Joel E. Denny [Sun, 21 May 2006 06:22:36 +0000 (06:22 +0000)]
* data/glr.c (yyresolveLocations): Remove bogus YYASSERT that the
state before an empty RHS is always resolved here. Only the location
of that state is guaranteed to be resolved, and that's enough. This
fixes the remaining bug reported by Derek M. Jones in
<http://lists.gnu.org/archive/html/bug-bison/2006-05/msg00027.html>.
* tests/glr-regression.at (Uninitialized location when reporting
ambiguity): Test the above case.
Also, the embedded comments in this test case claim it checks the case
of an empty RHS that has inherited the initial location. However, the
corresponding LHS was already resolved, so yyresolveLocations didn't
actually have reason to modify it. Fix this by forcing
nondeterministic operation at the beginning of the parse.
Paul Eggert [Sun, 21 May 2006 04:48:47 +0000 (04:48 +0000)]
* data/c.m4 (b4_yy_symbol_print_generate):
(b4_yy_symbol_print_generate): Use 'YYSTYPE const' rather than
'const YYSTYPE', and similarly for YYLTYPE. This fixes one
of the bugs reported today by Derek M Jones in
<http://lists.gnu.org/archive/html/bug-bison/2006-05/msg00027.html>.
* doc/bison.texinfo (Value Type): Document that YYSTYPE must be
defined to be a type name without parens or brackets.
(Location Type): Similarly for YYLTYPE.
* tests/regression.at (Trivial grammars): Put in a test for this
bug that will be caught by 'make maintainer-check' (though not,
alas, by 'make check' unless your compiler is picky).
Joel E. Denny [Wed, 17 May 2006 16:39:40 +0000 (16:39 +0000)]
* data/glr.c (yyreportTree): Make room in yystates for the state
preceding the RHS. This fixes the segmentation fault reported by Derek
M. Jones in
<http://lists.gnu.org/archive/html/help-bison/2006-05/msg00035.html>.
(yyreportTree, yypdumpstack): Subtract 1 from yyrule before printing
to the user. Reported for yyreportTree by Derek M. Jones later in the
same thread.
* THANKS: Add Derek M. Jones.
Update my email address.
Fix typo in Steve Murphy's name.
Paul Eggert [Mon, 15 May 2006 06:13:53 +0000 (06:13 +0000)]
* data/glr.c (yyreportSyntaxError): Fix off-by-one error in
checking against YYLAST that caused the parser to miss a potential
alternative in its diagnostic.
Problem reported by Maria Jose Moron Fernandez in
<http://lists.gnu.org/archive/html/bug-bison/2006-05/msg00024.html>.
* data/lalr1.cc (yysyntax_error_): Likewise.
* data/yacc.c (yysyntax_error): Likewise.
* tests/regression.at (_AT_DATA_DANCER_Y): Use static array for
tokens, in case we run into an older C compiler.
(_AT_DATA_EXPECT2_Y, AT_CHECK_EXPECT2): New macros.
Use them to check for the off-by-one error fixed above.
Paul Eggert [Mon, 15 May 2006 05:10:06 +0000 (05:10 +0000)]
* data/yacc.c (yytnamerr): Fix typo: local var should be of type
YYSIZE_T, not size_t.
* tests/regression.at (Trivial grammars): New test, to catch
the error fixed by the above patch.
Akim Demaille [Sun, 14 May 2006 20:40:35 +0000 (20:40 +0000)]
Implement --trace=m4.
* src/getargs.c (trace_types, trace_args): Accept trace_m4.
* src/output.c (output_skeleton): When set, pass -dV to m4.
Factor the handling of flags in m4.
* src/output.c (prepare): Rename the muscle names debug, defines,
error_verbose to debug_flag, defines_flag, error_verbose_flag.
* data/c.m4: Adjust.
(_b4_define_flag_if, b4_define_flag_if, b4_defines_if): New.
Use b4_define_flag_if to define other b4_FLAG_if macros.
(b4_location_if): As a consequence, rename as...
