Testsuite failures on Debian PowerPC

Hello,

today I tried installing GNU Smalltalk 3.2.4 on a Debian PowerPC machine.
Some of the tests failed and I can't figure out why.

I've attached two files: the output of the testsuite and the testsuite.log.

I would be really glad if someone can give me a hint :-)

Thanks!
Marc

--
★ lernu Esperanton! Esperanto - la lingvo internacia - www.esperanto.net ★

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

On 06/09/2011 10:53 PM, Marc Hanisch wrote:
> today I tried installing GNU Smalltalk 3.2.4 on a Debian PowerPC machine.
> Some of the tests failed and I can't figure out why.
>
> I've attached two files: the output of the testsuite and the testsuite.log.
>
> I would be really glad if someone can give me a hint:-)

Please try executing this from the REPL and cut/paste the output:

a := 1 sinh
b := (a squared + 1) sqrt.
c := (a * a + 1) sqrt.
d := 1 cosh.
e := 1 exp.
f := -1 exp.
g := (e + f) / 2.
h := (e - f) / 2.
i := FloatD emin - FloatD precision - 1.
j := 1.0d timesTwoPower: i.
j examine


Here is the output I get on an x86 machine by the way:

st> a := 1 sinh
1.1752011936438014
st> b := (a squared + 1) sqrt.
1.5430806348152437
st> c := (a * a + 1) sqrt.
1.5430806348152437
st> d := 1 cosh.
1.5430806348152437
st> e := 1 exp.
2.718281828459045
st> f := -1 exp.
0.36787944117144233
st> g := (e + f) / 2.
1.5430806348152437
st> h := (e - f) / 2.
1.1752011936438014
st> i := FloatD emin - FloatD precision - 1.
-1075
st> j := 1.0d timesTwoPower: i.
0.0
st> j examine
An instance of FloatD
   [1]: 0
   [2]: 0
   [3]: 0
   [4]: 0
   [5]: 0
   [6]: 0
   [7]: 0
   [8]: 0
0.0

Paolo

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

Hello Paolo,

2011/6/10 Paolo Bonzini <[hidden email]>:
The output at first looks the same as yours, but it seems the
precision is not exact:

st> a := 1 sinh
1.175201193643801
st> b := (a squared + 1) sqrt.
1.5430806348152435
st> c := (a * a + 1) sqrt.
1.5430806348152435
st> d := 1 cosh.
1.5430806348152437
st> e := 1 exp.
2.718281828459045
st> f := -1 exp.
0.3678794411714426
st> g := (e + f) / 2.
"Global garbage collection... done"
1.5430806348152437
st> h := (e - f) / 2.
1.175201193643801
st> i := FloatD emin - FloatD precision - 1.
-1075
st> j := 1.0d timesTwoPower: i.
0.0
st> j examine
An instance of FloatD
  [1]: 0
  [2]: 0
  [3]: 0
  [4]: 0
  [5]: 0
  [6]: 0
  [7]: 0
  [8]: 0
0.0

Can I though install Gnu Smalltalk on the PowerPC machine?
Marc

--
★ lernu Esperanton! Esperanto - la lingvo internacia - www.esperanto.net ★

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

> Can I though install Gnu Smalltalk on the PowerPC machine?

Yes, it's just a small imprecision in floating point operations. I'll
look at it in a couple of weeks.

Paolo

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

2011/6/14 Paolo Bonzini <[hidden email]>:
>> Can I though install Gnu Smalltalk on the PowerPC machine?
>
> Yes, it's just a small imprecision in floating point operations. I'll
> look at it in a couple of weeks.
>

Thank you :-)
Marc


--
★ lernu Esperanton! Esperanto - la lingvo internacia - www.esperanto.net ★

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk