1 /* Bison Action Scanner -*- C -*-
3 Copyright (C) 2006-2012 Free Software Foundation, Inc.
5 This file is part of Bison, the GNU Compiler Compiler.
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/>. */
20 %option debug nodefault noinput nounput noyywrap never-interactive
21 %option prefix="code_" outfile="lex.yy.c"
24 /* Work around a bug in flex 2.5.31. See Debian bug 333231
25 <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=333231>. */
29 #define FLEX_PREFIX(Id) code_ ## Id
30 #include <src/flex-scanner.h>
32 #include <src/complain.h>
33 #include <src/reader.h>
34 #include <src/getargs.h>
35 #include <src/muscle-tab.h>
36 #include <src/scan-code.h>
37 #include <src/symlist.h>
40 #include <get-errno.h>
43 /* The current calling start condition: SC_RULE_ACTION or
45 # define YY_DECL static char *code_lex (code_props *self, int sc_context)
48 #define YY_USER_ACTION location_compute (loc, &loc->end, yytext, yyleng);
50 static void handle_action_dollar (symbol_list *rule, char *cp,
52 static void handle_action_at (symbol_list *rule, char *cp, location at_loc);
54 /* A string to be pushed to obstack after dollar/at has been handled. */
55 static char *ref_tail_fields;
57 static location the_location;
58 static location *loc = &the_location;
60 /* A string representing the most recent translation. */
61 static char *last_string;
63 /* True if an untyped $$ or $n was seen. */
64 static bool untyped_var_seen;
67 /* C and C++ comments in code. */
68 %x SC_COMMENT SC_LINE_COMMENT
69 /* Strings and characters in code. */
70 %x SC_STRING SC_CHARACTER
71 /* Whether in a rule or symbol action. Specifies the translation
73 %x SC_RULE_ACTION SC_SYMBOL_ACTION
76 /* POSIX says that a tag must be both an id and a C union member, but
77 historically almost any character is allowed in a tag. We disallow
78 NUL and newline, as this simplifies our implementation. */
81 /* Zero or more instances of backslash-newline. Following GCC, allow
82 white space between the backslash and the newline. */
83 splice (\\[ \f\t\v]*\n)*
85 /* C style identifier. Must start with letter. Will be used for
86 named symbol references. Shall be kept synchronized with
87 scan-gram.l "letter" and "id". */
88 letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
89 id {letter}({letter}|[-0-9])*
90 ref -?[0-9]+|{id}|"["{id}"]"|"$"
95 /* Nesting level of the current code in braces. */
98 /* Whether a semicolon is probably needed.
99 The heuristic is that a semicolon is not needed after '{', '}', ';',
100 or a C preprocessor directive, and that whitespaces and comments
101 do not affect this flag.
102 Note that '{' does not need a semicolon because of '{}'.
103 A semicolon may be needed before a cpp direcive, but don't bother. */
104 bool need_semicolon = false;
106 /* Whether in a C preprocessor directive. Don't use a start condition
107 for this because, at the end of strings and comments, we still need
108 to know whether we're in a directive. */
111 /* This scanner is special: it is invoked only once, henceforth
112 is expected to return only once. This initialization is
113 therefore done once per action to translate. */
114 aver (sc_context == SC_SYMBOL_ACTION
115 || sc_context == SC_RULE_ACTION
116 || sc_context == INITIAL);
120 /*------------------------------------------------------------.
121 | Scanning a C comment. The initial '/ *' is already eaten. |
122 `------------------------------------------------------------*/
126 "*"{splice}"/" STRING_GROW; BEGIN sc_context;
130 /*--------------------------------------------------------------.
131 | Scanning a line comment. The initial '//' is already eaten. |
132 `--------------------------------------------------------------*/
136 "\n" STRING_GROW; BEGIN sc_context;
137 {splice} STRING_GROW;
141 /*--------------------------------------------.
