[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH] Do not allow identifiers that start with a negative number.
|
From: |
Alex Rozenman |
|
Subject: |
Re: [PATCH] Do not allow identifiers that start with a negative number. |
|
Date: |
Wed, 26 Jan 2011 20:44:40 +0200 |
Hi Paul, Hi All.
> How about the following patch? It tries to simplify things, at some
> cost in niceties in diagnostics. I pushed it into the trunk. Further
> comments are welcome.
>
This is no doubt that from the *code* point of view, your patch simplifies
things and makes the code much more reasonable than it was before. I also
don't like the "ref_tail_fields" trick, ad-hoc scanning in C code and
"find_prefix_end". However, in my opinion, there is too much functionality
that's missing now. I've just taken some examples from the test-suite:
start: foo foo.bar { $foo.bar; }
*** branch-2.5:
test1.y:2.22-29: warning: misleading reference: `$foo.bar'
test1.y:2.8-10: warning: refers to: $foo at $1
test1.y:2.12-18: warning: possibly meant: $[foo.bar] at $2
*** master:
(compiles silently)
I agree that the user will probably get an error in C compiler, but the
existing warning is very usable and may show the user what is the real
problem, IMHO.
if_stmt6: IF expr[cond] THEN stmt.list[then] ELSE stmt.list[else] FI
{ $if_stmt6 = new IfStmt($cond, $stmt.list.field, $else); };
*** branch-2.5:
test2.y:27.43-58: invalid reference: `$stmt.list.field'
test2.y:26.11-27.69: symbol not found in production: stmt
test2.y:26.40-43: possibly meant: $then.field, hiding $[stmt.list].field
at $4
test2.y:26.61-64: possibly meant: $else.field, hiding $[stmt.list].field
at $6
*** master:
test2.y:27.43-47: invalid reference: `$stmt'
test2.y:26.11-27.69: symbol not found in production: stmt
Here, the additional piece of information (about hiding) is missing. It is
also not clear, which 'stmt' is actually invalid.
If user's eye will always "parse" named references as $c_id.fields, shorter
error messages will be enough, and the overall picture will be clearer. But
we'll understand it only from user's feedbacks.
It is not clear for me, which part of the tricky code would go away if it
will be decided to re-implement id[id] parsing in parse-gram.y; it seems
that "ref_tail_fields" will still be there (if we would like to keep the
error messages untouched).
Note, that Joel's modification '5c9efc755' is redundant because of the
patch. I also will need to review the code again and remove additional
redundant sections, if it will be decided to keep it. Sorry for the long
delays and the short response - I am still very limited in time.
Alex
>
> From 5ca9ccffbc145bbd8d7cba00fae0cfd2d318a0c3 Mon Sep 17 00:00:00 2001
> From: Paul Eggert <address@hidden>
> Date: Sun, 9 Jan 2011 23:26:12 -0800
> Subject: [PATCH] Simplify handling of '.' and '-' after unbracketed named
> references.
>
> * doc/bison.texinfo (Mid-Rule Actions): Mention that periods and
> dashes make symbol names less convenient for named references.
> * src/scan-code.l:
> (handle_action_dollar): New arg textlen. All callers changed.
> (handle_action_at): Likewise. Also, args are pointers to const.
> (ref_tail_fields): Remove; no longer used.
> (letter): Now includes '-' and '.', since this is for Bison
> identifiers.
> (id): Now the simpler traditional defn, since letters now include
> '-' and '.'.
> (c_letter, c_id): New defns.
> (ref): Use c_id for unbracketed IDs.
> (<SC_RULE_ACTION>): Simplify, now that the distinction between
> Bison and unbracketed IDs are now in the regular expressions.
> (VARIANT_BAD_BRACKETING): Remove.
> (VARIANT_NOT_VISIBLE_FROM_MIDRULE): Renumber.
> (find_prefix_end): Remove, replacing with ....
> (identifier_matches): New function.
> (variant_add): Use it. Omit EXPLICIT_BRACKETING arg; no longer
> needed. CP arg is pointer to constant. All callers changed.
> (show_sub_messages): Remove args CP, EXPLICIT_BRACKETING, DOLLAR_OR_AT.
> New arg TEXT. All callers changed. Do not worry about showing
> trailing context.
