%symbol-default instead of an empty symbol list, and start working on
new per-type %destructor/%printer. Discussed at
<http://lists.gnu.org/archive/html/bison-patches/2006-09/msg00007.html>.
* NEWS (2.3+): Add %symbol-default to example.
* bison.texinfo (Freeing Discarded Symbols): Likewise.
(Bison Symbols): Add entry for %symbol-default.
* src/parse-gram.y (PERCENT_SYMBOL_DEFAULT): New token.
(generic_symlist, generic_symlist_item): New nonterminals for creating
a list in which each item is a symbol, semantic type, or
%symbol-default.
(grammar_declaration): Use generic_symlist in %destructor and %printer
declarations instead of symbols.1 or an empty list.
(symbol_declaration, precedence_declaration, symbols.1): Update actions
for changes to symbol_list.
* src/reader.c: Update for changes to symbol_list.
* src/scan-code.l: Likewise.
* src/scan-gram.l: Scan new PERCENT_SYMBOL_DEFAULT token.
* src/symlist.c, src/symlist.h: Extend such that a list node may
represent a semantic type or a %symbol-default in addition to just an
ordinary symbol. Add switched functions for setting %destructor's and
%printer's.
* tests/actions.at, tests/input.at: Add %symbol-default to all default
%destructor/%printer declarations.
isn't a question of whether the user *declares* that symbol (in %token,
for example). It's a question of whether the user by any means
*defines* the symbol at all (by simply using a char token, for
example). $end is defined by Bison whereas any other token with token
number 0 is defined by the user. The error token is always defined by
Bison regardless of whether the user declares it with %token, but we
may one day let the user define error as a nonterminal instead.
* NEWS (2.3+): Say "user-defined" instead of "user-declared".
* doc/bison.texinfo (Freeing Discarded Symbols): Likewise, and document
the meaning of "user-defined".
* tests/actions.at (Default %printer and %destructor for user-declared
end token): Rename to...
(Default %printer and %destructor for user-defined end token): ...
this.
* src/symtab.c (symbol_destructor_get, symbol_printer_get): In the
computation of whether to apply the default, don't maintain a list of
every Bison-defined symbol. Instead, just check for a first character
of '$', which a user symbol cannot have, and check for the error token.
$undefined, or $accept. This change fits the general rule that the
default %destructor and %printer are only for user-declared symbols,
and it solves several difficulties that are described in the new test
cases listed below.
* src/symtab.c (symbol_destructor_get, symbol_printer_get): Implement.
* tests/actions.at (Default %printer and %destructor are not for error
or $undefined, Default %printer and %destructor are not for $accept):
New test cases.
* src/reader.c (grammar_current_rule_begin): Don't set the start symbol
when parsing the first rule.
(check_and_convert_grammar): Search for it here after all grammar
declarations have been parsed. Skip midrules, which have dummy LHS
nonterminals.
* src/symtab.c (symbol_is_dummy): New function.
* src/symtab.h (symbol_is_dummy): Declare it.
* tests/input.at (%start after first rule): New test.
non-aliased/undeclared string literals since it might be useful to
those declaring %token-table.
* src/reader.c (check_and_convert_grammar): Undo changes in previous
commit: don't worry about complaints from symbols_pack.
* src/symtab.c (symbol_new, symbol_class_set,
symbol_check_alias_consistency): Undo changes in previous commit: count each string literal as a new symbol and token, assign it a symbol
number, and don't complain about non-aliased string literals.
(symbols_pack): Since symbol_make_alias still does not decrement symbol
and token counts but does still set aliased tokens to the same number,
symbol_pack_processor now leaves empty slots in the symbols array.
Remove those slots.
* tests/regression.at (Undeclared string literal): Remove test case
added in previous commit since non-aliased string literals are allowed
again.
(Characters Escapes, Web2c Actions): Undo changes in previous commit:
remove unnecessary string literal declarations.
* tests/sets.at (Firsts): Likewise.
be used before its declaration.
* src/reader.c (check_and_convert_grammar): Don't invoke packgram if
symbols_pack complained.
* src/symtab.c (symbol_new): Don't count a string literal as a new
symbol.