142 | Scanning user-code characters and strings. |
143 `--------------------------------------------*/
145 <SC_CHARACTER,SC_STRING>
147 {splice}|\\{splice}. STRING_GROW;
152 "'" STRING_GROW; BEGIN sc_context;
157 "\"" STRING_GROW; BEGIN sc_context;
161 <SC_RULE_ACTION,SC_SYMBOL_ACTION>{
165 need_semicolon = true;
170 need_semicolon = true;
178 BEGIN SC_LINE_COMMENT;
184 "$"("<"{tag}">")?{ref} {
186 handle_action_dollar (self->rule, yytext, *loc);
187 if (ref_tail_fields) {
188 obstack_sgrow (&obstack_for_string, ref_tail_fields);
190 need_semicolon = true;
194 handle_action_at (self->rule, yytext, *loc);
195 if (ref_tail_fields) {
196 obstack_sgrow (&obstack_for_string, ref_tail_fields);
198 need_semicolon = true;
201 complain_at (*loc, Wother, _("stray '$'"));
202 obstack_sgrow (&obstack_for_string, "$][");
203 need_semicolon = true;
206 complain_at (*loc, Wother, _("stray '@'"));
207 obstack_sgrow (&obstack_for_string, "@@");
208 need_semicolon = true;
211 obstack_sgrow (&obstack_for_string, "@{");
212 need_semicolon = true;
215 obstack_sgrow (&obstack_for_string, "@}");
216 need_semicolon = true;
219 ";" STRING_GROW; need_semicolon = false;
220 "{" STRING_GROW; ++braces_level; need_semicolon = false;
222 bool outer_brace = --braces_level == 0;
224 /* As an undocumented Bison extension, append ';' before the last
225 brace in braced code, so that the user code can omit trailing
226 ';'. But do not append ';' if emulating Yacc, since Yacc does
228 if (outer_brace && !yacc_flag && language_prio == default_prio
229 && skeleton_prio == default_prio && need_semicolon && ! in_cpp)
231 complain_at (*loc, Wother,
232 _("a ';' might be needed at the end of action code"));
233 complain_at (*loc, Wother,
234 _("future versions of Bison will not add the ';'"));
235 obstack_1grow (&obstack_for_string, ';');
239 need_semicolon = false;
242 /* Preprocessing directives should only be recognized at the beginning
243 of lines, allowing whitespace including comments, but in C/C++,
244 '#' can only be the start of preprocessor directives or within
245 '#define' directives anyway, so don't bother with begin of line. */
246 "#" STRING_GROW; in_cpp = true;
248 {splice} STRING_GROW;
249 [\n\r] STRING_GROW; if (in_cpp) in_cpp = need_semicolon = false;
252 /* YYFAIL is undocumented and was formally deprecated in Bison
255 STRING_GROW; need_semicolon = true;
256 complain_at (*loc, Wother,
257 _("use of YYFAIL, which is deprecated and will be removed"));
260 /* The sole purpose of this is to make sure identifiers that merely
261 contain YYFAIL don't produce the above warning. */
262 [A-Za-z_][0-9A-Za-z_]* STRING_GROW; need_semicolon = true;
264 . STRING_GROW; need_semicolon = true;
270 obstack_sgrow (&obstack_for_string, "]b4_dollar_dollar[");
271 self->is_value_used = true;
274 obstack_sgrow (&obstack_for_string, "]b4_at_dollar[");
275 muscle_percent_define_ensure("locations", the_location, true);
280 /*-----------------------------------------.
281 | Escape M4 quoting characters in C code. |
282 `-----------------------------------------*/
286 \$ obstack_sgrow (&obstack_for_string, "$][");
287 \@ obstack_sgrow (&obstack_for_string, "@@");
288 \[ obstack_sgrow (&obstack_for_string, "@{");
289 \] obstack_sgrow (&obstack_for_string, "@}");
292 /*-----------------------------------------------------.
293 | By default, grow the string obstack with the input. |
294 `-----------------------------------------------------*/
298 /* End of processing. */
307 is_dot_or_dash (char ch)
309 return ch == '.' || ch == '-';
313 contains_dot_or_dash (const char* p)
316 if (is_dot_or_dash (*p))
321 /* Defines a variant of a symbolic name resolution. */
324 /* Index in symbol list. */
325 unsigned symbol_index;
327 /* Matched symbol id and loc. */
331 /* Hiding named reference. */
332 named_ref* hidden_by;
334 /* Error flags. May contain zero (no errors) or
335 a combination of VARIANT_* values. */
339 /* Set when the variant refers to a symbol hidden
340 by an explicit symbol reference. */
341 #define VARIANT_HIDDEN (1 << 0)
343 /* Set when the variant refers to a symbol containing
344 dots or dashes. Will require explicit bracketing. */
345 #define VARIANT_BAD_BRACKETING (1 << 1)
347 /* Set when the variant refers to a symbol which is
348 not visible from current midrule. */
349 #define VARIANT_NOT_VISIBLE_FROM_MIDRULE (1 << 2)
351 static variant *variant_table = 0;
352 static unsigned variant_table_size = 0;
353 static unsigned variant_count = 0;
356 variant_table_grow (void)