(b4_locations_if): this, for consistency.
Adjust all the skeletons.
Paul Eggert [Sun, 14 May 2006 07:00:18 +0000 (07:00 +0000)]
* data/c.m4 (b4_copyright): Put the special exception for Bison
skeletons here, so we don't have to put it in each skeleton. All
uses changed. Wrap the copyright notice, in case it is longer
than 80 columns. Replace comma by newline after title.
Paul Eggert [Tue, 9 May 2006 06:40:13 +0000 (06:40 +0000)]
* data/glr.c: Switch to Bison 2.2 special-exception language in
the copyright notice. Use more-regular format for titles and
copyright notices.
* data/glr.cc: Likewise.
* data/location.cc: Likewise.
* data/yacc.cc: Likewise.
* doc/bison.texinfo (Conditions): Document this.
* NEWS: likewise. Upgrade version to 2.2.
Paul Eggert [Mon, 13 Mar 2006 23:21:29 +0000 (23:21 +0000)]
Remove @shorttitlepage stuff; it wasn't used
and didn't really work.
For the index, use @ifnotinfo, not @iftex.
Minor cleanups of spacing and terminology.
Akim Demaille [Sun, 12 Mar 2006 14:45:27 +0000 (14:45 +0000)]
Apply --prefix to C++ skeletons too: they change the namespace.
The test suite already exercize these cases.
* data/c++.m4 (b4_namespace): New.
* data/lalr1.cc, data/glr.cc, data/location.cc: Use it instead of `yy'.
* data/lalr1.cc (yytnameerr_): Move its definition into the namespace.
* data/yacc.c, data/glr.c: Remove a useless `[]'.
* doc/bison.texinfo: Document it.
(Option Cross Key): Use @multitable in all formats. It looks
nicer, even in TeX outputs.
(Rules): Use the same code whatever the output type is.
* tests/local.at (_AT_BISON_OPTION_PUSHDEFS)
(_AT_BISON_OPTION_POPDEFS): Support AT_NAME_PREFIX.
* tests/calc.at: Use it, instead of hard coding `yy'.
Akim Demaille [Fri, 10 Mar 2006 13:30:19 +0000 (13:30 +0000)]
* data/c.m4 (b4_token_enum): Always define the enum of tokens,
even if empty.
* data/lalrl1.cc, data/glr.cc (parser::token_type): New.
* doc/bison.texinfo (Calc++ Scanner): Use it.
Paul Eggert [Wed, 8 Mar 2006 19:45:10 +0000 (19:45 +0000)]
Fix some nits reported by twlevo.
* doc/FAQ: Remove ancient Y2K FAQ, replacing it with "secure"
and "POSIX". Use more-modern syntax for URLs. Mention C++
and ask for Java. Don't hardwire OS version numbers. Add
copyright notice.
* m4/.cvsignore: Add unistd_h.m4, for latest gnulib.
* src/conflicts.c (solved_conflicts_obstack): Now static.
Joel E. Denny [Wed, 8 Mar 2006 19:34:56 +0000 (19:34 +0000)]
* doc/bison.texinfo (Introduction): Mention GLR and C++ as on the web
page. Say "you can use it" not "you may use it" as on the web page;
we're describing capabilities not granting permission.
Paul Eggert [Tue, 7 Mar 2006 06:42:23 +0000 (06:42 +0000)]
* data/glr.c (yyresolveLocations): Rename local variables to avoid
shadowing warnings. Use usual patter for iterating through RHS.
* tests/glr-regression.at
(Uninitialized location when reporting ambiguity):
Modify yylex so that it uses its argument, rather than trying
to rely on ARGSUSED (which doesn't work for gcc with warnings).
const char -> char const.
Joel E. Denny [Mon, 6 Mar 2006 07:39:11 +0000 (07:39 +0000)]
* tests/glr-regression.at (Uninitialized location when reporting
ambiguity): New test case.
* data/glr.c (yyresolveLocations): New function, which uses
YYLLOC_DEFAULT.
