /* Parse a tuple type. */
1676 rust_parser::parse_tuple_type ()
1678 std::vector
<struct type
*> types
= parse_maybe_type_list ();
1680 auto_obstack obstack
;
1681 obstack_1grow (&obstack
, '(');
1682 for (int i
= 0; i
< types
.size (); ++i
)
1684 std::string type_name
= type_to_string (types
[i
]);
1687 obstack_1grow (&obstack
, ',');
1688 obstack_grow_str (&obstack
, type_name
.c_str ());
1691 obstack_grow_str0 (&obstack
, ")");
1692 const char *name
= (const char *) obstack_finish (&obstack
);
1694 /* We don't allow creating new tuple types (yet), but we do allow
1695 looking up existing tuple types. */
1696 struct type
*result
= rust_lookup_type (name
);
1697 if (result
== nullptr)
1698 error (_("could not find tuple type '%s'"), name
);
1706 rust_parser::parse_type ()
1708 switch (current_token
)
1711 return parse_array_type ();
1713 return parse_slice_type ();
1715 return parse_pointer_type ();
1717 return parse_function_type ();
1719 return parse_tuple_type ();
1726 std::string path
= parse_path (false);
1727 struct type
*result
= rust_lookup_type (path
.c_str ());
1728 if (result
== nullptr)
1729 error (_("No type name '%s' in current context"), path
.c_str ());
1733 error (_("type expected"));
1740 rust_parser::parse_path (bool for_expr
)
1742 unsigned n_supers
= 0;
1743 int first_token
= current_token
;
1745 switch (current_token
)
1749 if (current_token
!= COLONCOLON
)
1754 while (current_token
== KW_SUPER
)
1758 if (current_token
!= COLONCOLON
)
1759 error (_("'::' expected"));
1769 /* This is a gdb extension to make it possible to refer to items
1770 in other crates. It just bypasses adding the current crate
1771 to the front of the name. */
1776 if (current_token
!= IDENT
)
1777 error (_("identifier expected"));
1778 std::string path
= get_string ();
1779 bool saw_ident
= true;
1782 /* The condition here lets us enter the loop even if we see
1784 while (current_token
== COLONCOLON
|| current_token
== '<')
1786 if (current_token
== COLONCOLON
)
1791 if (current_token
== IDENT
)
1793 path
= path
+ "::" + get_string ();
1797 else if (current_token
== COLONCOLON
)
1799 /* The code below won't detect this scenario. */
1800 error (_("unexpected '::'"));
1804 if (current_token
!= '<')
1807 /* Expression use name::<...>, whereas types use name<...>. */
1810 /* Expressions use "name::<...>", so if we saw an identifier
1811 after the "::", we ignore the "<" here. */
1817 /* Types use "name<...>", so we need to have seen the
1824 std::vector
<struct type
*> types
= parse_type_list ();
1825 if (current_token
== '>')
1827 else if (current_token
== RSH
)
1833 error (_("'>' expected"));
1836 for (int i
= 0; i
< types
.size (); ++i
)
1840 path
+= type_to_string (types
[i
]);
1846 switch (first_token
)
1850 return super_name (path
, n_supers
);
1853 return crate_name (path
);
1862 gdb_assert_not_reached ("missing case in path parsing");
1866 /* Handle the parsing for a string expression. */
1869 rust_parser::parse_string ()
1871 gdb_assert (current_token
== STRING
);
1873 /* Wrap the raw string in the &str struct. */
1874 struct type
*type
= rust_lookup_type ("&str");