(symbol_class_set): Don't count a string literal as a new token, and
don't assign it a symbol number since symbol_make_alias does that.
(symbol_make_alias): It's not necessary to decrement the symbol and
token counts anymore. Don't assume that an alias declaration occurs
before any uses of the identifier or string, and thus don't assert that
one of them has the highest symbol number so far.
(symbol_check_alias_consistency): Complain if there's a string literal
that wasn't declared as an alias.
(symbols_pack): Bail if symbol_check_alias_consistency failed since
symbol_pack asserts that every token has been assigned a symbol number
although undeclared string literals have not.
* tests/regression.at (String alias declared after use, Undeclared
string literal): New test case.
(Characters Escapes, Web2c Actions): Declare string literals as
aliases.
* tests/sets.at (Firsts): Likewise.
them to the parser in order to improve error messages.
* src/scan-gram.l (SC_ESCAPED_STRING, SC_ESCAPED_CHARACTER,
SC_BRACED_CODE, SC_PROLOGUE): Implement.
* tests/input.at (Unclosed constructs): New test case.
* tests/regression.at (Invalid inputs): Update now that unclosed %{ is
seen.
* src/scan-gram.h, src/scan-gram.l (gram_last_braced_code_loc): Remove
unused global.
* src/symtab.c (symbol_user_token_number_set): If the token has an
alias, check and set its alias's user token number instead of its own,
which is set to indicate the alias. Previously, every occurrence of
the character token in the grammar overwrote that alias indicator with
the character code.
* tests/input.at (String aliases for character tokens): New test.
to prevent failures when building on older platforms.
Check for autopoint failure.
Set XGETTEXT_OPTIONS to values that check for C format strings,
so that translators are warned about them (this also helps
prevent core dumps).
rather than -W, so we don't get bogus warnings about sign comparisons.
Add -Wpointer-arith, since that warning is useful (it reports code
that does not conform to C89 and that some compilers reject).
* data/c.m4, data/glr.c, data/lalr1.cc, data/yacc.c: Undo latest change,
since it's no longer needed.
* src/flex-scanner.h (FLEX_NO_OBSTACK): New macro that blocks obstack
definitions so gcc won't warn when obstack_for_string is unused.
* src/scan-code.l: config.h and system.h are already #include'd by
scan-code-c.c, so get rid of them here.
* src/scan-gram.l: Likewise.
* src/scan-skel.l: Likewise, and use flex-scanner.h without obstack
definitions rather than duplicating the rest of it.
* src/scan-gram-c.c, scan-skel-c.c: #include "system.h".
* data/c.m4 (b4_safest_int_type): New macro.
* data/glr.c, data/lalr1.cc: Wherever you compare yycheck[i] against
a signed int type, cast it to b4_safest_int_type first.
* data/yacc.c: Likewise.
(b4_safest_int_type): Overwrite the one from c.m4 since b4_int_type is
also overwritten.
po.m4, stdint_h.m4, uintmax_t.m4, ulonglong.m4, warning.m4.
* bootstrap (gnulib_tool): Stop using --assume-autoconf;
the latest gnulib does this a different way.
(get_translations): Sharuzzaman Ahmat Raslan reported that the ms
translation was patched, so stop omitting it.
use these for all user-declared grammar symbols for which the user does
not declare a specific %printer/%destructor. Thus, the parser uses it
for token 0 if the user declares it but not if Bison generates it as
$end. Discussed starting at
<http://lists.gnu.org/archive/html/bison-patches/2006-02/msg00064.html>,
<http://lists.gnu.org/archive/html/bison-patches/2006-06/msg00091.html>,
and
<http://lists.gnu.org/archive/html/bison-patches/2006-07/msg00019.html>.
* NEWS (2.3+): Mention.
* doc/bison.texinfo (Actions in Mid-Rule): It's no longer impossible to
declare a %destructor for a mid-rule's semantic value. It's just
impossible to declare one specific to it.
(Freeing Discarded Symbols): Mention that @$ can be used in %destructor
code. Describe default %destructor form.
* src/parse-gram.y (grammar_declaration): Parse default
%printer/%destructor declarations.
* src/output.c (symbol_destructors_output): Use symbol_destructor_get
and symbol_destructor_location_get rather than accessing the destructor
and destructor_location members of struct symbol.
(symbol_printers_output): Likewise but for %printer's.
* src/reader.c (symbol_should_be_used): Likewise but for %destructor's
again.
* src/symtab.c (default_destructor, default_destructor_location,
default_printer, default_printer_location): New static global
variables to record the default %destructor and %printer.
(symbol_destructor_get, symbol_destructor_location_get,
symbol_printer_get, symbol_printer_location_get): New functions to
compute the appropriate %destructor and %printer for a symbol.
(default_destructor_set, default_printer_set): New functions to set the
default %destructor and %printer.
* src/symtab.h: Prototype all those new functions.
* tests/actions.at (Default %printer and %destructor): New test to
check that the right %printer and %destructor are called, that they're
not called for $end, and that $$ and @$ work correctly.
(Default %printer and %destructor for user-declared end token): New
test to check that the default %printer and %destructor are called for
a user-declared end token.
* tests/input.at (Default %printer and %destructor redeclared, Unused
values with default %destructor): New tests to check related grammar
warnings and errors.
Discussed starting at
<http://lists.gnu.org/archive/html/bison-patches/2006-07/msg00019.html>
and
<http://lists.gnu.org/archive/html/help-bison/2006-07/msg00013.html>.
Make the skeletons consistent in how they pop the end token and invoke
its %destructor.
* data/glr.c (yyrecoverSyntaxError, yyparse): Don't pop the start
state, which has token number 0, since this would invoke the
%destructor for the end token.
* data/lalr1.cc (yy::parser::parse): Don't check for the final state
until after shifting the end token, or else it won't be popped.
* data/yacc.c (yyparse): Likewise.
* data/glr.c (yyparse): Clear the lookahead after shifting it even when
it's the end token. Upon termination, destroy an unshifted lookahead
even when it's the end token.
* data/lalr1.cc (yy::parser::parse): Likewise.
* data/yacc.c (yyparse): Likewise.
* src/reader.c (packgram): Don't check rule 0. This suppresses unused
value warnings for the end token when the user gives the end token a
%destructor.
* tests/actions.at (Printers and Destructors): Test all the above.
* bootstrap (gnulib-modules): Remove hard-locale, stdio-safer.
Add fopen-safer.
(gnulib_files): Add m4/warning.m4. Don't worry about files
overwritten by autopoint.
Replace gt_INTL_SUBDIR_CORE with an empty body in m4/gettext_gl.m4.
Suppress "id", "ms", "tr" translations for now, since gettext 0.15
rejects them.
Don't use autoreconf; instead, invoke autopoint etc. by hand,
so that we can remove the intl files at a better time.
(intl_files_to_remove): Remove aclocal.m4, since it gets
rebuilt anyway. Remove m4/inttypes_h.m4, m4/inttypes.m4,
m4/isc-posix.m4, m4/lib-ld.m4, m4/lib-prefix.m4, m4/po.m4,
m4/stdint_h.m4, m4/uintmax_t.m4, m4/ulonglong.m4.
Add m4/inttypes-h.m4, m4/lock.m4, m4/visibility.m4.
Remove datarootdir hack; no longer needed.
* configure.ac: Use gl_WARNING_CFLAGS rather than BISON_WARNING.
(AM_GNU_GETTEXT_VERSION): Bump from 0.12 to 0.15.
* lib/.cvsignore: Remove hard-locale.c, hard-locale.h, strdup.c,
strdup.h.
* m4/.cvsignore: Remove hard-locale.m4, strdup.m4.
* m4/warning.m4: Remove from CVS, since we now use gnulib's version.
least 1.4.7 and 1.4.4) by putting a #line between `typedef union
YYSTYPE' and `{'.
* src/muscle_tab.h (muscle_grow): Replace the header comments with
those from muscle_tab.c since the old ones are misleading.
* src/scan-code.h, src/scan-code.l (translate_action)
(translate_rule_action, translate_symbol_action, translate_code):
Return char *, not const char *.
