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 "flex-scanner.h"
35 #include "scan-code.h"
39 #include <get-errno.h>
42 /* The current calling start condition: SC_RULE_ACTION or
44 # define YY_DECL static char *code_lex (code_props *self, int sc_context)
47 #define YY_USER_ACTION location_compute (loc, &loc->end, yytext, yyleng);
49 static char *fetch_type_name (char *cp, char const **type_name,
52 static void handle_action_dollar (symbol_list *rule, char *cp,
54 static void handle_action_at (symbol_list *rule, char *cp, location at_loc);
56 /* A string to be pushed to obstack after dollar/at has been handled. */
57 static char *ref_tail_fields;
59 static location the_location;
60 static location *loc = &the_location;
62 /* A string representing the most recent translation. */
63 static char *last_string;
65 /* True if an untyped $$ or $n was seen. */
66 static bool untyped_var_seen;
69 /* C and C++ comments in code. */
70 %x SC_COMMENT SC_LINE_COMMENT
71 /* Strings and characters in code. */
72 %x SC_STRING SC_CHARACTER
73 /* Whether in a rule or symbol action. Specifies the translation
75 %x SC_RULE_ACTION SC_SYMBOL_ACTION
78 /* POSIX says that a tag must be both an id and a C union member, but
79 historically almost any character is allowed in a tag. We disallow
80 NUL and newline, as this simplifies our implementation. */
83 /* Zero or more instances of backslash-newline. Following GCC, allow
84 white space between the backslash and the newline. */
85 splice (\\[ \f\t\v]*\n)*
87 /* C style identifier. Must start with letter. Will be used for
88 named symbol references. Shall be kept synchronized with
89 scan-gram.l "letter" and "id". */
90 letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
91 id {letter}({letter}|[-0-9])*
92 ref -?[0-9]+|{id}|"["{id}"]"|"$"
97 /* Nesting level of the current code in braces. */
100 /* Whether a semicolon is probably needed.
102 The heuristic is that a semicolon is not needed after '{', '}',
103 ';', or a C preprocessor directive, and that whitespaces and
104 comments do not affect this flag. Note that '{' does not need a
105 semicolon because of '{}'. A semicolon may be needed before a
106 cpp directive, but don't bother.
108 While it is maintained in several start-conditions (factoring
109 opportunities), it is meaningful only for SC_RULE_ACTION. */
110 bool need_semicolon = false;
112 /* Whether in a C preprocessor directive. Don't use a start condition
113 for this because, at the end of strings and comments, we still need
114 to know whether we're in a directive. */
117 /* This scanner is special: it is invoked only once, henceforth
118 is expected to return only once. This initialization is
119 therefore done once per action to translate. */
120 aver (sc_context == SC_SYMBOL_ACTION
121 || sc_context == SC_RULE_ACTION
122 || sc_context == INITIAL);
126 /*------------------------------------------------------------.
127 | Scanning a C comment. The initial '/ *' is already eaten. |
128 `------------------------------------------------------------*/
132 "*"{splice}"/" STRING_GROW; BEGIN sc_context;
136 /*--------------------------------------------------------------.
137 | Scanning a line comment. The initial '//' is already eaten. |
138 `--------------------------------------------------------------*/
142 "\n" STRING_GROW; BEGIN sc_context;
143 {splice} STRING_GROW;
147 /*--------------------------------------------.