359 if (variant_count > variant_table_size)
361 while (variant_count > variant_table_size)
362 variant_table_size = 2 * variant_table_size + 3;
363 variant_table = xnrealloc (variant_table, variant_table_size,
364 sizeof *variant_table);
366 return &variant_table[variant_count - 1];
370 variant_table_free (void)
372 free (variant_table);
374 variant_table_size = variant_count = 0;
378 find_prefix_end (const char *prefix, char *begin, char *end)
382 for (; *prefix && ptr != end; ++prefix, ++ptr)
393 variant_add (uniqstr id, location id_loc, unsigned symbol_index,
394 char *cp, char *cp_end, bool explicit_bracketing)
398 prefix_end = find_prefix_end (id, cp, cp_end);
400 (prefix_end == cp_end ||
401 (!explicit_bracketing && is_dot_or_dash (*prefix_end))))
403 variant *r = variant_table_grow ();
404 r->symbol_index = symbol_index;
416 get_at_spec(unsigned symbol_index)
418 static char at_buf[20];
419 if (symbol_index == 0)
420 strcpy (at_buf, "$$");
422 snprintf (at_buf, sizeof at_buf, "$%u", symbol_index);
427 show_sub_messages (const char* cp, bool explicit_bracketing,
428 int midrule_rhs_index, char dollar_or_at,
429 bool is_warning, unsigned indent)
433 for (i = 0; i < variant_count; ++i)
435 const variant *var = &variant_table[i];
436 const char *at_spec = get_at_spec (var->symbol_index);
441 complain_at_indent (var->loc, Wother, &indent,
442 _("refers to: %c%s at %s"), dollar_or_at,
445 complain_at_indent (var->loc, complaint, &indent,
446 _("refers to: %c%s at %s"), dollar_or_at,
451 static struct obstack msg_buf;
452 const char *tail = explicit_bracketing ? "" :
453 cp + strlen (var->id);
454 const char *id = var->hidden_by ? var->hidden_by->id :
456 location id_loc = var->hidden_by ? var->hidden_by->loc :
459 /* Create the explanation message. */
460 obstack_init (&msg_buf);
462 obstack_fgrow1 (&msg_buf, _("possibly meant: %c"), dollar_or_at);
463 if (contains_dot_or_dash (id))
464 obstack_fgrow1 (&msg_buf, "[%s]", id);
466 obstack_sgrow (&msg_buf, id);
467 obstack_sgrow (&msg_buf, tail);
469 if (var->err & VARIANT_HIDDEN)
471 obstack_fgrow1 (&msg_buf, _(", hiding %c"), dollar_or_at);
472 if (contains_dot_or_dash (var->id))
473 obstack_fgrow1 (&msg_buf, "[%s]", var->id);
475 obstack_sgrow (&msg_buf, var->id);
476 obstack_sgrow (&msg_buf, tail);
479 obstack_fgrow1 (&msg_buf, _(" at %s"), at_spec);
481 if (var->err & VARIANT_NOT_VISIBLE_FROM_MIDRULE)
484 _(", cannot be accessed from mid-rule action at $%d");
485 obstack_fgrow1 (&msg_buf, format, midrule_rhs_index);
488 obstack_1grow (&msg_buf, '\0');
490 complain_at_indent (id_loc, Wother, &indent, "%s",
491 (char *) obstack_finish (&msg_buf));
493 complain_at_indent (id_loc, complaint, &indent, "%s",
494 (char *) obstack_finish (&msg_buf));
495 obstack_free (&msg_buf, 0);
500 /* Returned from "parse_ref" when the reference
502 #define INVALID_REF (INT_MIN)
504 /* Returned from "parse_ref" when the reference
505 points to LHS ($$) of the current rule or midrule. */
506 #define LHS_REF (INT_MIN + 1)
508 /* Sub-messages indent. */
509 #define SUB_INDENT (4)
511 /* Parse named or positional reference. In case of positional
512 references, can return negative values for $-n "deep" stack
515 parse_ref (char *cp, symbol_list *rule, int rule_length,
516 int midrule_rhs_index, char *text, location text_loc,
521 bool explicit_bracketing;
523 unsigned valid_variants = 0;
524 unsigned valid_variant_index = 0;
529 if (c_isdigit (*cp) || (*cp == '-' && c_isdigit (* (cp + 1))))
531 long int num = strtol (cp, &cp, 10);
532 if (1 - INT_MAX + rule_length <= num && num <= rule_length)
536 complain_at (text_loc, complaint, _("integer out of range: %s"),
544 /* Ignore the brackets. */
546 for (p = ++cp; *p != ']'; ++p)
550 explicit_bracketing = true;
554 /* Take all characters of the name. */
556 for (p = cp; *p; ++p)
557 if (is_dot_or_dash (*p))
562 for (p = cp; *p; ++p)
566 explicit_bracketing = false;
569 /* Add all relevant variants. */
571 unsigned symbol_index;
573 for (symbol_index = 0, l = rule; !symbol_list_null (l);
574 ++symbol_index, l = l->next)
577 if (l->content_type != SYMLIST_SYMBOL)
580 var = variant_add (l->content.sym->tag, l->sym_loc,
581 symbol_index, cp, cp_end, explicit_bracketing);
582 if (var && l->named_ref)
583 var->hidden_by = l->named_ref;
586 variant_add (l->named_ref->id, l->named_ref->loc,
587 symbol_index, cp, cp_end, explicit_bracketing);
592 for (i = 0; i < variant_count; ++i)
594 variant *var = &variant_table[i];
595 unsigned symbol_index = var->symbol_index;
