* data/lalr1-fusion.cc (yydestruct_): Invoke the variant's
destructor.
Display the value only if yymsg is nonnull.
(yyreduce): Invoke yydestruct_ when popping lhs symbols.
This is used to help the user catch cases where some value gets
ovewritten by a new one. This should not happen, as this will
probably leak.
Unfortunately this uncovered a bug in the C++ parser itself: the
lookahead value was not destroyed between two calls to yylex. For
instance if the previous lookahead was a std::string, and then an int,
then the value of the std::string was correctly taken (i.e., the
lookahead was now an empty string), but std::string structure itself
was not reclaimed.
This is now done in variant::build(other&) (which is used to take the
value of the lookahead): other is not only stolen from its value, it
is also destroyed. This incurs a new performance penalty of a few
percent, and union becomes faster again.
* data/lalr1-fusion.cc (variant::build(other&)): Destroy other.
(b4_variant_if): New.
(variant::built): New.
Use it whereever the status of the variant changes.
* etc/bench.pl.in: Check the penalty of %define assert.
* data/lalr1-fusion.cc (b4_symbol_variant): Adjust additional
arguments.
(variant::build): New overload for
copy-construction-that-destroys.
(variant::swap): New.
(parser::yypush_): Use it in variant mode.
* data/bison.m4 (b4_copyright): Fix the indentation of the
copyright year paragraph.
Use b4_copyright_years when no years are given.
* data/lalr1.cc, data/lalr1-fusion.cc, data/location.cc
(b4_copyright_years): New.
Use it.
In order to make it easy to perform benchmarks to ensure that there are no
performance loss, lalr1.cc is forked into lalr1-fusion.cc. Eventually,
lalr1-fusion.cc will replace lalr1.cc.
Meanwhile, to make sure that lalr1-fusion.cc is correctly exercized by the
test suite, the user must install a symbolic link from lalr1.cc to it.
Instead of having three stacks (state, value, location), use a stack
of triples. This considerably simplifies the code (and it will be
easier not to require locations as currently does the C++ parser),
and also gives a 10% speedup according to etc/bench (probably mainly since
memory allocation is done once instead of three times).
Another motivation is to make it easier to destruct properly
semantic values: now that they are bound to their state (hence
symbol type) it will be easier to call the appropriate destructor.
These changes should probably benefit the C parser too.
* data/lalr1.cc: Copy as... * data/lalr1-fusion.cc: this new
file.
(b4_rhs_value, b4_rhs_location): New definitions overriding those
from c++.m4.
(state_stack_type, semantic_stack_type, location_stack_type)
(yystate_stack_, yysemantic_stack_, yylocation_stack_): Remove.
(data_type, stack_type, yystack_): New.
(YYLLOC_DEFAULT, yypush_): Adjust.
(yyerror_range): Now based on data_type, not location_type.
This is needed to prepare a forthcoming patch that fuses the three
stacks into one.
* data/lalr1.cc (parser::yypush_): New.
(parser::yynewstate): Change the semantics: instead of arriving to
this label when value and location have been pushed, but yystate
is to be pushed on the state stack, now the three of them must
have been pushed before. yystate still must be the new state.
This allows to use yypush_ everywhere instead of individual
handling of the stacks.
* data/lalr1.cc (b4_symbol_actions): New, overrides the default C
definition to use references instead of pointers.
(yy_symbol_value_print_, yy_symbol_print_, yydestruct_):
Take the value and location as references.
Adjust callers.
This patch was inspired by work by Michiel De Wilde. But he used Boost
variants which (i) requires Boost on the user side, (ii) is slow, and
(iii) has useless overhead (the parser knows the type of the semantic value
there is no reason to duplicate this information as Boost.Variants do).
This implementation reserves a buffer large enough to store the largest
objects. yy::variant implements this buffer. It was implemented with
Quentin Hocquet.
* src/output.c (type_names_output): New.
(output_skeleton): Invoke it.
* data/c++.m4 (b4_variant_if): New.
(b4_symbol_value): If needed, provide a definition for variants.
* data/lalr1.cc (b4_symbol_value, b4_symbol_action_)
(b4_symbol_variant, _b4_char_sizeof_counter, _b4_char_sizeof_dummy)
(b4_char_sizeof, yy::variant): New.
(parser::parse): If variants are requested, define
parser::union_type, parser::variant, change the definition of
semantic_type, construct $$ before running the user action instead
of performing a default $$ = $1.
* examples/variant.yy: New.
Based on an example by Michiel De Wilde.
To make future changes easier, no longer rely on ".TYPE" being the
way to get a semantic value.
* data/c.m4 (b4_symbol_value): New.
Use it.
* data/c++.m4, data/yacc.c: Use it.
* data/glr.c: Use b4_symbol_value.
(b4_rhs_data): New.
Use it.
2008-11-01 Di-an Jan <dianj@freeshell.org>
Paolo Bonzini <bonzini@gnu.org>
Support all Java parser class modifiers.
* data/java.m4 (b4_percent_define_get3): New.
(b4_final_if, b4_strictfp_if): New.
* data/lalr1.java (final, strictfp, extends, implements): Support.
* doc/bison.texinfo (final, strictfp, extends, implements): Add
documentation.
* tests/java.at (AT_CHECK_JAVA_MINIMAL): New.
(AT_CHECK_JAVA_MINIMAL_W_LEXER): New.
(AT_CHECK_JAVA_GREP): New.
(Java parser class modifiers): New test.
(Java parser class extends and implements): New test.
Model exception propagation better with throws and lex_throws.
* data/java.m4 (b4_list2): New.
(throws): Change default.
* data/lalr1.java (yyaction): Add throws.
