* BZ#2153: cacosh bug
@ 2006-01-15 15:44 Andreas Jaeger
2006-01-15 17:25 ` Ulrich Drepper
0 siblings, 1 reply; 4+ messages in thread
From: Andreas Jaeger @ 2006-01-15 15:44 UTC (permalink / raw)
To: Glibc hackers
[-- Attachment #1: Type: text/plain, Size: 12448 bytes --]
The appended patch fixes BZ#2153. Ok to commit?
I only updated the ulps file for x86-64, other archs might need a
similar change.
Andreas
2006-01-15 Andreas Jaeger <aj@suse.de>
* sysdeps/x86_64/fpu/libm-test-ulps: Updated.
BZ #2153
* math/s_cacosh.c (__cacosh): Do not return a negative
value. Patch by Wes Loewer <wjltemp-temp01@yahoo.com>.
* math/s_cacoshl.c (__cacoshl): Likewise.
* math/s_cacoshf.c (__cacoshf): Likewise.
* math/libm-test.inc (cacosh_test): Add new testcases and fix
existing ones.
============================================================
Index: math/s_cacosh.c
--- math/s_cacosh.c 14 Dec 2005 10:12:27 -0000 1.1
+++ math/s_cacosh.c 15 Jan 2006 15:43:35 -0000
@@ -1,5 +1,5 @@
/* Return arc hyperbole cosine for double value.
- Copyright (C) 1997 Free Software Foundation, Inc.
+ Copyright (C) 1997, 2006 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
@@ -73,6 +73,9 @@ __cacosh (__complex__ double x)
y = __csqrt (y);
+ if (__real__ x < 0.0)
+ y = -y;
+
__real__ y += __real__ x;
__imag__ y += __imag__ x;
============================================================
Index: math/s_cacoshf.c
--- math/s_cacoshf.c 14 Dec 2005 10:12:45 -0000 1.1
+++ math/s_cacoshf.c 15 Jan 2006 15:43:35 -0000
@@ -1,5 +1,5 @@
/* Return arc hyperbole cosine for float value.
- Copyright (C) 1997 Free Software Foundation, Inc.
+ Copyright (C) 1997, 2006 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
@@ -75,6 +75,9 @@ __cacoshf (__complex__ float x)
y = __csqrtf (y);
+ if (__real__ x < 0.0)
+ y = -y;
+
__real__ y += __real__ x;
__imag__ y += __imag__ x;
============================================================
Index: math/s_cacoshl.c
--- math/s_cacoshl.c 14 Dec 2005 10:13:05 -0000 1.1
+++ math/s_cacoshl.c 15 Jan 2006 15:43:35 -0000
@@ -1,5 +1,5 @@
/* Return arc hyperbole cosine for long double value.
- Copyright (C) 1997, 1998 Free Software Foundation, Inc.
+ Copyright (C) 1997, 1998, 2006 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
@@ -73,6 +73,9 @@ __cacoshl (__complex__ long double x)
y = __csqrtl (y);
+ if (__real__ x < 0.0)
+ y = -y;
+
__real__ y += __real__ x;
__imag__ y += __imag__ x;
============================================================
Index: math/libm-test.inc
--- math/libm-test.inc 13 Oct 2005 19:07:12 -0000 1.66
+++ math/libm-test.inc 15 Jan 2006 15:43:36 -0000
@@ -1,4 +1,4 @@
-/* Copyright (C) 1997-2002, 2003, 2004, 2005 Free Software Foundation, Inc.
+/* Copyright (C) 1997-2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Andreas Jaeger <aj@suse.de>, 1997.
