1 /* Separate lexical analyzer for GNU C++.
2 Copyright (C) 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
3 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
4 Hacked by Michael Tiemann (tiemann@cygnus.com)
6 This file is part of GCC.
8 GCC is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 2, or (at your option)
13 GCC is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with GCC; see the file COPYING. If not, write to
20 the Free Software Foundation, 59 Temple Place - Suite 330,
21 Boston, MA 02111-1307, USA. */
24 /* This file is the lexical analyzer for GNU C++. */
28 #include "coretypes.h"
41 static int interface_strcmp (const char *);
42 static void init_cp_pragma (void);
44 static tree parse_strconst_pragma (const char *, int);
45 static void handle_pragma_vtable (cpp_reader *);
46 static void handle_pragma_unit (cpp_reader *);
47 static void handle_pragma_interface (cpp_reader *);
48 static void handle_pragma_implementation (cpp_reader *);
49 static void handle_pragma_java_exceptions (cpp_reader *);
51 static void init_operators (void);
52 static void copy_lang_type (tree);
54 /* A constraint that can be tested at compile time. */
55 #define CONSTRAINT(name, expr) extern int constraint_##name [(expr) ? 1 : -1]
57 /* Functions and data structures for #pragma interface.
59 `#pragma implementation' means that the main file being compiled
60 is considered to implement (provide) the classes that appear in
61 its main body. I.e., if this is file "foo.cc", and class `bar'
62 is defined in "foo.cc", then we say that "foo.cc implements bar".
64 All main input files "implement" themselves automagically.
66 `#pragma interface' means that unless this file (of the form "foo.h"
67 is not presently being included by file "foo.cc", the
68 CLASSTYPE_INTERFACE_ONLY bit gets set. The effect is that none
69 of the vtables nor any of the inline functions defined in foo.h
72 There are cases when we want to link files such as "defs.h" and
73 "main.cc". In this case, we give "defs.h" a `#pragma interface',
74 and "main.cc" has `#pragma implementation "defs.h"'. */
79 struct impl_files *next;
82 static struct impl_files *impl_file_chain;
91 /* A mapping from tree codes to operator name information. */
92 operator_name_info_t operator_name_info[(int) LAST_CPLUS_TREE_CODE];
93 /* Similar, but for assignment operators. */
94 operator_name_info_t assignment_operator_name_info[(int) LAST_CPLUS_TREE_CODE];
96 /* Initialize data structures that keep track of operator names. */
98 #define DEF_OPERATOR(NAME, C, M, AR, AP) \
99 CONSTRAINT (C, sizeof "operator " + sizeof NAME <= 256);
100 #include "operators.def"
104 init_operators (void)
108 struct operator_name_info_t *oni;
110 #define DEF_OPERATOR(NAME, CODE, MANGLING, ARITY, ASSN_P) \
111 sprintf (buffer, ISALPHA (NAME[0]) ? "operator %s" : "operator%s", NAME); \
112 identifier = get_identifier (buffer); \
113 IDENTIFIER_OPNAME_P (identifier) = 1; \
116 ? &assignment_operator_name_info[(int) CODE] \
117 : &operator_name_info[(int) CODE]); \
118 oni->identifier = identifier; \
120 oni->mangled_name = MANGLING; \
123 #include "operators.def"
126 operator_name_info[(int) ERROR_MARK].identifier
127 = get_identifier ("<invalid operator>");
129 /* Handle some special cases. These operators are not defined in
130 the language, but can be produced internally. We may need them
131 for error-reporting. (Eventually, we should ensure that this
132 does not happen. Error messages involving these operators will
133 be confusing to users.) */
135 operator_name_info [(int) INIT_EXPR].name