1875 if (type
== nullptr)
1876 error (_("Could not find type '&str'"));
1878 std::vector
<std::pair
<std::string
, operation_up
>> field_v
;
1880 size_t len
= current_string_val
.length
;
1881 operation_up str
= make_operation
<string_operation
> (get_string ());
1883 = make_operation
<rust_unop_addr_operation
> (std::move (str
));
1884 field_v
.emplace_back ("data_ptr", std::move (addr
));
1886 struct type
*valtype
= get_type ("usize");
1887 operation_up lenop
= make_operation
<long_const_operation
> (valtype
, len
);
1888 field_v
.emplace_back ("length", std::move (lenop
));
1890 return make_operation
<rust_aggregate_operation
> (type
,
1892 std::move (field_v
));
1895 /* Parse a tuple struct expression. */
1898 rust_parser::parse_tuple_struct (struct type
*type
)
1900 std::vector
<operation_up
> args
= parse_paren_args ();
1902 std::vector
<std::pair
<std::string
, operation_up
>> field_v (args
.size ());
1903 for (int i
= 0; i
< args
.size (); ++i
)
1904 field_v
[i
] = { string_printf ("__%d", i
), std::move (args
[i
]) };
1906 return (make_operation
<rust_aggregate_operation
>
1907 (type
, operation_up (), std::move (field_v
)));
1910 /* Parse a path expression. */
1913 rust_parser::parse_path_expr ()
1915 std::string path
= parse_path (true);
1917 if (current_token
== '{')
1919 struct type
*type
= rust_lookup_type (path
.c_str ());
1920 if (type
== nullptr)
1921 error (_("Could not find type '%s'"), path
.c_str ());
1923 return parse_struct_expr (type
);
1925 else if (current_token
== '(')
1927 struct type
*type
= rust_lookup_type (path
.c_str ());
1928 /* If this is actually a tuple struct expression, handle it
1929 here. If it is a call, it will be handled elsewhere. */
1930 if (type
!= nullptr)
1932 if (!rust_tuple_struct_type_p (type
))
1933 error (_("Type %s is not a tuple struct"), path
.c_str ());
1934 return parse_tuple_struct (type
);
1938 return name_to_operation (path
);
1941 /* Parse an atom. "Atom" isn't a Rust term, but this refers to a
1942 single unitary item in the grammar; but here including some unary
1943 prefix and postfix expressions. */
1946 rust_parser::parse_atom (bool required
)
1948 operation_up result
;
1950 switch (current_token
)
1953 result
= parse_tuple ();
1957 result
= parse_array ();
1961 case DECIMAL_INTEGER
:
1962 result
= make_operation
<long_const_operation
> (current_int_val
.type
,
1963 current_int_val
.val
);
1968 result
= make_operation
<float_const_operation
> (current_float_val
.type
,
1969 current_float_val
.val
);
1974 result
= parse_string ();
1978 result
= make_operation
<string_operation
> (get_string ());
1984 result
= make_operation
<bool_operation
> (current_token
== KW_TRUE
);
1989 /* This is kind of a hacky approach. */
1991 pstate
->push_dollar (current_string_val
);
1992 result
= pstate
->pop ();
2002 result
= parse_path_expr ();
2007 result
= make_operation
<rust_unop_ind_operation
> (parse_atom (true));
2011 result
= make_operation
<unary_plus_operation
> (parse_atom (true));
2015 result
= make_operation
<unary_neg_operation
> (parse_atom (true));
2019 result
= make_operation
