yacc [-dltv] [-b file_prefix] [-p sym_prefix] grammar
The form and meaning of the grammar are described in the EXTENDED DESCRIPTION section.
The C source code and header file shall be produced in a form suitable as input for the C compiler (see c99).
The following options shall be supported:
The LANG and LC_* variables affect the execution of the yacc utility as stated. The main() function defined in Yacc Library shall call:
setlocale(LC_ALL, "")
and thus the program generated by yacc shall also be affected by the contents of these variables at runtime.
Standard error shall also be used for diagnostic messages.
This file shall contain the C source code for the yyparse() function. It shall contain code for the various semantic actions with macro substitution performed on them as described in the EXTENDED DESCRIPTION section. It also shall contain a copy of the #define statements in the header file. If a %union declaration is used, the declaration for YYSTYPE shall also be included in this file.
The header file shall contain #define statements that associate the token numbers with the token names. This allows source files other than the code file to access the token codes. If a %union declaration is used, the declaration for YYSTYPE and an extern YYSTYPE yylval declaration shall also be included in this file.
The description file shall be a text file containing a description of the state machine corresponding to the parser, using an unspecified format. Limits for internal tables (see Limits) shall also be reported, in an implementation-defined manner. (Some implementations may use dynamic allocation techniques and have no specific limit values to report.)
The input grammar includes rules describing the input structure of the target language and code to be invoked when these rules are recognized to provide the associated semantic action. The code to be executed shall appear as bodies of text that are intended to be C-language code. These bodies of text shall not contain C-language trigraphs. The C-language inclusions are presumed to form a correct function when processed by yacc into its output files. The code included in this way shall be executed during the recognition of the target language.
Given a grammar, the yacc utility generates the files described in the OUTPUT FILES section. The code file can be compiled and linked using c99. If the declaration and programs sections of the grammar file did not include definitions of main(), yylex(), and yyerror(), the compiled output requires linking with externally supplied versions of those functions. Default versions of main() and yyerror() are supplied in the yacc library and can be linked in by using the -l y operand to c99. The yacc library interfaces need not support interfaces with other than the default yy symbol prefix. The application provides the lexical analyzer function, yylex(); the lex utility is specifically designed to generate such a routine.
The application shall ensure that every specification file consists of three sections in order: declarations, grammar rules, and programs, separated by double <percent-sign> characters ("%%"). The declarations and programs sections can be empty. If the latter is empty, the preceding "%%" mark separating it from the rules section can be omitted.
The input is free form text following the structure of the grammar defined below.
The <blank>, <newline>, and <form-feed> character shall be ignored, except that the application shall ensure that they do not appear in names or multi-character reserved symbols. Comments shall be enclosed in "/* ... */", and can appear wherever a name is valid.
Names are of arbitrary length, made up of letters, periods ('.'), underscores ('_'), and non-initial digits. Uppercase and lowercase letters are distinct. Conforming applications shall not use names beginning in yy or YY since the yacc parser uses such names. Many of the names appear in the final output of yacc, and thus they should be chosen to conform with any additional rules created by the C compiler to be used. In particular they appear in #define statements.
A literal shall consist of a single character enclosed in single-quote characters. All of the escape sequences supported for character constants by the ISO C standard shall be supported by yacc.
The relationship with the lexical analyzer is discussed in detail below.
The application shall ensure that the NUL character is not used in grammar rules or literals.
The declarations section is used to define the symbols used to define the target language and their relationship with each other. In particular, much of the additional information required to resolve ambiguities in the context-free grammar for the target language is provided here.
Usually yacc assigns the relationship between the symbolic names it generates and their underlying numeric value. The declarations section makes it possible to control the assignment of these values.
It is also possible to keep semantic information associated with the tokens currently on the parse stack in a user-defined C-language union, if the members of the union are associated with the various names in the grammar. The declarations section provides for this as well.
The first group of declarators below all take a list of names as arguments. That list can optionally be preceded by the name of a C union member (called a tag below) appearing within '<' and '>'. (As an exception to the typographical conventions of the rest of this volume of POSIX.1-2008, in this case <tag> does not represent a metavariable, but the literal angle bracket characters surrounding a symbol.) The use of tag specifies that the tokens named on this line shall be of the same C type as the union member referenced by tag. This is discussed in more detail below.
