2012-08-27 16:01:27 +00:00
|
|
|
/* Generate table of tests in tst-strtod-round.c from
|
|
|
|
tst-strtod-round-data.
|
2022-01-01 18:54:23 +00:00
|
|
|
Copyright (C) 2012-2022 Free Software Foundation, Inc.
|
2012-08-27 16:01:27 +00:00
|
|
|
This file is part of the GNU C Library.
|
|
|
|
|
|
|
|
The GNU C Library is free software; you can redistribute it and/or
|
|
|
|
modify it under the terms of the GNU Lesser General Public
|
|
|
|
License as published by the Free Software Foundation; either
|
|
|
|
version 2.1 of the License, or (at your option) any later version.
|
|
|
|
|
|
|
|
The GNU C Library is distributed in the hope that it will be useful,
|
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
Lesser General Public License for more details.
|
|
|
|
|
|
|
|
You should have received a copy of the GNU Lesser General Public
|
|
|
|
License along with the GNU C Library; if not, see
|
Prefer https to http for gnu.org and fsf.org URLs
Also, change sources.redhat.com to sourceware.org.
This patch was automatically generated by running the following shell
script, which uses GNU sed, and which avoids modifying files imported
from upstream:
sed -ri '
s,(http|ftp)(://(.*\.)?(gnu|fsf|sourceware)\.org($|[^.]|\.[^a-z])),https\2,g
s,(http|ftp)(://(.*\.)?)sources\.redhat\.com($|[^.]|\.[^a-z]),https\2sourceware.org\4,g
' \
$(find $(git ls-files) -prune -type f \
! -name '*.po' \
! -name 'ChangeLog*' \
! -path COPYING ! -path COPYING.LIB \
! -path manual/fdl-1.3.texi ! -path manual/lgpl-2.1.texi \
! -path manual/texinfo.tex ! -path scripts/config.guess \
! -path scripts/config.sub ! -path scripts/install-sh \
! -path scripts/mkinstalldirs ! -path scripts/move-if-change \
! -path INSTALL ! -path locale/programs/charmap-kw.h \
! -path po/libc.pot ! -path sysdeps/gnu/errlist.c \
! '(' -name configure \
-execdir test -f configure.ac -o -f configure.in ';' ')' \
! '(' -name preconfigure \
-execdir test -f preconfigure.ac ';' ')' \
-print)
and then by running 'make dist-prepare' to regenerate files built
from the altered files, and then executing the following to cleanup:
chmod a+x sysdeps/unix/sysv/linux/riscv/configure
# Omit irrelevant whitespace and comment-only changes,
# perhaps from a slightly-different Autoconf version.
git checkout -f \
sysdeps/csky/configure \
sysdeps/hppa/configure \
sysdeps/riscv/configure \
sysdeps/unix/sysv/linux/csky/configure
# Omit changes that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/powerpc/powerpc64/ppc-mcount.S: trailing lines
git checkout -f \
sysdeps/powerpc/powerpc64/ppc-mcount.S \
sysdeps/unix/sysv/linux/s390/s390-64/syscall.S
# Omit change that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S: last line does not end in newline
git checkout -f sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S
2019-09-07 05:40:42 +00:00
|
|
|
<https://www.gnu.org/licenses/>. */
|
2012-08-27 16:01:27 +00:00
|
|
|
|
2016-05-09 16:49:11 +00:00
|
|
|
/* Compile this program as:
|
|
|
|
|
2016-05-09 22:59:37 +00:00
|
|
|
gcc -std=gnu11 -O2 -Wall -Wextra gen-tst-strtod-round.c -lmpfr \
|
2016-05-09 16:49:11 +00:00
|
|
|
-o gen-tst-strtod-round
|
|
|
|
|
|
|
|
(use of current MPFR version recommended) and run it as:
|
|
|
|
|
|
|
|
gen-tst-strtod-round tst-strtod-round-data tst-strtod-round-data.h
|
|
|
|
|
|
|
|
The output file will be generated as tst-strtod-round-data.h
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
2012-08-27 16:01:27 +00:00
|
|
|
#define _GNU_SOURCE
|
2012-09-06 15:33:06 +00:00
|
|
|
#include <assert.h>
|
2012-08-27 16:01:27 +00:00
|
|
|
#include <stdbool.h>
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <stdlib.h>
|
2016-05-09 16:49:11 +00:00
|
|
|