136 = operator_name_info [(int) MODIFY_EXPR].name;
137 operator_name_info [(int) EXACT_DIV_EXPR].name = "(ceiling /)";
138 operator_name_info [(int) CEIL_DIV_EXPR].name = "(ceiling /)";
139 operator_name_info [(int) FLOOR_DIV_EXPR].name = "(floor /)";
140 operator_name_info [(int) ROUND_DIV_EXPR].name = "(round /)";
141 operator_name_info [(int) CEIL_MOD_EXPR].name = "(ceiling %)";
142 operator_name_info [(int) FLOOR_MOD_EXPR].name = "(floor %)";
143 operator_name_info [(int) ROUND_MOD_EXPR].name = "(round %)";
144 operator_name_info [(int) ABS_EXPR].name = "abs";
145 operator_name_info [(int) TRUTH_AND_EXPR].name = "strict &&";
146 operator_name_info [(int) TRUTH_OR_EXPR].name = "strict ||";
147 operator_name_info [(int) RANGE_EXPR].name = "...";
148 operator_name_info [(int) CONVERT_EXPR].name = "+";
150 assignment_operator_name_info [(int) EXACT_DIV_EXPR].name
152 assignment_operator_name_info [(int) CEIL_DIV_EXPR].name
154 assignment_operator_name_info [(int) FLOOR_DIV_EXPR].name
156 assignment_operator_name_info [(int) ROUND_DIV_EXPR].name
158 assignment_operator_name_info [(int) CEIL_MOD_EXPR].name
160 assignment_operator_name_info [(int) FLOOR_MOD_EXPR].name
162 assignment_operator_name_info [(int) ROUND_MOD_EXPR].name
166 /* The reserved keyword table. */
169 const char *const word;
170 ENUM_BITFIELD(rid) const rid : 16;
171 const unsigned int disable : 16;
174 /* Disable mask. Keywords are disabled if (reswords[i].disable & mask) is
176 #define D_EXT 0x01 /* GCC extension */
177 #define D_ASM 0x02 /* in C99, but has a switch to turn it off */
179 CONSTRAINT(ridbits_fit, RID_LAST_MODIFIER < sizeof(unsigned long) * CHAR_BIT);
181 static const struct resword reswords[] =
183 { "_Complex", RID_COMPLEX, 0 },
184 { "__FUNCTION__", RID_FUNCTION_NAME, 0 },
185 { "__PRETTY_FUNCTION__", RID_PRETTY_FUNCTION_NAME, 0 },
186 { "__alignof", RID_ALIGNOF, 0 },
187 { "__alignof__", RID_ALIGNOF, 0 },
188 { "__asm", RID_ASM, 0 },
189 { "__asm__", RID_ASM, 0 },
190 { "__attribute", RID_ATTRIBUTE, 0 },
191 { "__attribute__", RID_ATTRIBUTE, 0 },
192 { "__builtin_offsetof", RID_OFFSETOF, 0 },
193 { "__builtin_va_arg", RID_VA_ARG, 0 },
194 { "__complex", RID_COMPLEX, 0 },
195 { "__complex__", RID_COMPLEX, 0 },
196 { "__const", RID_CONST, 0 },
197 { "__const__", RID_CONST, 0 },
198 { "__extension__", RID_EXTENSION, 0 },
199 { "__func__", RID_C99_FUNCTION_NAME, 0 },
200 { "__imag", RID_IMAGPART, 0 },
201 { "__imag__", RID_IMAGPART, 0 },
202 { "__inline", RID_INLINE, 0 },
203 { "__inline__", RID_INLINE, 0 },
204 { "__label__", RID_LABEL, 0 },
205 { "__null", RID_NULL, 0 },
206 { "__real", RID_REALPART, 0 },
207 { "__real__", RID_REALPART, 0 },
208 { "__restrict", RID_RESTRICT, 0 },
209 { "__restrict__", RID_RESTRICT, 0 },
210 { "__signed", RID_SIGNED, 0 },
211 { "__signed__", RID_SIGNED, 0 },
212 { "__thread", RID_THREAD, 0 },
213 { "__typeof", RID_TYPEOF, 0 },
214 { "__typeof__", RID_TYPEOF, 0 },
215 { "__volatile", RID_VOLATILE, 0 },
216 { "__volatile__", RID_VOLATILE, 0 },
217 { "asm", RID_ASM, D_ASM },
218 { "auto", RID_AUTO, 0 },
219 { "bool", RID_BOOL, 0 },
220 { "break", RID_BREAK, 0 },
221 { "case", RID_CASE, 0 },
222 { "catch", RID_CATCH, 0 },
223 { "char", RID_CHAR, 0 },
224 { "class", RID_CLASS, 0 },
225 { "const", RID_CONST, 0 },
226 { "const_cast", RID_CONSTCAST, 0 },
227 { "continue", RID_CONTINUE, 0 },
228 { "default", RID_DEFAULT, 0 },
229 { "delete", RID_DELETE, 0 },
231 { "double", RID_DOUBLE, 0 },
232 { "dynamic_cast", RID_DYNCAST, 0 },
233 { "else", RID_ELSE, 0 },
234 { "enum", RID_ENUM, 0 },
