Can not reproduce.

Does your distribution provide packages with full debug information? If so, please install them and you should be able to get a slightly more detailed backtrace - the extra information should be useful

*** Bug 140993 has been marked as a duplicate of this bug. ***

Confirmed on r628k. Debug-enabled backtrace:

#14 0x29852c2f in __gmp_invalid_operation () from /usr/local/lib/libgmp.so.7
#15 0x29854513 in __gmpf_set_d () from /usr/local/lib/libgmp.so.7
#16 0x298405ae in _knumfloat (this=0x8255c68, num=inf) at knumber_priv.h:264
#17 0x29843d17 in _knumfloat::power (this=0xbfbfda10, exponent=@0xbfbfda30)
    at /home/phil/kdesrc/kdeutils/kcalc/knumber/knumber_priv.cpp:858
#18 0x29843914 in _knuminteger::power (this=0xbfbfda80, exponent=@0x8258e20)
    at /home/phil/kdesrc/kdeutils/kcalc/knumber/knumber_priv.cpp:796
#19 0x29843c01 in _knumfraction::power (this=0x8259f70, exponent=@0x8258e20)
    at /home/phil/kdesrc/kdeutils/kcalc/knumber/knumber_priv.cpp:845
#20 0x2983f5dc in KNumber::power (this=0xbfbfdb40, exp=@0xbfbfdb50)
    at /home/phil/kdesrc/kdeutils/kcalc/knumber/knumber.cpp:505
#21 0x29826307 in ExecPower (left_op=@0xbfbfdb40, right_op=@0xbfbfdb50)
    at /home/phil/kdesrc/kdeutils/kcalc/kcalc_core.cpp:150
#22 0x2982bb35 in CalcEngine::evalOperation (this=0x8179190, arg1=

Analysis (thanks to tsdgeos for pointing out some of this):

1. For some reason, _fraction_input is set to true when creating the KNumber's for e and pi, so they're created as fractions that look like:
2718281828459045235360287471352662497757247093699959574966967627724076630353547594571382178525166427/1000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
(ie, one very large number / another very large number)

2. KCalc/GMP calculates the power as numerator1^numerator1 (or whatever it is), and overflows.

I can work around it by manually setting _fraction_input=false in the KNumber constructor. Then the numbers are created as floats and everything works. I don't know the best way to solve it properly though

*** Bug 141544 has been marked as a duplicate of this bug. ***

I was also able to reproduce this with KCalc 2.0.6 (KDE 3.5.6) on SuSE Linux 10.2 (86_64).

Stack trace looks similar.

Settings: Maximum number of digits: 14
Set decimal precision is off
Beep on error set
Show result in window title off
Group digits enabled

Only Logic buttons are enabled.

Steps to reproduce:

[Constants->Mathematics->Euler number] [X^y] [Constants->Mathematics->Pi] [+/-] [=]

It then immediately crashes.

System configuration startup check disabled.

Using host libthread_db library "/lib64/libthread_db.so.1".
[Thread debugging using libthread_db enabled]
[New Thread 47505908478016 (LWP 6647)]
[KCrash handler]
#5  0x00002b34d0702535 in raise () from /lib64/libc.so.6
#6  0x00002b34d0703990 in abort () from /lib64/libc.so.6
#7  0x00002b34d5616dc3 in __gmp_invalid_operation ()
   from /usr/lib64/libgmp.so.3
#8  0x00002b34d5618b80 in __gmpf_set_d () from /usr/lib64/libgmp.so.3
#9  0x00002b34d53a658f in _knumfloat::power ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#10 0x00002b34d53aa3ce in _knuminteger::power ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#11 0x00002b34d53aa55e in _knumfraction::power ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#12 0x00002b34d53a8f89 in KNumber::power ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#13 0x00002b34d5378769 in ExecPower () from /opt/kde3/lib64/libkdeinit_kcalc.so
#14 0x00002b34d5378619 in CalcEngine::evalOperation ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#15 0x00002b34d5387b34 in CalcEngine::evalStack ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#16 0x00002b34d5387d8a in CalcEngine::enterOperation ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#17 0x00002b34d5397f41 in KCalculator::EnterEqual ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#18 0x00002b34d539ac7e in KCalculator::qt_invoke ()
   from /opt/kde3/lib64/libkdeinit_kcalc.so
#19 0x00002b34cf73b57c in QObject::activate_signal ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#20 0x00002b34cf73c253 in QObject::activate_signal ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#21 0x00002b34cf76df27 in QWidget::event ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#22 0x00002b34cf6e4405 in QApplication::internalNotify ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#23 0x00002b34cf6e5221 in QApplication::notify ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#24 0x00002b34cf125f88 in KApplication::notify ()
   from /opt/kde3/lib64/libkdecore.so.4
#25 0x00002b34cf68c2f9 in QETWidget::translateMouseEvent ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#26 0x00002b34cf68b4e3 in QApplication::x11ProcessEvent ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#27 0x00002b34cf69a4ef in QEventLoop::processEvents ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#28 0x00002b34cf6f8fb3 in QEventLoop::enterLoop ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#29 0x00002b34cf6f8e62 in QEventLoop::exec ()
   from /usr/lib/qt3/lib64/libqt-mt.so.3
#30 0x00002b34d539e3f7 in kdemain () from /opt/kde3/lib64/libkdeinit_kcalc.so
#31 0x00000000004074a1 in launch ()
#32 0x0000000000407dd4 in handle_launcher_request ()
#33 0x0000000000408195 in handle_requests ()
#34 0x00000000004091ea in main ()