#include <string.h>
|
2012-08-27 16:01:27 +00:00
|
|
|
#include <mpfr.h>
|
|
|
|
|
|
|
|
/* Work around incorrect ternary value from mpfr_strtofr
|
|
|
|
<https://sympa.inria.fr/sympa/arc/mpfr/2012-08/msg00005.html>. */
|
|
|
|
#define WORKAROUND
|
|
|
|
|
|
|
|
static int
|
|
|
|
string_to_fp (mpfr_t f, const char *s, mpfr_rnd_t rnd)
|
|
|
|
{
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
mpfr_clear_overflow ();
|
2012-08-27 16:01:27 +00:00
|
|
|
#ifdef WORKAROUND
|
|
|
|
mpfr_t f2;
|
|
|
|
mpfr_init2 (f2, 100000);
|
|
|
|
int r0 = mpfr_strtofr (f2, s, NULL, 0, rnd);
|
|
|
|
int r = mpfr_set (f, f2, rnd);
|
2016-10-28 00:40:23 +00:00
|
|
|
r |= mpfr_subnormalize (f, r, rnd);
|
2012-08-27 16:01:27 +00:00
|
|
|
mpfr_clear (f2);
|
|
|
|
return r0 | r;
|
|
|
|
#else
|
|
|
|
int r = mpfr_strtofr (f, s, NULL, 0, rnd);
|
2016-10-28 00:40:23 +00:00
|
|
|
r |= mpfr_subnormalize (f, r, rnd);
|
2012-08-27 16:01:27 +00:00
|
|
|
return r;
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2016-05-09 22:59:37 +00:00
|
|
|
void
|
|
|
|
print_fp (FILE *fout, mpfr_t f, const char *suffix)
|
2012-08-27 16:01:27 +00:00
|
|
|
{
|
|
|
|
if (mpfr_inf_p (f))
|
2016-05-09 22:59:37 +00:00
|
|
|
mpfr_fprintf (fout, "\t%sINF%s", mpfr_signbit (f) ? "-" : "", suffix);
|
2012-08-27 16:01:27 +00:00
|
|
|
else
|
2016-05-09 22:59:37 +00:00
|
|
|
mpfr_fprintf (fout, "\t%Ra%s", f, suffix);
|
2012-08-27 16:01:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2016-05-09 22:59:37 +00:00
|
|
|
round_str (FILE *fout, const char *s, int prec, int emin, int emax,
|
|
|
|
bool ibm_ld)
|
2012-08-27 16:01:27 +00:00
|
|
|
{
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
mpfr_t max_value;
|
2012-08-27 16:01:27 +00:00
|
|
|
mpfr_t f;
|
|
|
|
mpfr_set_default_prec (prec);
|
|
|
|
mpfr_set_emin (emin);
|
|
|
|
mpfr_set_emax (emax);
|
|
|
|
mpfr_init (f);
|
|
|
|
int r = string_to_fp (f, s, MPFR_RNDD);
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
bool overflow = mpfr_overflow_p () != 0;
|
2013-06-13 15:41:58 +00:00
|
|
|
if (ibm_ld)
|
2012-09-06 15:33:06 +00:00
|
|
|
{
|
|
|
|
assert (prec == 106 && emin == -1073 && emax == 1024);
|
|
|
|
/* The maximum value in IBM long double has discontiguous
|
|
|
|
mantissa bits. */
|
|
|
|
mpfr_init2 (max_value, 107);
|
|
|
|
mpfr_set_str (max_value, "0x1.fffffffffffff7ffffffffffffcp+1023", 0,
|
|
|
|
MPFR_RNDN);
|
|
|
|
if (mpfr_cmpabs (f, max_value) > 0)
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
{
|
|
|
|
r = 1;
|
|
|
|
overflow = true;
|
|
|
|
}
|
2012-09-06 15:33:06 +00:00
|
|
|
}
|
2016-05-09 16:49:11 +00:00
|
|
|
mpfr_fprintf (fout, "\t%s,\n", r ? "false" : "true");
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
print_fp (fout, f, overflow ? ", true,\n" : ", false,\n");
|
2012-08-27 16:01:27 +00:00
|
|
|
string_to_fp (f, s, MPFR_RNDN);
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
overflow = (mpfr_overflow_p () != 0
|
|
|
|
|| (ibm_ld && mpfr_cmpabs (f, max_value) > 0));
|
|
|
|
print_fp (fout, f, overflow ? ", true,\n" : ", false,\n");
|
2012-08-27 16:01:27 +00:00
|
|
|
string_to_fp (f, s, MPFR_RNDZ);
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
overflow = (mpfr_overflow_p () != 0
|
|
|
|
|| (ibm_ld && mpfr_cmpabs (f, max_value) > 0));
|
|
|
|
print_fp (fout, f, overflow ? ", true,\n" : ", false,\n");
|
2012-08-27 16:01:27 +00:00
|
|
|
string_to_fp (f, s, MPFR_RNDU);
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
overflow = (mpfr_overflow_p () != 0
|
|
|
|
|| (ibm_ld && mpfr_cmpabs (f, max_value) > 0));
|
|
|
|
print_fp (fout, f, overflow ? ", true" : ", false");
|
2012-08-27 16:01:27 +00:00
|
|
|
mpfr_clear (f);
|
Fix strtod overflow detection (bug 23279).