148 | Scanning user-code characters and strings. |
149 `--------------------------------------------*/
151 <SC_CHARACTER,SC_STRING>
153 {splice}|\\{splice}. STRING_GROW;
158 "'" STRING_GROW; BEGIN sc_context;
163 "\"" STRING_GROW; BEGIN sc_context;
167 <SC_RULE_ACTION,SC_SYMBOL_ACTION>
172 need_semicolon = true;
177 need_semicolon = true;
185 BEGIN SC_LINE_COMMENT;
188 warn_at (*loc, _("stray '%s'"), yytext);
189 obstack_escape (&obstack_for_string, yytext);
190 need_semicolon = true;
193 obstack_escape (&obstack_for_string, yytext);
194 need_semicolon = true;
200 "$"("<"{tag}">")?{ref} {
201 ref_tail_fields = NULL;
202 handle_action_dollar (self->rule, yytext, *loc);
204 obstack_sgrow (&obstack_for_string, ref_tail_fields);
205 need_semicolon = true;
208 ref_tail_fields = NULL;
209 handle_action_at (self->rule, yytext, *loc);
211 obstack_sgrow (&obstack_for_string, ref_tail_fields);
212 need_semicolon = true;
215 ";" STRING_GROW; need_semicolon = false;
216 "{" STRING_GROW; ++braces_level; need_semicolon = false;
218 bool outer_brace = --braces_level == 0;
220 /* As an undocumented Bison extension, append ';' before the last
221 brace in braced code, so that the user code can omit trailing
222 ';'. But do not append ';' if emulating Yacc, since Yacc does
224 if (outer_brace && !yacc_flag && language_prio == default_prio
225 && skeleton_prio == default_prio && need_semicolon && ! in_cpp)
227 warn_at (*loc, _("a ';' might be needed at the end of action code"));
228 warn_at (*loc, _("future versions of Bison will not add the ';'"));
229 obstack_1grow (&obstack_for_string, ';');
233 need_semicolon = false;
236 /* Preprocessing directives should only be recognized at the beginning
237 of lines, allowing whitespace including comments, but in C/C++,
238 '#' can only be the start of preprocessor directives or within
239 '#define' directives anyway, so don't bother with begin of line. */
240 "#" STRING_GROW; in_cpp = true;
242 {splice} STRING_GROW;
243 [\n\r] STRING_GROW; if (in_cpp) in_cpp = need_semicolon = false;
246 /* YYFAIL is undocumented and was formally deprecated in Bison
249 STRING_GROW; need_semicolon = true;
250 warn_at (*loc, _("use of YYFAIL, which is deprecated and will be"
254 /* The sole purpose of this is to make sure identifiers that merely
255 contain YYFAIL don't produce the above warning. */
256 [A-Za-z_][0-9A-Za-z_]* STRING_GROW; need_semicolon = true;
258 . STRING_GROW; need_semicolon = true;
263 "$"("<"{tag}">")?"$" {
264 const char *type_name = NULL;
265 fetch_type_name (yytext + 1, &type_name, *loc)[-1] = 0;
266 obstack_sgrow (&obstack_for_string, "]b4_dollar_dollar(");
267 obstack_quote (&obstack_for_string, type_name);
268 obstack_sgrow (&obstack_for_string, ")[");
269 self->is_value_used = true;
272 obstack_sgrow (&obstack_for_string, "]b4_at_dollar[");
273 locations_flag = true;
280 /* Escape M4 quoting characters in C code. */
281 [$@\[\]] obstack_escape (&obstack_for_string, yytext);
283 /* By default, grow the string obstack with the input. */
286 /* End of processing. */
287 <<EOF>> STRING_FINISH; return last_string;
293 is_dot_or_dash (char ch)
295 return ch == '.' || ch == '-';
299 contains_dot_or_dash (const char* p)
302 if (is_dot_or_dash (*p))
307 /* Defines a variant of a symbolic name resolution. */
310 /* Index in symbol list. */
311 unsigned symbol_index;
313 /* Matched symbol id and loc. */
317 /* Hiding named reference. */
318 named_ref* hidden_by;
320 /* Error flags. May contain zero (no errors) or
321 a combination of VARIANT_* values. */
325 /* Set when the variant refers to a symbol hidden
326 by an explicit symbol reference. */
327 #define VARIANT_HIDDEN (1 << 0)
329 /* Set when the variant refers to a symbol containing
330 dots or dashes. Will require explicit bracketing. */
331 #define VARIANT_BAD_BRACKETING (1 << 1)
333 /* Set when the variant refers to a symbol which is
334 not visible from current midrule. */
335 #define VARIANT_NOT_VISIBLE_FROM_MIDRULE (1 << 2)
337 static variant *variant_table = NULL;
338 static unsigned variant_table_size = 0;
339 static unsigned variant_count = 0;
342 variant_table_grow (void)
345 if (variant_count > variant_table_size)
347 while (variant_count > variant_table_size)
348 variant_table_size = 2 * variant_table_size + 3;
349 variant_table = xnrealloc (variant_table, variant_table_size,
350 sizeof *variant_table);
352 return &variant_table[variant_count - 1];
356 variant_table_free (void)
358 free (variant_table);
359 variant_table = NULL;
360 variant_table_size = variant_count = 0;
364 find_prefix_end (const char *prefix, char *begin, char *end)