For lists used to define tokens, the first appearance of a given token can be followed by a positive integer (as a string of decimal digits). If this is done, the underlying value assigned to it for lexical purposes shall be taken to be that number.
The following declares name to be a token:
%token [<tag>] name [number] [name [number]]...
If tag is present, the C type for all tokens on this line shall be declared to be the type referenced by tag. If a positive integer, number, follows a name, that value shall be assigned to the token.
The following declares name to be a token, and assigns precedence to it:
%left [<tag>] name [number] [name [number]]... %right [<tag>] name [number] [name [number]]...
One or more lines, each beginning with one of these symbols, can appear in this section. All tokens on the same line have the same precedence level and associativity; the lines are in order of increasing precedence or binding strength. %left denotes that the operators on that line are left associative, and %right similarly denotes right associative operators. If tag is present, it shall declare a C type for names as described for %token.
The following declares name to be a token, and indicates that this cannot be used associatively:
%nonassoc [<tag>] name [number] [name [number]]...
If the parser encounters associative use of this token it reports an error. If tag is present, it shall declare a C type for names as described for %token.
The following declares that union member names are non-terminals, and thus it is required to have a tag field at its beginning:
%type <tag> name...
Because it deals with non-terminals only, assigning a token number or using a literal is also prohibited. If this construct is present, yacc shall perform type checking; if this construct is not present, the parse stack shall hold only the int type.
Every name used in grammar not defined by a %token, %left, %right, or %nonassoc declaration is assumed to represent a non-terminal symbol. The yacc utility shall report an error for any non-terminal symbol that does not appear on the left side of at least one grammar rule.
Once the type, precedence, or token number of a name is specified, it shall not be changed. If the first declaration of a token does not assign a token number, yacc shall assign a token number. Once this assignment is made, the token number shall not be changed by explicit assignment.
The following declarators do not follow the previous pattern.
The following declares the non-terminal name to be the start symbol, which represents the largest, most general structure described by the grammar rules:
%start name
By default, it is the left-hand side of the first grammar rule; this default can be overridden with this declaration.
The following declares the yacc value stack to be a union of the various types of values desired.
%union { body of union (in C) }
The body of the union shall not contain unbalanced curly brace preprocessing tokens.
By default, the values returned by actions (see below) and the lexical analyzer shall be of type int. The yacc utility keeps track of types, and it shall insert corresponding union member names in order to perform strict type checking of the resulting parser.
Alternatively, given that at least one <tag> construct is used, the union can be declared in a header file (which shall be included in the declarations section by using a #include construct within %{ and %}), and a typedef used to define the symbol YYSTYPE to represent this union. The effect of %union is to provide the declaration of YYSTYPE directly from the yacc input.
C-language declarations and definitions can appear in the declarations section, enclosed by the following marks:
%{ ... %}
These statements shall be copied into the code file, and have global scope within it so that they can be used in the rules and program sections. The statements shall not contain "%}" outside a comment, string literal, or multi-character constant.
The application shall ensure that the declarations section is terminated by the token %%.
The rules section defines the context-free grammar to be accepted by the function yacc generates, and associates with those rules C-language actions and additional precedence information. The grammar is described below, and a formal definition follows.
The rules section is comprised of one or more grammar rules. A grammar rule has the form:
A : BODY ;
The symbol A represents a non-terminal name, and BODY represents a sequence of zero or more names, literals, and semantic actions that can then be followed by optional precedence rules. Only the names and literals participate in the formation of the grammar; the semantic actions and precedence rules are used in other ways. The <colon> and the <semicolon> are yacc punctuation. If there are several successive grammar rules with the same left-hand side, the <vertical-line> ('|') can be used to avoid rewriting the left-hand side; in this case the <semicolon> appears only after the last rule. The BODY part can be empty (or empty of names and literals) to indicate that the non-terminal symbol matches the empty string.
The yacc utility assigns a unique number to each rule. Rules using the vertical bar notation are distinct rules. The number assigned to the rule appears in the description file.
The elements comprising a BODY are:
These actions can return values and can obtain the values returned by previous actions. These values are kept in objects of type YYSTYPE (see %union). The result value of the action shall be kept on the parse stack with the left-hand side of the rule, to be accessed by other reductions as part of their right-hand side. By using the <tag> information provided in the declarations section, the code generated by yacc can be strictly type checked and contain arbitrary information. In addition, the lexical analyzer can provide the same kinds of values for tokens, if desired.
An action is an arbitrary C statement and as such can do input or output, call subprograms, and alter external variables. An action is one or more C statements enclosed in curly braces '{' and '}'. The statements shall not contain unbalanced curly brace preprocessing tokens.
Certain pseudo-variables can be used in the action. These are macros for access to data structures known internally to yacc.
Actions can occur anywhere in a rule (not just at the end); an action can access values returned by actions to its left, and in turn the value it returns can be accessed by actions to its right. An action appearing in the middle of a rule shall be equivalent to replacing the action with a new non-terminal symbol and adding an empty rule with that non-terminal symbol on the left-hand side. The semantic action associated with the new rule shall be equivalent to the original action. The use of actions within rules might introduce conflicts that would not otherwise exist.
By default, the value of a rule shall be the value of the first element in it. If the first element does not have a type (particularly in the case of a literal) and type checking is turned on by %type, an error message shall result.
If a program section follows, the application shall ensure that the grammar rules are terminated by %%.
The programs section can include the definition of the lexical analyzer yylex(), and any other functions; for example, those used in the actions specified in the grammar rules. It is unspecified whether the programs section precedes or follows the semantic actions in the output file; therefore, if the application contains any macro definitions and declarations intended to apply to the code in the semantic actions, it shall place them within "%{ ... %}" in the declarations section.
The following input to yacc yields a parser for the input to yacc. This formal syntax takes precedence over the preceding text syntax description.
The lexical structure is defined less precisely; Lexical Structure of the Grammar defines most terms. The correspondence between the previous terms and the tokens below is as follows.
/* Grammar for the input to yacc. */
/* Basic entries. */
/* The following are recognized by the lexical analyzer. */
%token IDENTIFIER /* Includes identifiers and literals */
%token C_IDENTIFIER /* identifier (but not literal)
followed by a :. */
%token NUMBER /* [0-9][0-9]* */
/* Reserved words : %type=>TYPE %left=>LEFT, and so on */
%token LEFT RIGHT NONASSOC TOKEN PREC TYPE START UNION
%token MARK /* The %% mark. */
%token LCURL /* The %{ mark. */
%token RCURL /* The %} mark. */
/* 8-bit character literals stand for themselves; */
/* tokens have to be defined for multi-byte characters. */
%start spec
%%
spec : defs MARK rules tail
;
tail : MARK
{
/* In this action, set up the rest of the file. */
}
| /* Empty; the second MARK is optional. */
;
defs : /* Empty. */
| defs def
;
def : START IDENTIFIER
| UNION
{
/* Copy union definition to output. */
}
| LCURL
{
/* Copy C code to output file. */
}
RCURL
| rword tag nlist
;
rword : TOKEN
| LEFT
| RIGHT
| NONASSOC
| TYPE
;
tag : /* Empty: union tag ID optional. */
| '<' IDENTIFIER '>'
;
nlist : nmno
| nlist nmno
;
nmno : IDENTIFIER /* Note: literal invalid with % type. */
| IDENTIFIER NUMBER /* Note: invalid with % type. */
;
/* Rule section */
rules : C_IDENTIFIER rbody prec
| rules rule
;
rule : C_IDENTIFIER rbody prec
| '|' rbody prec
;
rbody : /* empty */
| rbody IDENTIFIER
| rbody act
;
act : '{'
{
/* Copy action, translate $$, and so on. */
}
'}'
;
prec : /* Empty */
| PREC IDENTIFIER
| PREC IDENTIFIER act
| prec ';'
;
The parser produced for an input grammar may contain states in which conflicts occur. The conflicts occur because the grammar is not LALR(1). An ambiguous grammar always contains at least one LALR(1) conflict. The yacc utility shall resolve all conflicts, using either default rules or user-specified precedence rules.
Conflicts are either shift/reduce conflicts or reduce/reduce conflicts. A shift/reduce conflict is where, for a given state and lookahead symbol, both a shift action and a reduce action are possible. A reduce/reduce conflict is where, for a given state and lookahead symbol, reductions by two different rules are possible.
The rules below describe how to specify what actions to take when a conflict occurs. Not all shift/reduce conflicts can be successfully resolved this way because the conflict may be due to something other than ambiguity, so incautious use of these facilities can cause the language accepted by the parser to be much different from that which was intended. The description file shall contain sufficient information to understand the cause of the conflict. Where ambiguity is the reason either the default or explicit rules should be adequate to produce a working parser.
The declared precedences and associativities (see Declarations Section) are used to resolve parsing conflicts as follows:
Conflicts resolved by precedence or associativity shall not be counted in the shift/reduce and reduce/reduce conflicts reported by yacc on either standard error or in the description file.
The token error shall be reserved for error handling. The name error can be used in grammar rules. It indicates places where the parser can recover from a syntax error. The default value of error shall be 256. Its value can be changed using a %token declaration. The lexical analyzer should not return the value of error.
The parser shall detect a syntax error when it is in a state where the action associated with the lookahead symbol is error. A semantic action can cause the parser to initiate error handling by executing the macro YYERROR. When YYERROR is executed, the semantic action passes control back to the parser. YYERROR cannot be used outside of semantic actions.
When the parser detects a syntax error, it normally calls yyerror() with the character string "syntax error" as its argument. The call shall not be made if the parser is still recovering from a previous error when the error is detected. The parser is considered to be recovering from a previous error until the parser has shifted over at least three normal input symbols since the last error was detected or a semantic action has executed the macro yyerrok. The parser shall not call yyerror() when YYERROR is executed.
The macro function YYRECOVERING shall return 1 if a syntax error has been detected and the parser has not yet fully recovered from it. Otherwise, zero shall be returned.
When a syntax error is detected by the parser, the parser shall check if a previous syntax error has been detected. If a previous error was detected, and if no normal input symbols have been shifted since the preceding error was detected, the parser checks if the lookahead symbol is an endmarker (see Interface to the Lexical Analyzer). If it is, the parser shall return with a non-zero value. Otherwise, the lookahead symbol shall be discarded and normal parsing shall resume.
When YYERROR is executed or when the parser detects a syntax error and no previous error has been detected, or at least one normal input symbol has been shifted since the previous error was detected, the parser shall pop back one state at a time until the parse stack is empty or the current state allows a shift over error. If the parser empties the parse stack, it shall return with a non-zero value. Otherwise, it shall shift over error and then resume normal parsing. If the parser reads a lookahead symbol before the error was detected, that symbol shall still be the lookahead symbol when parsing is resumed.
The macro yyerrok in a semantic action shall cause the parser to act as if it has fully recovered from any previous errors. The macro yyclearin shall cause the parser to discard the current lookahead token. If the current lookahead token has not yet been read, yyclearin shall have no effect.
The macro YYACCEPT shall cause the parser to return with the value zero. The macro YYABORT shall cause the parser to return with a non-zero value.
The yylex() function is an integer-valued function that returns a token number representing the kind of token read. If there is a value associated with the token returned by yylex() (see the discussion of tag above), it shall be assigned to the external variable yylval.
If the parser and yylex() do not agree on these token numbers, reliable communication between them cannot occur. For (single-byte character) literals, the token is simply the numeric value of the character in the current character set. The numbers for other tokens can either be chosen by yacc, or chosen by the user. In either case, the #define construct of C is used to allow yylex() to return these numbers symbolically. The #define statements are put into the code file, and the header file if that file is requested. The set of characters permitted by yacc in an identifier is larger than that permitted by C. Token names found to contain such characters shall not be included in the #define declarations.
If the token numbers are chosen by yacc, the tokens other than literals shall be assigned numbers greater than 256, although no order is implied. A token can be explicitly assigned a number by following its first appearance in the declarations section with a number. Names and literals not defined this way retain their default definition. All token numbers assigned by yacc shall be unique and distinct from the token numbers used for literals and user-assigned tokens. If duplicate token numbers cause conflicts in parser generation, yacc shall report an error; otherwise, it is unspecified whether the token assignment is accepted or an error is reported.
The end of the input is marked by a special token called the endmarker, which has a token number that is zero or negative. (These values are invalid for any other token.) All lexical analyzers shall return zero or negative as a token number upon reaching the end of their input. If the tokens up to, but excluding, the endmarker form a structure that matches the start symbol, the parser shall accept the input. If the endmarker is seen in any other context, it shall be considered an error.
In addition to yyparse() and yylex(), the functions yyerror() and main() are required to make a complete program. The application can supply main() and yyerror(), or those routines can be obtained from the yacc library.
The following functions shall appear only in the yacc library accessible through the -l y operand to c99; they can therefore be redefined by a conforming application:
The order of the -l y and -l l operands given to c99 is significant; the application shall either provide its own main() function or ensure that -l y precedes -l l.
The parser generated by yacc shall have diagnostic facilities in it that can be optionally enabled at either compile time or at runtime (if enabled at compile time). The compilation of the runtime debugging code is under the control of YYDEBUG, a preprocessor symbol. If YYDEBUG has a non-zero value, the debugging code shall be included. If its value is zero, the code shall not be included.
In parsers where the debugging code has been included, the external int yydebug can be used to turn debugging on (with a non-zero value) and off (zero value) at runtime. The initial value of yydebug shall be zero.
When -t is specified, the code file shall be built such that, if YYDEBUG is not already defined at compilation time (using the c99 -D YYDEBUG option, for example), YYDEBUG shall be set explicitly to 1. When -t is not specified, the code file shall be built such that, if YYDEBUG is not already defined, it shall be set explicitly to zero.
The format of the debugging output is unspecified but includes at least enough information to determine the shift and reduce actions, and the input symbols. It also provides information about error recovery.
The parser constructed by yacc implements an LALR(1) parsing algorithm as documented in the literature. It is unspecified whether the parser is table-driven or direct-coded.
A parser generated by yacc shall never request an input symbol from yylex() while in a state where the only actions other than the error action are reductions by a single rule.
The literature of parsing theory defines these concepts.
The
yacc
utility may have several internal tables. The minimum maximums for
these tables are shown in the following table. The exact meaning of
these values is implementation-defined. The implementation shall
define the relationship between these values and between them and any
error messages that the implementation may generate should it run out
of space for any internal structure. An implementation may combine
groups of these resources into a single pool as long as the total
available to the user does not fall below the sum of the sizes
specified by this section.
| |||||||||||||||||||||||||||
The following sections are informative.
The description of the -p option specifies the minimal set of function and variable names that cause conflict when multiple parsers are linked together. YYSTYPE does not need to be changed. Instead, the programmer can use -b to give the header files for different parsers different names, and then the file with the yylex() for a given parser can include the header for that parser. Names such as yyclearerr do not need to be changed because they are used only in the actions; they do not have linkage. It is possible that an implementation has other names, either internal ones for implementing things such as yyclearerr, or providing non-standard features that it wants to change with -p.
Unary operators that are the same token as a binary operator in general need their precedence adjusted. This is handled by the %prec advisory symbol associated with the particular grammar rule defining that unary operator. (See Grammar Rules in yacc.) Applications are not required to use this operator for unary operators, but the grammars that do not require it are rare.
c99 y.tab.c -l y
Both the lex library and the yacc library contain main(). To access the yacc main():
c99 y.tab.c lex.yy.c -l y -l l
This ensures that the yacc library is searched first, so that its main() is used.
The historical yacc libraries have contained two simple functions that are normally coded by the application programmer. These functions are similar to the following code:
#include <locale.h> int main(void) { extern int yyparse(); setlocale(LC_ALL, ""); /* If the following parser is one created by lex, the application must be careful to ensure that LC_CTYPE and LC_COLLATE are set to the POSIX locale. */ (void) yyparse(); return (0); } #include <stdio.h> int yyerror(const char *msg) { (void) fprintf(stderr, "%s\n", msg); return (0); }
There has been confusion between the class of grammars, the algorithms needed to generate parsers, and the algorithms needed to parse the languages. They are all reasonably orthogonal. In particular, a parser generator that accepts the full range of LR(1) grammars need not generate a table any more complex than one that accepts SLR(1) (a relatively weak class of LR grammars) for a grammar that happens to be SLR(1). Such an implementation need not recognize the case, either; table compression can yield the SLR(1) table (or one even smaller than that) without recognizing that the grammar is SLR(1). The speed of an LR(1) parser for any class is dependent more upon the table representation and compression (or the code generation if a direct parser is generated) than upon the class of grammar that the table generator handles.
