Euler constant

[klaus-brockhaus@t-online.de (Klaus Brockhaus)]

I am using PARI/GP 2.0.17 for Windows.

For realprecision = 32000 or higher, Euler leads to the message
"  ***   division by zero in R/S".

For several reasons I became sceptical as to the reliability of the Euler value for higher realprecision, so I compared the digits produced by Pari for realprecision 25000 and the digits given by J.
Borwein, 170000 digits of Euler or
gamma constant:
<
http://www.worldwideschool.org/library/books/sci/math/MiscellaneousMathematicalConstants/chap35.html
>
and found that from digit 19292 onward there is disagreement. Digits 19281 to 19300:

05228370342108957865 on the web page,
05228370342566205373 by PARI.

Can someone confirm this observation? Is it a known bug?

Klaus