* src/parse-gram.y (declaration): Rename as...
(prologue_declaration): this.
(string_content): Remove this nonterminal, use STRING.
(braceless, content, content.opt): New nonterminal.
Use them.
(%define): Now accept content.opt, i.e., accept also BRACED_CODE
as value.
* src/scan-gram.l (getargs.h): Don't include it.
YYSTYPE if your grammar contains at least one <type> tag.
Your YYSTYPE need not be a macro; it can be a typedef.
* doc/bison.texinfo (Value Type, Multiple Types, Location Type):
(Union Decl, Decl Summary): Document this.
* data/glr.c (YYSTYPE): Implement this.
* data/glr.cc (YYSTYPE): Likewise.
* data/lalr1.cc (YYSTYPE): Likewise.
* data/yacc.c (YYSTYPE): Likewise.
* src/output.c (prepare): Output tag_seen_flag.
* src/parse-gram.y (declaration, grammar_declaration):
Use 'union_seen' rather than 'typed' to determine whether
%union has been seen, since grammars can now be typed without
%union.
(symbol_declaration, type.opt, symbol_def):
Keep track of whether a tag has been seen.
* src/reader.c (union_seen, tag_seen): New vars.
(typed): remove.
* src/reader.h (union_seen, tag_seen, typed): Likewise.
* src/scan-code.l (untyped_var_seen): New variable.
(handle_action_dollar): Adjust to above changes.
(handle_action_dollar, handle_action_at):
Improve overflow checking for outlandish numbers.
* tests/input.at (AT_CHECK_UNUSED_VALUES): Redo test to
avoid new diagnostics generated by above changes.
* tests/regression.at (YYSTYPE typedef): Add test to check
for type tags without %union.
* src/getargs.c (report_argmatch, trace_argmatch): Remove,
replaced by...
(flags_argmatch, FLAGS_ARGMATCH): this new function and macro.
Adjust callers.
* src/getargs.h, src/getargs.c (warnings, warnings_flags)
(warnings_args, warnings_types): New.
(getargs, short_options, long_options): Accept -W/--warnings.
Sort the options by alphabetical order, upper case letter right
before its lower case.
so they don't collide with user-defined macros.
(yy_stack_print): Don't assume that yytype_int16 promotes to int;
this was never guaranteed, and now that we're using gnulib stdint,
which defines int_fast16_t to long int, the problem is exposed.
* bootstrap (gnulib_modules): Add stdint.
Remove special case for m4/onceonly_2_57.m4, since gnulib-tool
no longer copies it.
(intl_files_to_remove): Remove m4/longlong.m4 and m4/wchar_t.m4,
since stdint needs the former and wcwidth (which is now required
by mbswidth) needs the latter.
Append 'datarootdir = @datarootdir@' to po/Makefile.in.in, to
work around a compatibility glitch between gettext 0.14.6 and
Autoconf 2.60.
* configure.ac (AC_PREREQ): Require Autoconf 2.60.
Do not check for uintptr_t, since new stdint module does the right
thing.
* lib/.cvsignore: Remove alloca.c, alloca.h, alloca_.h.
Add stdint.h, stdint_.h, wcwidth.h.
* m4/.cvsignore: Remove alloca.m4, onceonly.m4.
Add absolute-header.m4, double-slash-root.m4, longlong.m4,
stdint.m4, wchar_t.m4, wcwidth.m4.
* src/files.c: Include <dirname.h> and <stdio-safer.h> in the
usual order for ../lib/*.h files.
(file_name_split): Use last_component, not base_name, to adjust
to gnulib changes.
* src/parse-gram.h: Include <strverscmp.h> in the usual order
for ../lib/*.h files.
(YYTYPE_INT16, YYTYPE_INT8, YYTYPE_UINT16, YYTYPE_UINT8):
Define unconditionally, since we now assume the stdint module.
* src/scan-skel.l: Include <dirname.h>.
(BASE_QPUTS): Use last_component, not base_name.
* src/system.h: Include <unlocked-io.h> in the usual order
for ../lib/*.h files. Include <stdint.h> unconditionally,
since we now use the stdint module.