235 { "explicit", RID_EXPLICIT, 0 },
236 { "export", RID_EXPORT, 0 },
237 { "extern", RID_EXTERN, 0 },
238 { "false", RID_FALSE, 0 },
239 { "float", RID_FLOAT, 0 },
240 { "for", RID_FOR, 0 },
241 { "friend", RID_FRIEND, 0 },
242 { "goto", RID_GOTO, 0 },
244 { "inline", RID_INLINE, 0 },
245 { "int", RID_INT, 0 },
246 { "long", RID_LONG, 0 },
247 { "mutable", RID_MUTABLE, 0 },
248 { "namespace", RID_NAMESPACE, 0 },
249 { "new", RID_NEW, 0 },
250 { "operator", RID_OPERATOR, 0 },
251 { "private", RID_PRIVATE, 0 },
252 { "protected", RID_PROTECTED, 0 },
253 { "public", RID_PUBLIC, 0 },
254 { "register", RID_REGISTER, 0 },
255 { "reinterpret_cast", RID_REINTCAST, 0 },
256 { "return", RID_RETURN, 0 },
257 { "short", RID_SHORT, 0 },
258 { "signed", RID_SIGNED, 0 },
259 { "sizeof", RID_SIZEOF, 0 },
260 { "static", RID_STATIC, 0 },
261 { "static_cast", RID_STATCAST, 0 },
262 { "struct", RID_STRUCT, 0 },
263 { "switch", RID_SWITCH, 0 },
264 { "template", RID_TEMPLATE, 0 },
265 { "this", RID_THIS, 0 },
266 { "throw", RID_THROW, 0 },
267 { "true", RID_TRUE, 0 },
268 { "try", RID_TRY, 0 },
269 { "typedef", RID_TYPEDEF, 0 },
270 { "typename", RID_TYPENAME, 0 },
271 { "typeid", RID_TYPEID, 0 },
272 { "typeof", RID_TYPEOF, D_ASM|D_EXT },
273 { "union", RID_UNION, 0 },
274 { "unsigned", RID_UNSIGNED, 0 },
275 { "using", RID_USING, 0 },
276 { "virtual", RID_VIRTUAL, 0 },
277 { "void", RID_VOID, 0 },
278 { "volatile", RID_VOLATILE, 0 },
279 { "wchar_t", RID_WCHAR, 0 },
280 { "while", RID_WHILE, 0 },
289 int mask = ((flag_no_asm ? D_ASM : 0)
290 | (flag_no_gnu_keywords ? D_EXT : 0));
292 ridpointers = ggc_calloc ((int) RID_MAX, sizeof (tree));
293 for (i = 0; i < ARRAY_SIZE (reswords); i++)
295 id = get_identifier (reswords[i].word);
296 C_RID_CODE (id) = reswords[i].rid;
297 ridpointers [(int) reswords[i].rid] = id;
298 if (! (reswords[i].disable & mask))
299 C_IS_RESERVED_WORD (id) = 1;
304 init_cp_pragma (void)
306 c_register_pragma (0, "vtable", handle_pragma_vtable);
307 c_register_pragma (0, "unit", handle_pragma_unit);
308 c_register_pragma (0, "interface", handle_pragma_interface);
309 c_register_pragma (0, "implementation", handle_pragma_implementation);
310 c_register_pragma ("GCC", "interface", handle_pragma_interface);
311 c_register_pragma ("GCC", "implementation", handle_pragma_implementation);
312 c_register_pragma ("GCC", "java_exceptions", handle_pragma_java_exceptions);
315 /* Initialize the C++ front end. This function is very sensitive to
316 the exact order that things are done here. It would be nice if the
317 initialization done by this routine were moved to its subroutines,
318 and the ordering dependencies clarified and reduced. */
322 static const enum tree_code stmt_codes[] = {
326 INIT_STATEMENT_CODES (stmt_codes);
328 /* We cannot just assign to input_filename because it has already
329 been initialized and will be used later as an N_BINCL for stabs+
331 #ifdef USE_MAPPED_LOCATION
332 push_srcloc (BUILTINS_LOCATION);
334 push_srcloc ("<built-in>", 0);
339 init_cp_semantics ();
344 current_function_decl = NULL;
346 class_type_node = ridpointers[(int) RID_CLASS];
348 cxx_init_decl_processing ();
350 /* The fact that G++ uses COMDAT for many entities (inline
351 functions, template instantiations, virtual tables, etc.) mean
352 that it is fundamentally unreliable to try to make decisions
353 about whether or not to output a particular entity until the end
354 of the compilation. However, the inliner requires that functions
355 be provided to the back end if they are to be inlined.
356 Therefore, we always use unit-at-a-time mode; in that mode, we
357 can provide entities to the back end and it will decide what to
358 emit based on what is actually needed. */
359 flag_unit_at_a_time = 1;
361 if (c_common_init () == false)
375 /* Return nonzero if S is not considered part of an
376 INTERFACE/IMPLEMENTATION pair. Otherwise, return 0. */
379 interface_strcmp (const char* s)
381 /* Set the interface/implementation bits for this scope. */
382 struct impl_files *ifiles;
385 for (ifiles = impl_file_chain; ifiles; ifiles = ifiles->next)
387 const char *t1 = ifiles->filename;
390 if (*s1 != *t1 || *s1 == 0)
393 while (*s1 == *t1 && *s1 != 0)
400 /* Don't get faked out by xxx.yyy.cc vs xxx.zzz.cc. */
401 if (strchr (s1, '.') || strchr (t1, '.'))
404 if (*s1 == '\0' || s1[-1] != '.' || t1[-1] != '.')
417 /* Parse a #pragma whose sole argument is a string constant.
418 If OPT is true, the argument is optional. */
420 parse_strconst_pragma (const char* name, int opt)
429 if (c_lex (&x) != CPP_EOF)
430 warning (0, "junk at end of #pragma %s", name);
434 if (t == CPP_EOF && opt)
437 error ("invalid #pragma %s", name);
442 handle_pragma_vtable (cpp_reader* dfile ATTRIBUTE_UNUSED )
444 parse_strconst_pragma ("vtable", 0);
445 sorry ("#pragma vtable no longer supported");
449 handle_pragma_unit (cpp_reader* dfile ATTRIBUTE_UNUSED )
451 /* Validate syntax, but don't do anything. */
452 parse_strconst_pragma ("unit", 0);
456 handle_pragma_interface (cpp_reader* dfile ATTRIBUTE_UNUSED )
458 tree fname = parse_strconst_pragma ("interface", 1);
459 struct c_fileinfo *finfo;
460 const char *filename;
462 if (fname == (tree)-1)
465 filename = lbasename (input_filename);
467 filename = ggc_strdup (TREE_STRING_POINTER (fname));
469 finfo = get_fileinfo (filename);
471 if (impl_file_chain == 0)
473 /* If this is zero at this point, then we are
474 auto-implementing. */
475 if (main_input_filename == 0)
476 main_input_filename = input_filename;
479 finfo->interface_only = interface_strcmp (filename);
480 /* If MULTIPLE_SYMBOL_SPACES is set, we cannot assume that we can see
481 a definition in another file. */
482 if (!MULTIPLE_SYMBOL_SPACES || !finfo->interface_only)
483 finfo->interface_unknown = 0;
486 /* Note that we have seen a #pragma implementation for the key MAIN_FILENAME.
487 We used to only allow this at toplevel, but that restriction was buggy
488 in older compilers and it seems reasonable to allow it in the headers
489 themselves, too. It only needs to precede the matching #p interface.
491 We don't touch finfo->interface_only or finfo->interface_unknown;
492 the user must specify a matching #p interface for this to have
496 handle_pragma_implementation (cpp_reader* dfile ATTRIBUTE_UNUSED )
498 tree fname = parse_strconst_pragma ("implementation", 1);
499 const char *filename;
500 struct impl_files *ifiles = impl_file_chain;
502 if (fname == (tree)-1)
507 if (main_input_filename)
508 filename = main_input_filename;
510 filename = input_filename;
511 filename = lbasename (filename);
515 filename = ggc_strdup (TREE_STRING_POINTER (fname));
517 /* We currently cannot give this diagnostic, as we reach this point
518 only after cpplib has scanned the entire translation unit, so
519 cpp_included always returns true. A plausible fix is to compare
520 the current source-location cookie with the first source-location
521 cookie (if any) of the filename, but this requires completing the
522 --enable-mapped-location project first. See PR 17577. */
523 if (cpp_included (parse_in, filename))
524 warning (0, "#pragma implementation for %qs appears after "
525 "file is included", filename);
529 for (; ifiles; ifiles = ifiles->next)
531 if (! strcmp (ifiles->filename, filename))
536 ifiles = xmalloc (sizeof (struct impl_files));
537 ifiles->filename = filename;
538 ifiles->next = impl_file_chain;
539 impl_file_chain = ifiles;
543 /* Indicate that this file uses Java-personality exception handling. */
545 handle_pragma_java_exceptions (cpp_reader* dfile ATTRIBUTE_UNUSED )
548 if (c_lex (&x) != CPP_EOF)
549 warning (0, "junk at end of #pragma GCC java_exceptions");
551 choose_personality_routine (lang_java);
554 /* Issue an error message indicating that the lookup of NAME (an
555 IDENTIFIER_NODE) failed. Returns the ERROR_MARK_NODE. */
558 unqualified_name_lookup_error (tree name)
560 if (IDENTIFIER_OPNAME_P (name))
562 if (name != ansi_opname (ERROR_MARK))
563 error ("%qD not defined", name);
567 error ("%qD was not declared in this scope", name);
568 /* Prevent repeated error messages by creating a VAR_DECL with
569 this NAME in the innermost block scope. */
570 if (current_function_decl)
573 decl = build_decl (VAR_DECL, name, error_mark_node);
574 DECL_CONTEXT (decl) = current_function_decl;
575 push_local_binding (name, decl, 0);
576 /* Mark the variable as used so that we do not get warnings
577 about it being unused later. */
578 TREE_USED (decl) = 1;
582 return error_mark_node;
585 /* Like unqualified_name_lookup_error, but NAME is an unqualified-id
586 used as a function. Returns an appropriate expression for
590 unqualified_fn_lookup_error (tree name)
592 if (processing_template_decl)
594 /* In a template, it is invalid to write "f()" or "f(3)" if no
595 declaration of "f" is available. Historically, G++ and most
596 other compilers accepted that usage since they deferred all name
597 lookup until instantiation time rather than doing unqualified
598 name lookup at template definition time; explain to the user what
601 Note that we have the exact wording of the following message in
602 the manual (trouble.texi, node "Name lookup"), so they need to
604 pedwarn ("there are no arguments to %qD that depend on a template "
605 "parameter, so a declaration of %qD must be available",
608 if (!flag_permissive)
613 error ("(if you use %<-fpermissive%>, G++ will accept your "
614 "code, but allowing the use of an undeclared name is "
622 return unqualified_name_lookup_error (name);
626 build_lang_decl (enum tree_code code, tree name, tree type)
630 t = build_decl (code, name, type);