> (parse_ref): Args CP, RULE, TEXT are now pointers to const. New
> arg TEXTLEN. Remove arg DOLLAR_OR_AT. All callers changed.
> Simplify code now that the regular expressions capture the
> restrictions.
> * src/scan-gram.l (letter, id): Adjust to match scan-code.l.
> * src/symlist.c (symbol_list_null): Arg is now pointer to const.
> * src/symlist.h: Likewise.
> * tests/named-refs.at (Misleading references): These are now caught
> by the C compiler, not by Bison; that's good enough. Adjust test
> to reflect this.
> (Many kinds of errors, Unresolved references): Adjust expected
> diagnostics to match new behavior. The same errors are caught,
> though the diagnostics are not quite as fancy.
> ($ or @ followed by . or -): Likewise. Also, Make the grammar
> unambiguous, so that diagnostics are not complicated by ambiguity
> warnings.
> ---
> ChangeLog | 41 ++++++++++++
> doc/bison.texinfo | 4 +-
> src/scan-code.l | 182
> ++++++++++++++++++++-------------------------------
> src/scan-gram.l | 5 +-
> src/symlist.c | 2 +-
> src/symlist.h | 2 +-
> tests/named-refs.at | 70 +++++++------------
> 7 files changed, 147 insertions(+), 159 deletions(-)
>
> diff --git a/ChangeLog b/ChangeLog
> index 9255aab..1ed26ad 100644
> --- a/ChangeLog
> +++ b/ChangeLog
> @@ -1,5 +1,46 @@
> 2011-01-09 Paul Eggert <address@hidden>
>
> + Simplify handling of '.' and '-' after unbracketed named
> references.
> + * doc/bison.texinfo (Mid-Rule Actions): Mention that periods and
> + dashes make symbol names less convenient for named references.
> + * src/scan-code.l:
> + (handle_action_dollar): New arg textlen. All callers changed.
> + (handle_action_at): Likewise. Also, args are pointers to const.
> + (ref_tail_fields): Remove; no longer used.
> + (letter): Now includes '-' and '.', since this is for Bison
> + identifiers.
> + (id): Now the simpler traditional defn, since letters now include
> + '-' and '.'.
> + (c_letter, c_id): New defns.
> + (ref): Use c_id for unbracketed IDs.
> + (<SC_RULE_ACTION>): Simplify, now that the distinction between
> + Bison and unbracketed IDs are now in the regular expressions.
> + (VARIANT_BAD_BRACKETING): Remove.
> + (VARIANT_NOT_VISIBLE_FROM_MIDRULE): Renumber.
> + (find_prefix_end): Remove, replacing with ....
> + (identifier_matches): New function.
> + (variant_add): Use it. Omit EXPLICIT_BRACKETING arg; no longer
> + needed. CP arg is pointer to constant. All callers changed.
> + (show_sub_messages): Remove args CP, EXPLICIT_BRACKETING,
> DOLLAR_OR_AT.
> + New arg TEXT. All callers changed. Do not worry about showing
> + trailing context.
> + (parse_ref): Args CP, RULE, TEXT are now pointers to const. New
> + arg TEXTLEN. Remove arg DOLLAR_OR_AT. All callers changed.
> + Simplify code now that the regular expressions capture the
> + restrictions.
> + * src/scan-gram.l (letter, id): Adjust to match scan-code.l.
> + * src/symlist.c (symbol_list_null): Arg is now pointer to const.
> + * src/symlist.h: Likewise.
> + * tests/named-refs.at (Misleading references): These are now
> caught
> + by the C compiler, not by Bison; that's good enough. Adjust test
> + to reflect this.
> + (Many kinds of errors, Unresolved references): Adjust expected
> + diagnostics to match new behavior. The same errors are caught,
> + though the diagnostics are not quite as fancy.
> + ($ or @ followed by . or -): Likewise. Also, Make the grammar
> + unambiguous, so that diagnostics are not complicated by ambiguity
> + warnings.
> +
> Fix minor problems encountered by a fresh bootstrap.
> * data/glr.c, data/yacc.c: Do not use apostrophes in '#' comments,
> as they confuse xgettext, which tries to parse them as C character
> diff --git a/doc/bison.texinfo b/doc/bison.texinfo
> index cc1e064..29d6364 100644
> --- a/doc/bison.texinfo
> +++ b/doc/bison.texinfo
> @@ -3125,7 +3125,9 @@ By convention, it should be all lower case.