@@ -1116,7 +1116,11 @@ cacosh_test (void)
TEST_c_c (cacosh, nan_value, nan_value, nan_value, nan_value);
TEST_c_c (cacosh, 0.75L, 1.25L, 1.13239363160530819522266333696834467L, 1.11752014915610270578240049553777969L);
- TEST_c_c (cacosh, -2, -3, -1.9833870299165354323470769028940395L, 2.1414491111159960199416055713254211L);
+ TEST_c_c (cacosh, -2, -3, 1.9833870299165354323470769028940395L, -2.1414491111159960199416055713254211L);
+ TEST_c_c (cacosh, 0.3, 0.4, 0.405112337178030872507338405200229579L, 1.29016676450309079292712537938218583L);
+ TEST_c_c (cacosh, -0.3, 0.4, 0.405112337178030872507338405200229579L, 1.85142588908670244553551800389731705L);
+ TEST_c_c (cacosh, 0.3, -0.4, 0.405112337178030872507338405200229579L, -1.29016676450309079292712537938218583L);
+ TEST_c_c (cacosh, -0.3, -0.4, 0.405112337178030872507338405200229579L, -1.85142588908670244553551800389731705L);
END (cacosh, complex);
}
============================================================
Index: sysdeps/x86_64/fpu/libm-test-ulps
--- sysdeps/x86_64/fpu/libm-test-ulps 20 Jul 2005 18:15:21 -0000 1.10
+++ sysdeps/x86_64/fpu/libm-test-ulps 15 Jan 2006 15:43:36 -0000
@@ -48,20 +48,54 @@ ildouble: 2
ldouble: 2
# cacosh
-Test "Real part of: cacosh (-2 - 3 i) == -1.9833870299165354323470769028940395 + 2.1414491111159960199416055713254211 i":
+Test "Real part of: cacosh (-0.3 + 0.4 i) == 0.405112337178030872507338405200229579 + 1.85142588908670244553551800389731705 i":
+float: 1
+ifloat: 1
+ildouble: 735
+ldouble: 735
+Test "Imaginary part of: cacosh (-0.3 + 0.4 i) == 0.405112337178030872507338405200229579 + 1.85142588908670244553551800389731705 i":
+ildouble: 119
+ldouble: 119
+Test "Real part of: cacosh (-0.3 - 0.4 i) == 0.405112337178030872507338405200229579 - 1.85142588908670244553551800389731705 i":
+float: 1
+ifloat: 1
+ildouble: 735
+ldouble: 735
+Test "Imaginary part of: cacosh (-0.3 - 0.4 i) == 0.405112337178030872507338405200229579 - 1.85142588908670244553551800389731705 i":
+ldouble: 119
+Test "Real part of: cacosh (-2 - 3 i) == 1.9833870299165354323470769028940395 + 2.1414491111159960199416055713254211 i":
double: 1
float: 7
idouble: 1
ifloat: 7
ildouble: 6
ldouble: 6
-Test "Imaginary part of: cacosh (-2 - 3 i) == -1.9833870299165354323470769028940395 + 2.1414491111159960199416055713254211 i":
+Test "Imaginary part of: cacosh (-2 - 3 i) == 1.9833870299165354323470769028940395 - 2.1414491111159960199416055713254211 i":
double: 1
float: 3
idouble: 1
ifloat: 3
ildouble: 1
ldouble: 1
+Test "Real part of: cacosh (0.3 + 0.4 i) == 0.405112337178030872507338405200229579 + 1.29016676450309079292712537938218583 i":
+float: 1
+ifloat: 1
+ildouble: 735
+ldouble: 735
+Test "Imaginary part of: cacosh (0.3 + 0.4 i) == 0.405112337178030872507338405200229579 + 1.29016676450309079292712537938218583 i":
+ildouble: 118
+ldouble: 118
+Test "Real part of: cacosh (0.3 - 0.4 i) == 0.405112337178030872507338405200229579 - 1.29016676450309079292712537938218583 i":
+float: 1
+ifloat: 1
+ildouble: 735
+ldouble: 735
+Test "Imaginary part of: cacosh (0.3 - 0.4 i) == 0.405112337178030872507338405200229579 - 1.29016676450309079292712537938218583 i":
+ildouble: 118
+ldouble: 118
+Test "Imaginary part of: cacosh (-0.3 - 0.4 i) == 0.405112337178030872507338405200229579 - 1.85142588908670244553551800389731705 i":
+ildouble: 119
+ldouble: 119
Test "Real part of: cacosh (0.75 + 1.25 i) == 1.13239363160530819522266333696834467 + 1.11752014915610270578240049553777969 i":
ildouble: 1
ldouble: 1
@@ -103,8 +137,8 @@ double: 1
float: 1
idouble: 1
ifloat: 1
-ldouble: 1
ildouble: 1
+ldouble: 1
# catan
Test "Real part of: catan (-2 - 3 i) == -1.4099210495965755225306193844604208 - 0.22907268296853876629588180294200276 i":
@@ -168,10 +202,10 @@ ifloat: 1
ildouble: 1
ldouble: 1
Test "Imaginary part of: ccos (0.75 + 1.25 i) == 1.38173873063425888530729933139078645 - 1.09193013555397466170919531722024128 i":