597 /* Check visibility from mid-rule actions. */
598 if (midrule_rhs_index != 0
599 && (symbol_index == 0 || midrule_rhs_index < symbol_index))
600 var->err |= VARIANT_NOT_VISIBLE_FROM_MIDRULE;
602 /* Check correct bracketing. */
603 if (!explicit_bracketing && contains_dot_or_dash (var->id))
604 var->err |= VARIANT_BAD_BRACKETING;
606 /* Check using of hidden symbols. */
608 var->err |= VARIANT_HIDDEN;
612 valid_variant_index = i;
617 switch (valid_variants)
621 unsigned len = (explicit_bracketing || !ref_tail_fields) ?
622 cp_end - cp : ref_tail_fields - cp;
625 complain_at_indent (text_loc, complaint, &indent,
626 _("invalid reference: %s"), quote (text));
627 indent += SUB_INDENT;
630 location sym_loc = text_loc;
631 sym_loc.start.column += 1;
632 sym_loc.end = sym_loc.start;
634 _("syntax error after '%c', expecting integer, letter,"
635 " '_', '[', or '$'");
636 complain_at_indent (sym_loc, complaint, &indent, format,
639 else if (midrule_rhs_index)
642 _("symbol not found in production before $%d: %.*s");
643 complain_at_indent (rule->location, complaint, &indent, format,
644 midrule_rhs_index, len, cp);
649 _("symbol not found in production: %.*s");
650 complain_at_indent (rule->location, complaint, &indent, format,
654 if (variant_count > 0)
655 show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
656 dollar_or_at, false, indent);
662 if (variant_count > 1)
664 complain_at_indent (text_loc, Wother, &indent,
665 _("misleading reference: %s"), quote (text));
666 show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
667 dollar_or_at, true, indent + SUB_INDENT);
670 unsigned symbol_index =
671 variant_table[valid_variant_index].symbol_index;
672 return (symbol_index == midrule_rhs_index) ? LHS_REF : symbol_index;
679 complain_at_indent (text_loc, complaint, &indent,
680 _("ambiguous reference: %s"), quote (text));
681 show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
682 dollar_or_at, false, indent + SUB_INDENT);
691 /* Keeps track of the maximum number of semantic values to the left of
692 a handle (those referenced by $0, $-1, etc.) are required by the
693 semantic actions of this grammar. */
694 int max_left_semantic_context = 0;
697 /*------------------------------------------------------------------.
698 | TEXT is pointing to a wannabee semantic value (i.e., a '$'). |
700 | Possible inputs: $[<TYPENAME>]($|integer) |
702 | Output to OBSTACK_FOR_STRING a reference to this semantic value. |
703 `------------------------------------------------------------------*/
706 handle_action_dollar (symbol_list *rule, char *text, location dollar_loc)
708 char const *type_name = NULL;
711 symbol_list *effective_rule;
712 int effective_rule_length;
715 if (rule->midrule_parent_rule)
717 effective_rule = rule->midrule_parent_rule;
718 effective_rule_length = rule->midrule_parent_rhs_index - 1;
722 effective_rule = rule;
723 effective_rule_length = symbol_list_length (rule->next);
726 /* Get the type name if explicit. */
733 /* The '>' symbol will be later replaced by '\0'. Original
734 'text' is needed for error messages. */
737 if (untyped_var_seen)
738 complain_at (dollar_loc, complaint,
739 _("explicit type given in untyped grammar"));
743 n = parse_ref (cp, effective_rule, effective_rule_length,
744 rule->midrule_parent_rhs_index, text, dollar_loc, '$');
756 type_name = symbol_list_n_type_name_get (rule, dollar_loc, 0);
760 if (union_seen | tag_seen)
762 if (rule->midrule_parent_rule)
763 complain_at (dollar_loc, complaint,
764 _("$$ for the midrule at $%d of %s"
765 " has no declared type"),
766 rule->midrule_parent_rhs_index,
767 quote (effective_rule->content.sym->tag));
769 complain_at (dollar_loc, complaint,
770 _("$$ of %s has no declared type"),
771 quote (rule->content.sym->tag));
774 untyped_var_seen = true;
778 obstack_fgrow1 (&obstack_for_string,
779 "]b4_lhs_value([%s])[", type_name);
780 rule->action_props.is_value_used = true;
784 if (max_left_semantic_context < 1 - n)
785 max_left_semantic_context = 1 - n;
786 if (!type_name && 0 < n)
788 symbol_list_n_type_name_get (effective_rule, dollar_loc, n);
791 if (union_seen | tag_seen)
792 complain_at (dollar_loc, complaint,
793 _("$%s of %s has no declared type"), cp,
794 quote (effective_rule->content.sym->tag));
796 untyped_var_seen = true;
800 obstack_fgrow3 (&obstack_for_string,
801 "]b4_rhs_value(%d, %d, [%s])[",
802 effective_rule_length, n, type_name);
804 symbol_list_n_get (effective_rule, n)->action_props.is_value_used =
811 /*------------------------------------------------------.
