- This patch uses gdtoa imported from OpenBSD if newlib configure
option "--enable-newlib-use-gdtoa=no" is NOT specified. gdtoa
provides more accurate output and faster conversion than legacy
ldtoa, while it requires more heap memory.
- If the number has large integer part and small fraction part is
specified in output format, e.g. printf("%.3f", sqrt(2)*1e60);,
valid output digits were insufficient. This patch fixes the issue.
https://cygwin.com/pipermail/cygwin/2021-November/249930.html
reported a regression introduce by using a dynamically sized local
char array in favor of a statically sized array.
Fix this by reverting to a statically sized array, using a small
buffer on the stack for a reasonable number of requested digits, a
big mallocated buffer otherwise. This should work for small targets
as well, given that malloc is used in printf anyway right now.
This is *still* hopefully just a temporary measure, unless somebody
actually provides a new ldtoa.
Fixes: 4d90e53359 ("ldtoa: fix dropping too many digits from output")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
ldtoa cuts the number of digits it returns based on a computation of
number of supported bits (144) divide by log10(2). Not only is the
integer approximation of log10(2) ~= 8/27 missing a digit here, it
also fails to take really small double and long double values into
account.
Allow for the full potential precision of long double values. At the
same time, change the local string array allocation to request only as
much bytes as necessary to support the caller-requested number of
digits, to keep the stack size low on small targets.
In the long run a better fix would be to switch to gdtoa, as the BSD
variants, as well as Mingw64 do.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* libc/stdlib/ldtoa (_ldcheck): Make sure the setting of
rnd is done after the last local variable declaration.
2007-06-06 Jeff Johnston <jjohnstn@redhat.com>
* libc/stdlib/mprec.h[Just_16]: Make sure that Pack_16 is defined.
(MATH_ERRNO, MATH_ERREXCEPT, math_errhandling): Add macros
required by POSIX.
* libc/stdlib/ldtoa.c (USE_INFINITY): Rename from INFINITY, to
avoid clash with <math.h>.
* libc/stdio/vfprintf.c (cvt): Don't rely on pointer aliasing to
determine characteristics of long double. Use a union instead.
* ldtoa.c (_ldtoa_r): Ditto.
(_ldcheck): Ditto.
(_strtold): Ditto.
(union uconv): New union.
* libc/stdlib/ldtoa.c (_ldtoa_r): Fix code to allocate the format
buffer based on the precision, after we have processed the input value
in a local buffer and know its relative magnitude.
* configure.host: Support long double I/O for x86-linux.
* libc/stdlib/ldtoa.c (_ldtoa_r): Fix code to allocate a buffer
large enough to hold formatted result.
* libc/machine/powerpc/simdldtoa.c (_simdldtoa_r): Ditto.
* libc/stdlib/ldtoa.c (_ldcheck): New routine
that categorizes a long double as NaN, Infinity, or other.
* libc/stdio/vfprintf.c [WANT_IO_LONG_DBL](_VFPRINTF_R): Removed
isinfl and isnanl static routines which were i386-specific. Changed
calls to the two removed routines to a single _ldcheck call.
* libc/stdio/vfieeefp.h (ldieee): Fixed missing semi-colons.
* libc/stdlib/Makefile.am: Added ldtoa.c to list of sources.
* libc/stdlib/Makefile.in: Regenerated.
* libc/stdio/floatio.h: Added suitable MAXEXP for long double.
* libc/stdio/vfieeefp.h: Added long double bit structures.
* libc/stdio/vfprintf.c[WANT_IO_LONG_DBL]: Added long double support.
[WANT_IO_LONG_DBL](isinfl, isnanl): New static long double routines.
(exponent): Changed expbuf to reasonable maximum instead of MAXEXP.
* libc/stdio/vfscanf.c[WANT_IO_LONG_DBL]: Added long double support.
* libc/stdlib/ldtoa.c: New file containing _ldtoa_r and
_strtold routines used for conversions between character
and long double.
Takashi Yano
Corinna Vinschen
Jeff Johnston
Yaakov Selkowitz
Eric Blake
Matthew Green