(parse): Add lex_throws in addition to throws.
* doc/bison.texinfo (throws, lex_throws): Add documentation.
* tests/java.at (Java throws specifications): New test.
Improve documentation for Java parsers.
* doc/bison.texinfo (Java Parsers): Add subsections.
Don't quote first argument of %define.
(Java Bison Interface): Document output files. Move documentation
of parser class and merge into Java Parser Interface. Document
features that error out. Document directives with no effect.
Move note about Javadoc higher.
(Java Semantic Values): Explicitly mention stype.
Document that generic types cannot be used.
(Java Location Values): Use @deftypeivar. Document constructors.
Correct return value for toString.
(Java Parser Interface): List undocumented constants/fields.
Move documentation of fields added by %parse-param closer to list
of members. Document that token names are added as fields.
Document constructors accurately. Remove error method.
(Java Scanner Interface): Move note on %pure-parser to Java Bison
Interface. Describe %code lexer and yylex accutately.
Remove documentation that does not match the code.
(Java Action Features): New.
(Java Differences): Add reference. Add item on semantic values.
Add note about @{ ... @}. Clarify %% epilogue placement.
(Java Declarations Summary): New.
Fix Java skeleton.
* data/java.m4 (b4_prefix): Correct quoting for m4_define_default.
(b4_remove_comma): Quote test argument.
* tests/java.at (Java parser class and package names): New test.
(Java %parse-param and %lex-param): New test.
(Java stype, position_class and location_class): New test.
2008-10-27 Di-an Jan <dianj@freeshell.org>
* data/lalr1.jave: Update copyright years.
(YYParser): Correct name of "generated from" file in Javadoc:
use b4_file_name instead of @ofile@.
(Location constructor): Correct Javadoc parameter name.
(yylloc): Add missing opening m4 quote after b4_location_if.
This removes a stray [ in the Javadoc of Lexer.getStartPos.
(Lexer.yyerror): Fix incorrect m4 and Javadoc.
(YYParser constructor): Correct Javadoc parameter name.
* data/lalr1.cc (debug_stream, set_debug_stream)
(debug_level_type, debug_level, set_debug_level): Don't
declare them when YYDEBUG is not defined.
The implementation are already YYDEBUG-aware.
2008-10-08 Di-an Jan <dianj@freeshell.org> (tiny change)
Small but important bugfixes for the Java skeleton.
* data/lalr1.java (yyerror): Change Location to b4_location_type.
(yy_symbol_print): Call toString on yyvaluep.
* data/m4sugar/m4sugar.m4: Copy entire file from autoconf.
* data/m4sugar/foreach.m4: New file, copied from autoconf.
* data/Makefile.am (dist_m4sugar_DATA): Distribute it.
* src/output.c (output_skeleton): Tell m4 how to find it.
Signed-off-by: Eric Blake <ebb9@byu.net>
* data/m4sugar/m4sugar.m4 (m4_fst): Delete.
(m4_map, m4_map_sep, _m4_map): Rewrite more efficiently.
(m4_apply, _m4_apply, m4_mapall, m4_mapall_sep): New macros.
* data/java.m4 (b4_token_enums): Use more efficient short-circuit
for empty list.
* data/c.m4 (b4_token_defines, b4_token_enums, b4_c_ansi_formals):
Likewise.
(b4_parse_param_for): Avoid m4_fst, now that autoconf no longer
declares it.
Signed-off-by: Eric Blake <ebb9@byu.net>
* data/m4sugar/m4sugar.m4 (m4_unquote, m4_wrap_lifo): New macros.
(m4_wrap): Guarantee FIFO order, in spite of m4 1.6.
(m4_init): Consolidate wrapped text into single m4_wrap.
* data/bison.m4 (b4_check_user_names_wrap): Stick with LIFO order
in wrapped text.
Signed-off-by: Eric Blake <ebb9@byu.net>
* data/m4sugar/m4sugar.m4 (changeword): Nuke.
(m4_prepend): Remove, as it is unused and inherently quadratic,
whereas m4_append is linear in newer m4.
(m4_mkstemp): New builtin.
(m4_symbols): Make rename conditional.
(m4_version_prereq): Ensure fatal error if used in bison, which
intentionally lacks version.m4.
Signed-off-by: Eric Blake <ebb9@byu.net>
* ChangeLog: For the sake of people still using CVS, don't use dollars
when mentioning Id.
* data/xslt/bison.xsl: Remove Id from header comments, where it was
unusual anyway.
* data/xslt/xml2dot.xsl: Likewise.
* data/xslt/xml2text.xsl: Likewise.
* data/xslt/xml2xhtml.xsl: Likewise.
* doc/Doxyfile.in (PROJECT_NUMBER): Don't use ID.
* doc/Makefile.am (neutralize): Remove, no longer needed.
(.x.1): Don't use neutralize.
(edit): Don't substitute for ID.
(Doxyfile): Don't define Id, and thus don't depend on ChangeLog.
<http://lists.gnu.org/archive/html/help-bison/2008-02/msg00023.html>.
* data/yacc.c: Clean up whitespace in the output a little.
(yypstate_allocated): Define for impure push parsers regardless of
whether the pull interface is also requested.
* tests/push.at (Push Parsing: Multiple impure instances): Extend to
check impure push parsers without the pull interface.
* data/yacc.c (yypstate_new): Don't try to invoke yyerror since
yyerror takes arguments specified by %parse-param while yypstate_new
does not.
* doc/bison.texinfo (Parser Create Function): Document that
yypstate_new returns 0 for multiple impure parser instances.
* tests/push.at (Push Parsing: Multiple impure instances): Update
expected stderr output.
