Prev: NautyTools`NautyTools
Next: Remote kernel only works with "raw mathlink connection" selected
From: Szabolcs Horvát on 30 Sep 2007 04:08
Evaluating the following three statements with a fresh kernel, in this
exact order, results in an error message in Mathematica 6.0 (not in 5.2).

In[1]:= Limit[1/x, x -> Infinity]
Out[1]= 0

In[2]:= Exp[Sqrt[x]] + O[x]^2
Out[2]= 1+Sqrt[x]+x/2+x^(3/2)/6+O[x]^2

In[3]:= Sum[x^n/n!, {n, 0, Infinity}]
During evaluation of In[3]:= SeriesData::sdatn: Order specification ___
in SeriesData[_,_,{},___] is not a machine-size integer. >>
Out[3]= E^x

Can anyone explain this? It is very strange that all three (seemingly
independent) statements need to be evaluated to trigger the error.

--
Szabolcs

From: Jean-Marc Gulliet on 1 Oct 2007 05:09
Szabolcs Horv=C3=A1t wrote:

> Evaluating the following three statements with a fresh kernel, in this =

> exact order, results in an error message in Mathematica 6.0 (not in 5.2=
).
<snip>

It also appears in version 6.0.1.

In[1]:= Limit[1/x, x -> Infinity]

Out[1]= 0

In[2]:= Exp[Sqrt[x]] + O[x]^2

Out[2]= (SeriesData[$CellContext`x, 0, {
1, 1, Rational[1, 2], Rational[1, 6]}, 0, 4, 2])

In[3]:= Sum[x^n/n!, {n, 0, Infinity}]

During evaluation of In[3]:= SeriesData::sdatn: Order specification \
___ in SeriesData[_,_,{},___] is not a machine-size integer. >>

Out[3]= \[ExponentialE]^x

In[4]:= $Version

Out[4]= "6.0 for Microsoft Windows (32-bit) (June 19, 2007)"

Regards,
Jean-Marc


From: Bhuvanesh on 2 Oct 2007 05:42
Yes, this was reported both internally and later via tech support. In ordinary circumstances this only happens when using the frontend, and it has already been fixed both in the development version and for the next maintenance release.

Thanks for the report, and sorry for the inconvenience.

Bhuvanesh,
Wolfram Research

 | 
Pages: 1
Prev: NautyTools`NautyTools
Next: Remote kernel only works with "raw mathlink connection" selected