>
> Symbol names can contain letters, underscores, periods, dashes, and (not
> at the beginning) digits. Dashes in symbol names are a GNU
> -extension, incompatible with POSIX Yacc. Terminal symbols
> +extension, incompatible with POSIX Yacc. Periods and dashes make symbol
> +names less convenient to use with named references, which require brackets
> +around such names (@pxref{Named References}). Terminal symbols
> that contain periods or dashes make little sense: since they are not
> valid symbols (in most programming languages) they are not exported as
> token names.
> diff --git a/src/scan-code.l b/src/scan-code.l
> index 3dd1044..9a08af7 100644
> --- a/src/scan-code.l
> +++ b/src/scan-code.l
> @@ -47,12 +47,9 @@ YY_DECL;
>
> #define YY_USER_ACTION location_compute (loc, &loc->end, yytext, yyleng);
>
> -static void handle_action_dollar (symbol_list *rule, char *cp,
> - location dollar_loc);
> -static void handle_action_at (symbol_list *rule, char *cp, location
> at_loc);
> -
> -/* A string to be pushed to obstack after dollar/at has been handled. */
> -static char *ref_tail_fields;
> +static void handle_action_dollar (symbol_list *, char *, unsigned,
> location);
> +static void handle_action_at (symbol_list const *, char const *, unsigned,
> + location);
>
> static location the_location;
> static location *loc = &the_location;
> @@ -82,12 +79,16 @@ tag [^\0\n>]+
> white space between the backslash and the newline. */
> splice (\\[ \f\t\v]*\n)*
>
> -/* C style identifier. Must start with letter. Will be used for
> - named symbol references. Shall be kept synchronized with
> - scan-gram.l "letter" and "id". */
> -letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
> -id -*(-|{letter}({letter}|[-0-9])*)
> -ref -?[0-9]+|{id}|"["{id}"]"|"$"
> +/* A Bison identifier. Keep this synchronized with scan-gram.l "id". */
> +letter [-.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
> +id {letter}({letter}|[0-9])*
> +
> +/* An identifier that can appear unbracketed in a reference.
> + This happens to be the same as a C-language identifier. */
> +c_letter [abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
> +c_id {c_letter}({c_letter}|[0-9])*
> +
> +ref -?[0-9]+|{c_id}|"["{id}"]"|"$"
>
> %%
>
> @@ -182,19 +183,11 @@ ref -?[0-9]+|{id}|"["{id}"]"|"$"
> <SC_RULE_ACTION>
> {
> "$"("<"{tag}">")?{ref} {
> - ref_tail_fields = 0;
> - handle_action_dollar (self->rule, yytext, *loc);
> - if (ref_tail_fields) {
> - obstack_sgrow (&obstack_for_string, ref_tail_fields);
> - }
> + handle_action_dollar (self->rule, yytext, yyleng, *loc);
> need_semicolon = true;
> }
> "@"{ref} {
> - ref_tail_fields = 0;
> - handle_action_at (self->rule, yytext, *loc);
> - if (ref_tail_fields) {
> - obstack_sgrow (&obstack_for_string, ref_tail_fields);
> - }
> + handle_action_at (self->rule, yytext, yyleng, *loc);
> need_semicolon = true;
> }
> "$" {
> @@ -338,13 +331,9 @@ typedef struct
> by an explicit symbol reference. */
> #define VARIANT_HIDDEN (1 << 0)
>
> -/* Set when the variant refers to a symbol containing
> - dots or dashes. Will require explicit bracketing. */
> -#define VARIANT_BAD_BRACKETING (1 << 1)
> -
> /* Set when the variant refers to a symbol which is
> not visible from current midrule. */
> -#define VARIANT_NOT_VISIBLE_FROM_MIDRULE (1 << 2)
> +#define VARIANT_NOT_VISIBLE_FROM_MIDRULE (1 << 1)
>
> static variant *variant_table = 0;
> static unsigned variant_table_size = 0;
> @@ -372,31 +361,25 @@ variant_table_free (void)
> variant_table_size = variant_count = 0;
> }
>
> -static char *
> -find_prefix_end (const char *prefix, char *begin, char *end)
> +/* Return TRUE if ID matches the string from CP up to CP_END.
