On the CI, both GCC and Clang report:
src/parse-gram.c: In function 'yy_lac':
src/parse-gram.c:1479:29: error: format '%hd' expects argument of type 'int',
but argument 3 has type 'yytype_int16 {aka long int}' [-Werror=format=]
YYDPRINTF ((stderr, " G%hd", yystate));
^
Although yytype_int16 is supposed to be a short int, not a long int.
This must be explored.
* data/yacc.c (yy_lac): Work around typing issue.
* data/glr.c: Don't use `foo |= bar` with foo and bar being yybool:
the result appears to be an int, not a yybool.
Use yybool where appropriate.
Add casts where needed.
Reported by Derek Clegg.
http://lists.gnu.org/archive/html/bug-bison/2018-10/msg00018.html
Rather than adding casts, we should be more careful with types. For
instance yystate should be a yytype_int16. But currently we can't: it
is also used sometimes for storing other things that state numbers.
* data/yacc.c (yyparse): Add missing casts.
The yacc.c skeleton is old, and was using many tricks to save
registers. Today's register allocators can do this themselves. Let's
keep the code simpler to read and let compilers do their job.
* data/yacc.c: Avoid using yystate for different types of content.
An inline function would be better, but doing this portably will be
a problem.
* data/lalr1.cc: Fix oldish comment.
* data/stack.hh: Prefer typename for type names.
Use size() instead of duplicating it.
* examples/variant-11.yy, examples/variant.yy (yylex): Use int,
as this is the type of the semantic value.
Currently we emit useless code for places where we might issue user
content, but there is none. This commit avoids this. Besides, some
of the comments looked like implementation details ("Copy the first
part of user declarations"), rather than made for the reader of the
result ("First part of user prologue").
On Bison's parse-gram.c we get:
@@ -76,10 +76,6 @@
#define yynerrs gram_nerrs
-/* Copy the first part of user declarations. */
-
-#line 82 "src/parse-gram.c" /* yacc.c:339 */
-
* data/bison.m4 (b4_define_user_code): Accept a comment to document
the section.
Do not emit any code if the content is empty.
Adjust callers to not emit the comment.
Do not
* data/glr.c, data/glr.cc, data/lalr1.cc, data/lalr1.java, data/yacc.c:
Adjust.
When we introduced variants in Bison, C++ did not have the 'emplace'
functions, and we chose 'build'. Let's align with modern C++ and
promote 'emplace' rather than 'build'.
* data/lalr1.cc, data/variant.hh (emplace): New.
(build): Deprecate in favor of emplace.
* doc/bison.texi: Adjust.
In some casing, once we moved a stack symbol, we forget to mark the
source stack symbol as emptied. As a consequence, it may be destroyed
a second time.
This happens when the stack has to be resized.
* data/lalr1.cc (stack_symbol_type::stack_symbol_type): Record that
the source was emptied.
(stack_symbol_type::operator=): Likewise.
* tests/c++.at (C++ Variant-based Symbols Unit Tests): Force the stack
to be resized. Check its content.
The files stack.hh and position.hh are deprecated. Rather than
devoting specify %define variables to discard them (api.position.file
and api.stack.file), and rather than having to use special rules when
api.location.file is used, let's simply decide that from %require
"3.2" onwards, these files will not be generated.
The only noticeable thing here is that, in order to be able to check
the behavior of %require "3.2", to have this version (which is still
3.1-*) to accept %require "3.2".
* src/gram.h, src/gram.c (required_version): New.
* src/parse-gram.y (version_check): Set it.
* src/output.c (prepare): Pass it m4.
* data/bison.m4 (b4_required_version_if): Receive it and use it.
* data/location.cc, data/stack.hh: Replace the api.*.file with only
required version comparison.
* tests/input.at: No longer check api.stack.file and api.position.file.
* NEWS, doc/bison.texi: Don't mention them.
Document the %require 3.2 behavior.
* tests/output.at: Use %require 3.2 instead.
Currently we use "<dir><api.location.file>" as \file argument, and as
base for the CPP guard. This is not nice when <dir> is absolute, in
which case it is expected that the user will use api.location.include
to get something nicer. If defined, use that name instead.
* data/location.cc (b4_location_path): New.
Use it.
* tests/c++.at (Shared locations): Check the guard and Doxygen doc.
Users may want to generate the location file elsewhere, say
$top_srcdir/include/ast/location.hh. Yet, we should not generate
`#include "$top_srcdir/include/ast/location.hh"` but probably
something like `#include <ast/location.hh>`, or `#include
"ast/location.hh", or `#include <location.hh>`. It entirely depends
on the compiler flags (-I/-isystem) that are used. Bison cannot guess
what is expected, so let's give the user a means to tell how the
location file should be included.
