Future/bison
1
0
Fork 0
mirror of https://git.savannah.gnu.org/git/bison.git synced 2026-03-23 03:03:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

doc: reformat grammar snippets.

Browse Source 
* doc/bison.texinfo: Convert the grammar examples to
use a narrower style.  This helps fitting into the
@smallbook constraints.
http://lists.gnu.org/archive/html/bison-patches/2012-03/msg00011.html
(cherry picked from commit 5e9b6624d7)

Conflicts:

	doc/bison.texinfo
This commit is contained in:
Akim Demaille
2012-03-16 14:26:00 +01:00
parent ea118b7262
commit de6be119ca
1 changed files with 423 additions and 381 deletions
Download Patch File Download Diff File Expand all files Collapse all files

320
doc/bison.texinfo
View File

@@ -637,8 +637,7 @@ the statement; the naked semicolon, and the colon, are Bison punctuation
used in every rule. used in every rule.
@example @example
stmt: RETURN expr ';' stmt: RETURN expr ';' ;
;
@end example @end example
@noindent @noindent
@@ -710,8 +709,7 @@ For example, here is a rule that says an expression can be the sum of
two subexpressions: two subexpressions:
@example @example
expr: expr '+' expr @{ $$ = $1 + $3; @} expr: expr '+' expr @{ $$ = $1 + $3; @} ;
;
@end example @end example
@noindent @noindent
@@ -890,24 +888,26 @@ parses a vastly simplified form of Pascal type declarations.
%% %%
@group @group
type_decl : TYPE ID '=' type ';' type_decl: TYPE ID '=' type ';' ;
;
@end group @end group
@group @group
type : '(' id_list ')' type:
'(' id_list ')'
| expr DOTDOT expr | expr DOTDOT expr
; ;
@end group @end group
@group @group
id_list : ID id_list:
ID
| id_list ',' ID | id_list ',' ID
; ;
@end group @end group
@group @group
expr : '(' expr ')' expr:
'(' expr ')'
| expr '+' expr | expr '+' expr
| expr '-' expr | expr '-' expr
| expr '*' expr | expr '*' expr
@@ -994,27 +994,32 @@ Let's consider an example, vastly simplified from a C++ grammar.
%% %%
prog: prog:
/* Nothing. */
| prog stmt @{ printf ("\n"); @} | prog stmt @{ printf ("\n"); @}
; ;
stmt : expr ';' %dprec 1 stmt:
expr ';' %dprec 1
| decl %dprec 2 | decl %dprec 2
; ;
expr : ID @{ printf ("%s ", $$); @} expr:
ID @{ printf ("%s ", $$); @}
| TYPENAME '(' expr ')' | TYPENAME '(' expr ')'
@{ printf ("%s <cast> ", $1); @} @{ printf ("%s <cast> ", $1); @}
| expr '+' expr @{ printf ("+ "); @} | expr '+' expr @{ printf ("+ "); @}
| expr '=' expr @{ printf ("= "); @} | expr '=' expr @{ printf ("= "); @}
; ;
decl : TYPENAME declarator ';' decl:
TYPENAME declarator ';'
@{ printf ("%s <declare> ", $1); @} @{ printf ("%s <declare> ", $1); @}
| TYPENAME declarator '=' expr ';' | TYPENAME declarator '=' expr ';'
@{ printf ("%s <init-declare> ", $1); @} @{ printf ("%s <init-declare> ", $1); @}
; ;
declarator : ID @{ printf ("\"%s\" ", $1); @} declarator:
ID @{ printf ("\"%s\" ", $1); @}
| '(' declarator ')' | '(' declarator ')'
; ;
@end example @end example
@@ -1085,7 +1090,8 @@ other. To do so, you could change the declaration of @code{stmt} as
follows: follows:
@example @example
stmt : expr ';' %merge <stmtMerge> stmt:
expr ';' %merge <stmtMerge>
| decl %merge <stmtMerge> | decl %merge <stmtMerge>
; ;
@end example @end example
@@ -1494,19 +1500,22 @@ Here are the grammar rules for the reverse polish notation calculator.