> + The string does not contain null bytes. */
> +static bool
> +identifier_matches (char const *id, char const *cp, char const *cp_end)
> {
> - char *ptr = begin;
> -
> - for (; *prefix && ptr != end; ++prefix, ++ptr)
> - if (*prefix != *ptr)
> - return 0;
> -
> - if (*prefix)
> - return 0;
> -
> - return ptr;
> + while (cp != cp_end)
> + if (*id++ != *cp++)
> + return false;
> + return !*id;
> }
>
> +/* If scanning ID, return a new variant with that ID, at location
> + ID_LOC with index SYMBOL_INDEX. Otherwise, return NULL. The
> + currently scanned identifier starts at CP and ends at CP_END. */
> static variant *
> variant_add (uniqstr id, location id_loc, unsigned symbol_index,
> - char *cp, char *cp_end, bool explicit_bracketing)
> + char const *cp, char const *cp_end)
> {
> - char *prefix_end;
> -
> - prefix_end = find_prefix_end (id, cp, cp_end);
> - if (prefix_end &&
> - (prefix_end == cp_end ||
> - (!explicit_bracketing && is_dot_or_dash (*prefix_end))))
> + if (identifier_matches (id, cp, cp_end))
> {
> variant *r = variant_table_grow ();
> r->symbol_index = symbol_index;
> @@ -421,11 +404,15 @@ get_at_spec(unsigned symbol_index)
> return at_buf;
> }
>
> +/* Show a subsidiary message for a problem with a grammar rule. TEXT
> + points to the problematic reference. MIDRULE_RHS_INDEX is the rhs
> + index (1-origin) in the rule. If IS_WARNING, it is a warning,
> + otherwise a complaint. Indent the message INDENT spaces. */
> static void
> -show_sub_messages (const char* cp, bool explicit_bracketing,
> - int midrule_rhs_index, char dollar_or_at,
> +show_sub_messages (char const *text, int midrule_rhs_index,
> bool is_warning, unsigned indent)
> {
> + char dollar_or_at = *text;
> unsigned i;
>
> for (i = 0; i < variant_count; ++i)
> @@ -445,8 +432,6 @@ show_sub_messages (const char* cp, bool
> explicit_bracketing,
> else
> {
> static struct obstack msg_buf;
> - const char *tail = explicit_bracketing ? "" :
> - cp + strlen (var->id);
> const char *id = var->hidden_by ? var->hidden_by->id :
> var->id;
> location id_loc = var->hidden_by ? var->hidden_by->loc :
> @@ -460,7 +445,6 @@ show_sub_messages (const char* cp, bool
> explicit_bracketing,
> obstack_fgrow1 (&msg_buf, "[%s]", id);
> else
> obstack_sgrow (&msg_buf, id);
> - obstack_sgrow (&msg_buf, tail);
>
> if (var->err & VARIANT_HIDDEN)
> {
> @@ -469,7 +453,6 @@ show_sub_messages (const char* cp, bool
> explicit_bracketing,
> obstack_fgrow1 (&msg_buf, "[%s]", var->id);
> else
> obstack_sgrow (&msg_buf, var->id);
> - obstack_sgrow (&msg_buf, tail);
> }
>
> obstack_fgrow1 (&msg_buf, _(" at %s"), at_spec);
> @@ -504,16 +487,22 @@ show_sub_messages (const char* cp, bool
> explicit_bracketing,
> /* Sub-messages indent. */
> #define SUB_INDENT (4)
>
> -/* Parse named or positional reference. In case of positional
> - references, can return negative values for $-n "deep" stack
> - accesses. */
> +/* Return the index of a named or positional reference starting at CP
> + for a rule RULE of length RULE_LENGTH. If MIDRULE_RHS_INDEX is
> + nonzero, this is a generated midrule whose rhs index (1-origin) is
> + MIDRULE_RHS_INDEX in the parent rule. The entire semantic value
> + containing the reference is TEXT, of length TEXTLEN. Its location
> + is TEXT_LOC.
> +
> + In case of positional references, this can return negative values
> + for $-n "deep" stack accesses. */
> static long int
> -parse_ref (char *cp, symbol_list *rule, int rule_length,
> - int midrule_rhs_index, char *text, location text_loc,
> - char dollar_or_at)
> +parse_ref (char const *cp, symbol_list const *rule, int rule_length,
> + int midrule_rhs_index, char const *text, unsigned textlen,
> + location text_loc)
> {
> - symbol_list *l;
> - char *cp_end;
> + symbol_list const *l;
> + char const *cp_end;
> bool explicit_bracketing;
> unsigned i;
> unsigned valid_variants = 0;
> @@ -522,9 +511,9 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> if ('$' == *cp)
> return LHS_REF;
>
> - if (c_isdigit (*cp) || (*cp == '-' && c_isdigit (* (cp + 1))))
> + if (c_isdigit (*cp) || *cp == '-')
> {
> - long int num = strtol (cp, &cp, 10);
> + long int num = strtol (cp, NULL, 10);
> if (1 - INT_MAX + rule_length <= num && num <= rule_length)
> return num;
> else
> @@ -535,32 +524,9 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> }
> }
>
> - if ('[' == *cp)
> - {
> - /* Ignore the brackets. */
> - char *p;
> - for (p = ++cp; *p != ']'; ++p)
> - continue;
> - cp_end = p;
> -
> - explicit_bracketing = true;
> - }
> - else
> - {
> - /* Take all characters of the name. */
> - char* p;
> - for (p = cp; *p; ++p)
> - if (is_dot_or_dash (*p))
> - {
> - ref_tail_fields = p;
> - break;
> - }
> - for (p = cp; *p; ++p)
> - continue;
> - cp_end = p;
> -
> - explicit_bracketing = false;
> - }
> + explicit_bracketing = (*cp == '[');
> + cp += explicit_bracketing;
> + cp_end = text + textlen - explicit_bracketing;
>
> /* Add all relevant variants. */
> {
> @@ -574,13 +540,13 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> continue;
>
> var = variant_add (l->content.sym->tag, l->sym_loc,
> - symbol_index, cp, cp_end, explicit_bracketing);
> + symbol_index, cp, cp_end);
> if (var && l->named_ref)
> var->hidden_by = l->named_ref;
>
> if (l->named_ref)
> variant_add (l->named_ref->id, l->named_ref->loc,
> - symbol_index, cp, cp_end, explicit_bracketing);
> + symbol_index, cp, cp_end);
> }
> }
>
> @@ -595,10 +561,6 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> && (symbol_index == 0 || midrule_rhs_index < symbol_index))
> var->err |= VARIANT_NOT_VISIBLE_FROM_MIDRULE;
>
> - /* Check correct bracketing. */
> - if (!explicit_bracketing && contains_dot_or_dash (var->id))
> - var->err |= VARIANT_BAD_BRACKETING;
> -
> /* Check using of hidden symbols. */
> if (var->hidden_by)
> var->err |= VARIANT_HIDDEN;
> @@ -614,8 +576,7 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> {
> case 0:
> {
> - unsigned len = (explicit_bracketing || !ref_tail_fields) ?
> - cp_end - cp : ref_tail_fields - cp;
> + unsigned len = cp_end - cp;
> unsigned indent = 0;
>
> complain_at_indent (text_loc, &indent, _("invalid reference: %s"),
> @@ -629,7 +590,7 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> const char *format =
> _("syntax error after `%c', expecting integer, letter,"
> " `_', `[', or `$'");
> - complain_at_indent (sym_loc, &indent, format, dollar_or_at);
> + complain_at_indent (sym_loc, &indent, format, *text);
> }
> else if (midrule_rhs_index)
> {
> @@ -647,8 +608,7 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> }
>
> if (variant_count > 0)
> - show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
> - dollar_or_at, false, indent);
> + show_sub_messages (text, midrule_rhs_index, false, indent);
> return INVALID_REF;
> }
> case 1:
> @@ -658,8 +618,8 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> {
> warn_at_indent (text_loc, &indent, _("misleading reference:
> %s"),
> quote (text));
> - show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
> - dollar_or_at, true, indent + SUB_INDENT);
> + show_sub_messages (text, midrule_rhs_index, true,
> + indent + SUB_INDENT);
> }
> {
> unsigned symbol_index =
> @@ -673,8 +633,8 @@ parse_ref (char *cp, symbol_list *rule, int
> rule_length,
> unsigned indent = 0;
> complain_at_indent (text_loc, &indent, _("ambiguous reference:
> %s"),
> quote (text));
> - show_sub_messages (cp, explicit_bracketing, midrule_rhs_index,
> - dollar_or_at, false, indent + SUB_INDENT);
> + show_sub_messages (text, midrule_rhs_index, false,
> + indent + SUB_INDENT);
> return INVALID_REF;
> }
> }
> @@ -692,13 +652,14 @@ int max_left_semantic_context = 0;
> /*------------------------------------------------------------------.