The speed of the parser generator is somewhat dependent upon the class of grammar it handles. However, the original Knuth article algorithms for constructing LR parsers were judged by its author to be impractically slow at that time. Although full LR is more complex than LALR(1), as computer speeds and algorithms improve, the difference (in terms of acceptable wall-clock execution time) is becoming less significant.
Potential authors are cautioned that the referenced DeRemer and Pennello article previously cited identifies a bug (an over-simplification of the computation of LALR(1) lookahead sets) in some of the LALR(1) algorithm statements that preceded it to publication. They should take the time to seek out that paper, as well as current relevant work, particularly Aho's.
The -b option was added to provide a portable method for permitting yacc to work on multiple separate parsers in the same directory. If a directory contains more than one yacc grammar, and both grammars are constructed at the same time (by, for example, a parallel make program), conflict results. While the solution is not historical practice, it corrects a known deficiency in historical implementations. Corresponding changes were made to all sections that referenced the filenames y.tab.c (now ``the code file''), y.tab.h (now ``the header file''), and y.output (now ``the description file'').
The grammar for yacc input is based on System V documentation. The textual description shows there that the ';' is required at the end of the rule. The grammar and the implementation do not require this. (The use of C_IDENTIFIER causes a reduce to occur in the right place.)
Also, in that implementation, the constructs such as %token can be terminated by a <semicolon>, but this is not permitted by the grammar. The keywords such as %token can also appear in uppercase, which is again not discussed. In most places where '%' is used, <backslash> can be substituted, and there are alternate spellings for some of the symbols (for example, %LEFT can be "%<" or even "\<").
Historically, <tag> can contain any characters except '>', including white space, in the implementation. However, since the tag must reference an ISO C standard union member, in practice conforming implementations need to support only the set of characters for ISO C standard identifiers in this context.
Some historical implementations are known to accept actions that are terminated by a period. Historical implementations often allow '$' in names. A conforming implementation does not need to support either of these behaviors.
Deciding when to use %prec illustrates the difficulty in specifying the behavior of yacc. There may be situations in which the grammar is not, strictly speaking, in error, and yet yacc cannot interpret it unambiguously. The resolution of ambiguities in the grammar can in many instances be resolved by providing additional information, such as using %type or %union declarations. It is often easier and it usually yields a smaller parser to take this alternative when it is appropriate.
The size and execution time of a program produced without the runtime debugging code is usually smaller and slightly faster in historical implementations.
Statistics messages from several historical implementations include the following types of information:
n/512 terminals, n/300 non-terminals n/600 grammar rules, n/1500 states n shift/reduce, n reduce/reduce conflicts reported n/350 working sets used Memory: states, etc. n/15000, parser n/15000 n/600 distinct lookahead sets n extra closures n shift entries, n exceptions n goto entries n entries saved by goto default Optimizer space used: input n/15000, output n/15000 n table entries, n zero Maximum spread: n, Maximum offset: n
The report of internal tables in the description file is left implementation-defined because all aspects of these limits are also implementation-defined. Some implementations may use dynamic allocation techniques and have no specific limit values to report.
The format of the y.output file is not given because specification of the format was not seen to enhance applications portability. The listing is primarily intended to help human users understand and debug the parser; use of y.output by a conforming application script would be unusual. Furthermore, implementations have not produced consistent output and no popular format was apparent. The format selected by the implementation should be human-readable, in addition to the requirement that it be a text file.
Standard error reports are not specifically described because they are seldom of use to conforming applications and there was no reason to restrict implementations.
Some implementations recognize "={" as equivalent to '{' because it appears in historical documentation. This construction was recognized and documented as obsolete as long ago as 1978, in the referenced Yacc: Yet Another Compiler-Compiler. This volume of POSIX.1-2008 chose to leave it as obsolete and omit it.
Multi-byte characters should be recognized by the lexical analyzer and returned as tokens. They should not be returned as multi-byte character literals. The token error that is used for error recovery is normally assigned the value 256 in the historical implementation. Thus, the token value 256, which is used in many multi-byte character sets, is not available for use as the value of a user-defined token.
The Base Definitions volume of POSIX.1-2008, Chapter 8, Environment Variables, Section 12.2, Utility Syntax Guidelines
Any typographical or formatting errors that appear in this page are most likely to have been introduced during the conversion of the source files to man page format. To report such errors, see https://www.kernel.org/doc/man-pages/reporting_bugs.html .