язык точной арифметики (precision arithmetic language)
Расширенное описание (Extended description)
Grammar
The grammar in this section and the lexical conventions in the
following section shall together describe the syntax for bc
programs. The general conventions for this style of grammar are
described in Section 1.3, Grammar Conventions. A valid program
can be represented as the non-terminal symbol program
in the
grammar. This formal syntax shall take precedence over the text
syntax description.
%token EOF NEWLINE STRING LETTER NUMBER
%token MUL_OP
/* '*', '/', '%' */
%token ASSIGN_OP
/* '=', '+=', '-=', '*=', '/=', '%=', '^=' */
%token REL_OP
/* '==', '<=', '>=', '!=', '<', '>' */
%token INCR_DECR
/* '++', '--' */
%token Define Break Quit Length
/* 'define', 'break', 'quit', 'length' */
%token Return For If While Sqrt
/* 'return', 'for', 'if', 'while', 'sqrt' */
%token Scale Ibase Obase Auto
/* 'scale', 'ibase', 'obase', 'auto' */
%start program
%%
program : EOF
| input_item program
;
input_item : semicolon_list NEWLINE
| function
;
semicolon_list : /* empty */
| statement
| semicolon_list ';' statement
| semicolon_list ';'
;
statement_list : /* empty */
| statement
| statement_list NEWLINE
| statement_list NEWLINE statement
| statement_list ';'
| statement_list ';' statement
;
statement : expression
| STRING
| Break
| Quit
| Return
| Return '(' return_expression ')'
| For '(' expression ';'
relational_expression ';'
expression ')' statement
| If '(' relational_expression ')' statement
| While '(' relational_expression ')' statement
| '{' statement_list '}'
;
function : Define LETTER '(' opt_parameter_list ')'
'{' NEWLINE opt_auto_define_list
statement_list '}'
;
opt_parameter_list : /* empty */
| parameter_list
;
parameter_list : LETTER
| define_list ',' LETTER
;
opt_auto_define_list : /* empty */
| Auto define_list NEWLINE
| Auto define_list ';'
;
define_list : LETTER
| LETTER '[' ']'
| define_list ',' LETTER
| define_list ',' LETTER '[' ']'
;
opt_argument_list : /* empty */
| argument_list
;
argument_list : expression
| LETTER '[' ']' ',' argument_list
;
relational_expression : expression
| expression REL_OP expression
;
return_expression : /* empty */
| expression
;
expression : named_expression
| NUMBER
| '(' expression ')'
| LETTER '(' opt_argument_list ')'
| '-' expression
| expression '+' expression
| expression '-' expression
| expression MUL_OP expression
| expression '^' expression
| INCR_DECR named_expression
| named_expression INCR_DECR
| named_expression ASSIGN_OP expression
| Length '(' expression ')'
| Sqrt '(' expression ')'
| Scale '(' expression ')'
;
named_expression : LETTER
| LETTER '[' expression ']'
| Scale
| Ibase
| Obase
;
Lexical Conventions in bc
The lexical conventions for bc programs, with respect to the
preceding grammar, shall be as follows:
1. Except as noted, bc shall recognize the longest possible
token or delimiter beginning at a given point.
2. A comment shall consist of any characters beginning with the
two adjacent characters "/*"
and terminated by the next
occurrence of the two adjacent characters "*/"
. Comments
shall have no effect except to delimit lexical tokens.
3. The <newline> shall be recognized as the token NEWLINE
.
4. The token STRING
shall represent a string constant; it shall
consist of any characters beginning with the double-quote
character ('"'
) and terminated by another occurrence of the
double-quote character. The value of the string is the
sequence of all characters between, but not including, the
two double-quote characters. All characters shall be taken
literally from the input, and there is no way to specify a
string containing a double-quote character. The length of the
value of each string shall be limited to {BC_STRING_MAX}
bytes.
5. A <blank> shall have no effect except as an ordinary
character if it appears within a STRING
token, or to delimit
a lexical token other than STRING
.
6. The combination of a <backslash> character immediately
followed by a <newline> shall have no effect other than to
delimit lexical tokens with the following exceptions:
* It shall be interpreted as the character sequence
"\<newline>"
in STRING
tokens.
* It shall be ignored as part of a multi-line NUMBER
token.
7. The token NUMBER
shall represent a numeric constant. It shall
be recognized by the following grammar:
NUMBER : integer
| '.' integer
| integer '.'
| integer '.' integer
;
integer : digit
| integer digit
;
digit : 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7
| 8 | 9 | A | B | C | D | E | F
;
8. The value of a NUMBER
token shall be interpreted as a numeral
in the base specified by the value of the internal register
ibase
(described below). Each of the digit
characters shall
have the value from 0 to 15 in the order listed here, and the
<period> character shall represent the radix point. The
behavior is undefined if digits greater than or equal to the
value of ibase
appear in the token. However, note the
exception for single-digit values being assigned to ibase
and
obase
themselves, in Operations in bc.