(uintptr_t): Declare if UINTPTR_MAX is not defined, not
HAVE_UINTPTR_T, since we now use the stdint module.
(base_name): Remove decl, since files now include <dirname.h>
to get the decl.
New, default to 1.
* data/yacc.c, data/glr.c, data/location.cc: Use them.
* NEWS, doc/bison.texinfo: The initial column and line are 1 by
default.
* tests/calc.at: Adjust.
(b4_syncline): Also output the location of its invocation (from
the skeleton).
(b4_user_action, b4_define_user_action, b4_user_actions)
(b4_user_initial_action, b4_user_post_prologue, b4_user_start_header)
(b4_user_stype): New.
* data/yacc.c, data/glr.c, data/lalr1.cc, data/glr.cc: Use them.
on every other line.
* src/parse-gram.y (%initial-action): Initialize @$ correctly.
* tests/input.at (Torturing the Scanner): Update output.
* src/scan-gram.l (scanner_cursor): Declare it static.
declaration".
* src/symtab.c (redeclaration): Do that here.
* src/reader.c (record_merge_function_type): In the case of a result
type clash, report the previous declaration rather than the very first
one in the grammar file.
* tests/glr-regression.at (Missed %merge type warnings when LHS type is
declared later): Add a third declaration to check this behavior.
* tests/input.at (Incompatible Aliases): Update output.
declared after the %merge.
* src/reader.c (get_merge_function): Don't set the merger type yet.
(record_merge_function_type): New function for setting the merger type
and checking for clashes.
(grammar_current_rule_merge_set): Set the location of the %merge for
the current rule.
(packgram): Invoke record_merge_function_type for each rule now that
all symbol type declarations have been parsed.
* src/reader.h (merger_list.type_declaration_location): New member
storing the location of the first %merge from which the type for this
merging function was derived.
* src/symlist.h (symbol_list.merger_declaration_location): New member
storing the location of a rule's %merge, if any.
* tests/glr-regression.at (Missed %merge type warnings when LHS type is
declared later): New test to catch the error fixed by the above patch.
declarations appear after the rules. Don't mistake the type of $$ in
a midrule to be that of its parent rule's $$.
* src/reader.c (grammar_current_rule_end): Don't invoke
grammar_rule_check yet since not all symbol declarations may have been
parsed yet.
(grammar_midrule_action): Likewise.
Don't record whether the midrule's $$ has been used yet since actions
haven't been translated yet.
Record the midrule's parent rule and its RHS index within the parent
rule.
(grammar_current_rule_action_append): Don't translate the action yet
since not all symbol declarations may have been parsed yet and, thus,
warnings about types for $$, $n, @$, and @n can't be reported yet.
(packgram): Translate the action and invoke grammar_rule_check now that
all symbol declarations have been parsed.
* src/scan-code.l (handle_action_dollar): Now that this is invoked
after parsing the entire grammar file, the symbol list here in the case
of a midrule is actually the midrule's empty RHS, so reference its
parent rule's RHS where necessary.
On the other hand, now that you can already know it's a midrule, you
aren't forced to think $$ has the same type as its parent rule's $$.
(handle_action_at): In the case of a midrule, reference the parent rule
where necessary.
* src/symlist.c (symbol_list_new): Initialize new midrule-related
members.
(symbol_list_length): Now that this is invoked after all rules have
been parsed, a NULL symbol (rather than a NULL symbol list node)
terminates a rule. symbol_list_print already does this correctly.
* src/symlist.h (symbol_list.midrule_parent_rule,
symbol_list.midrule_parent_rhs_index): New members so that midrules can
remember their relationships with their parents.
* tests/input.at (Type Clashes): Extend to catch the midrule $$ error
fixed by the above patch.
(_AT_UNUSED_VALUES_DECLARATIONS, AT_CHECK_UNUSED_VALUES): New m4 macros
implementing...
(Unused values): ... this old test case and...
(Unused values before symbol declarations): ... this new test case.
This one is the same as `Unused values' except that all symbol
declarations appear after the rules in order to catch the rest of the
errors fixed by the above patch.
