[Date Index] [Thread Index] [Author Index]

new bug in Limit

• To: mathgroup at yoda.physics.unc.edu
• Subject: new bug in Limit
• From: brigham!rprice at physics.att.com
• Date: Tue, 8 Jun 93 16:29:48 -0400

Dear mathgroup,

One of my old mma codes written under 1.2 now fails to work under 2.1, and  
here's the reason why:

In[1]:=
f[x_] = Erfc[x]/x;

In[2]:=
Limit[f[x], x->0]
Out[2]=
ComplexInfinity

This is a perfectly well-defined limit, as you can see by forcing mma to  
apply l'Hopital's rule directly,

In[3]:=
D[Numerator[f[x]], x]/D[Denominator[f[x]], x] /. x->0
Out[3]=
   -2
--------
Sqrt[Pi]

Presumably there is some reason this new bug crept into mma version 2.  Idle  
curiosity would have me ask what it could possibly be, but since I need to  
get some work done, all I'm really interested in is a fix or workaround.

Rod Price
rprice at physics.att.com