Re: float scanner status, upcoming release

[Szabolcs Nagy <nsz@port70.net>]

* Rich Felker <dalias@aerifal.cx> [2012-04-09 20:48:19 -0400]:
> Revised version of the code, in preparation to integrate with musl.
> Still needs hex float support, and I want to eliminate the fscanf
> dependency. Comments welcome.
> 

this one fails if the number starts with a .
./floatscan .3

large exponent is slow:
./floatscan 1e+1000000
./floatscan 1e-1000000

subnormal tests still fail

here are some test cases where many digits are needed for correct rounding:

long double (ld80, min subnormal is 2^-16445, min normal is 2^-16382):

// 2^-16445 * 0.5 - eps (should be rounded to 0)
.1822599765941237301264202966809709908199525407846781671860490243514185844316698e-4950
// 2^-16445 * 0.5 + eps (should be rounded to 0x1p-16445L)
.1822599765941237301264202966809709908199525407846781671860490243514185844316699e-4950
// 2^-16445 * 1.5 - eps (should be rounded to 0x1p-16445L)
.5467799297823711903792608900429129724598576223540345015581470730542557532950096e-4950
// 2^-16445 * 1.5 + eps (should be rounded to 0x1p-16444L)
.5467799297823711903792608900429129724598576223540345015581470730542557532950097e-4950
// 2^-16382 + 2^-16446 - eps (should be rounded to 0x1p-16382L)
.3362103143112093506444937793915876332724499641527442230928779770593420866576777e-4931
// 2^-16382 + 2^-16446 + eps (should be rounded to 0x1.0000000000000002p-16382)
.3362103143112093506444937793915876332724499641527442230928779770593420866576778e-4931

double (min subnormal is 2^-1074, min normal is 2^-1022):

// 2^-1074 * 0.5 - eps
.2470328229206232720882843964341106861825299013071623822127928412503377536351043e-323
// 2^-1074 * 0.5 + epsA
.2470328229206232720882843964341106861825299013071623822127928412503377536351044e-323
// 2^-1074 * 1.5 - eps
.7410984687618698162648531893023320585475897039214871466383785237510132609053131e-323
// 2^-1074 * 1.5 + eps
.7410984687618698162648531893023320585475897039214871466383785237510132609053132e-323
// 2^-1022 + 2^-1075 - eps
.2225073858507201630123055637955676152503612414573018013083228724049586647606759e-307
// 2^-1022 + 2^-1075 + eps
.2225073858507201630123055637955676152503612414573018013083228724049586647606760e-307

float (min subnormal is 2^-149, min normal is 2^-126):
// 2^-149 * 0.5 - eps
.7006492321624085354618647916449580656401309709382578858785341419448955413429303e-45
// 2^-149 * 0.5 + eps
.7006492321624085354618647916449580656401309709382578858785341419448955413429304e-45
// 2^-149 * 0.5 - eps
.2101947696487225606385594374934874196920392912814773657635602425834686624028790e-44
// 2^-149 * 0.5 + eps
.2101947696487225606385594374934874196920392912814773657635602425834686624028791e-44
// 2^-126 + 2^-150 - eps
.1175494420887210724209590083408724842314472120785184615334540294131831453944281e-37
// 2^-126 + 2^-150 + eps
.1175494420887210724209590083408724842314472120785184615334540294131831453944282e-37