intdoc.in

gcc-2.95.3 Linux下最常用的C编译器

@xref{IBClr Intrinsic}, for how to set a bit in a variable to 0.

")

DEFDOC (BTEST, "Test bit.", "\
Returns @code{.TRUE.} if bit @var{@2@} in @var{@1@} is
1, @code{.FALSE.} otherwise.

(Bit 0 is the low-order (rightmost) bit, adding the value 
@ifinfo
2**0,
@end ifinfo
@iftex
@tex
$2^0$,
@end tex
@end iftex
or 1,
to the number if set to 1;
bit 1 is the next-higher-order bit, adding 
@ifinfo
2**1,
@end ifinfo
@iftex
@tex
$2^1$,
@end tex
@end iftex
or 2;
bit 2 adds 
@ifinfo
2**2,
@end ifinfo
@iftex
@tex
$2^2$,
@end tex
@end iftex
or 4; and so on.)

@xref{Bit_Size Intrinsic}, for how to obtain the number of bits
in a type.
The leftmost bit of @var{@1@} is @samp{BIT_SIZE(@var{@1@}-1)}.
")

DEFDOC (CMPLX, "Construct @code{COMPLEX(KIND=1)} value.", "\
If @var{@1@} is not type @code{COMPLEX},
constructs a value of type @code{COMPLEX(KIND=1)} from the
real and imaginary values specified by @var{@1@} and
@var{@2@}, respectively.
If @var{@2@} is omitted, @samp{0.} is assumed.

If @var{@1@} is type @code{COMPLEX},
converts it to type @code{COMPLEX(KIND=1)}.

@xref{Complex Intrinsic}, for information on easily constructing
a @code{COMPLEX} value of arbitrary precision from @code{REAL}
arguments.
")

DEFDOC (DCMPLX, "Construct @code{COMPLEX(KIND=2)} value.", "\
If @var{@1@} is not type @code{COMPLEX},
constructs a value of type @code{COMPLEX(KIND=2)} from the
real and imaginary values specified by @var{@1@} and
@var{@2@}, respectively.
If @var{@2@} is omitted, @samp{0D0} is assumed.

If @var{@1@} is type @code{COMPLEX},
converts it to type @code{COMPLEX(KIND=2)}.

Although this intrinsic is not standard Fortran,
it is a popular extension offered by many compilers
that support @code{DOUBLE COMPLEX}, since it offers
the easiest way to convert to @code{DOUBLE COMPLEX}
without using Fortran 90 features (such as the @samp{KIND=}
argument to the @code{CMPLX()} intrinsic).

(@samp{CMPLX(0D0, 0D0)} returns a single-precision
@code{COMPLEX} result, as required by standard FORTRAN 77.
That's why so many compilers provide @code{DCMPLX()}, since
@samp{DCMPLX(0D0, 0D0)} returns a @code{DOUBLE COMPLEX}
result.
Still, @code{DCMPLX()} converts even @code{REAL*16} arguments
to their @code{REAL*8} equivalents in most dialects of
Fortran, so neither it nor @code{CMPLX()} allow easy
construction of arbitrary-precision values without
potentially forcing a conversion involving extending or
reducing precision.
GNU Fortran provides such an intrinsic, called @code{COMPLEX()}.)

@xref{Complex Intrinsic}, for information on easily constructing
a @code{COMPLEX} value of arbitrary precision from @code{REAL}
arguments.
")

DEFDOC (CONJG, "Complex conjugate.", "\
Returns the complex conjugate:

@example
COMPLEX(REALPART(@var{@1@}), -IMAGPART(@var{@1@}))
@end example
")

DEFDOC (DCONJG, "Complex conjugate (archaic).", ARCHAIC (CONJG, Conjg))

DEFDOC (COS, "Cosine.", "\
Returns the cosine of @var{@1@}, an angle measured
in radians.