@example @example
@group @group
input: /* empty */ input:
/* empty */
| input line | input line
; ;
@end group @end group
@group @group
line: '\n' line:
| exp '\n' @{ printf ("\t%.10g\n", $1); @} '\n'
| exp '\n' @{ printf ("%.10g\n", $1); @}
; ;
@end group @end group
@group @group
exp: NUM @{ $$ = $1; @} exp:
NUM @{ $$ = $1; @}
| exp exp '+' @{ $$ = $1 + $2; @} | exp exp '+' @{ $$ = $1 + $2; @}
| exp exp '-' @{ $$ = $1 - $2; @} | exp exp '-' @{ $$ = $1 - $2; @}
| exp exp '*' @{ $$ = $1 * $2; @} | exp exp '*' @{ $$ = $1 * $2; @}
@@ -1548,7 +1557,8 @@ rule are referred to as @code{$1}, @code{$2}, and so on.
Consider the definition of @code{input}: Consider the definition of @code{input}:
@example @example
input: /* empty */ input:
/* empty */
| input line | input line
; ;
@end example @end example
@@ -1582,8 +1592,9 @@ input tokens; we will arrange for the latter to happen at end-of-input.
Now consider the definition of @code{line}: Now consider the definition of @code{line}:
@example @example
line: '\n' line:
| exp '\n' @{ printf ("\t%.10g\n", $1); @} '\n'
| exp '\n' @{ printf ("%.10g\n", $1); @}
; ;
@end example @end example
@@ -1610,7 +1621,8 @@ The second handles an addition-expression, which looks like two expressions
followed by a plus-sign. The third handles subtraction, and so on. followed by a plus-sign. The third handles subtraction, and so on.
@example @example
exp: NUM exp:
NUM
| exp exp '+' @{ $$ = $1 + $2; @} | exp exp '+' @{ $$ = $1 + $2; @}
| exp exp '-' @{ $$ = $1 - $2; @} | exp exp '-' @{ $$ = $1 - $2; @}
@dots{} @dots{}
@@ -1652,7 +1664,8 @@ exp : NUM | exp exp '+' @{$$ = $1 + $2; @} | @dots{} ;
means the same thing as this: means the same thing as this:
@example @example
exp: NUM exp:
NUM
| exp exp '+' @{ $$ = $1 + $2; @} | exp exp '+' @{ $$ = $1 + $2; @}
| @dots{} | @dots{}
; ;
@@ -1899,19 +1912,22 @@ parentheses nested to arbitrary depth. Here is the Bison code for
%% /* The grammar follows. */ %% /* The grammar follows. */
@group @group
input: /* empty */ input:
/* empty */
| input line | input line
; ;
@end group @end group
@group @group
line: '\n' line:
'\n'
| exp '\n' @{ printf ("\t%.10g\n", $1); @} | exp '\n' @{ printf ("\t%.10g\n", $1); @}
; ;
@end group @end group
@group @group
exp: NUM @{ $$ = $1; @} exp:
NUM @{ $$ = $1; @}
| exp '+' exp @{ $$ = $1 + $3; @} | exp '+' exp @{ $$ = $1 + $3; @}
| exp '-' exp @{ $$ = $1 - $3; @} | exp '-' exp @{ $$ = $1 - $3; @}
| exp '*' exp @{ $$ = $1 * $3; @} | exp '*' exp @{ $$ = $1 * $3; @}
@@ -1981,7 +1997,8 @@ been added to one of the alternatives for @code{line}:
@example @example
@group @group
line: '\n' line:
'\n'
| exp '\n' @{ printf ("\t%.10g\n", $1); @} | exp '\n' @{ printf ("\t%.10g\n", $1); @}
| error '\n' @{ yyerrok; @} | error '\n' @{ yyerrok; @}
; ;
@@ -2074,19 +2091,22 @@ wrong expressions or subexpressions.