(yyresolveValue): Invoke yyresolveLocations before reporting an
ambiguity.
* doc/bison.texinfo (Default Action for Locations): Note
YYLLOC_DEFAULT's usage for ambiguity locations.
(GLR Semantic Actions): Cross-reference those notes.
Joel E. Denny [Sat, 4 Mar 2006 03:29:03 +0000 (03:29 +0000)]
* tests/glr-regression.at (Leaked semantic values when reporting
ambiguity): Remove unnecessary union and type declarations.
(Leaked lookahead after nondeterministic parse syntax error): New test
case.
* data/glr.c (yyparse): Check for zero stacks remaining before
attempting to shift the lookahead so that you don't lose it.
Joel E. Denny [Thu, 2 Mar 2006 06:18:09 +0000 (06:18 +0000)]
Avoid memory leaks by not invoking longjmp in yyreportAmbiguity.
* tests/glr-regression.at (Leaked semantic values when reporting
ambiguity): New test case.
* data/glr.c (yyreportAmbiguity): Invoke yyyerror directly and return
yyabort rather than invoking yyFail, which invokes longjmp. Remove the
now unnecessary yystackp parameter.
(yyresolveValue): Return yyreportAmbiguity's result. Now the necessary
destructors can be called.
* tests/glr-regression.at: Don't invoke bison with `-t' unnecessarily
in existing testcases.
Joel E. Denny [Thu, 2 Mar 2006 01:35:15 +0000 (01:35 +0000)]
Don't leak semantic values for parent RHS when a user action cuts the
parser, and clean up related code a bit.
* tests/glr-regression.at (Leaked merged semantic value if user action
cuts parse) Rename to...
(Leaked semantic values if user action cuts parse) ... this. Add check
for leaked parent RHS values.
* data/glr.c (yydestroyGLRState): In debugging output, distinguish
between an unresolved state (non-empty chain of semantic options) and
an incomplete one (signaled by an empty chain).
(yyresolveStates): Document the interface. Move all manipulation of an
successfully or unsuccessfully resolved yyGLRState to...
(yyresolveValue): ... here so that yyresolveValue always leaves a
yyGLRState with consistent data and thus is easier to understand.
Remove the yyvalp and yylocp parameters since they are always just
taken from the yys parameter. When reporting a discarded merged value
in debugging output, note that it is incompletely merged. Document the
interface.
(yyresolveAction): If resolving any of the RHS states fails, destroy
them all rather than leaking them. Thus, as long as user actions are
written to clean up the RHS correctly, yyresolveAction always cleans up
the RHS of a semantic option. Document the interface.
Joel E. Denny [Tue, 21 Feb 2006 21:01:57 +0000 (21:01 +0000)]
* doc/bison.texinfo (Mid-Rule Actions): Explain how to bury a
mid-rule action inside a nonterminal symbol in order to declare a
destructor for its semantic value.
Paul Eggert [Fri, 17 Feb 2006 00:08:17 +0000 (00:08 +0000)]
* data/yacc.c [(! defined yyoverflow || YYERROR_VERBOSE) && !
YYSTACK_USE_ALLOCA && ! defined YYSTACK_ALLOC && defined
__cplusplus && ! defined _STDLIB_H && !
((defined YYMALLOC || defined malloc) && (defined YYFREE ||
defined free))]: Include <stdlib.h> rather than rolling our own
declarations of malloc and free, to avoid problems with
incompatible declarations (using 'throw') C++'s stdlib.h. This
should fix Debian bug 340012
<http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=340012>,
reported by Guillaume Melquiond.
Paul Eggert [Fri, 10 Feb 2006 08:00:30 +0000 (08:00 +0000)]
* m4/cxx.m4 (BISON_TEST_FOR_WORKING_CXX_COMPILER): Catch a
configuration screwup "./configure CC=/opt/SUNWspro/bin/c89
CFLAGS="-xarch=generic64" CXX=/opt/SUNWspro/bin/CC
LDFLAGS="-xarch=generic64" with Sun C 5.7 on Solaris 10; this
messes up because C++ code is compiled in 32-bit mode but linked
in 64-bit mode.
Paul Eggert [Wed, 8 Feb 2006 18:56:01 +0000 (18:56 +0000)]
* examples/calc++/Makefile.am (check_PROGRAMS): Renamed from
nodist_PROGRAMS, since we don't need to actually compile the
example if we're just doing a plain 'make'. This avoids bothering
the installer unnecessarily about problems due to weird C++
compilers.
Joel E. Denny [Mon, 30 Jan 2006 11:15:15 +0000 (11:15 +0000)]
During deterministic GLR operation, user actions should be able to
influence the parse by changing yychar. To make this easier to fix and
to make glr.c easier to evolve in general, don't maintain yytoken in
parallel with yychar; just compute yytoken when needed.
* tests/glr-regression.at (Incorrect lookahead during deterministic
GLR): Check that setting yychar in a user action has the intended
effect.
* data/glr.c (yyGLRStack): Remove yytokenp member.
(yyclearin): Don't set *yytokenp.
(yyprocessOneStack, yyreportSyntaxError, yyrecoverSyntaxError): Examine
yychar rather than *yytokenp to determine the current lookahead.
Compute yytoken locally when needed.
(yyparse): Likewise. Remove the local yytoken that yytokenp used to
point to.
* doc/bison.texinfo (Bison Options): Remove stray sentence fragment
after `--report' documentation.
Paul Eggert [Mon, 30 Jan 2006 09:00:40 +0000 (09:00 +0000)]
* src/parse-gram.y (grammar_declaration): Location of printer
symbol is @1, not list->location.
* tests/input.at (Incompatible Aliases): Adjust to above change.
Paul Eggert [Mon, 30 Jan 2006 07:26:00 +0000 (07:26 +0000)]
Fix a longstanding bug uncovered by bro-0.9a9/src/parse.y, which I
got from <http://bro-ids.org/download.html>. The bug is that
when two actions appeared in succession, the second one was
scanned before the first one was added to the grammar rule
as a midrule action. Bison then output the incorrect warning
"parse.y:905.17-906.36: warning: unused value: $3".
* src/parse-gram.y (BRACED_CODE, action): These are no longer
associated with a value.
(rhs): Don't invoke grammar_current_rule_action_append.
(action): Invoke it here instead.
* src/reader.c (grammar_midrule_action): Now extern.
(grammar_current_rule_action_append): Don't invoke
grammar_midrule_action; that is now the scanner's job.
* src/reader.h (last_string, last_braced_code_loc):
(grammar_midrule_action): New decls.
* src/scan-gram.l (last_string): Now extern, sigh.
(last_braced_code_loc): New extern variable.
(<INITIAL>"{"): Invoke grammar_midrule_action if the current
rule already has an action.
(<SC_BRACED_CODE>"}"): Set last_braced_code_loc before returning.
* tests/input.at (AT_CHECK_UNUSED_VALUES):
Add some tests to check that the above changes fixed the bug.
Paul Eggert [Fri, 27 Jan 2006 22:51:00 +0000 (22:51 +0000)]
* src/reader.c (symbol_should_be_used): Renamed from symbol_typed_p.
All used changed. Check whether the symbol has a destructor,
not whether it is typed.
* tests/input.at (AT_CHECK_UNUSED_VALUES): Add a destructor, so
that the values are still reported as unused. All line numbers
adjusted.
Paul Eggert [Mon, 23 Jan 2006 08:39:52 +0000 (08:39 +0000)]
Work around a bug in bro 0.8, which underparenthesizes its
definition of YYLLOC_DEFAULT.
* data/glr.c: Change all uses of YYLLOC_DEFAULT to parenthesize
their arguments.
* data/lalr1.cc: Likewise.
* data/yacc.cc: Likewise.
Paul Eggert [Mon, 23 Jan 2006 07:59:42 +0000 (07:59 +0000)]
* data/yacc.c (b4_int_type): Use yytype_uint8, etc., so that the
user can override the types.