> | TEXT is pointing to a wannabee semantic value (i.e., a `$'). |
> | |
> -| Possible inputs: $[<TYPENAME>]($|integer) |
> +| Possible inputs: $(<TYPENAME>|)($|integer|c_id|[id]) |
> | |
> | Output to OBSTACK_FOR_STRING a reference to this semantic value. |
> `------------------------------------------------------------------*/
>
> static void
> -handle_action_dollar (symbol_list *rule, char *text, location dollar_loc)
> +handle_action_dollar (symbol_list *rule, char *text, unsigned textlen,
> + location dollar_loc)
> {
> char const *type_name = NULL;
> char *cp = text + 1;
> @@ -735,7 +696,7 @@ handle_action_dollar (symbol_list *rule, char *text,
> location dollar_loc)
> }
>
> n = parse_ref (cp, effective_rule, effective_rule_length,
> - rule->midrule_parent_rhs_index, text, dollar_loc, '$');
> + rule->midrule_parent_rhs_index, text, textlen,
> dollar_loc);
>
> if (gt_ptr)
> *gt_ptr = '\0';
> @@ -806,10 +767,11 @@ handle_action_dollar (symbol_list *rule, char *text,
> location dollar_loc)
> `------------------------------------------------------*/
>
> static void
> -handle_action_at (symbol_list *rule, char *text, location at_loc)
> +handle_action_at (symbol_list const *rule, char const *text, unsigned
> textlen,
> + location at_loc)
> {
> - char *cp = text + 1;
> - symbol_list *effective_rule;
> + char const *cp = text + 1;
> + symbol_list const *effective_rule;
> int effective_rule_length;
> int n;
>
> @@ -827,7 +789,7 @@ handle_action_at (symbol_list *rule, char *text,
> location at_loc)
> muscle_percent_define_ensure("locations", at_loc, true);
>
> n = parse_ref (cp, effective_rule, effective_rule_length,
> - rule->midrule_parent_rhs_index, text, at_loc, '@');
> + rule->midrule_parent_rhs_index, text, textlen, at_loc);
> switch (n)
> {
> case INVALID_REF:
> diff --git a/src/scan-gram.l b/src/scan-gram.l
> index 4129181..a40aade 100644
> --- a/src/scan-gram.l
> +++ b/src/scan-gram.l
> @@ -118,8 +118,9 @@ static void unexpected_newline (boundary, char const
> *);
> /* Bracketed identifiers support. */
> %x SC_BRACKETED_ID SC_RETURN_BRACKETED_ID
>
> -letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
> -id -*(-|{letter}({letter}|[-0-9])*)
> +/* A Bison identifier. Keep this synchronized with scan-code.l "id". */
> +letter [-.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_]
> +id {letter}({letter}|[0-9])*
> directive %{id}
> int [0-9]+
>
> diff --git a/src/symlist.c b/src/symlist.c
> index e717c3e..5483f07 100644
> --- a/src/symlist.c
> +++ b/src/symlist.c
> @@ -217,7 +217,7 @@ symbol_list_n_type_name_get (symbol_list *l, location
> loc, int n)
> }
>
> bool
> -symbol_list_null (symbol_list *node)
> +symbol_list_null (symbol_list const *node)
> {
> return !node ||
> (node->content_type == SYMLIST_SYMBOL && !(node->content.sym));
> diff --git a/src/symlist.h b/src/symlist.h
> index 297d414..e5a0714 100644
> --- a/src/symlist.h
> +++ b/src/symlist.h
> @@ -116,7 +116,7 @@ symbol_list *symbol_list_n_get (symbol_list *l, int n);
> uniqstr symbol_list_n_type_name_get (symbol_list *l, location loc, int n);
>
> /* Check whether the node is a border element of a rule. */
> -bool symbol_list_null (symbol_list *node);
> +bool symbol_list_null (symbol_list const *node);
>
> /** Set the \c \%destructor for \c node as \c code at \c loc. */
> void symbol_list_destructor_set (symbol_list *node, char const *code,