As shown by bug 23279, strtod's round_and_return has an off-by-one
error in its overflow detection, only counting an exponent greater
than MAX_EXP as overflowing when an exponent of MAX_EXP also means
overflow (recall the ISO C definition of DBL_MAX_EXP etc. is based on
a floating-point model where 2^exp is multiplied by a value in the
interval [0.5, 1), so 2^MAX_EXP is not representable).
For decimal arguments to strtod, a separate overflow check in the main
implementation covers the case where the integer part of the argument
(truncated to the nearest integer towards zero) has more than MAX_EXP
bits, meaning that this issue in round_and_return only affects cases
(arguments with absolute value strictly between the maximum
representable value and 2^MAX_EXP) where overflow depends on the
rounding mode; in such cases, the returned value would still have been
correct on overflow but without the overflow exception being raised or
errno being set to ERANGE. For hex float arguments, however, other
cases can arise, as shown in bug 23279, where a value with exponent
already set to MAX_EXP is passed into round_and_return and a result
can wrongly end up being NaN, or infinity instead of the largest
finite value.
This patch fixes the off-by-one error, adds testing of overflow
exceptions to the tst-strtod-round framework, and adds tests of these
issues.
Tested for x86_64. Also ran the tst-strtod-round tests for powerpc to
make sure the new tests didn't introduce any new failures for IBM long
double.
[BZ #23279]
* stdlib/strtod_l.c (round_and_return): Handle an exponent of
MAX_EXP as overflowing.
* stdlib/gen-tst-strtod-round.c (string_to_fp): Clear MPFR
overflow flag.
(round_str): Output also whether result overflows in each rounding
mode.
* stdlib/tst-strtod-round-data: Add more tests.
* stdlib/tst-strtod-round-data.h: Regenerated.
* stdlib/tst-strtod-round-skeleton.c (_XNTRY): Update comment.
(TEST): Handle extra arguments for overflow flags.
(struct test_overflow): New type.
[!FE_OVERFLOW] (FE_OVERFLOW): Define to 0.
(GEN_ONE_TEST): Clear all exceptions. Test overflow flag.
(test_in_one_mode): Take argument with overflow information.
(do_test): Update calls to test_in_one_mode.
