X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/56ca3d8fce0c9db2f2829d7a7a07a9ad06cdf6a8..db8ab2be3337ef87629b761078ebc46a01e7f876:/NEWS diff --git a/NEWS b/NEWS index ae8b297c..a20898fc 100644 --- a/NEWS +++ b/NEWS @@ -2,45 +2,271 @@ GNU Bison NEWS * Noteworthy changes in release ?.? (????-??-??) [?] -** Future changes: +** Bug fixes + + Bugs and portability issues in the test suite have been fixed. + + Some errors in translations have been addressed, and --help now directs + users to the appropriate place to report them. + + Stray Info files shipped by accident are removed. + + Incorrect definitions of YY_, issued by yacc.c when no parser header is + generated, are removed. + + All the generated headers are self-contained. + +** Changes in the format of error messages + + This used to be the format of many error reports: + + foo.y:5.10-24: result type clash on merge function 'merge': != + foo.y:4.13-27: previous declaration + + It is now: + + foo.y:5.10-25: result type clash on merge function 'merge': != + foo.y:4.13-27: previous declaration + +** Header guards (yacc.c, glr.c, glr.cc) + + In order to avoid collisions, the header guards are now + YY___INCLUDED, instead of merely _. + For instance the header generated from + + %define api.prefix "calc" + %defines "lib/parse.h" + + will use YY_CALC_LIB_PARSE_H_INCLUDED as guard. + +** Exception safety (lalr1.cc) + + The parse function now catches exceptions, uses the %destructors to + release memory (the lookahead symbol and the symbols pushed on the stack) + before rethrowing the exception. + + This feature is somewhat experimental. User feedback would be + appreciated. + +** Fix compiler warnings in the generated parser (yacc.c, glr.c) + + The compilation of pure parsers (%define api.pure) can trigger GCC + warnings such as: + + input.c: In function 'yyparse': + input.c:1503:12: warning: 'yylval' may be used uninitialized in this + function [-Wmaybe-uninitialized] + *++yyvsp = yylval; + ^ + + This is now fixed; pragmas to avoid these warnings are no longer needed. + + Warnings from clang ("equality comparison with extraneous parentheses" and + "function declared 'noreturn' should not return") have also been + addressed. + +** New %define variable: api.location.type (glr.cc, lalr1.cc) + + The %define variable api.location.type defines the name of the type to use + for locations. When defined, Bison no longer generates the position.hh + and location.hh files, nor does the parser will include them: the user is + then responsible to define her type. + + This can be used in programs with several parsers to factor their location + and position files: let one of them generate them, and let the others + simply resue these types and files. + + This feature was actually introduced, but not documented, in Bison 2.5, + under the name "location_type" (which is maintained for backward + compatibility). + +* Noteworthy changes in release 2.6.2 (2012-08-03) [stable] + +** Bug fixes + + Buffer overruns, complaints from Flex, and portability issues in the test + suite have been fixed. + +** Spaces in %lex- and %parse-param (lalr1.cc, glr.cc) + + Trailing end-of-lines in %parse-param or %lex-param would result in + invalid C++. This is fixed. + +** Spurious spaces and end-of-lines + + The generated files no longer end (nor start) with empty lines. + +* Noteworthy changes in release 2.6.1 (2012-07-30) [stable] + + Bison no longer executes user-specified M4 code when processing a grammar. + +** Future Changes + + In addition to the removal of the features announced in Bison 2.6, the + next major release will remove the "Temporary hack for adding a semicolon + to the user action", as announced in the release 2.5. Instead of: + + exp: exp "+" exp { $$ = $1 + $3 }; + + write: + + exp: exp "+" exp { $$ = $1 + $3; }; + +** Bug fixes + +*** Type names are now properly escaped. + +*** glr.cc: set_debug_level and debug_level work as expected. + +*** Stray @ or $ in actions + + While Bison used to warn about stray $ or @ in action rules, it did not + for other actions such as printers, destructors, or initial actions. It + now does. + +** Type names in actions + + For consistency with rule actions, it is now possible to qualify $$ by a + type-name in destructors, printers, and initial actions. For instance: + + %printer { fprintf (yyo, "(%d, %f)", $$, $$); } <*> <>; + + will display two values for each typed and untyped symbol (provided + that YYSTYPE has both "ival" and "fval" fields). - The next major release will drop support for generating parsers in K&R C, - and remove the definition of yystype (removal announced since Bison - 1.875). +* Noteworthy changes in release 2.6 (2012-07-19) [stable] -** The generated header is included (yacc.c) +** Future Changes + + The next major release of Bison will drop support for the following + deprecated features. Please report disagreements to bug-bison@gnu.org. + +*** K&R C parsers + + Support for generating parsers in K&R C will be removed. Parsers + generated for C support ISO C90, and are tested with ISO C99 and ISO C11 + compilers. + +*** Features deprecated since Bison 1.875 + + The definitions of yystype and yyltype will be removed; use YYSTYPE and + YYLTYPE. + + YYPARSE_PARAM and YYLEX_PARAM, deprecated in favor of %parse-param and + %lex-param, will no longer be supported. + + Support for the preprocessor symbol YYERROR_VERBOSE will be removed, use + %error-verbose. + +*** The generated header will be included (yacc.c) Instead of duplicating the content of the generated header (definition of - YYSTYPE, yyltype etc.), the generated parser now includes it, as was - already the case for GLR or C++ parsers. + YYSTYPE, yyparse declaration etc.), the generated parser will include it, + as is already the case for GLR or C++ parsers. This change is deferred + because existing versions of ylwrap (e.g., Automake 1.12.1) do not support + it. -** Headers (yacc.c, glr.c, glr.cc) +** Generated Parser Headers -*** Guards +*** Guards (yacc.c, glr.c, glr.cc) The generated headers are now guarded, as is already the case for C++ - parsers (lalr1.cc). For intance, with --defines=foo.h: + parsers (lalr1.cc). For instance, with --defines=foo.h: - #ifndef YY_FOO_H - # define YY_FOO_H - ... - #endif /* !YY_FOO_H */ + #ifndef YY_FOO_H + # define YY_FOO_H + ... + #endif /* !YY_FOO_H */ -*** New declarations +*** New declarations (yacc.c, glr.c) The generated header now declares yydebug and yyparse. Both honor --name-prefix=bar_, and yield - int bar_parse (void); + int bar_parse (void); rather than - #define yyparse bar_parse - int yyparse (void); + #define yyparse bar_parse + int yyparse (void); in order to facilitate the inclusion of several parser headers inside a single compilation unit. +*** Exported symbols in C++ + + The symbols YYTOKEN_TABLE and YYERROR_VERBOSE, which were defined in the + header, are removed, as they prevent the possibility of including several + generated headers from a single compilation unit. + +*** YYLSP_NEEDED + + For the same reasons, the undocumented and unused macro YYLSP_NEEDED is no + longer defined. + +** New %define variable: api.prefix + + Now that the generated headers are more complete and properly protected + against multiple inclusions, constant names, such as YYSTYPE are a + problem. While yyparse and others are properly renamed by %name-prefix, + YYSTYPE, YYDEBUG and others have never been affected by it. Because it + would introduce backward compatibility issues in projects not expecting + YYSTYPE to be renamed, instead of changing the behavior of %name-prefix, + it is deprecated in favor of a new %define variable: api.prefix. + + The following examples compares both: + + %name-prefix "bar_" | %define api.prefix "bar_" + %token FOO %token FOO + %union { int ival; } %union { int ival; } + %% %% + exp: 