I'm able to reproduce this in Slackware Linux current branch (KDE 3.5.7/KCalc 2.0.6) next way:
1. Run kcalc.
2. Menu: Constants->Math->Pi
3. Button: M+
4. Button: x^y
5. Button: MR
6. Button: =
7. I see next crash report:

(no debugging symbols found)
Using host libthread_db library "/lib/libthread_db.so.1".
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
[New Thread -1233537328 (LWP 8067)]
[KCrash handler]
#5  0xb688fcd7 in raise () from /lib/libc.so.6
#6  0xb6891548 in abort () from /lib/libc.so.6
#7  0xb6646fd3 in __gmp_invalid_operation () from /usr/lib/libgmp.so.3
#8  0xb6648beb in __gmpf_set_d () from /usr/lib/libgmp.so.3
#9  0xb66c7d18 in _knumfloat::power () from /usr/lib/libkdeinit_kcalc.so
#10 0xb66c8e31 in _knuminteger::power () from /usr/lib/libkdeinit_kcalc.so
#11 0xb66c90c7 in _knumfraction::power () from /usr/lib/libkdeinit_kcalc.so
#12 0xb66c3d03 in KNumber::power () from /usr/lib/libkdeinit_kcalc.so
#13 0xb66a8de0 in CalcEngine::evalOperation ()
   from /usr/lib/libkdeinit_kcalc.so
#14 0xb66a8d44 in CalcEngine::evalOperation ()
   from /usr/lib/libkdeinit_kcalc.so
#15 0xb66adf5e in CalcEngine::evalStack () from /usr/lib/libkdeinit_kcalc.so
#16 0xb66ae208 in CalcEngine::enterOperation ()
   from /usr/lib/libkdeinit_kcalc.so
#17 0xb668f9f2 in KCalculator::EnterEqual () from /usr/lib/libkdeinit_kcalc.so
#18 0xb668fa4d in KCalculator::slotEqualclicked ()
   from /usr/lib/libkdeinit_kcalc.so
#19 0xb66961c6 in KCalculator::qt_invoke () from /usr/lib/libkdeinit_kcalc.so
#20 0xb71370c9 in QObject::activate_signal ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#21 0xb7137cdd in QObject::activate_signal ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#22 0xb745206c in QButton::clicked () from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#23 0xb71c9b3c in QButton::mouseReleaseEvent ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#24 0xb71700de in QWidget::event () from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#25 0xb70d8817 in QApplication::internalNotify ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#26 0xb70d96e9 in QApplication::notify ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#27 0xb77068ee in KApplication::notify () from /usr/lib/libkdecore.so.4
#28 0xb7079cf9 in QETWidget::translateMouseEvent ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#29 0xb7078bcc in QApplication::x11ProcessEvent ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#30 0xb7088ee1 in QEventLoop::processEvents ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#31 0xb70ef141 in QEventLoop::enterLoop ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#32 0xb70eefe6 in QEventLoop::exec () from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#33 0xb70d847f in QApplication::exec ()
   from /usr/lib/qt-3.3.8/lib/libqt-mt.so.3
#34 0xb66a0681 in kdemain () from /usr/lib/libkdeinit_kcalc.so
#35 0xb7617654 in kdeinitmain () from /usr/lib/kde3/kcalc.so
#36 0x0804e457 in ?? ()
#37 0x00000007 in ?? ()
#38 0x08094f80 in ?? ()
#39 0x00000001 in ?? ()
#40 0x00000000 in ?? ()

*** Bug 152538 has been marked as a duplicate of this bug. ***

Confirmed even on KCalc 2.1 (KDE 4, SVN revision: 747166)

Confirmed on kcalc 2.0.6 (ubuntu gutsy x86).
For some reason, pi^e works ...
also, a workaround for e^pi is to calculate (e^(pi/2))^2 :)

can confirm on Mandriva 2007.1 with kcalc 2.0.6
select: Euler no.
select: x^y
select: pi
select: =
crash

Confirmed, kcalc 2.0.6, up-to-date Gentoo stable, on two different machines:

1. CHOST="i686-pc-linux-gnu" CFLAGS="-march=prescott -O2 -pipe -fomit-frame-pointer"
2. CHOST="i686-pc-linux-gnu" CFLAGS="-march=pentium4 -O2 -pipe -fomit-frame-pointer -fforce-addr"

BTW, an easier work-around for people whom this is annoying.

1. Settings->Constants Buttons (make sure it's on).
2. Right click two of the buttons and set them to pi and the Euler number (at this point, the buttons have the ratio-based versions that will still trigger the bug).
3. Settings->Configure KCalc. Hit Ok. (the "Constansts" settings panels, which you don't actually have to visit, has the decimalized versions of the numbers. Hitting Ok saves the decimalized into those constants buttons).

Now, just be sure to use the constant buttons instead of choosing the constants from the Constants menu.

This bug is there with kcalc 2.2, KDE 4.00.67.

Some more notes:

Calculating PI^PI causes kcalc to calculate the following:
*- snip -*
785398163397448309615660845819875721049292349843776455243736148076954101571552249657008706335529267/250000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 ^ 785398163397448309615660845819875721049292349843776455243736148076954101571552249657008706335529267/250000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
*- snap -*

This can be calculated as (n1/d1)^(n2/d2) where n1/d1 and n2/d2 are the nominators/denominators of the above big numbers. As GMP has only limited capabilities of doing a^x with fractional or float numbers, one can calculate the expression a^(n/d) with dth_root_of(a^n). There are also limitations in GMP, doing this. Instead, one can write dth_root_of(X) = exp(ln(X) / d). Finished. Doing this and replacing the method 

_knumber * _knumfraction::power(_knumber const & exponent) const {...}

with this, the bug can be closed but the calculation taked VERY_VERY long and is consuming lot of memory. 

Any ideas?

GnuShi

I want to share another test case with you:
http://xkcd.com/217/

SVN commit 844132 by aucahuasi:

CCBUG: 140630

With this patch the bug can be closed, I tested this on KDE 4.1 version: 4.1.60 (KDE 4.1.60 (KDE 4.2 >= 20080709)) and on KCalc 2.4. Too long values of some constants (PI and e) made that kcalc-core and gmp throws exeptions of NAN (HUGE_VAL issues).


 M  +2 -2      scienceconstants.xml  


WebSVN link: http://websvn.kde.org/?view=rev&revision=844132

Reducing the accuracty of the PI and e constant as propsed in comment #18
to solve a crash bug seem like a change with a lot of side effects.
Does this mean that the system will still crash if the old values are used, ie

  2.718281828459045235360287471352662497757247093699959574966967627724076630353547594571382178525166427^3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628034825342117068

About the comment #19: 
Hi there, I have made the change in the svn repository I tested it and solved the issue ... I think that have constants with too long values is unnecesary becouse KCalc works very fine now and is a tool design to the commun user, besides the pression of the operations with this constants are fairly acceptable.

>Does this mean that the system will still 
>crash if the old values are used, ie 
I guess yes, I -simply- changed the values of PI and e becouse it was the source of the problem.

Percy

*** Bug 169047 has been marked as a duplicate of this bug. ***

I think this bug should be closed as a fixed becouse the crash is gone now and beside is known the source of the problem.

Percy

Should be fixed per revision 853717.

http://websvn.kde.org/?view=rev&revision=853717

*** Bug 177099 has been marked as a duplicate of this bug. ***

*** Bug 180841 has been marked as a duplicate of this bug. ***

*** Bug 202255 has been marked as a duplicate of this bug. ***