> diff --git a/tests/named-refs.at b/tests/named-refs.at
> index 74549c6..17cbef3 100644
> --- a/tests/named-refs.at
> +++ b/tests/named-refs.at
> @@ -279,11 +279,9 @@ start: foo foo.bar { $foo.bar; }
> foo: '1'
> foo.bar: '2'
> ]])
> -AT_BISON_CHECK([-o test.c test.y], 0, [],
> -[[test.y:11.22-29: warning: misleading reference: `$foo.bar'
> -test.y:11.8-10: warning: refers to: $foo at $1
> -test.y:11.12-18: warning: possibly meant: $[foo.bar] at $2
> -]])
> +AT_BISON_CHECK([-o test.c test.y])
> +AT_CHECK([if $CC $CFLAGS $CPPFLAGS test.c; then false; else true; fi],
> + [0], [ignore], [ignore])
> AT_CLEANUP
>
> #######################################################################
> @@ -358,43 +356,33 @@ factor: '(' expr ')' { $$ = $2; }
> AT_BISON_CHECK([-o test.c test.y], 1, [],
> [[test.y:24.36-41: invalid reference: `$cond1'
> test.y:23.11-24.62: symbol not found in production: cond1
> -test.y:26.43-53: invalid reference: `$stmt.field'
> +test.y:26.43-47: invalid reference: `$stmt'
> test.y:25.11-26.60: symbol not found in production: stmt
> -test.y:25.35-38: possibly meant: $then.field, hiding $stmt.field at $4
> -test.y:28.43-52: invalid reference: `$stmt.list'
> +test.y:25.35-38: possibly meant: $then, hiding $stmt at $4
> +test.y:28.43-47: invalid reference: `$stmt'
> test.y:27.11-28.59: symbol not found in production: stmt
> -test.y:27.30-38: possibly meant: $[stmt.list] at $4
> test.y:30.43-46: ambiguous reference: `$xyz'
> test.y:29.35-37: refers to: $xyz at $4
> test.y:29.50-52: refers to: $xyz at $6
> -test.y:32.43-52: invalid reference: `$stmt.list'
> +test.y:32.43-47: invalid reference: `$stmt'
> test.y:31.11-32.63: symbol not found in production: stmt
> -test.y:31.40-43: possibly meant: $then, hiding $[stmt.list] at $4
> -test.y:31.61-64: possibly meant: $else, hiding $[stmt.list] at $6
> -test.y:34.43-58: invalid reference: `$stmt.list.field'
> +test.y:34.43-47: invalid reference: `$stmt'
> test.y:33.11-34.69: symbol not found in production: stmt
> -test.y:33.40-43: possibly meant: $then.field, hiding
> $[stmt.list].field at $4
> -test.y:33.61-64: possibly meant: $else.field, hiding
> $[stmt.list].field at $6
> test.y:36.43-54: invalid reference: `$[stmt.list]'
> test.y:35.11-36.71: symbol not found in production: stmt.list
> test.y:35.40-43: possibly meant: $then, hiding $[stmt.list] at $4
> test.y:35.61-64: possibly meant: $else, hiding $[stmt.list] at $6
> -test.y:38.43-49: invalid reference: `$then.1'
> +test.y:38.43-47: invalid reference: `$then'
> test.y:37.11-38.60: symbol not found in production: then
> -test.y:37.40-45: possibly meant: $[then.1] at $4
> -test.y:40.43-55: invalid reference: `$then.1.field'
> +test.y:40.43-47: invalid reference: `$then'
> test.y:39.11-40.66: symbol not found in production: then
> -test.y:39.40-45: possibly meant: $[then.1].field at $4
> -test.y:42.44-50: invalid reference: `$stmt.x'
> +test.y:42.44-48: invalid reference: `$stmt'
> test.y:41.12-42.57: symbol not found in production: stmt
> -test.y:41.36-41: possibly meant: $[stmt.x].x, hiding $stmt.x at $4
> -test.y:41.36-41: possibly meant: $[stmt.x] at $4
> -test.y:44.13-22: invalid reference: `$if-stmt-a'
> +test.y:41.36-41: possibly meant: $[stmt.x], hiding $stmt at $4
> +test.y:44.13-15: invalid reference: `$if'
> test.y:43.12-44.59: symbol not found in production: if
> -test.y:43.1-9: possibly meant: $[if-stmt-a] at $$