9. The following keywords shall be recognized as tokens:
auto ibase length return while
break if obase scale
define for quit sqrt
10. Any of the following characters occurring anywhere except
within a keyword shall be recognized as the token LETTER
:
a b c d e f g h i j k l m n o p q r s t u v w x y z
11. The following single-character and two-character sequences
shall be recognized as the token ASSIGN_OP
:
= += -= *= /= %= ^=
12. If an '='
character, as the beginning of a token, is followed
by a '-'
character with no intervening delimiter, the
behavior is undefined.
13. The following single-characters shall be recognized as the
token MUL_OP
:
* / %
14. The following single-character and two-character sequences
shall be recognized as the token REL_OP
:
== <= >= != < >
15. The following two-character sequences shall be recognized as
the token INCR_DECR
:
++ --
16. The following single characters shall be recognized as tokens
whose names are the character:
<newline> ( ) , + - ; [ ] ^ { }
17. The token EOF
is returned when the end of input is reached.
Operations in bc
There are three kinds of identifiers: ordinary identifiers, array
identifiers, and function identifiers. All three types consist
of single lowercase letters. Array identifiers shall be followed
by square brackets ("[]"
). An array subscript is required except
in an argument or auto list. Arrays are singly dimensioned and
can contain up to {BC_DIM_MAX} elements. Indexing shall begin at
zero so an array is indexed from 0 to {BC_DIM_MAX}-1. Subscripts
shall be truncated to integers. The application shall ensure that
function identifiers are followed by parentheses, possibly
enclosing arguments. The three types of identifiers do not
conflict.
The following table summarizes the rules for precedence and
associativity of all operators. Operators on the same line shall
have the same precedence; rows are in order of decreasing
precedence.
Table: Operators in
bc
┌──────────────────────────┬───────────────┐
│ Operator
│ Associativity
│
├──────────────────────────┼───────────────┤
│++, -- │ N/A │
│unary - │ N/A │
│^ │ Right to left │
│*, /, % │ Left to right │
│+, binary - │ Left to right │
│=, +=, -=, *=, /=, %=, ^= │ Right to left │
│==, <=, >=, !=, <, > │ None │
└──────────────────────────┴───────────────┘
Each expression or named expression has a scale, which is the
number of decimal digits that shall be maintained as the
fractional portion of the expression.
Named expressions are places where values are stored. Named
expressions shall be valid on the left side of an assignment. The
value of a named expression shall be the value stored in the
place named. Simple identifiers and array elements are named
expressions; they have an initial value of zero and an initial
scale of zero.
The internal registers scale
, ibase
, and obase
are all named
expressions. The scale of an expression consisting of the name of
one of these registers shall be zero; values assigned to any of
these registers are truncated to integers. The scale
register
shall contain a global value used in computing the scale of
expressions (as described below). The value of the register scale
is limited to 0 ≤ scale
≤ {BC_SCALE_MAX} and shall have a default
value of zero. The ibase
and obase
registers are the input and
output number radix, respectively. The value of ibase
shall be
limited to:
2 ≤ ibase ≤ 16
The value of obase
shall be limited to:
2 ≤ obase ≤ {BC_BASE_MAX}
When either ibase
or obase
is assigned a single digit
value from
the list in Lexical Conventions in bc, the value shall be assumed
in hexadecimal. (For example, ibase
=A sets to base ten,
regardless of the current ibase
value.) Otherwise, the behavior
is undefined when digits greater than or equal to the value of
ibase
appear in the input. Both ibase
and obase
shall have
initial values of 10.
Internal computations shall be conducted as if in decimal,
regardless of the input and output bases, to the specified number
of decimal digits. When an exact result is not achieved (for
example, scale
=0; 3.2/1),
the result shall be truncated.
For all values of obase
specified by this volume of POSIX.1‐2017,
bc shall output numeric values by performing each of the
following steps in order:
1. If the value is less than zero, a <hyphen-minus> ('-'
)
character shall be output.
2. One of the following is output, depending on the numerical
value:
* If the absolute value of the numerical value is greater
than or equal to one, the integer portion of the value
shall be output as a series of digits appropriate to
obase
(as described below), most significant digit first.
The most significant non-zero digit shall be output next,
followed by each successively less significant digit.
* If the absolute value of the numerical value is less than
one but greater than zero and the scale of the numerical
value is greater than zero, it is unspecified whether the
character 0 is output.
* If the numerical value is zero, the character 0 shall be
output.
3. If the scale of the value is greater than zero and the
numeric value is not zero, a <period> character shall be
output, followed by a series of digits appropriate to obase
(as described below) representing the most significant
portion of the fractional part of the value. If s represents
the scale of the value being output, the number of digits
output shall be s if obase
is 10, less than or equal to s if
obase
is greater than 10, or greater than or equal to s if
obase
is less than 10. For obase
values other than 10, this
should be the number of digits needed to represent a
precision of 10s.
For obase
values from 2 to 16, valid digits are the first obase
of the single characters:
0 1 2 3 4 5 6 7 8 9 A B C D E F
which represent the values zero to 15, inclusive, respectively.
For bases greater than 16, each digit shall be written as a
separate multi-digit decimal number. Each digit except the most
significant fractional digit shall be preceded by a single
<space>. For bases from 17 to 100, bc shall write two-digit
decimal numbers; for bases from 101 to 1000, three-digit decimal
strings, and so on. For example, the decimal number 1024 in base
25 would be written as:
01 15 24
and in base 125, as:
008 024
Very large numbers shall be split across lines with 70 characters
per line in the POSIX locale; other locales may split at
different character boundaries. Lines that are continued shall
end with a <backslash>.
A function call shall consist of a function name followed by
parentheses containing a <comma>-separated list of expressions,
which are the function arguments. A whole array passed as an
argument shall be specified by the array name followed by empty
square brackets. All function arguments shall be passed by value.
As a result, changes made to the formal parameters shall have no
effect on the actual arguments. If the function terminates by
executing a return
statement, the value of the function shall be
the value of the expression in the parentheses of the return
statement or shall be zero if no expression is provided or if
there is no return
statement.
The result of sqrt
(expression) shall be the square root of the
expression. The result shall be truncated in the least
significant decimal place. The scale of the result shall be the
scale of the expression or the value of scale
, whichever is
larger.
The result of length
(expression) shall be the total number of
significant decimal digits in the expression. The scale of the
result shall be zero.
The result of scale
(expression) shall be the scale of the
expression. The scale of the result shall be zero.
A numeric constant shall be an expression. The scale shall be the
number of digits that follow the radix point in the input
representing the constant, or zero if no radix point appears.
The sequence ( expression ) shall be an expression with the same
value and scale as expression. The parentheses can be used to
alter the normal precedence.
The semantics of the unary and binary operators are as follows:
-expression
The result shall be the negative of the expression. The
scale of the result shall be the scale of expression.
The unary increment and decrement operators shall not modify the
scale of the named expression upon which they operate. The scale
of the result shall be the scale of that named expression.
++named-expression
The named expression shall be incremented by one. The
result shall be the value of the named expression after
incrementing.
--named-expression
The named expression shall be decremented by one. The
result shall be the value of the named expression after
decrementing.
named-expression++
The named expression shall be incremented by one. The
result shall be the value of the named expression before
incrementing.
named-expression--
The named expression shall be decremented by one. The
result shall be the value of the named expression before
decrementing.
The exponentiation operator, <circumflex> ('^'
), shall bind right
to left.
expression^expression
The result shall be the first expression raised to the
power of the second expression. If the second expression
is not an integer, the behavior is undefined. If a is the
scale of the left expression and b is the absolute value of
the right expression, the scale of the result shall be:
if b >= 0 min(a * b, max(scale, a)) if b < 0 scale
The multiplicative operators ('*'
, '/'
, '%'
) shall bind left to
right.
expression*expression
The result shall be the product of the two expressions. If
a and b are the scales of the two expressions, then the
scale of the result shall be:
min(a+b,max(scale,a,b))
expression/expression
The result shall be the quotient of the two expressions.
The scale of the result shall be the value of scale
.
expression%expression
For expressions a and b, a%b shall be evaluated equivalent
to the steps:
1. Compute a/b to current scale.
2. Use the result to compute:
a - (a / b) * b
to scale:
max(scale + scale(b), scale(a))
The scale of the result shall be:
max(scale + scale(b), scale(a))
When scale
is zero, the '%'
operator is the mathematical
remainder operator.
The additive operators ('+'
, '-'
) shall bind left to right.
expression+expression
The result shall be the sum of the two expressions. The
scale of the result shall be the maximum of the scales of
the expressions.
expression-expression
The result shall be the difference of the two expressions.
The scale of the result shall be the maximum of the scales
of the expressions.
The assignment operators ('='
, "+="
, "-="
, "*="
, "/="
, "%="
,
"^="
) shall bind right to left.
named-expression=expression
This expression shall result in assigning the value of the
expression on the right to the named expression on the
left. The scale of both the named expression and the result
shall be the scale of expression.