2018-06-13 16:06:14 +00:00
|
|
|
if (ibm_ld)
|
|
|
|
mpfr_clear (max_value);
|
2012-08-27 16:01:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2016-05-09 16:49:11 +00:00
|
|
|
round_for_all (FILE *fout, const char *s)
|
2012-08-27 16:01:27 +00:00
|
|
|
{
|
|
|
|
static const struct fmt {
|
|
|
|
int prec;
|
|
|
|
int emin;
|
|
|
|
int emax;
|
2013-06-13 15:41:58 +00:00
|
|
|
bool ibm_ld;
|
2016-05-09 22:59:37 +00:00
|
|
|
} formats[] = {
|
|
|
|
{ 24, -148, 128, false },
|
|
|
|
{ 53, -1073, 1024, false },
|
2012-09-05 10:37:16 +00:00
|
|
|
/* This is the Intel extended float format. */
|
2016-05-09 22:59:37 +00:00
|
|
|
{ 64, -16444, 16384, false },
|
2012-09-05 10:37:16 +00:00
|
|
|
/* This is the Motorola extended float format. */
|
2016-05-09 22:59:37 +00:00
|
|
|
{ 64, -16445, 16384, false },
|
|
|
|
{ 106, -1073, 1024, true },
|
|
|
|
{ 113, -16493, 16384, false },
|
2012-08-27 16:01:27 +00:00
|
|
|
};
|
2016-05-09 16:49:11 +00:00
|
|
|
mpfr_fprintf (fout, " TEST (\"");
|
2012-08-27 16:01:27 +00:00
|
|
|
const char *p;
|
|
|
|
for (p = s; *p; p++)
|
|
|
|
{
|
2016-05-09 16:49:11 +00:00
|
|
|
fputc (*p, fout);
|
2012-08-27 16:01:27 +00:00
|
|
|
if ((p - s) % 60 == 59 && p[1])
|
2016-05-09 16:49:11 +00:00
|
|
|
mpfr_fprintf (fout, "\"\n\t\"");
|
2012-08-27 16:01:27 +00:00
|
|
|
}
|
2016-05-09 16:49:11 +00:00
|
|
|
mpfr_fprintf (fout, "\",\n");
|
2012-08-27 16:01:27 +00:00
|
|
|
int i;
|
2016-05-09 22:59:37 +00:00
|
|
|
int n_formats = sizeof (formats) / sizeof (formats[0]);
|
|
|
|
for (i = 0; i < n_formats; i++)
|
2012-08-27 16:01:27 +00:00
|
|
|
{
|
2016-05-09 22:59:37 +00:00
|
|
|
round_str (fout, s, formats[i].prec, formats[i].emin,
|
|
|
|
formats[i].emax, formats[i].ibm_ld);
|
|
|
|
if (i < n_formats - 1)
|
2016-05-09 16:49:11 +00:00
|
|
|
mpfr_fprintf (fout, ",\n");
|
2012-08-27 16:01:27 +00:00
|
|
|
}
|
2016-05-09 16:49:11 +00:00
|
|
|
mpfr_fprintf (fout, "),\n");
|
2012-08-27 16:01:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
int
|
2016-05-09 16:49:11 +00:00
|
|
|
main (int argc, char **argv)
|
2012-08-27 16:01:27 +00:00
|
|
|
{
|
|
|
|
char *p = NULL;
|
|
|
|
size_t len;
|
|
|
|
ssize_t nbytes;
|
2016-05-09 16:49:11 +00:00
|
|
|
FILE *fin, *fout;
|
|
|
|
char *fin_name, *fout_name;
|
|
|
|
|
|
|
|
if (argc < 3)
|
|
|
|
{
|
|
|
|
fprintf (stderr, "Usage: %s <input> <output>\n", basename (argv[0]));
|
|
|
|
return EXIT_FAILURE;
|
|
|
|
}
|
|
|
|
|
|
|
|
fin_name = argv[1];
|
|
|
|
fout_name = argv[2];
|
|
|
|
|
|
|
|
fin = fopen (fin_name, "r");
|
|
|
|
if (fin == NULL)
|
|
|
|
{
|
|
|
|
perror ("Could not open input for reading");
|
|
|
|
return EXIT_FAILURE;
|
|
|
|
}
|
|
|
|
|
|
|
|
fout = fopen (fout_name, "w");
|
|
|
|
if (fout == NULL)
|
|
|
|
{
|
|
|
|
perror ("Could not open output for writing");
|
|
|
|
return EXIT_FAILURE;
|
|
|
|
}
|
|
|
|
|
|
|
|
fprintf (fout, "/* This file was generated by %s from %s. */\n",
|
|
|
|
__FILE__, fin_name);
|
|
|
|
fputs ("static const struct test tests[] = {\n", fout);
|
|
|
|
while ((nbytes = getline (&p, &len, fin)) != -1)
|
2012-08-27 16:01:27 +00:00
|
|
|
{
|
|
|
|
if (p[nbytes - 1] == '\n')
|
|
|
|
p[nbytes - 1] = 0;
|
2016-05-09 16:49:11 +00:00
|
|
|
round_for_all (fout, p);
|
2012-08-27 16:01:27 +00:00
|
|
|
free (p);
|
|
|
|
p = NULL;
|
|
|
|
}
|
2016-05-09 16:49:11 +00:00
|
|
|
fputs ("};\n", fout);
|
|
|
|
|
|
|
|
return EXIT_SUCCESS;
|
2012-08-27 16:01:27 +00:00
|
|
|
}
|