If we link several glr2.cc parsers together, we get linking failures
because of duplicate symbols.
* data/skeletons/glr2.cc (semantic_option::indexIn)
(semantic_option::next): Remove the useless overloads.
We don't need them in the header file.
* data/skeletons/glr2.cc (YY_EXCEPTIONS): Define only in the
implementation file.
* tests/headers.at (Several Parsers): Also check glr2.cc.
Let's use c++/glr to demonstrate custom error messages in C++ (not
just in glr2.cc).
* examples/c++/glr/c++-types.yy (report_syntax_error): New.
* examples/c++/glr/c++-types.test: Adjust.
* examples/c/bistromathic/parse.y: Comment changes.
* tests/local.at (AT_YYERROR_DEFINE(c++)): Use a nicer way to print
the lookakead's name.
Currently glr2.cc uses three variables/struct members to denote the
symbols' kind (or state), value and location. lalr1.cc has two types
for "complete" symbols: symbol_type and stack_symbol_type. Let's use
that model in glr2.cc too.
For a start use yyla (a symbol_type) to denote the lookahead, instead
of the triple yytoken, yylval and yylloc. This will make easier the
introduction of the "context" subclass, used in parse.error=custom.
It simplifies the code in several places. For instance from:
symbol_kind_type yytoken_current = this->yytoken;]b4_variant_if([[
value_type yylval_current;
]b4_symbol_variant([this->yytoken],
[yylval_current], [move], [this->yylval])], [[
value_type yylval_current = this->yylval;]])[]b4_locations_if([
location_type yylloc_current = this->yylloc;])[
to:
symbol_type yyla_current = std::move (this->yyla);
* data/skeletons/glr2.cc (yytoken, yylval, yylloc): Replace by...
(yyla): this.
Adjust all dependencies.
(yyloc_default): Remove, unused.
* tests/c++.at, tests/glr-regression.at, tests/types.at: C++11 is
required for glr2.cc.
Adjust to changes in glr2.cc.
* data/skeletons/c++.m4: Don't define obsolete identifiers in the case
of glr2.cc. Let's not start with technical debt.
* data/skeletons/glr2.cc, data/skeletons/lalr1.cc,
* data/skeletons/variant.hh: Use token_kind_type, not token_type.
* tests/c++.at, tests/local.at: Use value_type, not semantic_type.
It is not available on HP-UX 11i.
Reported by Larkin Nickle <[email protected]>.
<https://lists.gnu.org/r/bug-bison/2021-09/msg00012.html>
Gnulib provides no replacement, but anyway we should be using doubles,
since difftime uses doubles.
* bootstrap.conf: We want portability on stdtod.
* src/counterexample.c: Use double, not float, for time measurements.
The recent changes to comply with POSIX are breaking Automake's test
suite.
Reported by Kiyoshi Kanazawa.
<https://lists.gnu.org/r/bug-bison/2021-09/msg00005.html>
To limit the impact of POSIX changes, bind them to $POSIXLY_CORRECT.
Suggested by Karl Berry.
<https://lists.gnu.org/r/bug-bison/2021-09/msg00009.html>
The existing `maintainer-check-posix` Make target checks these
changes.
* src/getargs.h, src/getargs.c (set_yacc): New.
Use it.
* data/skeletons/bison.m4 (b4_posix_if): New.
* data/skeletons/yacc.c (b4_declare_yyerror_and_yylex): Use it.
* doc/bison.texi, tests/local.at: Adjust.
For some reason this test fails on Solaris/x86. But multistart is not
part of 3.8, so we can postpone the debugging of this issue.
Reported by Dagobert Michelsen.
<https://lists.gnu.org/r/bug-bison/2021-08/msg00027.html>
* tests/report.at (Multistart reports): Comment out.
Contrary to what commit d8cc6b073e "c++:
shorten the assertions that check whether tokens are correct"
believed, MS Visual C++'s preprocessor limitation is not on the input
line length, but on the size of the line holding the full C++
statement.
Reported by Vince Huffaker <[email protected]>
<https://lists.gnu.org/r/help-bison/2021-08/msg00003.html>
* data/skeletons/variant.hh (_b4_symbol_constructor_define): Hide the
assertion from Visual C++.
On Solaris, sed throws away the NUL bytes from the stream, even in C
locale. So instead of postprocessing bison's stderr to neutralize
changes in value of `argv[0]`, use an envvar to actually neutralize
variations of `argv[0]` during tests.
Reported by Dagobert Michelsen.
<https://lists.gnu.org/r/bug-bison/2021-08/msg00025.html>
* src/main.c (main): Change `argv[0]` if BISON_PROGRAM_NAME is
defined.
* tests/bison.in: No longer mess with stderr, just pass
the expected BISON_PROGRAM_NAME value.
Besides, for mysterious reasons, this fails on some environment.
Reported by Dagobert Michelsen.
<https://lists.gnu.org/archive/html/bug-bison/2021-08/msg00008.html>
* examples/c/lexcalc/lexcalc.test, examples/c/lexcalc/parse.y,
* examples/c/lexcalc/scan.l: Revert to a single-start example.
Reported by Dagobert Michelsen.
https://lists.gnu.org/r/bug-bison/2021-08/msg00006.html
* m4/bison-cxx-std.m4 (_BISON_CXXSTD_98_snippet): We don't need
vector::data, it was only for glr2.cc, which is C++11 anyway.
(_BISON_CXXSTD_11_snippet): We need vector::data and std::swap on
arrays.
* m4/cxx.m4 (BISON_TEST_FOR_WORKING_CXX_COMPILER): We don't need
vector::data.
* tests/local.at (AT_COMPILE_CXX): Skip when glr2.cc and no support
for C++11.
In c2ba260487 ("glr.c: fix signature
when using custom error messages"), I meant to add a test case for C,
not C++. It does not work in C++.
* tests/calc.at: Run for glr.c, not glr.cc.
GCC 4.2 on macOS claims to support C++98, but does not feature it.
input.cc: In member function 'void state_stack::yycompressStack()':
input.cc:1774: error: 'class std::vector<glr_stack_item, std::allocator<glr_stack_item> >' has no member named 'data'
Reported by Christopher Nielsen <[email protected]>.
<https://trac.macports.org/raw-attachment/ticket/59927/bison-test-results-20210811-95b72.log.xz>.
* m4/bison-cxx-std.m4 (_BISON_CXXSTD_98_snippet): Check for it.
* m4/cxx.m4 (BISON_TEST_FOR_WORKING_CXX_COMPILER): Ditto.
* tests/atlocal.in: It was an error for tests to depend on gnulib:
they must not, as gnulib would hide portability issues that we want to
catch. So this piece of code is no longer useful, and must not be
useful.
In POSIX Yacc mode, declare yyerror and yylex unless already #defined,
or if YYERROR_IS_DECLARED/YYLEX_IS_DECLARED are defined (for
consistency with Bison's YYSTYPE_IS_DECLARED/YYLTYPE_IS_DECLARED).
See <https://austingroupbugs.net/view.php?id=1388#c5220>.
* data/skeletons/c.m4 (b4_function_declare): Resurect.
(b4_lex_formals): Since we will possibly expose this prototype
in the header, take the prefix into account.
* data/skeletons/yacc.c (b4_declare_yyerror_and_yylex): New.
(b4_shared_declarations): Use it.
* tests/local.at (AT_YACC_IF): New.
When in Yacc mode, set the `yacc` Autotest keyword.
(AT_YYERROR_DECLARE(c)): Don't declare in Yacc mode,
to avoid clashes (since this signature is static).
(AT_YYERROR_DEFINE(c)): Don't define as static in Yacc mode.