631 retrofit_lang_decl (t);
633 /* All nesting of C++ functions is lexical; there is never a "static
634 chain" in the sense of GNU C nested functions. */
635 if (code == FUNCTION_DECL)
636 DECL_NO_STATIC_CHAIN (t) = 1;
641 /* Add DECL_LANG_SPECIFIC info to T. Called from build_lang_decl
642 and pushdecl (for functions generated by the backend). */
645 retrofit_lang_decl (tree t)
647 struct lang_decl *ld;
650 if (CAN_HAVE_FULL_LANG_DECL_P (t))
651 size = sizeof (struct lang_decl);
653 size = sizeof (struct lang_decl_flags);
655 ld = GGC_CNEWVAR (struct lang_decl, size);
657 ld->decl_flags.can_be_full = CAN_HAVE_FULL_LANG_DECL_P (t) ? 1 : 0;
658 ld->decl_flags.u1sel = TREE_CODE (t) == NAMESPACE_DECL ? 1 : 0;
659 ld->decl_flags.u2sel = 0;
660 if (ld->decl_flags.can_be_full)
661 ld->u.f.u3sel = TREE_CODE (t) == FUNCTION_DECL ? 1 : 0;
663 DECL_LANG_SPECIFIC (t) = ld;
664 if (current_lang_name == lang_name_cplusplus
665 || decl_linkage (t) == lk_none)
666 SET_DECL_LANGUAGE (t, lang_cplusplus);
667 else if (current_lang_name == lang_name_c)
668 SET_DECL_LANGUAGE (t, lang_c);
669 else if (current_lang_name == lang_name_java)
670 SET_DECL_LANGUAGE (t, lang_java);
674 #ifdef GATHER_STATISTICS
675 tree_node_counts[(int)lang_decl] += 1;
676 tree_node_sizes[(int)lang_decl] += size;
681 cxx_dup_lang_specific_decl (tree node)
684 struct lang_decl *ld;
686 if (! DECL_LANG_SPECIFIC (node))
689 if (!CAN_HAVE_FULL_LANG_DECL_P (node))
690 size = sizeof (struct lang_decl_flags);
692 size = sizeof (struct lang_decl);
693 ld = GGC_NEWVAR (struct lang_decl, size);
694 memcpy (ld, DECL_LANG_SPECIFIC (node), size);
695 DECL_LANG_SPECIFIC (node) = ld;
697 #ifdef GATHER_STATISTICS
698 tree_node_counts[(int)lang_decl] += 1;
699 tree_node_sizes[(int)lang_decl] += size;
703 /* Copy DECL, including any language-specific parts. */
706 copy_decl (tree decl)
710 copy = copy_node (decl);
711 cxx_dup_lang_specific_decl (copy);
715 /* Replace the shared language-specific parts of NODE with a new copy. */
718 copy_lang_type (tree node)
721 struct lang_type *lt;
723 if (! TYPE_LANG_SPECIFIC (node))
726 if (TYPE_LANG_SPECIFIC (node)->u.h.is_lang_type_class)
727 size = sizeof (struct lang_type);
729 size = sizeof (struct lang_type_ptrmem);
730 lt = GGC_NEWVAR (struct lang_type, size);
731 memcpy (lt, TYPE_LANG_SPECIFIC (node), size);
732 TYPE_LANG_SPECIFIC (node) = lt;
734 #ifdef GATHER_STATISTICS
735 tree_node_counts[(int)lang_type] += 1;
736 tree_node_sizes[(int)lang_type] += size;
740 /* Copy TYPE, including any language-specific parts. */
743 copy_type (tree type)
747 copy = copy_node (type);
748 copy_lang_type (copy);
753 cxx_make_type (enum tree_code code)
755 tree t = make_node (code);
757 /* Create lang_type structure. */
758 if (IS_AGGR_TYPE_CODE (code)
759 || code == BOUND_TEMPLATE_TEMPLATE_PARM)
761 struct lang_type *pi = GGC_CNEW (struct lang_type);
763 TYPE_LANG_SPECIFIC (t) = pi;
764 pi->u.c.h.is_lang_type_class = 1;
766 #ifdef GATHER_STATISTICS
767 tree_node_counts[(int)lang_type] += 1;
768 tree_node_sizes[(int)lang_type] += sizeof (struct lang_type);
772 /* Set up some flags that give proper default behavior. */
773 if (IS_AGGR_TYPE_CODE (code))
775 struct c_fileinfo *finfo = get_fileinfo (lbasename (input_filename));
776 SET_CLASSTYPE_INTERFACE_UNKNOWN_X (t, finfo->interface_unknown);
777 CLASSTYPE_INTERFACE_ONLY (t) = finfo->interface_only;
784 make_aggr_type (enum tree_code code)
786 tree t = cxx_make_type (code);
788 if (IS_AGGR_TYPE_CODE (code))
789 SET_IS_AGGR_TYPE (t, 1);