<rust_unop_compl_operation
> (parse_atom (true));
2022 result
= parse_sizeof ();
2025 result
= parse_addr ();
2031 error (_("unexpected token"));
2034 /* Now parse suffixes. */
2037 switch (current_token
)
2040 result
= parse_field (std::move (result
));
2044 result
= parse_index (std::move (result
));
2048 result
= parse_call (std::move (result
));
2059 /* The parser as exposed to gdb. */
2062 rust_language::parser (struct parser_state
*state
) const
2064 rust_parser
parser (state
);
2066 operation_up result
;
2069 result
= parser
.parse_entry_point ();
2071 catch (const gdb_exception
&exc
)
2073 if (state
->parse_completion
)
2075 result
= std::move (parser
.completion_op
);
2076 if (result
== nullptr)
2083 state
->set_operation (std::move (result
));
2092 /* A test helper that lexes a string, expecting a single token. */
2095 rust_lex_test_one (rust_parser
*parser
, const char *input
, int expected
)
2099 parser
->reset (input
);
2101 token
= parser
->lex_one_token ();
2102 SELF_CHECK (token
== expected
);
2106 token
= parser
->lex_one_token ();
2107 SELF_CHECK (token
== 0);
2111 /* Test that INPUT lexes as the integer VALUE. */
2114 rust_lex_int_test (rust_parser
*parser
, const char *input
,
2115 ULONGEST value
, int kind
)
2117 rust_lex_test_one (parser
, input
, kind
);
2118 SELF_CHECK (parser
->current_int_val
.val
== value
);
2121 /* Test that INPUT throws an exception with text ERR. */
2124 rust_lex_exception_test (rust_parser
*parser
, const char *input
,
2129 /* The "kind" doesn't matter. */
2130 rust_lex_test_one (parser
, input
, DECIMAL_INTEGER
);
2133 catch (const gdb_exception_error
&except
)
2135 SELF_CHECK (strcmp (except
.what (), err
) == 0);
2139 /* Test that INPUT lexes as the identifier, string, or byte-string
2140 VALUE. KIND holds the expected token kind. */
2143 rust_lex_stringish_test (rust_parser
*parser
, const char *input
,
2144 const char *value
, int kind
)
2146 rust_lex_test_one (parser
, input
, kind
);
2147 SELF_CHECK (parser
->get_string () == value
);
2150 /* Helper to test that a string parses as a given token sequence. */
2153 rust_lex_test_sequence (rust_parser
*parser
, const char *input
, int len
,
2154 const int expected
[])
2158 parser
->reset (input
);
2160 for (i
= 0; i
< len
; ++i
)
2162 int token
= parser
->lex_one_token ();
2163 SELF_CHECK (token
== expected
[i
]);
2167 /* Tests for an integer-parsing corner case. */
2170 rust_lex_test_trailing_dot (rust_parser
*parser
)
2172 const int expected1
[] = { DECIMAL_INTEGER
, '.', IDENT
, '(', ')', 0 };
2173 const int expected2
[] = { INTEGER
, '.', IDENT
, '(', ')', 0 };
2174 const int expected3
[] = { FLOAT
, EQEQ
, '(', ')', 0 };
2175 const int expected4
[] = { DECIMAL_INTEGER
, DOTDOT
, DECIMAL_INTEGER
, 0 };
2177 rust_lex_test_sequence (parser
, "23.g()", ARRAY_SIZE (expected1
), expected1
);
2178 rust_lex_test_sequence (parser
, "23_0.g()", ARRAY_SIZE (expected2
),
2180 rust_lex_test_sequence (parser
, "23.==()", ARRAY_SIZE (expected3
),
2182 rust_lex_test_sequence (parser
, "23..25", ARRAY_SIZE (expected4
), expected4
);
2185 /* Tests of completion. */