* tests/regression.at (Early token definitions with --yacc): Specify
that we are in Yacc mode.
Unfortunately it seems to be quite difficult to have "LAC: Exploratory
stack" run for D.
* data/skeletons/lalr1.d: We need File when traces are enabled.
* tests/local.at (AT_YYLEX_DEFINE(d)): New.
* tests/regression.at: Prepare for D, but don't run it, it does not
work.
* tests/local.at (AT_FOR_EACH_SKEL): New.
Use where appropriate.
* data/skeletons/lalr1.d: Reject -d.
* tests/input.at, tests/scanner.at: Also check D.
Check in m4's output if there are sequences such as m4_foo or b4_foo,
which are probably resulting from incorrect m4 processing.
It actually already is useful:
- it caught a leaking b4_lac_if leaking from glr.c, where LAC is not
supported, hence b4_lac_if is not defined.
- it also caught references to location.hh in position.hh when
location.hh does not exist.
- while making "Code injection" robust to these new warnings (it is
its very purpose to let b4_canary pass unevaluated), I saw that it
did not check lalr1.d, and when adding lalr1.d, it revealed it did
underquote ocurrences of token value types.
* src/scan-skel.l (macro): New abbreviation.
Use it.
* data/skeletons/glr.c: Don't use b4_lac_if, we don't have it.
* data/skeletons/location.cc: Don't generate position.hh when we don't
generate location.hh.
* data/skeletons/d.m4 (b4_basic_symbol_constructor_define): Fix
underquotation.
* data/skeletons/bison.m4 (b4_canary): New.
* tests/input.at (Code injection): Use it, and check lalr1.d too.
* examples/c++/Makefile, examples/c++/calc++/Makefile,
* examples/c++/glr/Makefile, examples/c/bistromathic/Makefile,
* examples/c/calc/Makefile, examples/c/glr/Makefile,
* examples/c/lexcalc/Makefile, examples/c/mfcalc/Makefile,
* examples/c/pushcalc/Makefile, examples/c/reccalc/Makefile,
* examples/c/rpcalc/Makefile, examples/d/calc/Makefile,
* examples/d/simple/Makefile, examples/java/calc/Makefile,
* examples/java/simple/Makefile:
Use --html to generate *.html directly.
No longer demonstrate --xml.
No longer show rules for xml to html.
Use --header, not --defines.
Use --graph without specifying the output file now that we
generate *.gv by default.
We had:
```
-mbchar ...|\xF0[\x\90-\xBF]([\x80-\xBF]{2})|...
+mbchar ...|\xF0[\x90-\xBF]([\x80-\xBF]{2})|...
```
so a precise sequence that matches the incorrect regex can let NUL
bytes pass through, which triggers an assertion violation downstream.
It is a pity that Flex does not report an error for such input.
Reported by Ahcheong Lee <[email protected]>.
<https://lists.gnu.org/r/bug-bison/2021-04/msg00003.html>
* src/scan-gram.l (mbchar): Fix the bad regex.
* tests/input.at (Invalid inputs): Check that case.
* doc/bison.texi: Use @samp{...}, not "..." for pieces of code.
Use @samp{...}, not @command{...} for command lines.
Promote %header/--header over %defines/--defines.
Spellcheck.
The union of the values is handled by the backend.
In D, unions can hold classes, structs, etc., so this is more similar
to the C++ api.value.type variant.
* data/skeletons/d.m4, data/skeletons/lalr1.d: Here.
* tests/calc.at, tests/local.at: Test it.
When a pstate is used for multiple successive parses, some state may
leak from one run into the following one. That was introduced in
330552ea49 "yacc.c: push: don't clear
the parser state when accepting/rejecting".
Reported by Ryan <[email protected]>
https://lists.gnu.org/r/bug-bison/2021-03/msg00000.html
* data/skeletons/yacc.c (yypush_parse): We reusing a pstate from a
previous run, do behave as if it were the first run.