812 | TEXT is a location token (i.e., a '@...'). Output to |
813 | OBSTACK_FOR_STRING a reference to this location. |
814 `------------------------------------------------------*/
817 handle_action_at (symbol_list *rule, char *text, location at_loc)
820 symbol_list *effective_rule;
821 int effective_rule_length;
824 if (rule->midrule_parent_rule)
826 effective_rule = rule->midrule_parent_rule;
827 effective_rule_length = rule->midrule_parent_rhs_index - 1;
831 effective_rule = rule;
832 effective_rule_length = symbol_list_length (rule->next);
835 muscle_percent_define_ensure("locations", at_loc, true);
837 n = parse_ref (cp, effective_rule, effective_rule_length,
838 rule->midrule_parent_rhs_index, text, at_loc, '@');
845 obstack_sgrow (&obstack_for_string, "]b4_lhs_location[");
849 obstack_fgrow2 (&obstack_for_string, "]b4_rhs_location(%d, %d)[",
850 effective_rule_length, n);
856 /*-------------------------.
857 | Initialize the scanner. |
858 `-------------------------*/
860 /* Translate the dollars and ats in \a self, in the context \a sc_context
861 (SC_RULE_ACTION, SC_SYMBOL_ACTION, INITIAL). */
864 translate_action (code_props *self, int sc_context)
867 static bool initialized = false;
870 obstack_init (&obstack_for_string);
875 loc->start = loc->end = self->location.start;
876 yy_switch_to_buffer (yy_scan_string (self->code));
877 res = code_lex (self, sc_context);
878 yy_delete_buffer (YY_CURRENT_BUFFER);
883 /*------------------------------------------------------------------------.
884 | Implementation of the public interface as documented in "scan-code.h". |
885 `------------------------------------------------------------------------*/
888 code_props_none_init (code_props *self)
890 *self = code_props_none;
893 code_props code_props_none = CODE_PROPS_NONE_INIT;
896 code_props_plain_init (code_props *self, char const *code,
899 code_props_none_init (self);
900 self->kind = CODE_PROPS_PLAIN;
902 self->location = code_loc;
906 code_props_symbol_action_init (code_props *self, char const *code,
909 code_props_none_init (self);
910 self->kind = CODE_PROPS_SYMBOL_ACTION;
912 self->location = code_loc;
916 code_props_rule_action_init (code_props *self, char const *code,
917 location code_loc, symbol_list *rule,
918 named_ref *name, bool is_predicate)
920 code_props_none_init (self);
921 self->kind = CODE_PROPS_RULE_ACTION;
923 self->location = code_loc;
925 self->named_ref = name;
926 self->is_predicate = is_predicate;
930 code_props_translate_code (code_props *self)
934 case CODE_PROPS_NONE:
936 case CODE_PROPS_PLAIN:
937 self->code = translate_action (self, INITIAL);
939 case CODE_PROPS_SYMBOL_ACTION:
940 self->code = translate_action (self, SC_SYMBOL_ACTION);
942 case CODE_PROPS_RULE_ACTION:
943 self->code = translate_action (self, SC_RULE_ACTION);
949 code_scanner_last_string_free (void)
955 code_scanner_free (void)
957 obstack_free (&obstack_for_string, 0);
958 variant_table_free ();
960 /* Reclaim Flex's buffers. */