* data/location.cc (b4_location_file): New.
* data/glr.cc, data/lalr1.cc: Use it.
In the case a user wants to create location.hh elsewhere, it can be
helpful to define api.location.file to some possibly absolute path
such as -Dapi.location.file='"$(top_srcdir)/include/ast/location.hh"'.
Currently this does not work with `-o foo/parser.cc`, as we join foo/
and $(top_srcdir) together, the latter starting with slash.
We should not try to do that in m4, manipulating file names is quite
complex when you through Windows file name in. Let m4 delegate this
to gnulib.
* src/scan-skel.l (at_output): Accept up to two arguments.
* data/bison.m4 (b4_output): Adjust.
* tests/skeletons.at (Fatal errors but M4 continues producing output):
Adjust to keep the error.
* data/location.cc, data/stack.hh: Leave the concatenation to @output.
* tests/output.at: Exercise api.location.file with an absolute path.
Make it easier to have fewer files.
* data/stack.hh: Don't generate stack.hh when api.location.file is
specified.
* tests/calc++.at, tests/output.at: Adjust tests.
Let's put the definition of position into location.hh, there's no real
value in keeping them separate: they are small, and share the same
requirements.
To help users transition to this new model, still generate position.hh
by default, but as a simple include to location.hh.
* data/location.cc (api.position.file): Accept only 'none' as possible
value.
(position.hh): Make it a stub.
(location.hh): Adjust.
(b4_position_define): Merge into...
(b4_location_define): this.
* data/glr.cc, data/lalr1.cc, tests/input.at, tests/output.at: Adjust.
* data/location.cc: Sort includes.
(b4_position_file, b4_location_file): New.
When there's a file for locations but not for positions, include the
definition of position in the location file.
* data/lalr1.cc (b4_shared_declarations): Include the
position/location file when it exists.
Otherwise, define the class.
* data/glr.cc: Likewise.
* tests/input.at (%define file variables): Check them.
* tests/output.at (C++ output): Check various cases with
api.position.file and api.location.file.
It was not a good idea to generate the file stack.hh. It never was.
But now we have to deal with backward compatibility: if we stop
generating it, the build system of some build system will probably
break.
So offer the user a means to (i) decide what the name of the output
file should be, and (ii) not generate this file at all (its content
will be inline where the parser is defined).
* data/lalr1.cc (b4_percent_define_check_file_complain)
(b4_percent_define_check_file): New.
* data/stack.hh: Generate the file only if api.stack.file is not
empty.
In that case, use it as file name.
* data/lalr1.cc: Adjust to include the right file, or to include
the definition of stack.
* tests/calc.at, tests/output.at: Exercise api.stack.file.
We use both styles, let's stick to a single one. Autoconf uses the
prefix one, let's do the same.
* data/bison.m4, data/c++.m4, data/c-like.m4, data/lalr1.cc,
* data/variant.hh, data/yacc.c: Rename all the b4_*_ macros
as _b4_*.
The CI is littered with
# -*- compilation -*-
423. regression.at:907: testing Dancer %glr-parser ...
./regression.at:907: bison -fno-caret -o dancer.c dancer.y
./regression.at:907: $BISON_C_WORKS
stderr:
stdout:
./regression.at:907: $CC $CFLAGS $CPPFLAGS $LDFLAGS -o dancer dancer.c $LIBS
stderr:
icc: command line warning #10006: ignoring unknown option '-Wcast-align'
icc: command line warning #10006: ignoring unknown option '-fparse-all-comments'
icc: command line warning #10006: ignoring unknown option '-Wdocumentation'
icc: command line warning #10006: ignoring unknown option '-Wnull-dereference'
icc: command line warning #10006: ignoring unknown option '-Wbad-function-cast'
icc: command line warning #10006: ignoring unknown option '-fno-color-diagnostics'
icc: command line warning #10006: ignoring unknown option '-Wno-keyword-macro'
dancer.c(755): error #1628: function declared with "noreturn" does return
}
^
dancer.c(761): error #1628: function declared with "noreturn" does return
}
^
compilation aborted for dancer.c (code 2)
ICC sees that `longjmp(buf, 1);` does not return, it sees that
`abort();` does not either, but fails to see it for
`longjmp(buf, 1); abort();`
* data/glr.c (YYLONGJMP): Be even clearer on the fact this does not
return.
In C++ pre C++11 it is standard practice to use 0 for the null pointer.
But GCC pre 8 -std=c++98 with -Wzero-as-null-pointer-constant warns about
this.
So disable -Wzero-as-null-pointer-constant when compiling C++ pre 11.
Let's do this in AT_DATA_SOURCE_PROLOGUE (which is pasted on top of
all the test grammar files). Unfortunately, that shifts all the
locations in the expected error messages, which would be too noisy.