@xref{ACos Intrinsic}, for the inverse of this function.
")

DEFDOC (CCOS, "Cosine (archaic).", ARCHAIC (COS, Cos))

DEFDOC (DCOS, "Cosine (archaic).", ARCHAIC (COS, Cos))

DEFDOC (CDCOS, "Cosine (archaic).", ARCHAIC (COS, Cos))

DEFDOC (COSH, "Hyperbolic cosine.", "\
Returns the hyperbolic cosine of @var{@1@}.
")

DEFDOC (DCOSH, "Hyperbolic cosine (archaic).", ARCHAIC (COSH, CosH))

DEFDOC (SQRT, "Square root.", "\
Returns the square root of @var{@1@}, which must
not be negative.

To calculate and represent the square root of a negative
number, complex arithmetic must be used.
For example, @samp{SQRT(COMPLEX(@var{@1@}))}.

The inverse of this function is @samp{SQRT(@var{@1@}) * SQRT(@var{@1@})}.
")

DEFDOC (CSQRT, "Square root (archaic).", ARCHAIC (SQRT, SqRt))

DEFDOC (DSQRT, "Square root (archaic).", ARCHAIC (SQRT, SqRt))

DEFDOC (CDSQRT, "Square root (archaic).", ARCHAIC (SQRT, SqRt))

DEFDOC (DBLE, "Convert to double precision.", "\
Returns @var{@1@} converted to double precision
(@code{REAL(KIND=2)}).
If @var{@1@} is @code{COMPLEX}, the real part of
@var{@1@} is used for the conversion
and the imaginary part disregarded.

@xref{Sngl Intrinsic}, for the function that converts
to single precision.

@xref{Int Intrinsic}, for the function that converts
to @code{INTEGER}.

@xref{Complex Intrinsic}, for the function that converts
to @code{COMPLEX}.
")

DEFDOC (DIM, "Difference magnitude (non-negative subtract).", "\
Returns @samp{@var{@1@}-@var{@2@}} if @var{@1@} is greater than
@var{@2@}; otherwise returns zero.
")

DEFDOC (DDIM, "Difference magnitude (archaic).", ARCHAIC_2 (DIM, DiM))
DEFDOC (IDIM, "Difference magnitude (archaic).", ARCHAIC_2 (DIM, DiM))

DEFDOC (DPROD, "Double-precision product.", "\
Returns @samp{DBLE(@var{@1@})*DBLE(@var{@2@})}.
")

DEFDOC (EXP, "Exponential.", "\
Returns @samp{@var{e}**@var{@1@}}, where
@var{e} is approximately 2.7182818.

@xref{Log Intrinsic}, for the inverse of this function.
")

DEFDOC (CEXP, "Exponential (archaic).", ARCHAIC (EXP, Exp))

DEFDOC (DEXP, "Exponential (archaic).", ARCHAIC (EXP, Exp))

DEFDOC (CDEXP, "Exponential (archaic).", ARCHAIC (EXP, Exp))

DEFDOC (FLOAT, "Conversion (archaic).", ARCHAIC (REAL, Real))
DEFDOC (DFLOAT, "Conversion (archaic).", ARCHAIC (REAL, Real))

DEFDOC (IFIX, "Conversion (archaic).", ARCHAIC (INT, Int))

DEFDOC (LONG, "Conversion to @code{INTEGER(KIND=1)} (archaic).", "\
Archaic form of @code{INT()} that is specific
to one type for @var{@1@}.
@xref{Int Intrinsic}.

The precise meaning of this intrinsic might change
in a future version of the GNU Fortran language,
as more is learned about how it is used.
")

DEFDOC (SHORT, "Convert to @code{INTEGER(KIND=6)} value@99@truncated to whole number.", "\
Returns @var{@1@} with the fractional portion of its
magnitude truncated and its sign preserved, converted
to type @code{INTEGER(KIND=6)}.

If @var{@1@} is type @code{COMPLEX}, its real part
is truncated and converted, and its imaginary part is disgregarded.

@xref{Int Intrinsic}.

The precise meaning of this intrinsic might change
in a future version of the GNU Fortran language,
as more is learned about how it is used.
")

DEFDOC (INT2, "Convert to @code{INTEGER(KIND=6)} value@99@truncated to whole number.", "\
Returns @var{@1@} with the fractional portion of its
magnitude truncated and its sign preserved, converted
to type @code{INTEGER(KIND=6)}.

If @var{@1@} is type @code{COMPLEX}, its real part
is truncated and converted, and its imaginary part is disgregarded.

@xref{Int Intrinsic}.

The precise meaning of this intrinsic might change
in a future version of the GNU Fortran language,
as more is learned about how it is used.
")

DEFDOC (INT8, "Convert to @code{INTEGER(KIND=2)} value@99@truncated to whole number.", "\
Returns @var{@1@} with the fractional portion of its
magnitude truncated and its sign preserved, converted
to type @code{INTEGER(KIND=2)}.

If @var{@1@} is type @code{COMPLEX}, its real part
is truncated and converted, and its imaginary part is disgregarded.

@xref{Int Intrinsic}.

The precise meaning of this intrinsic might change
in a future version of the GNU Fortran language,
as more is learned about how it is used.
")

DEFDOC (LEN, "Length of character entity.", "\
Returns the length of @var{@1@}.

If @var{@1@} is an array, the length of an element
of @var{@1@} is returned.

Note that @var{@1@} need not be defined when this
intrinsic is invoked, since only the length, not
the content, of @var{@1@} is needed.

@xref{Bit_Size Intrinsic}, for the function that determines
the size of its argument in bits.
")

DEFDOC (TAN, "Tangent.", "\
Returns the tangent of @var{@1@}, an angle measured
in radians.

@xref{ATan Intrinsic}, for the inverse of this function.
")

DEFDOC (DTAN, "Tangent (archaic).", ARCHAIC (TAN, Tan))

DEFDOC (TANH, "Hyperbolic tangent.", "\
Returns the hyperbolic tangent of @var{@1@}.
")

DEFDOC (DTANH, "Hyperbolic tangent (archaic).", ARCHAIC (TANH, TanH))

DEFDOC (SNGL, "Convert (archaic).", ARCHAIC (REAL, Real))

DEFDOC (SIN, "Sine.", "\
Returns the sine of @var{@1@}, an angle measured
in radians.

@xref{ASin Intrinsic}, for the inverse of this function.
")

DEFDOC (CSIN, "Sine (archaic).", ARCHAIC (SIN, Sin))

DEFDOC (DSIN, "Sine (archaic).", ARCHAIC (SIN, Sin))

DEFDOC (CDSIN, "Sine (archaic).", ARCHAIC (SIN, Sin))

DEFDOC (SINH, "Hyperbolic sine.", "\
Returns the hyperbolic sine of @var{@1@}.
")

DEFDOC (DSINH, "Hyperbolic sine (archaic).", ARCHAIC (SINH, SinH))

DEFDOC (LSHIFT, "Left-shift bits.", "\
Returns @var{@1@} shifted to the left
@var{@2@} bits.

Although similar to the expression
@samp{@var{@1@}*(2**@var{@2@})}, there
are important differences.
For example, the sign of the result is
not necessarily the same as the sign of
@var{@1@}.

Currently this intrinsic is defined assuming
the underlying representation of @var{@1@}
is as a two's-complement integer.
It is unclear at this point whether that
definition will apply when a different
representation is involved.

@xref{LShift Intrinsic}, for the inverse of this function.

@xref{IShft Intrinsic}, for information
on a more widely available left-shifting
intrinsic that is also more precisely defined.
")

DEFDOC (RSHIFT, "Right-shift bits.", "\
Returns @var{@1@} shifted to the right
@var{@2@} bits.

Although similar to the expression
@samp{@var{@1@}/(2**@var{@2@})}, there
are important differences.
For example, the sign of the result is
undefined.

Currently this intrinsic is defined assuming
the underlying representation of @var{@1@}
is as a two's-complement integer.
It is unclear at this point whether that
definition will apply when a different
representation is involved.

@xref{RShift Intrinsic}, for the inverse of this function.

@xref{IShft Intrinsic}, for information
on a more widely available right-shifting
intrinsic that is also more precisely defined.
")

DEFDOC (LGE, "Lexically greater than or equal.", "\
Returns @samp{.TRUE.} if @samp{@var{@1@}.GE.@var{@2@}},
@samp{.FALSE.} otherwise.
@var{@1@} and @var{@2@} are interpreted as containing
ASCII character codes.
If either value contains a character not in the ASCII
character set, the result is processor dependent.

If the @var{@1@} and @var{@2@} are not the same length,
the shorter is compared as if spaces were appended to
it to form a value that has the same length as the longer.

The lexical comparison intrinsics @code{LGe}, @code{LGt},
@code{LLe}, and @code{LLt} differ from the corresponding
intrinsic operators @code{.GE.}, @code{.GT.},
@code{.LE.}, @code{.LT.}.
Because the ASCII collating sequence is assumed,
the following expressions always return @samp{.TRUE.}:

@smallexample
LGE ('0', ' ')
LGE ('A', '0')
LGE ('a', 'A')
@end smallexample

The following related expressions do @emph{not} always
return @samp{.TRUE.}, as they are not necessarily evaluated
assuming the arguments use ASCII encoding:

@smallexample
'0' .GE. ' '
'A' .GE. '0'
'a' .GE. 'A'
@end smallexample

The same difference exists
between @code{LGt} and @code{.GT.};
between @code{LLe} and @code{.LE.}; and
between @code{LLt} and @code{.LT.}.
")

DEFDOC (LGT, "Lexically greater than.", "\
Returns @samp{.TRUE.} if @samp{@var{@1@}.GT.@var{@2@}},
@samp{.FALSE.} otherwise.
@var{@1@} and @var{@2@} are interpreted as containing
ASCII character codes.
If either value contains a character not in the ASCII
character set, the result is processor dependent.

If the @var{@1@} and @var{@2@} are not the same length,
the shorter is compared as if spaces were appended to
it to form a value that has the same length as the longer.

@xref{LGe Intrinsic}, for information on the distinction
between the @code{@0@} intrinsic and the @code{.GT.}
operator.
")

DEFDOC (LLE, "Lexically less than or equal.", "\
Returns @samp{.TRUE.} if @samp{@var{@1@}.LE.@var{@2@}},
@samp{.FALSE.} otherwise.
@var{@1@} and @var{@2@} are interpreted as containing
ASCII character codes.
If either value contains a character not in the ASCII
character set, the result is processor dependent.

If the @var{@1@} and @var{@2@} are not the same length,
the shorter is compared as if spaces were appended to
it to form a value that has the same length as the longer.

@xref{LGe Intrinsic}, for information on the distinction
between the @code{@0@} intrinsic and the @code{.LE.}
operator.
")

DEFDOC (LLT, "Lexically less than.", "\
Returns @samp{.TRUE.} if @samp{@var{@1@}.LT.@var{@2@}},
@samp{.FALSE.} otherwise.
@var{@1@} and @var{@2@} are interpreted as containing
ASCII character codes.
If either value contains a character not in the ASCII
character set, the result is processor dependent.

If the @var{@1@} and @var{@2@} are not the same length,
the shorter is compared as if spaces were appended to
it to form a value that has the same length as the longer.

@xref{LGe Intrinsic}, for information on the distinction
between the @code{@0@} intrinsic and the @code{.LT.}
operator.
")

DEFDOC (SIGN, "Apply sign to magnitude.", "\
Returns @samp{ABS(@var{@1@})*@var{s}}, where
@var{s} is +1 if @samp{@var{@2@}.GE.0},
-1 otherwise.

@xref{Abs Intrinsic}, for the function that returns
the magnitude of a value.
")