diff options
author | Shiro Kawai <shiro@acm.org> | 2024-03-05 23:41:33 -1000 |
---|---|---|
committer | Shiro Kawai <shiro@acm.org> | 2024-03-05 23:41:33 -1000 |
commit | e7aa04c764b7fb0969bdaa349bb1418845c81d8b (patch) | |
tree | 7cbcf0257c2e5211fb8ecfae0ce0787e095bdb2c | |
parent | b75b9a5e885c13eb4a4440abd08fedb0180ffa78 (diff) |
Update docs for greatest-positive-flonum etc.
-rw-r--r-- | doc/corelib.texi | 37 |
1 files changed, 23 insertions, 14 deletions
diff --git a/doc/corelib.texi b/doc/corelib.texi index 7e0690130..4a1b79947 100644 --- a/doc/corelib.texi +++ b/doc/corelib.texi @@ -3281,7 +3281,7 @@ number, @code{magnitude} is used in place of @code{abs}. When omitted, @var{relative-tolerance} is assumed to be @code{(flonum-epsilon)}, and @var{absolute-tolerance} is -@code{(flonum-min-denormalized)}. +@code{(least-positive-flonum)}. That is, by default, @code{approx=?} tolerates 1 ULP (unit in the last place) error. @@ -3311,7 +3311,7 @@ close to zero, in which case relative tolerance becomes too small. 省略された場合、@code{relative-tolerance}には @code{(flonum-epsilon)}の値が、@var{absolute-tolerance}には -@code{(flonum-min-denormalized)}の値が使われます。 +@code{(least-positive-flonum)}の値が使われます。 すなわち、デフォルトでは@code{approx=?}は 1 ULP (unit in the last place)のエラーを許容するということです。 @@ -3321,21 +3321,26 @@ close to zero, in which case relative tolerance becomes too small. @end defun @defun flonum-epsilon -@defunx flonum-min-normalized -@defunx flonum-min-denormalized +@defunx greatest-positive-flonum +@defunx least-positive-flonum +@defunx least-positive-normalized-flonum @c EN Returns flonums with the following characteristics, respectively: @table @code @item flonum-epsilon Returns the least positive flonum e such that, for a normalized flonum @code{x}, @code{x} and @code{(* x (+ 1.0 e))} are distinguishable. -@item flonum-min-normalized +@item greatest-positive-flonum +Returns the greatest positive finite flonum. It is +@code{DBL_MAX} in C. +@item least-positive-flonum +Returns the least positive flonum representable as floating-point +number. If the platform supports denormalized flonum, +it returns the least positive denormalized floating-point number. Otherwise, +it returns the least positive normalized floating-point number. +@item least-positive-normalized-flonum Returns the least positive flonum representable as normalized floating-point -number. -@item flonum-min-denormalized -Returns the least positive flonum representable as denormalized floating-point -number. If the platform doesn't support denormalized flonum, -it returns the least positive normalized floating number. +number. It is @code{DBL_MIN} in C. @end table @c JP それぞれ次の性質を持つ浮動小数点数を返します: @@ -3343,11 +3348,15 @@ it returns the least positive normalized floating number. @item flonum-epsilon 正規化浮動小数点数@code{x}に対し、@code{x}と@code{(* x (+ 1.0 e))}が 異なる値となるような最小の正の浮動小数点数。 -@item flonum-min-normalized +@item greatest-positive-flonum +有限な最大の正の浮動小数点数。Cで言うところの@code{DBL_MAX}。 +@item least-positive-flonum +最小の正の浮動小数点数。 +プラットフォームが非正規化浮動小数点数をサポートしている場合は最小の +正の非正規化浮動小数点数が、そうでなければ +最小の正の正規化浮動小数点数が返ります。 +@item least-positive-normalized-flonum 最小の正の正規化浮動小数点数。 -@item flonum-min-denormalized -最小の正の非正規化浮動小数点数。但しプラットフォームが非正規化浮動小数点数を -サポートしない場合は、最小の正の正規化浮動小数点数。 @end table @c COMMON @end defun |