(short): #undef, to work around a bug in Pike 7.0.
(yytype_uint8, yytype_int8, yytype_uint16, yytype_int16): New types.
(union yyalloc.yyss): Use yytype_int16 rather than short.
All uses changed.
(yysigned_char): Remove.
* src/parse-gram.y (YYTYPE_UINT8, YYTYPE_INT8, YYTYPE_UINT16):
(YYTYPE_INT16): New macros, to test the new facility in yacc.c.
* tests/regression.at (Web2c Actions): Adjust to above changes.
Paul Eggert [Mon, 23 Jan 2006 04:37:09 +0000 (04:37 +0000)]
(check_and_convert_grammar): New function.
(reader): Close the input file even if something went wrong during
parsing. Minor file descriptor leak reported by twlevo.
Paul Eggert [Sun, 22 Jan 2006 08:17:18 +0000 (08:17 +0000)]
* src/assoc.c (assoc_to_string): Use a default: abort (); case
to pacify gcc -Wswitch-default.
* src/scan-gram.l (adjust_location): Use a default: break; case
to pacify gcc -Wswitch-default.
* src/scan-skel.h (skel_get_lineno, skel_get_in, skel_get_out):
(skel_get_leng, skel_get_text, skel_set_lineno, skel_set_in):
(skel_set_out, skel_get_debug, skel_set_debug, skel_lex_destroy):
Move these decls to scan-skel.l, since they don't need to be
visible elsewhere.
* src/scan-skel.l: Accept the above decls.
(skel_lex): New decl, to pacify GCC when unpatched flex 2.5.31
is used.
Paul Eggert [Fri, 20 Jan 2006 22:23:44 +0000 (22:23 +0000)]
* doc/bison.texinfo (Prologue, Rules, Actions, Union Decl, Action Decl):
(Destructor Decl, Parser Function, Pure Calling):
Describe rules for braces inside C code more carefully.
Paul Eggert [Fri, 20 Jan 2006 01:59:30 +0000 (01:59 +0000)]
Fix some porting glitches found by Nelson H. F. Beebe.
* lib/abitset.c (abitset_resize): Rewrite to avoid warnings from
compilers that don't understand that abort () does not return.
* src/state.c (transitions_to): Likewise.
* m4/cxx.m4 (BISON_TEST_FOR_WORKING_CXX_COMPILER): Check
that '#include <cstdlib>' works.
* src/system.h (INT8_MIN, INT16_MIN, INT32_MIN, INT8_MAX):
(INT16_MAX, UINT8_MAX, INT32_MAX, UINT16_MAX, UINT32_MAX):
#undef if ! defined __STDC_VERSION__ || __STDC_VERSION__ < 199901,
for the benefit of some pre-C99 compilers.
Paul Eggert [Fri, 20 Jan 2006 00:08:30 +0000 (00:08 +0000)]
Minor fixups to get 'make maintainer-check' to work.
* configure.ac: Don't use -Wnested-externs, as it's incompatible
with the new verify.h implementation.
* data/c.m4 (b4_yy_symbol_print_generate): YYUSE (yyoutput).
* data/glr.c (YYUSE): Depend on __GNUC__ as well.
* data/yacc.c (YYUSE): Likewise.
* data/lalr1.cc (yysyntax_error_): YYUSE (yystate).
* lib/subpipe.c (end_of_output_subpipe): The args are unused.
* src/parse-gram.y (declaration): Don't pass a string constant
to a function that expects char *, since GCC might complain
about the constant value.
* src/reader.c (symbol_typed_p): Add parens to pacify GCC.
* tests/cxx-type.at (_AT_TEST_GLR_CXXTYPES): #undef yylloc and yylval
before #defining them.
* tests/glr-regression.at
(Incorrectly initialized location for empty right-hand side in GLR):
In yyerror, use the msg arg.
(Corrupted semantic options if user action cuts parse):
(Incorrect lookahead during deterministic GLR):
(Incorrect lookahead during nondeterministic GLR):
Don't name a local var 'index'; it shadows string.h's 'index'.