Instead, let's introduce testsuite.h, which can vary in length, and
include it in AT_DATA_SOURCE_PROLOGUE.
* tests/testsuite.h: New.
Disable -Wzero-as-null-pointer-constant's warning with GCC pre 8,
C++ pre 11.
* tests/local.at (AT_DATA_SOURCE_PROLOGUE): Use it.
* tests/atlocal.in (CPPFLAGS): Find it.
* tests/local.mk: Ship it.
* data/c.m4 (YY_NULLPTR): Prefer ((void*)0) to 0 in C.
Reported by Frank Heckenbach.
http://lists.gnu.org/archive/html/bug-bison/2018-03/msg00002.html
Actually the assignment operator should never be needed: the C++98
requirements for vector::push_back is CopyInsertable, which does not require
an assignment operator. However, libstdc++ shipped with GCC up to (and
including) 6 uses the assignment operator (which affects Clang on top of
libstdc++, but also ICC). So let's keep it for legacy C++.
See https://gcc.godbolt.org/z/q0XXmC.
* data/lalr1.cc (stack_symbol_type::operator=): Remove.
* data/c++.m4 (basic_symbol::operator=): Ditto.
* tests/c++.at (C++ Variant-based Symbols Unit Tests): Adjust.
Modern C++ (i.e., C++11 and later) introduced "move only" types: types such
as std::unique_ptr<T> that can never be duplicated. They must never be
copied (by assignments and constructors), they must be "moved". The
implementation of lalr1.cc used to copy symbols (including their semantic
values). This commit ensures that values are only moved in modern C++, yet
remain compatible with C++98/C++03.
Suggested by Frank Heckenbach, who provided a full implementation on
top of C++17's std::variant.
See http://lists.gnu.org/archive/html/bug-bison/2018-03/msg00002.html,
and https://lists.gnu.org/archive/html/bison-patches/2018-04/msg00002.html.
Symbols (terminal/non terminal) are handled by several functions that used
to take const-refs, which resulted eventually in a copy pushed on the stack.
With modern C++ (C++11 and later) the callers must use std::move, and the
callees must take their arguments as rvalue refs (foo&&). In order to avoid
duplicating these functions to support both legacy C++ and modern C++, let's
introduce macros (YY_MOVE, YY_RVREF, etc.) that rely on copy-semantics for
C++98/03, and move-semantics for modern C++.
That's easy for inner types, when the parser's functions pass arguments to
each other. Functions facing the user (make_NUMBER, make_STRING, etc.)
should support both rvalue-refs (for instance to support move-only types:
make_INT (std::make_unique<int> (1))), and lvalue-refs (so that we can pass
a variable: make_INT (my_int)). To avoid the multiplication of the
signatures (there is also the location), let's take the argument by value.
See:
https://lists.gnu.org/archive/html/bison-patches/2018-09/msg00024.html.
* data/c++.m4 (b4_cxx_portability): New.
(basic_symbol): In C++11, replace copy-ctors with move-ctors.
In C++11, replace copies with moves.
* data/lalr1.cc (stack_symbol_type, yypush_): Likewise.
Use YY_MOVE to avoid useless copies.
* data/variant.hh (variant): Support move-semantics.
(make_SYMBOL): In C++11, in order to support both read-only lvalues,
and rvalues, take the argument as a copy.
* data/stack.hh (yypush_): Use rvalue-refs in C++11.
* tests/c++.at: Use move semantics.
* tests/headers.at: Adjust to the new macros (YY_MOVE, etc.).
* configure.ac (CXX98_CXXFLAGS, CXX11_CXXFLAGS, CXX14_CXXFLAGS)
(CXX17_CXXFLAGS, ENABLE_CXX11): New.
* tests/atlocal.in: Receive them.
* examples/variant.yy: Don't define things in std.
* examples/variant-11.test, examples/variant-11.yy: New.
Check the support of move-only types.
* examples/README, examples/local.mk: Adjust.
This generates less code, which is nicer to read, but also takes less
chances with compilers such as G++ 4.8 that are too strict and check
"dead code" (templated code that is not instantiated).
* data/c++.m4 (b4_symbol_type_declare, b4_symbol_type_define): When
variants are used, don't generate code meant for non variants.
Currently, in glr.cc, we emit the definitions of basic_symbol and
symbol_type, although there are not used.
* data/c++.m4 (b4_public_types_declare): Extract these definitions from
here, and move them...
(b4_symbol_type_declare): here.
(b4_public_types_declare): Also remove the definition of the symbol
constructors.
* data/lalr1.cc (b4_shared_declarations): Adjust: call
b4_symbol_type_declare and b4_symbol_constructor_declare.
