(fleven? ix)   -- what's ix?  I infer a flonum with integer value, but

you didn't use the same parameter name in flodd?.

For fldenormalized,

issubnormal is not in C99.   Looks like it was put in to C11.

Presumably, this expression will suffice where issubnormal is not available?

(fpclassify(x) == FP_SUBNORMAL)

In the spec for fl+* you refer to fl-fast-fl+*

..but in the spec for the constant, you spell it "fl-fast-+*"

Also, the users of this SRFI would be better able to make use of fl-fast-+*

if it were known at compile-time.    Would it make more sense to

suggest this symbol be in (features) when relevant?

The constant fl-2/pi, you describe as bound to 2 ** fl-pi, but

I'm sure you meant 2 / fl-pi.

fl-greatest

fl-least

Perhaps even clearer, "...the largest positive/negative finite flonums..."

In the definitions for fl-integer-exponent-zero and

fl-integer-exponent-nan, you refer to flinteger-binary-log which

doesn't exist anymore and can just be deleted.

sign-bit -- did you intend to add this function without an fl-prefix?

The description of fllog10 is lacking a closing parenthesis.