> -test.y:46.46-54: invalid reference: `$then-a.f'
> +test.y:46.46-50: invalid reference: `$then'
> test.y:45.12-46.65: symbol not found in production: then
> -test.y:45.41-46: possibly meant: $[then-a].f at $4
> ]])
> AT_CLEANUP
>
> @@ -527,9 +515,9 @@ sym_a : 'a';
> sym_b : 'b';
> ]])
> AT_BISON_CHECK([-o test.c test.y], 1, [],
> -[[test.y:13.8-17: invalid reference: `$sym.field'
> +[[test.y:13.8-11: invalid reference: `$sym'
> test.y:12.1-13.21: symbol not found in production: sym
> -test.y:16.8-21: invalid reference: `$<aa>sym.field'
> +test.y:16.8-15: invalid reference: `$<aa>sym'
> test.y:15.1-16.25: symbol not found in production: sym
> test.y:19.8-19: invalid reference: `$[sym.field]'
> test.y:18.1-19.23: symbol not found in production: sym.field
> @@ -543,9 +531,9 @@ test.y:31.8-13: invalid reference: `$[sym]'
> test.y:30.1-33.21: symbol not found in production before $3: sym
> test.y:33.8-17: invalid reference: `$<aa>[sym]'
> test.y:30.1-33.21: symbol not found in production: sym
> -test.y:37.8-17: invalid reference: `$sym-field'
> +test.y:37.8-11: invalid reference: `$sym'
> test.y:36.1-37.21: symbol not found in production: sym
> -test.y:40.8-21: invalid reference: `$<aa>sym-field'
> +test.y:40.8-15: invalid reference: `$<aa>sym'
> test.y:39.1-40.25: symbol not found in production: sym
> test.y:43.8-19: invalid reference: `$[sym-field]'
> test.y:42.1-43.23: symbol not found in production: sym-field
> @@ -572,21 +560,15 @@ start:
> .field { $.field; }
> | -field { @-field; }
> | 'a' { @.field; }
> -| 'a' { $-field; }
> +| 'b' { $-field; }
> ;
> -.field: ;
> --field: ;
> +.field: 'c' ;
> +-field: 'd' ;
> ]])
> -AT_BISON_CHECK([[test.y]], [[1]], [],
> -[[test.y:4.12-18: invalid reference: `$.field'
> -test.y:4.13: syntax error after `$', expecting integer, letter,
> `_', `@<:@', or `$'
> -test.y:4.3-8: possibly meant: $[.field] at $1
> -test.y:5.12-18: invalid reference: address@hidden'
> -test.y:5.13: syntax error after `@', expecting integer, letter,
> `_', `@<:@', or `$'
> -test.y:5.3-8: possibly meant: @[-field] at $1
> -test.y:6.12-18: invalid reference: address@hidden'
> -test.y:6.13: syntax error after `@', expecting integer, letter,
> `_', `@<:@', or `$'
> -test.y:7.12-18: invalid reference: `$-field'
> -test.y:7.13: syntax error after `$', expecting integer, letter,
> `_', `@<:@', or `$'
> +AT_BISON_CHECK([[test.y]], [], [],
> +[[test.y:4.12: warning: stray `$'
> +test.y:5.12: warning: stray `@'
> +test.y:6.12: warning: stray `@'
> +test.y:7.12: warning: stray `$'
> ]])
> AT_CLEANUP
> --
> 1.7.3
>
>
>
--
Best regards,
Alex Rozenman (address@hidden).
- Re: [PATCH] Do not allow identifiers that start with a negative number., (continued)
- Re: [PATCH] Do not allow identifiers that start with a negative number., Akim Demaille, 2011/01/26
- Re: [PATCH] Do not allow identifiers that start with a negative number., Akim Demaille, 2011/01/26
- Re: [PATCH] Do not allow identifiers that start with a negative number., Akim Demaille, 2011/01/26
- Re: [PATCH] Do not allow identifiers that start with a negative number., Akim Demaille, 2011/01/26
- Re: [PATCH] Do not allow identifiers that start with a negative number., Joel E. Denny, 2011/01/27
- Re: [PATCH] Do not allow identifiers that start with a negative number., Joel E. Denny, 2011/01/29
- Re: [PATCH] Do not allow identifiers that start with a negative number.,
Alex Rozenman <=