@example @example
@group @group
input : /* empty */ input:
/* empty */
| input line | input line
; ;
@end group @end group
@group @group
line : '\n' line:
'\n'
| exp '\n' @{ printf ("%d\n", $1); @} | exp '\n' @{ printf ("%d\n", $1); @}
; ;
@end group @end group
@group @group
exp : NUM @{ $$ = $1; @} exp:
NUM @{ $$ = $1; @}
| exp '+' exp @{ $$ = $1 + $3; @} | exp '+' exp @{ $$ = $1 + $3; @}
| exp '-' exp @{ $$ = $1 - $3; @} | exp '-' exp @{ $$ = $1 - $3; @}
| exp '*' exp @{ $$ = $1 * $3; @} | exp '*' exp @{ $$ = $1 * $3; @}
@@ -2334,7 +2354,8 @@ those which mention @code{VAR} or @code{FNCT}, are new.
@comment file: mfcalc.y @comment file: mfcalc.y
@example @example
@group @group
input: /* empty */ input:
/* empty */
| input line | input line
; ;
@end group @end group
@@ -2342,13 +2363,14 @@ input: /* empty */
@group @group
line: line:
'\n' '\n'
| exp '\n' @{ printf ("\t%.10g\n", $1); @} | exp '\n' @{ printf ("%.10g\n", $1); @}
| error '\n' @{ yyerrok; @} | error '\n' @{ yyerrok; @}
; ;
@end group @end group
@group @group
exp: NUM @{ $$ = $1; @} exp:
NUM @{ $$ = $1; @}
| VAR @{ $$ = $1->value.var; @} | VAR @{ $$ = $1->value.var; @}
| VAR '=' exp @{ $$ = $3; $1->value.var = $3; @} | VAR '=' exp @{ $$ = $3; $1->value.var = $3; @}
| FNCT '(' exp ')' @{ $$ = (*($1->value.fnctptr))($3); @} | FNCT '(' exp ')' @{ $$ = (*($1->value.fnctptr))($3); @}
@@ -3261,8 +3283,7 @@ A Bison grammar rule has the following general form:
@example @example
@group @group
@var{result}: @var{components}@dots{} @var{result}: @var{components}@dots{};
;
@end group @end group
@end example @end example
@@ -3275,8 +3296,7 @@ For example,
@example @example
@group @group
exp: exp '+' exp exp: exp '+' exp;
;
@end group @end group
@end example @end example
@@ -3321,7 +3341,8 @@ be joined with the vertical-bar character @samp{|} as follows:
@example @example
@group @group
@var{result}: @var{rule1-components}@dots{} @var{result}:
@var{rule1-components}@dots{}
| @var{rule2-components}@dots{} | @var{rule2-components}@dots{}
@dots{} @dots{}
; ;
@@ -3337,13 +3358,15 @@ comma-separated sequence of zero or more @code{exp} groupings:
@example @example
@group @group
expseq: /* empty */ expseq:
/* empty */
| expseq1 | expseq1
; ;
@end group @end group
@group @group
expseq1: exp expseq1:
exp
| expseq1 ',' exp | expseq1 ',' exp
; ;
@end group @end group
@@ -3365,7 +3388,8 @@ comma-separated sequence of one or more expressions:
@example @example
@group @group
expseq1: exp expseq1:
exp
| expseq1 ',' exp | expseq1 ',' exp
; ;
@end group @end group
@@ -3380,7 +3404,8 @@ the same construct is defined using @dfn{right recursion}:
@example @example
@group @group
expseq1: exp expseq1:
exp
| exp ',' expseq1 | exp ',' expseq1
; ;
@end group @end group
@@ -3406,13 +3431,15 @@ For example:
@example @example
@group @group
expr: primary expr:
primary
| primary '+' primary | primary '+' primary
; ;
@end group @end group
@group @group
primary: constant primary:
constant
| '(' expr ')' | '(' expr ')'
; ;
@end group @end group
@@ -3536,9 +3563,9 @@ Here is a typical example:
@example @example
@group @group
exp: @dots{} exp:
| exp '+' exp @dots{}
@{ $$ = $1 + $3; @} | exp '+' exp @{ $$ = $1 + $3; @}
@end group @end group
@end example @end example
@@ -3546,9 +3573,9 @@ Or, in terms of named references:
@example @example
@group @group
exp[result]: @dots{} exp[result]:
| exp[left] '+' exp[right] @dots{}
@{ $result = $left + $right; @} | exp[left] '+' exp[right] @{ $result = $left + $right; @}
@end group @end group
@end example @end example
@@ -3595,14 +3622,15 @@ is a case in which you can use this reliably:
@example @example
@group @group
foo: expr bar '+' expr @{ @dots{} @} foo:
expr bar '+' expr @{ @dots{} @}
| expr bar '-' expr @{ @dots{} @} | expr bar '-' expr @{ @dots{} @}
; ;
@end group @end group
@group @group
bar: /* empty */ bar:
@{ previous_expr = $0; @} /* empty */ @{ previous_expr = $0; @}
; ;
@end group @end group
@end example @end example
@@ -3633,9 +3661,9 @@ in the rule. In this example,
@example @example
@group @group
exp: @dots{} exp:
| exp '+' exp @dots{}
@{ $$ = $1 + $3; @} | exp '+' exp @{ $$ = $1 + $3; @}
@end group @end group
@end example @end example
@@ -3702,11 +3730,11 @@ remove it afterward. Here is how it is done:
@example @example
@group @group
stmt: LET '(' var ')' stmt:
@{ $<context>$ = push_context (); LET '(' var ')'
declare_variable ($3); @} @{ $<context>$ = push_context (); declare_variable ($3); @}
stmt @{ $$ = $6; stmt
pop_context ($<context>5); @} @{ $$ = $6; pop_context ($<context>5); @}
@end group @end group
@end example @end example
@@ -3748,15 +3776,19 @@ declare a destructor for that symbol:
%% %%
stmt: let stmt stmt:
@{ $$ = $2; let stmt
pop_context ($1); @} @{
; $$ = $2;
pop_context ($1);
@};
let: LET '(' var ')' let:
@{ $$ = push_context (); LET '(' var ')'
declare_variable ($3); @} @{
; $$ = push_context ();
declare_variable ($3);
@};
@end group @end group
@end example @end example
@@ -3775,7 +3807,8 @@ declaration or not:
@example @example
@group @group
compound: '@{' declarations statements '@}' compound:
'@{' declarations statements '@}'
| '@{' statements '@}' | '@{' statements '@}'
; ;
@end group @end group
@@ -3786,7 +3819,8 @@ But when we add a mid-rule action as follows, the rules become nonfunctional:
@example @example
@group @group
compound: @{ prepare_for_local_variables (); @} compound:
@{ prepare_for_local_variables (); @}
'@{' declarations statements '@}' '@{' declarations statements '@}'
@end group @end group
@group @group
@@ -3808,7 +3842,8 @@ actions into the two rules, like this:
@example @example
@group @group
compound: @{ prepare_for_local_variables (); @} compound:
@{ prepare_for_local_variables (); @}
'@{' declarations statements '@}' '@{' declarations statements '@}'
| @{ prepare_for_local_variables (); @} | @{ prepare_for_local_variables (); @}
'@{' statements '@}' '@{' statements '@}'
@@ -3826,7 +3861,8 @@ does work is to put the action after the open-brace, like this:
@example @example
@group @group
compound: '@{' @{ prepare_for_local_variables (); @} compound:
'@{' @{ prepare_for_local_variables (); @}
declarations statements '@}' declarations statements '@}'
| '@{' statements '@}' | '@{' statements '@}'
; ;
@@ -3842,17 +3878,15 @@ serves as a subroutine:
@example @example
@group @group
subroutine: /* empty */ subroutine:
@{ prepare_for_local_variables (); @} /* empty */ @{ prepare_for_local_variables (); @}
; ;
@end group @end group
@group @group
compound: subroutine compound:
'@{' declarations statements '@}' subroutine '@{' declarations statements '@}'
| subroutine | subroutine '@{' statements '@}'
'@{' statements '@}'
; ;
@end group @end group
@end example @end example
@@ -3938,7 +3972,8 @@ Here is a basic example using the default data type for locations:
@example @example
@group @group
exp: @dots{} exp:
@dots{}
| exp '/' exp | exp '/' exp
@{ @{
@@$.first_column = @@1.first_column; @@$.first_column = @@1.first_column;
@@ -3969,7 +4004,8 @@ example above simply rewrites this way:
@example @example
@group @group
exp: @dots{} exp:
@dots{}
| exp '/' exp | exp '/' exp
@{ @{
if ($3) if ($3)
@@ -6355,13 +6391,15 @@ factorial operators (@samp{!}), and allow parentheses for grouping.
@example @example
@group @group
expr: term '+' expr expr:
term '+' expr
| term | term
; ;
@end group @end group
@group @group
term: '(' expr ')' term:
'(' expr ')'
| term '!' | term '!'
| NUMBER | NUMBER
; ;
@@ -6471,7 +6509,8 @@ the conflict:
%% %%
@end group @end group
@group @group
stmt: expr stmt:
expr
| if_stmt | if_stmt
; ;
@end group @end group
@@ -6483,7 +6522,8 @@ if_stmt:
; ;
@end group @end group
expr: variable expr:
variable
; ;
@end example @end example
@@ -6512,7 +6552,8 @@ input @w{@samp{1 - 2 * 3}} can be parsed in two different ways):
@example @example
@group @group
expr: expr '-' expr expr:
expr '-' expr
| expr '*' expr | expr '*' expr
| expr '<' expr | expr '<' expr
| '(' expr ')' | '(' expr ')'
@@ -6671,7 +6712,8 @@ Now the precedence of @code{UMINUS} can be used in specific rules:
@example @example
@group @group
exp: @dots{} exp:
@dots{}
| exp '-' exp | exp '-' exp
@dots{} @dots{}
| '-' exp %prec UMINUS | '-' exp %prec UMINUS
@@ -6740,19 +6782,17 @@ of zero or more @code{word} groupings.
@example @example
@group @group
sequence: /* empty */ sequence:
@{ printf ("empty sequence\n"); @} /* empty */ @{ printf ("empty sequence\n"); @}
| maybeword | maybeword
| sequence word | sequence word @{ printf ("added word %s\n", $2); @}
@{ printf ("added word %s\n", $2); @}
; ;
@end group @end group
@group @group
maybeword: /* empty */ maybeword:
@{ printf ("empty maybeword\n"); @} /* empty */ @{ printf ("empty maybeword\n"); @}
| word | word @{ printf ("single word %s\n", $1); @}
@{ printf ("single word %s\n", $1); @}
; ;
@end group @end group
@end example @end example
@@ -6781,26 +6821,28 @@ reduce/reduce conflict must be studied and usually eliminated. Here is the
proper way to define @code{sequence}: proper way to define @code{sequence}:
@example @example
sequence: /* empty */ sequence:
@{ printf ("empty sequence\n"); @} /* empty */ @{ printf ("empty sequence\n"); @}
| sequence word | sequence word @{ printf ("added word %s\n", $2); @}
@{ printf ("added word %s\n", $2); @}
; ;
@end example @end example
Here is another common error that yields a reduce/reduce conflict: Here is another common error that yields a reduce/reduce conflict:
@example @example
sequence: /* empty */ sequence:
/* empty */
| sequence words | sequence words
| sequence redirects | sequence redirects
; ;
words: /* empty */ words:
/* empty */
| words word | words word
; ;
redirects:/* empty */ redirects:
/* empty */
| redirects redirect | redirects redirect
; ;
@end example @end example
@@ -6821,7 +6863,8 @@ Here are two ways to correct these rules. First, to make it a single level
of sequence: of sequence:
@example @example
sequence: /* empty */ sequence:
/* empty */
| sequence word | sequence word
| sequence redirect | sequence redirect
; ;
@@ -6832,20 +6875,23 @@ from being empty:
@example @example
@group @group
sequence: /* empty */ sequence:
/* empty */
| sequence words | sequence words
| sequence redirects | sequence redirects
; ;
@end group @end group
@group @group
words: word words:
word
| words word | words word
; ;
@end group @end group
@group @group
redirects:redirect redirects:
redirect
| redirects redirect | redirects redirect
; ;
@end group @end group
@@ -6863,8 +6909,7 @@ Here is an example:
%token ID %token ID
%% %%
def: param_spec return_spec ',' def: param_spec return_spec ',';
;
param_spec: param_spec:
type type
| name_list ':' type | name_list ':' type
@@ -6877,12 +6922,10 @@ return_spec:
; ;
@end group @end group
@group @group
type: ID type: ID;
;
@end group @end group
@group @group
name: ID name: ID;
;
name_list: name_list:
name name
| name ',' name_list | name ',' name_list
@@ -6937,8 +6980,7 @@ distinct. In the above example, adding one rule to
return_spec: return_spec:
type type
| name ':' type | name ':' type
/* This rule is never used. */ | ID BOGUS /* This rule is never used. */
| ID BOGUS
; ;
@end group @end group
@end example @end example
@@ -7540,7 +7582,8 @@ in the current context, the parse can continue.
For example: For example:
@example @example
stmnts: /* empty string */ stmnts:
/* empty string */
| stmnts '\n' | stmnts '\n'
| stmnts exp '\n' | stmnts exp '\n'
| stmnts error '\n' | stmnts error '\n'
@@ -7584,7 +7627,8 @@ close-delimiter will probably appear to be unmatched, and generate another,
spurious error message: spurious error message:
@example @example
primary: '(' expr ')' primary:
'(' expr ')'
| '(' error ')' | '(' error ')'
@dots{} @dots{}
; ;
@@ -7709,16 +7753,14 @@ duplication, with actions omitted for brevity:
@example @example
@group @group
initdcl: initdcl:
declarator maybeasm '=' declarator maybeasm '=' init
init
| declarator maybeasm | declarator maybeasm
; ;
@end group @end group
@group @group
notype_initdcl: notype_initdcl:
notype_declarator maybeasm '=' notype_declarator maybeasm '=' init
init
| notype_declarator maybeasm | notype_declarator maybeasm
; ;
@end group @end group
@@ -7761,15 +7803,12 @@ as an identifier if it appears in that context. Here is how you can do it:
@dots{} @dots{}
@end group @end group
@group @group
expr: IDENTIFIER expr:
IDENTIFIER
| constant | constant
| HEX '(' | HEX '(' @{ hexflag = 1; @}
@{ hexflag = 1; @} expr ')' @{ hexflag = 0; $$ = $4; @}
expr ')' | expr '+' expr @{ $$ = make_sum ($1, $3); @}
@{ hexflag = 0;
$$ = $4; @}
| expr '+' expr
@{ $$ = make_sum ($1, $3); @}
@dots{} @dots{}
; ;
@end group @end group
@@ -7804,11 +7843,11 @@ For example, in C-like languages, a typical error recovery rule is to skip
tokens until the next semicolon, and then start a new statement, like this: tokens until the next semicolon, and then start a new statement, like this:
@example @example
stmt: expr ';' stmt:
expr ';'
| IF '(' expr ')' stmt @{ @dots{} @} | IF '(' expr ')' stmt @{ @dots{} @}
@dots{} @dots{}
error ';' | error ';' @{ hexflag = 0; @}
@{ hexflag = 0; @}
; ;
@end example @end example
@@ -7826,9 +7865,9 @@ and skips to the close-parenthesis:
@example @example
@group @group
expr: @dots{} expr:
| '(' expr ')' @dots{}
@{ $$ = $2; @} | '(' expr ')' @{ $$ = $2; @}
| '(' error ')' | '(' error ')'
@dots{} @dots{}
@end group @end group
@@ -7888,7 +7927,8 @@ The following grammar file, @file{calc.y}, will be used in the sequel:
%left '+' '-' %left '+' '-'
%left '*' %left '*'
%% %%
exp: exp '+' exp exp:
exp '+' exp
| exp '-' exp | exp '-' exp
| exp '*' exp | exp '*' exp
| exp '/' exp | exp '/' exp
@@ -9302,8 +9342,9 @@ The grammar itself is straightforward.
%start unit; %start unit;
unit: assignments exp @{ driver.result = $2; @}; unit: assignments exp @{ driver.result = $2; @};
assignments: assignments assignment @{@} assignments:
| /* Nothing. */ @{@}; /* Nothing. */ @{@}
| assignments assignment @{@};
assignment: assignment:
"identifier" ":=" exp "identifier" ":=" exp
@@ -10313,7 +10354,8 @@ real start-symbol:
@example @example
%token START_FOO START_BAR; %token START_FOO START_BAR;
%start start; %start start;
start: START_FOO foo start:
START_FOO foo
| START_BAR bar; | START_BAR bar;
@end example @end example