* tests/push.at (Pstate reuse): Check this.
The right-shift added in c22902e360
("tables: fix handling for useless tokens") is incorrect. In
particular, we need to reset the "new" bits.
Reported by Balázs Scheidler.
https://github.com/akimd/bison/issues/74
* src/tables.c (pos_set_set): Fix the right-shift.
Don't repeatedly call malloc/free for each call to map_file_name.
* bootstrap.conf: We need hash-map.
* src/files.h, src/files.c (map_file_name): The caller must not free
the result.
Adjust callers.
(mapped_dir_prefix, spec_mapped_header_file): Remove.
* src/files.c
(map_file_name): Rename as...
(map_file_name_alloc): this.
(mapped_files, map_file_name, string_equals, string_hash, string_free):
New.
* examples/c/bistromathic/parse.y, examples/c/glr/c++-types.y,
* examples/c/lexcalc/parse.y: Use const where appropriate.
Avoid `yy` prefixes where it does not make sense.
Avoid the `p` prefix for pointers.
* examples/c/bistromathic/parse.y, examples/c/glr/c++-types.y,
* examples/c/lexcalc/parse.y: Don't use the private internal detail
`YY_LOCATION_PRINT`, use `YYLOCATION_PRINT`.
It is very helpful to be able to look at the locations in the debug
traces, let's provide the users with (i) a means to do that for their
location types, and (ii) a public macro to print locations when debug
traces are enabled.
* data/skeletons/c.m4 (b4_yy_location_print_define): Rename as...
(b4_yylocation_print_define): this.
Define YYLOCATION_PRINT instead of YY_LOCATION_PRINT.
Ensure backward compatibility for those who might have defined/used
YY_LOCATION_PRINT in spite the warnings.
Adjust dependencies.
* data/skeletons/glr2.cc: We don't use YYLOCATION_PRINT here.
That change was started in Bison 3.4. The announcement for 3.7 stated
that in Bison 3.8 we would use *.gv by default.
* src/files.c (compute_output_file_names): spec_graph_file defaults
too *.gv.
* doc/bison.texi, examples/c++/calc++/local.mk, tests/output.at:
Adjust.
Its removal was annonced several times in NEWS (for Bison 3.5, 3.6, 3.7).
* data/skeletons/c.m4, data/skeletons/yacc.c: Remove support for YYPRINT.
* NEWS: Fix the mess introduced by the merge.
Document the removal of YYPRINT.
* doc/bison.texi (The YYPRINT Macro): Remove.
Using #define YYSTYPE has always been strongly discouraged in C++.
Macros are dangerous and can result in subtle bugs.
https://lists.gnu.org/r/bug-bison/2020-12/msg00007.html
Maybe some people are currently using #define YYSTYPE. Instead of
dropping support right now, first issue a warning. Bison can "see" if
YYDEBUG is defined (it could even be on the command line), only the
compiler knows. Unfortunately `#warning` is non-portable, and
actually GCC even dies on it when `-pedantic` is enabled. So we need
to use `#pragma message`. We must make it conditional as some
compilers might not support it, but it doesn't matter if only _some_
compilers emit the warning: it should be enough to catch the attention
of the developers.
* data/skeletons/c++.m4: Issue a warning when the user defined
YYSTYPE.
* tests/actions.at: Don't #define YYSTYPE.
* tests/headers.at (Several parsers): Ignore the YYSTYPE in the
warning.
We have been accepting this for years, but it is deprecated: people
are expecting to define api.value.type instead.
* doc/bison.texi: Make it clear that YYSTYPE and YYLTYPE are for C
only.
* data/skeletons/glr2.cc, data/skeletons/lalr1.cc (b4_lex): Don't
refer to YYSTYPE, but to value_type.
Note that this is "symbolic" in the sense that these values are never
used by b4_function_call.
Akim Demaille