368 for (; *prefix && ptr != end; ++prefix, ++ptr)
379 variant_add (uniqstr id, location id_loc, unsigned symbol_index,
380 char *cp, char *cp_end, bool explicit_bracketing)
384 prefix_end = find_prefix_end (id, cp, cp_end);
386 (prefix_end == cp_end ||
387 (!explicit_bracketing && is_dot_or_dash (*prefix_end))))
389 variant *r = variant_table_grow ();
390 r->symbol_index = symbol_index;
402 get_at_spec (unsigned symbol_index)
404 static char at_buf[20];
405 if (symbol_index == 0)
406 strcpy (at_buf, "$$");
408 snprintf (at_buf, sizeof at_buf, "$%u", symbol_index);
413 show_sub_messages (const char* cp, bool explicit_bracketing,
414 int midrule_rhs_index, char dollar_or_at,
415 bool is_warning, unsigned indent)
419 for (i = 0; i < variant_count; ++i)
421 const variant *var = &variant_table[i];
422 const char *at_spec = get_at_spec (var->symbol_index);
427 warn_at_indent (var->loc, &indent, _("refers to: %c%s at %s"),
428 dollar_or_at, var->id, at_spec);
430 complain_at_indent (var->loc, &indent, _("refers to: %c%s at %s"),
431 dollar_or_at, var->id, at_spec);
435 static struct obstack msg_buf;
436 const char *tail = explicit_bracketing ? "" :
437 cp + strlen (var->id);
438 const char *id = var->hidden_by ? var->hidden_by->id :
440 location id_loc = var->hidden_by ? var->hidden_by->loc :
443 /* Create the explanation message. */
444 obstack_init (&msg_buf);
446 obstack_fgrow1 (&msg_buf, _("possibly meant: %c"), dollar_or_at);
447 if (contains_dot_or_dash (id))
448 obstack_fgrow1 (&msg_buf, "[%s]", id);
450 obstack_sgrow (&msg_buf, id);
451 obstack_sgrow (&msg_buf, tail);
453 if (var->err & VARIANT_HIDDEN)
455 obstack_fgrow1 (&msg_buf, _(", hiding %c"), dollar_or_at);
456 if (contains_dot_or_dash (var->id))
457 obstack_fgrow1 (&msg_buf, "[%s]", var->id);
459 obstack_sgrow (&msg_buf, var->id);
460 obstack_sgrow (&msg_buf, tail);
463 obstack_fgrow1 (&msg_buf, _(" at %s"), at_spec);
465 if (var->err & VARIANT_NOT_VISIBLE_FROM_MIDRULE)
468 _(", cannot be accessed from mid-rule action at $%d");
469 obstack_fgrow1 (&msg_buf, format, midrule_rhs_index);
472 obstack_1grow (&msg_buf, '\0');
474 warn_at_indent (id_loc, &indent, "%s",
475 (char *) obstack_finish (&msg_buf));
477 complain_at_indent (id_loc, &indent, "%s",
478 (char *) obstack_finish (&msg_buf));
479 obstack_free (&msg_buf, 0);
484 /* Returned from "parse_ref" when the reference
486 #define INVALID_REF (INT_MIN)
488 /* Returned from "parse_ref" when the reference
489 points to LHS ($$) of the current rule or midrule. */
490 #define LHS_REF (INT_MIN + 1)
492 /* Sub-messages indent. */
493 #define SUB_INDENT (4)
495 /* Parse named or positional reference. In case of positional
496 references, can return negative values for $-n "deep" stack
499 parse_ref (char *cp, symbol_list *rule, int rule_length,
500 int midrule_rhs_index, char *text, location text_loc,
505 bool explicit_bracketing;
507 unsigned valid_variants = 0;
508 unsigned valid_variant_index = 0;
513 if (c_isdigit (*cp) || (*cp == '-' && c_isdigit (* (cp + 1))))
515 long int num = strtol (cp, &cp, 10);
516 if (1 - INT_MAX + rule_length <= num && num <= rule_length)
520 complain_at (text_loc, _("integer out of range: %s"),
528 /* Ignore the brackets. */
530 for (p = ++cp; *p != ']'; ++p)
534 explicit_bracketing = true;
538 /* Take all characters of the name. */
540 for (p = cp; *p; ++p)
541 if (is_dot_or_dash (*p))
546 for (p = cp; *p; ++p)
550 explicit_bracketing = false;
553 /* Add all relevant variants. */
555 unsigned symbol_index;
557 for (symbol_index = 0, l = rule; !symbol_list_null (l);
558 ++symbol_index, l = l->next)
561 if (l->content_type != SYMLIST_SYMBOL)
564 var = variant_add (l->content.sym->tag, l->sym_loc,
565 symbol_index, cp, cp_end, explicit_bracketing);
566 if (var && l->named_ref)
567 var->hidden_by = l->named_ref;
570 variant_add (l->named_ref->id, l->named_ref->loc,
571 symbol_index, cp, cp_end, explicit_bracketing);
576 for (i = 0; i < variant_count; ++i)
578 variant *var = &variant_table[i];
579 unsigned symbol_index = var->symbol_index;
581 /* Check visibility from mid-rule actions. */
582 if (midrule_rhs_index != 0
583 && (symbol_index == 0 || midrule_rhs_index < symbol_index))
584 var->err |= VARIANT_NOT_VISIBLE_FROM_MIDRULE;
586 /* Check correct bracketing. */
587 if (!explicit_bracketing && contains_dot_or_dash (var->id))
588 var->err |= VARIANT_BAD_BRACKETING;
590 /* Check using of hidden symbols. */
592 var->err |= VARIANT_HIDDEN;
596 valid_variant_index = i;
601 switch (valid_variants)
605 unsigned len = (explicit_bracketing || !ref_tail_fields) ?