2188 rust_lex_test_completion (rust_parser
*parser
)
2190 const int expected
[] = { IDENT
, '.', COMPLETE
, 0 };
2192 parser
->pstate
->parse_completion
= 1;
2194 rust_lex_test_sequence (parser
, "something.wha", ARRAY_SIZE (expected
),
2196 rust_lex_test_sequence (parser
, "something.", ARRAY_SIZE (expected
),
2199 parser
->pstate
->parse_completion
= 0;
2202 /* Test pushback. */
2205 rust_lex_test_push_back (rust_parser
*parser
)
2209 parser
->reset (">>=");
2211 token
= parser
->lex_one_token ();
2212 SELF_CHECK (token
== COMPOUND_ASSIGN
);
2213 SELF_CHECK (parser
->current_opcode
== BINOP_RSH
);
2215 parser
->push_back ('=');
2217 token
= parser
->lex_one_token ();
2218 SELF_CHECK (token
== '=');
2220 token
= parser
->lex_one_token ();
2221 SELF_CHECK (token
== 0);
2224 /* Unit test the lexer. */
2227 rust_lex_tests (void)
2231 /* Set up dummy "parser", so that rust_type works. */
2232 struct parser_state
ps (language_def (language_rust
), target_gdbarch (),
2233 nullptr, 0, 0, nullptr, 0, nullptr, false);
2234 rust_parser
parser (&ps
);
2236 rust_lex_test_one (&parser
, "", 0);
2237 rust_lex_test_one (&parser
, " \t \n \r ", 0);
2238 rust_lex_test_one (&parser
, "thread 23", 0);
2239 rust_lex_test_one (&parser
, "task 23", 0);
2240 rust_lex_test_one (&parser
, "th 104", 0);
2241 rust_lex_test_one (&parser
, "ta 97", 0);
2243 rust_lex_int_test (&parser
, "'z'", 'z', INTEGER
);
2244 rust_lex_int_test (&parser
, "'\\xff'", 0xff, INTEGER
);
2245 rust_lex_int_test (&parser
, "'\\u{1016f}'", 0x1016f, INTEGER
);
2246 rust_lex_int_test (&parser
, "b'z'", 'z', INTEGER
);
2247 rust_lex_int_test (&parser
, "b'\\xfe'", 0xfe, INTEGER
);
2248 rust_lex_int_test (&parser
, "b'\\xFE'", 0xfe, INTEGER
);
2249 rust_lex_int_test (&parser
, "b'\\xfE'", 0xfe, INTEGER
);
2251 /* Test all escapes in both modes. */
2252 rust_lex_int_test (&parser
, "'\\n'", '\n', INTEGER
);
2253 rust_lex_int_test (&parser
, "'\\r'", '\r', INTEGER
);
2254 rust_lex_int_test (&parser
, "'\\t'", '\t', INTEGER
);
2255 rust_lex_int_test (&parser
, "'\\\\'", '\\', INTEGER
);
2256 rust_lex_int_test (&parser
, "'\\0'", '\0', INTEGER
);
2257 rust_lex_int_test (&parser
, "'\\''", '\'', INTEGER
);
2258 rust_lex_int_test (&parser
, "'\\\"'", '"', INTEGER
);
2260 rust_lex_int_test (&parser
, "b'\\n'", '\n', INTEGER
);
2261 rust_lex_int_test (&parser
, "b'\\r'", '\r', INTEGER
);
2262 rust_lex_int_test (&parser
, "b'\\t'", '\t', INTEGER
);
2263 rust_lex_int_test (&parser
, "b'\\\\'", '\\', INTEGER
);
2264 rust_lex_int_test (&parser
, "b'\\0'", '\0', INTEGER
);
2265 rust_lex_int_test (&parser
, "b'\\''", '\'', INTEGER
);
2266 rust_lex_int_test (&parser
, "b'\\\"'", '"', INTEGER
);
2268 rust_lex_exception_test (&parser
, "'z", "Unterminated character literal");
2269 rust_lex_exception_test (&parser
, "b'\\x0'", "Not enough hex digits seen");
2270 rust_lex_exception_test (&parser
, "b'\\u{0}'",
2271 "Unicode escape in byte literal");
2272 rust_lex_exception_test (&parser
, "'\\x0'", "Not enough hex digits seen");
2273 rust_lex_exception_test (&parser
, "'\\u0'", "Missing '{' in Unicode escape");
2274 rust_lex_exception_test (&parser