-ildouble: 1
-ldouble: 1
float: 1
ifloat: 1
+ildouble: 1
+ldouble: 1
# ccosh
Test "Real part of: ccosh (-2 - 3 i) == -3.72454550491532256547397070325597253 + 0.511822569987384608834463849801875634 i":
@@ -304,6 +338,9 @@ idouble: 1
ifloat: 1
# cos
+Test "cos (0.80190127184058835) == 0.69534156199418473":
+double: 1
+idouble: 1
Test "cos (M_PI_6l * 2.0) == 0.5":
double: 1
float: 1
@@ -323,16 +360,13 @@ idouble: 1
ifloat: 1
ildouble: 1
ldouble: 1
-Test "cos (0.80190127184058835) == 0.69534156199418473":
-double: 1
-idouble: 1
# cpow
Test "Real part of: cpow (0.75 + 1.25 i, 0.0 + 1.0 i) == 0.331825439177608832276067945276730566 + 0.131338600281188544930936345230903032 i":
float: 1
ifloat: 1
-ldouble: 1
ildouble: 1
+ldouble: 1
Test "Imaginary part of: cpow (0.75 + 1.25 i, 0.0 + 1.0 i) == 0.331825439177608832276067945276730566 + 0.131338600281188544930936345230903032 i":
float: 1
ifloat: 1
@@ -380,15 +414,15 @@ ildouble: 1
ldouble: 1
# csin
+Test "Imaginary part of: csin (-2 - 3 i) == -9.15449914691142957346729954460983256 + 4.16890695996656435075481305885375484 i":
+double: 1
+idouble: 1
Test "Real part of: csin (0.75 + 1.25 i) == 1.28722291002649188575873510790565441 + 1.17210635989270256101081285116138863 i":
ildouble: 1
ldouble: 1
Test "Imaginary part of: csin (0.75 + 1.25 i) == 1.28722291002649188575873510790565441 + 1.17210635989270256101081285116138863 i":
float: 1
ifloat: 1
-Test "Imaginary part of: csin (-2 - 3 i) == -9.15449914691142957346729954460983256 + 4.16890695996656435075481305885375484 i":
-double: 1
-idouble: 1
# csinh
Test "Real part of: csinh (-2 - 3 i) == 3.59056458998577995201256544779481679 - 0.530921086248519805267040090660676560 i":
@@ -440,12 +474,12 @@ ldouble: 3
# ctanh
Test "Real part of: ctanh (-2 - 3 i) == -0.965385879022133124278480269394560686 + 0.988437503832249372031403430350121098e-2 i":
+double: 1
float: 2
+idouble: 1
ifloat: 2
ildouble: 5
ldouble: 5
-double: 1
-idouble: 1
Test "Imaginary part of: ctanh (-2 - 3 i) == -0.965385879022133124278480269394560686 + 0.988437503832249372031403430350121098e-2 i":
ildouble: 25
ldouble: 25
@@ -456,10 +490,10 @@ Test "Real part of: ctanh (0.75 + 1.25 i
double: 1
idouble: 1
Test "Imaginary part of: ctanh (0.75 + 1.25 i) == 1.37260757053378320258048606571226857 + 0.385795952609750664177596760720790220 i":
-ildouble: 1
-ldouble: 1
double: 1
idouble: 1
+ildouble: 1
+ldouble: 1
# erf
Test "erf (1.25) == 0.922900128256458230136523481197281140":
@@ -481,26 +515,26 @@ ldouble: 1
# exp10
Test "exp10 (-1) == 0.1":
+double: 2
+float: 1
+idouble: 2
+ifloat: 1
ildouble: 1
ldouble: 1
+Test "exp10 (0.75) == 5.62341325190349080394951039776481231":
+double: 1
float: 1
+idouble: 1
ifloat: 1
-double: 2
-idouble: 2
-Test "exp10 (0.75) == 5.62341325190349080394951039776481231":
ildouble: 2
ldouble: 2
-float: 1
-ifloat: 1
-double: 1
-idouble: 1
Test "exp10 (3) == 1000":
-ildouble: 8
-ldouble: 8
-float: 2
-ifloat: 2
double: 6
+float: 2
idouble: 6
+ifloat: 2
+ildouble: 8
+ldouble: 8
# expm1
Test "expm1 (0.75) == 1.11700001661267466854536981983709561":
@@ -714,12 +748,12 @@ ldouble: 1
# log10
Test "log10 (0.75) == -0.124938736608299953132449886193870744":
-ildouble: 1
-ldouble: 1
-float: 2
-ifloat: 2
double: 1
+float: 2
idouble: 1
+ifloat: 2
+ildouble: 1
+ldouble: 1
Test "log10 (e) == log10(e)":
float: 1
ifloat: 1
@@ -732,6 +766,9 @@ float: 1
ifloat: 1
# sincos
+Test "sincos (0.80190127184058835, &sin_res, &cos_res) puts 0.69534156199418473 in cos_res":
+double: 1
+idouble: 1
Test "sincos (M_PI_6l*2.0, &sin_res, &cos_res) puts 0.5 in cos_res":
double: 1
float: 1
@@ -754,9 +791,6 @@ ldouble: 1
Test "sincos (pi/6, &sin_res, &cos_res) puts 0.86602540378443864676372317075293616 in cos_res":
float: 1
ifloat: 1
-Test "sincos (0.80190127184058835, &sin_res, &cos_res) puts 0.69534156199418473 in cos_res":
-double: 1
-idouble: 1
# tan
Test "tan (pi/4) == 1":
@@ -970,16 +1004,16 @@ double: 1
float: 7
idouble: 1
ifloat: 7
-ildouble: 6
-ldouble: 6
+ildouble: 735
+ldouble: 735
Function: Imaginary part of "cacosh":
double: 1
float: 3
idouble: 1
ifloat: 3
-ildouble: 1
-ldouble: 1
+ildouble: 119
+ldouble: 119
Function: Real part of "casin":
double: 1
@@ -1178,12 +1212,12 @@ ildouble: 5
ldouble: 5
Function: Imaginary part of "ctanh":
+double: 1
float: 1
+idouble: 1
ifloat: 1
ildouble: 25
ldouble: 25
-double: 1
-idouble: 1
Function: "erf":
double: 1
@@ -1196,12 +1230,12 @@ ildouble: 1
ldouble: 1
Function: "exp10":
-ildouble: 8
-ldouble: 8
-float: 2
-ifloat: 2
double: 6
+float: 2
idouble: 6
+ifloat: 2
+ildouble: 8
+ldouble: 8
Function: "expm1":
double: 1
@@ -1250,12 +1284,12 @@ ildouble: 1
ldouble: 1
Function: "log10":
+double: 1
float: 2
+idouble: 1
ifloat: 2
ildouble: 1
ldouble: 1
-double: 1
-idouble: 1
Function: "log1p":
float: 1
--
Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj
SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: BZ#2153: cacosh bug
2006-01-15 15:44 BZ#2153: cacosh bug Andreas Jaeger
@ 2006-01-15 17:25 ` Ulrich Drepper
2006-01-15 17:40 ` Andreas Jaeger
0 siblings, 1 reply; 4+ messages in thread
From: Ulrich Drepper @ 2006-01-15 17:25 UTC (permalink / raw)
To: Andreas Jaeger; +Cc: Glibc hackers
[-- Attachment #1: Type: text/plain, Size: 539 bytes --]
Andreas Jaeger wrote:
>
> +ildouble: 735
> +ldouble: 735
>
> Function: Imaginary part of "cacosh":
> double: 1
> float: 3
> idouble: 1
> ifloat: 3
> -ildouble: 1
> -ldouble: 1
> +ildouble: 119
> +ldouble: 119
If there are error ranges like this there is clearly something wrong.
This has to be fixed.
Separate out the sign fix from adding the new test cases and file a bug
for the new tests, pointing to the high ulps.
--
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 251 bytes --]
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: BZ#2153: cacosh bug
2006-01-15 17:25 ` Ulrich Drepper
@ 2006-01-15 17:40 ` Andreas Jaeger
2006-01-15 17:47 ` Ulrich Drepper
0 siblings, 1 reply; 4+ messages in thread
From: Andreas Jaeger @ 2006-01-15 17:40 UTC (permalink / raw)
To: Ulrich Drepper; +Cc: Glibc hackers
[-- Attachment #1: Type: text/plain, Size: 896 bytes --]
Ulrich Drepper <drepper@redhat.com> writes:
> Andreas Jaeger wrote:
>>
>> +ildouble: 735
>> +ldouble: 735
>>
>> Function: Imaginary part of "cacosh":
>> double: 1
>> float: 3
>> idouble: 1
>> ifloat: 3
>> -ildouble: 1
>> -ldouble: 1
>> +ildouble: 119
>> +ldouble: 119
>
> If there are error ranges like this there is clearly something wrong.
> This has to be fixed.
>
> Separate out the sign fix from adding the new test cases and file a bug
> for the new tests, pointing to the high ulps.
So, the change to just make the sign fix in cacosh and libm-test is ok
to commit?
I'll then commit those changes and add the new tests to a new
bugreport,
Andreas
--
Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj/
SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: BZ#2153: cacosh bug
2006-01-15 17:40 ` Andreas Jaeger
@ 2006-01-15 17:47 ` Ulrich Drepper
0 siblings, 0 replies; 4+ messages in thread
From: Ulrich Drepper @ 2006-01-15 17:47 UTC (permalink / raw)
To: Andreas Jaeger; +Cc: Glibc hackers
[-- Attachment #1: Type: text/plain, Size: 210 bytes --]
Andreas Jaeger wrote:
> So, the change to just make the sign fix in cacosh and libm-test is ok
> to commit?
Yes.
--
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 251 bytes --]
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2006-01-15 17:47 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2006-01-15 15:44 BZ#2153: cacosh bug Andreas Jaeger
2006-01-15 17:25 ` Ulrich Drepper
2006-01-15 17:40 ` Andreas Jaeger
2006-01-15 17:47 ` Ulrich Drepper
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).