606 cp_end - cp : ref_tail_fields - cp;
609 complain_at_indent (text_loc, &indent, _("invalid reference: %s"),
611 indent += SUB_INDENT;
614 location sym_loc = text_loc;
615 sym_loc.start.column += 1;
616 sym_loc.end = sym_loc.start;
618 _("syntax error after '%c', expecting integer, letter,"
619 " '_', '[', or '$'");
620 complain_at_indent (sym_loc, &indent, format, dollar_or_at);
622 else if (midrule_rhs_index)
625 _("symbol not found in production before $%d: %.*s");
626 complain_at_indent (rule->location, &indent, format,
627 midrule_rhs_index, len, cp);
632 _("symbol not found in production: %.*s");
633 complain_at_indent (rule->location, &indent, format,
637 if (variant_count > 0)
638 show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
639 dollar_or_at, false, indent);
645 if (variant_count > 1)
647 warn_at_indent (text_loc, &indent, _("misleading reference: %s"),
649 show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
650 dollar_or_at, true, indent + SUB_INDENT);
653 unsigned symbol_index =
654 variant_table[valid_variant_index].symbol_index;
655 return (symbol_index == midrule_rhs_index) ? LHS_REF : symbol_index;
662 complain_at_indent (text_loc, &indent, _("ambiguous reference: %s"),
664 show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
665 dollar_or_at, false, indent + SUB_INDENT);
674 /* Keeps track of the maximum number of semantic values to the left of
675 a handle (those referenced by $0, $-1, etc.) are required by the
676 semantic actions of this grammar. */
677 int max_left_semantic_context = 0;
680 /* If CP points to a typename (i.e., <.*?>), set TYPE_NAME to its
681 beginning (i.e., after the opening "<", and return the pointer
682 immediately after it. */
686 fetch_type_name (char *cp, char const **type_name,
695 /* The '>' symbol will be later replaced by '\0'. Original
696 'text' is needed for error messages. */
698 if (untyped_var_seen)
699 complain_at (dollar_loc, _("explicit type given in untyped grammar"));
705 /*------------------------------------------------------------------.
706 | TEXT is pointing to a wannabee semantic value (i.e., a '$'). |
708 | Possible inputs: $[<TYPENAME>]($|integer) |
710 | Output to OBSTACK_FOR_STRING a reference to this semantic value. |
711 `------------------------------------------------------------------*/
714 handle_action_dollar (symbol_list *rule, char *text, location dollar_loc)
716 char const *type_name = NULL;
718 symbol_list *effective_rule;
719 int effective_rule_length;
722 if (rule->midrule_parent_rule)
724 effective_rule = rule->midrule_parent_rule;
725 effective_rule_length = rule->midrule_parent_rhs_index - 1;
729 effective_rule = rule;
730 effective_rule_length = symbol_list_length (rule->next);
733 /* Get the type name if explicit. */
734 cp = fetch_type_name (cp, &type_name, dollar_loc);
736 n = parse_ref (cp, effective_rule, effective_rule_length,
737 rule->midrule_parent_rhs_index, text, dollar_loc, '$');
750 type_name = symbol_list_n_type_name_get (rule, dollar_loc, 0);
754 if (union_seen | tag_seen)
756 if (rule->midrule_parent_rule)
757 complain_at (dollar_loc,
758 _("$$ for the midrule at $%d of %s"
759 " has no declared type"),
760 rule->midrule_parent_rhs_index,
761 quote (effective_rule->content.sym->tag));
763 complain_at (dollar_loc, _("$$ of %s has no declared type"),
764 quote (rule->content.sym->tag));
767 untyped_var_seen = true;
770 obstack_sgrow (&obstack_for_string, "]b4_lhs_value(");
771 obstack_quote (&obstack_for_string, type_name);
772 obstack_sgrow (&obstack_for_string, ")[");
773 rule->action_props.is_value_used = true;
777 if (max_left_semantic_context < 1 - n)
778 max_left_semantic_context = 1 - n;
779 if (!type_name && 0 < n)
781 symbol_list_n_type_name_get (effective_rule, dollar_loc, n);
784 if (union_seen | tag_seen)
785 complain_at (dollar_loc, _("$%s of %s has no declared type"),
786 cp, quote (effective_rule->content.sym->tag));
788 untyped_var_seen = true;
791 obstack_fgrow2 (&obstack_for_string,
792 "]b4_rhs_value(%d, %d, ", effective_rule_length, n);
793 obstack_quote (&obstack_for_string, type_name);
794 obstack_sgrow (&obstack_for_string, ")[");
796 symbol_list_n_get (effective_rule, n)->action_props.is_value_used =
803 /*------------------------------------------------------.