The compound assignment forms:
named-expression <operator>= expression
shall be equivalent to:
named-expression=named-expression <operator> expression
except that the named-expression shall be evaluated only once.
Unlike all other operators, the relational operators ('<'
, '>'
,
"<="
, ">="
, "=="
, "!="
) shall be only valid as the object of an
if
, while
, or inside a for
statement.
expression1<expression2
The relation shall be true if the value of expression1 is
strictly less than the value of expression2.
expression1>expression2
The relation shall be true if the value of expression1 is
strictly greater than the value of expression2.
expression1<=expression2
The relation shall be true if the value of expression1 is
less than or equal to the value of expression2.
expression1>=expression2
The relation shall be true if the value of expression1 is
greater than or equal to the value of expression2.
expression1==expression2
The relation shall be true if the values of expression1 and
expression2 are equal.
expression1!=expression2
The relation shall be true if the values of expression1 and
expression2 are unequal.
There are only two storage classes in bc: global and automatic
(local). Only identifiers that are local to a function need be
declared with the auto
command. The arguments to a function shall
be local to the function. All other identifiers are assumed to
be global and available to all functions. All identifiers, global
and local, have initial values of zero. Identifiers declared as
auto shall be allocated on entry to the function and released on
returning from the function. They therefore do not retain values
between function calls. Auto arrays shall be specified by the
array name followed by empty square brackets. On entry to a
function, the old values of the names that appear as parameters
and as automatic variables shall be pushed onto a stack. Until
the function returns, reference to these names shall refer only
to the new values.
References to any of these names from other functions that are
called from this function also refer to the new value until one
of those functions uses the same name for a local variable.
When a statement is an expression, unless the main operator is an
assignment, execution of the statement shall write the value of
the expression followed by a <newline>.
When a statement is a string, execution of the statement shall
write the value of the string.
Statements separated by <semicolon> or <newline> characters shall
be executed sequentially. In an interactive invocation of bc,
each time a <newline> is read that satisfies the grammatical
production:
input_item : semicolon_list NEWLINE
the sequential list of statements making up the semicolon_list
shall be executed immediately and any output produced by that
execution shall be written without any delay due to buffering.
In an if
statement (if
(relation) statement), the statement shall
be executed if the relation is true.
The while
statement (while
(relation) statement) implements a loop
in which the relation is tested; each time the relation is true,
the statement shall be executed and the relation retested. When
the relation is false, execution shall resume after statement.
A for
statement(for
(expression; relation; expression) statement)
shall be the same as:
first-expression
while (relation) {
statement
last-expression
}
The application shall ensure that all three expressions are
present.
The break
statement shall cause termination of a for
or while
statement.
The auto
statement (auto
identifier [
,identifier]
...) shall
cause the values of the identifiers to be pushed down. The
identifiers can be ordinary identifiers or array identifiers.
Array identifiers shall be specified by following the array name
by empty square brackets. The application shall ensure that the
auto
statement is the first statement in a function definition.
A define
statement:
define LETTER ( opt_parameter_list ) {
opt_auto_define_list
statement_list
}
defines a function named LETTER
. If a function named LETTER
was
previously defined, the define
statement shall replace the
previous definition. The expression:
LETTER ( opt_argument_list )
shall invoke the function named LETTER
. The behavior is
undefined if the number of arguments in the invocation does not
match the number of parameters in the definition. Functions shall
be defined before they are invoked. A function shall be
considered to be defined within its own body, so recursive calls
are valid. The values of numeric constants within a function
shall be interpreted in the base specified by the value of the
ibase
register when the function is invoked.
The return
statements (return
and return
(expression)) shall cause
termination of a function, popping of its auto variables, and
specification of the result of the function. The first form shall
be equivalent to return
(0). The value and scale of the result
returned by the function shall be the value and scale of the
expression returned.
The quit
statement (quit
) shall stop execution of a bc program at
the point where the statement occurs in the input, even if it
occurs in a function definition, or in an if
, for
, or while
statement.
The following functions shall be defined when the -l
option is
specified:
s
( expression )
Sine of argument in radians.
c
( expression )
Cosine of argument in radians.
a
( expression )
Arctangent of argument.
l
( expression )
Natural logarithm of argument.
e
( expression )
Exponential function of argument.
j
( expression1, expression2 )
Bessel function of expression2 of the first kind of integer
order expression1.
The scale of the result returned by these functions shall be the
value of the scale
register at the time the function is invoked.
The value of the scale
register after these functions have
completed their execution shall be the same value it had upon
invocation. The behavior is undefined if any of these functions
is invoked with an argument outside the domain of the
mathematical function.