'a'; exp: 'a'; + + bison generates: + + #ifndef BAR_FOO_H #ifndef BAR_FOO_H + # define BAR_FOO_H # define BAR_FOO_H + + /* Enabling traces. */ /* Enabling traces. */ + # ifndef YYDEBUG | # ifndef BAR_DEBUG + > # if defined YYDEBUG + > # if YYDEBUG + > # define BAR_DEBUG 1 + > # else + > # define BAR_DEBUG 0 + > # endif + > # else + # define YYDEBUG 0 | # define BAR_DEBUG 0 + > # endif + # endif | # endif + + # if YYDEBUG | # if BAR_DEBUG + extern int bar_debug; extern int bar_debug; + # endif # endif + + /* Tokens. */ /* Tokens. */ + # ifndef YYTOKENTYPE | # ifndef BAR_TOKENTYPE + # define YYTOKENTYPE | # define BAR_TOKENTYPE + enum yytokentype { | enum bar_tokentype { + FOO = 258 FOO = 258 + }; }; + # endif # endif + + #if ! defined YYSTYPE \ | #if ! defined BAR_STYPE \ + && ! defined YYSTYPE_IS_DECLARED | && ! defined BAR_STYPE_IS_DECLARED + typedef union YYSTYPE | typedef union BAR_STYPE + { { + int ival; int ival; + } YYSTYPE; | } BAR_STYPE; + # define YYSTYPE_IS_DECLARED 1 | # define BAR_STYPE_IS_DECLARED 1 + #endif #endif + + extern YYSTYPE bar_lval; | extern BAR_STYPE bar_lval; + + int bar_parse (void); int bar_parse (void); + + #endif /* !BAR_FOO_H */ #endif /* !BAR_FOO_H */ + * Noteworthy changes in release 2.5.1 (2012-06-05) [stable] ** Future changes: @@ -78,10 +304,10 @@ GNU Bison NEWS The header files such as "parser.hh", "location.hh", etc. used a constant name for preprocessor guards, for instance: - #ifndef BISON_LOCATION_HH - # define BISON_LOCATION_HH - ... - #endif // !BISON_LOCATION_HH + #ifndef BISON_LOCATION_HH + # define BISON_LOCATION_HH + ... + #endif // !BISON_LOCATION_HH The inclusion guard is now computed from "PREFIX/FILE-NAME", where lower case characters are converted to upper case, and series of @@ -89,10 +315,10 @@ GNU Bison NEWS With "bison -o lang++/parser.cc", "location.hh" would now include: - #ifndef YY_LANG_LOCATION_HH - # define YY_LANG_LOCATION_HH - ... - #endif // !YY_LANG_LOCATION_HH + #ifndef YY_LANG_LOCATION_HH + # define YY_LANG_LOCATION_HH + ... + #endif // !YY_LANG_LOCATION_HH *** C++ locations: @@ -323,33 +549,33 @@ GNU Bison NEWS to use it. If, for instance, your location structure has "first" and "last" members, instead of - # define YYLLOC_DEFAULT(Current, Rhs, N) \ - do \ - if (N) \ - { \ - (Current).first = (Rhs)[1].location.first; \ - (Current).last = (Rhs)[N].location.last; \ - } \ - else \ - { \ - (Current).first = (Current).last = (Rhs)[0].location.last; \ - } \ - while (false) + # define YYLLOC_DEFAULT(Current, Rhs, N) \ + do \ + if (N) \ + { \ + (Current).first = (Rhs)[1].location.first; \ + (Current).last = (Rhs)[N].location.last; \ + } \ + else \ + { \ + (Current).first = (Current).last = (Rhs)[0].location.last; \ + } \ + while (false) use: - # define YYLLOC_DEFAULT(Current, Rhs, N) \ - do \ - if (N) \ - { \ - (Current).first = YYRHSLOC (Rhs, 1).first; \ - (Current).last = YYRHSLOC (Rhs, N).last; \ - } \ - else \ - { \ - (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \ - } \ - while (false) + # define YYLLOC_DEFAULT(Current, Rhs, N) \ + do \ + if (N) \ + { \ + (Current).first = YYRHSLOC (Rhs, 1).first; \ + (Current).last = YYRHSLOC (Rhs, N).last; \ + } \ + else \ + { \ + (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \ + } \ + while (false) ** YYLLOC_DEFAULT in C++: @@ -968,26 +1194,26 @@ GNU Bison NEWS if the symbols have destructors. For instance: exp: exp "?" exp ":" exp { $1 ? $1 : $3; } - | exp "+" exp - ; + | exp "+" exp + ; will trigger a warning about $$ and $5 in the first rule, and $3 in the second ($1 is copied to $$ by the default rule). This example most likely contains three errors, and could be rewritten as: exp: exp "?" exp ":" exp - { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); } - | exp "+" exp - { $$ = $1 ? $1 : $3; if ($1) free ($3); } - ; + { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); } + | exp "+" exp + { $$ = $1 ? $1 : $3; if ($1) free ($3); } + ; However, if the original actions were really intended, memory leaks and all, the warnings can be suppressed by letting Bison believe the values are used, e.g.: exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); } - | exp "+" exp { $$ = $1; (void) $3; } - ; + | exp "+" exp { $$ = $1; (void) $3; } + ; If there are mid-rule actions, the warning is issued if no action uses it. The following triggers no warning: $1 and $3 are used. @@ -1202,6 +1428,33 @@ GNU Bison NEWS - "parsing stack overflow..." -> "parser stack overflow" GLR parsers now report "parser stack overflow" as per the Bison manual. +** %parse-param and %lex-param + The macros YYPARSE_PARAM and YYLEX_PARAM provide a means to pass + additional context to yyparse and yylex. They suffer from several + shortcomings: + + - a single argument only can be added, + - their types are weak (void *), + - this context is not passed to ancillary functions such as yyerror, + - only yacc.c parsers support them. + + The new %parse-param/%lex-param directives provide a more precise control. + For instance: + + %parse-param {int *nastiness} + %lex-param {int *nastiness} + %parse-param {int *randomness} + + results in the following signatures: + + int yylex (int *nastiness); + int yyparse (int *nastiness, int *randomness); + + or, if both %pure-parser and %locations are used: + + int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness); + int yyparse (int *nastiness, int *randomness); + ** Bison now warns if it detects conflicting outputs to the same file, e.g., it generates a warning for "bison -d -o foo.h foo.y" since that command outputs both code and header to foo.h. @@ -1231,16 +1484,16 @@ GNU Bison NEWS In agreement with POSIX and with other Yaccs, leaving a default action is valid when $$ is untyped, and $1 typed: - untyped: ... typed; + untyped: ... typed; but the converse remains an error: - typed: ... untyped; + typed: ... untyped; ** Values of mid-rule actions The following code: - foo: { ... } { $$ = $1; } ... + foo: { ... } { $$ = $1; } ... was incorrectly rejected: $1 is defined in the second mid-rule action, and is equal to the $$ of the first mid-rule action. @@ -1322,9 +1575,9 @@ GNU Bison NEWS ** Incorrect "Token not used" On a grammar such as - %token useless useful - %% - exp: '0' %prec useful; + %token useless useful + %% + exp: '0' %prec useful; where a token was used to set the precedence of the last rule, bison reported both "useful" and "useless" as useless tokens. @@ -1343,9 +1596,9 @@ GNU Bison NEWS the user symbol is used in the reports, the graphs, and the verbose error messages instead of "$end", which remains being the default. For instance - %token MYEOF 0 + %token MYEOF 0 or - %token MYEOF 0 "end of file" + %token MYEOF 0 "end of file" ** Semantic parser This old option, which has been broken for ages, is removed. @@ -1381,9 +1634,9 @@ GNU Bison NEWS Previous versions don't complain when there is a type clash on the default action if the rule has a mid-rule action, such as in: - %type bar - %% - bar: '0' {} '0'; + %type bar + %% + bar: '0' {} '0'; This is fixed. @@ -1680,9 +1933,11 @@ along with this program. If not, see . LocalWords: struct yystype DJGPP lex param Haible NUM alloca YYSTACK NUL goto LocalWords: YYMAXDEPTH Unescaped UCNs YYLTYPE's yyltype typedefs inline Yaccs LocalWords: Heriyanto Reenable dprec Hilfinger Eggert MYEOF Folle Menezes EOF - LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG + LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG YY LocalWords: namespaces strerror const autoconfiguration Dconst Autoconf's FDL - LocalWords: Automake TMPDIR LESSEQ + LocalWords: Automake TMPDIR LESSEQ ylwrap endif yydebug YYTOKEN YYLSP ival hh + LocalWords: extern YYTOKENTYPE TOKENTYPE yytokentype tokentype STYPE lval pdf + LocalWords: lang yyoutput dvi html ps POSIX lvalp llocp calc yyo fval Local Variables: mode: outline