804 | TEXT is a location token (i.e., a '@...'). Output to |
805 | OBSTACK_FOR_STRING a reference to this location. |
806 `------------------------------------------------------*/
809 handle_action_at (symbol_list *rule, char *text, location at_loc)
812 symbol_list *effective_rule;
813 int effective_rule_length;
816 if (rule->midrule_parent_rule)
818 effective_rule = rule->midrule_parent_rule;
819 effective_rule_length = rule->midrule_parent_rhs_index - 1;
823 effective_rule = rule;
824 effective_rule_length = symbol_list_length (rule->next);
827 locations_flag = true;
829 n = parse_ref (cp, effective_rule, effective_rule_length,
830 rule->midrule_parent_rhs_index, text, at_loc, '@');
837 obstack_sgrow (&obstack_for_string, "]b4_lhs_location[");
841 obstack_fgrow2 (&obstack_for_string, "]b4_rhs_location(%d, %d)[",
842 effective_rule_length, n);
848 /*-------------------------.
849 | Initialize the scanner. |
850 `-------------------------*/
852 /* Translate the dollars and ats in \a self, in the context \a sc_context
853 (SC_RULE_ACTION, SC_SYMBOL_ACTION, INITIAL). */
856 translate_action (code_props *self, int sc_context)
859 static bool initialized = false;
862 obstack_init (&obstack_for_string);
867 loc->start = loc->end = self->location.start;
868 yy_switch_to_buffer (yy_scan_string (self->code));
869 res = code_lex (self, sc_context);
870 yy_delete_buffer (YY_CURRENT_BUFFER);
875 /*------------------------------------------------------------------------.
876 | Implementation of the public interface as documented in "scan-code.h". |
877 `------------------------------------------------------------------------*/
880 code_props_none_init (code_props *self)
882 *self = code_props_none;
885 code_props const code_props_none = CODE_PROPS_NONE_INIT;
888 code_props_plain_init (code_props *self, char const *code,
891 self->kind = CODE_PROPS_PLAIN;
893 self->location = code_loc;
894 self->is_value_used = false;
896 self->named_ref = NULL;
900 code_props_symbol_action_init (code_props *self, char const *code,
903 self->kind = CODE_PROPS_SYMBOL_ACTION;
905 self->location = code_loc;
906 self->is_value_used = false;
908 self->named_ref = NULL;
912 code_props_rule_action_init (code_props *self, char const *code,
913 location code_loc, symbol_list *rule,
916 self->kind = CODE_PROPS_RULE_ACTION;
918 self->location = code_loc;
919 self->is_value_used = false;
921 self->named_ref = name;
925 code_props_translate_code (code_props *self)
929 case CODE_PROPS_NONE:
931 case CODE_PROPS_PLAIN:
932 self->code = translate_action (self, INITIAL);
934 case CODE_PROPS_SYMBOL_ACTION:
935 self->code = translate_action (self, SC_SYMBOL_ACTION);
937 case CODE_PROPS_RULE_ACTION:
938 self->code = translate_action (self, SC_RULE_ACTION);
944 code_scanner_last_string_free (void)
950 code_scanner_free (void)
952 obstack_free (&obstack_for_string, 0);
953 variant_table_free ();
955 /* Reclaim Flex's buffers. */