, "'\\u{0", "Missing '}' in Unicode escape");
2275 rust_lex_exception_test (&parser
, "'\\u{0000007}", "Overlong hex escape");
2276 rust_lex_exception_test (&parser
, "'\\u{}", "Not enough hex digits seen");
2277 rust_lex_exception_test (&parser
, "'\\Q'", "Invalid escape \\Q in literal");
2278 rust_lex_exception_test (&parser
, "b'\\Q'", "Invalid escape \\Q in literal");
2280 rust_lex_int_test (&parser
, "23", 23, DECIMAL_INTEGER
);
2281 rust_lex_int_test (&parser
, "2_344__29", 234429, INTEGER
);
2282 rust_lex_int_test (&parser
, "0x1f", 0x1f, INTEGER
);
2283 rust_lex_int_test (&parser
, "23usize", 23, INTEGER
);
2284 rust_lex_int_test (&parser
, "23i32", 23, INTEGER
);
2285 rust_lex_int_test (&parser
, "0x1_f", 0x1f, INTEGER
);
2286 rust_lex_int_test (&parser
, "0b1_101011__", 0x6b, INTEGER
);
2287 rust_lex_int_test (&parser
, "0o001177i64", 639, INTEGER
);
2288 rust_lex_int_test (&parser
, "0x123456789u64", 0x123456789ull
, INTEGER
);
2290 rust_lex_test_trailing_dot (&parser
);
2292 rust_lex_test_one (&parser
, "23.", FLOAT
);
2293 rust_lex_test_one (&parser
, "23.99f32", FLOAT
);
2294 rust_lex_test_one (&parser
, "23e7", FLOAT
);
2295 rust_lex_test_one (&parser
, "23E-7", FLOAT
);
2296 rust_lex_test_one (&parser
, "23e+7", FLOAT
);
2297 rust_lex_test_one (&parser
, "23.99e+7f64", FLOAT
);
2298 rust_lex_test_one (&parser
, "23.82f32", FLOAT
);
2300 rust_lex_stringish_test (&parser
, "hibob", "hibob", IDENT
);
2301 rust_lex_stringish_test (&parser
, "hibob__93", "hibob__93", IDENT
);
2302 rust_lex_stringish_test (&parser
, "thread", "thread", IDENT
);
2304 rust_lex_stringish_test (&parser
, "\"string\"", "string", STRING
);
2305 rust_lex_stringish_test (&parser
, "\"str\\ting\"", "str\ting", STRING
);
2306 rust_lex_stringish_test (&parser
, "\"str\\\"ing\"", "str\"ing", STRING
);
2307 rust_lex_stringish_test (&parser
, "r\"str\\ing\"", "str\\ing", STRING
);
2308 rust_lex_stringish_test (&parser
, "r#\"str\\ting\"#", "str\\ting", STRING
);
2309 rust_lex_stringish_test (&parser
, "r###\"str\\\"ing\"###", "str\\\"ing",
2312 rust_lex_stringish_test (&parser
, "b\"string\"", "string", BYTESTRING
);
2313 rust_lex_stringish_test (&parser
, "b\"\x73tring\"", "string", BYTESTRING
);
2314 rust_lex_stringish_test (&parser
, "b\"str\\\"ing\"", "str\"ing", BYTESTRING
);
2315 rust_lex_stringish_test (&parser
, "br####\"\\x73tring\"####", "\\x73tring",
2318 for (i
= 0; i
< ARRAY_SIZE (identifier_tokens
); ++i
)
2319 rust_lex_test_one (&parser
, identifier_tokens
[i
].name
,
2320 identifier_tokens
[i
].value
);
2322 for (i
= 0; i
< ARRAY_SIZE (operator_tokens
); ++i
)
2323 rust_lex_test_one (&parser
, operator_tokens
[i
].name
,
2324 operator_tokens
[i
].value
);
2326 rust_lex_test_completion (&parser
);
2327 rust_lex_test_push_back (&parser
);
2330 #endif /* GDB_SELF_TEST */
2334 void _initialize_rust_exp ();
2336 _initialize_rust_exp ()
2338 int code
= regcomp (&number_regex
, number_regex_text
, REG_EXTENDED
);
2339 /* If the regular expression was incorrect, it was a programming
2341 gdb_assert (code
== 0);
2344 selftests::register_test ("rust-lex", rust_lex_tests
);