ellap() bug?

[Iwao KIMURA <iwao@sci.toyama-u.ac.jp>]

Hi all,

I heard that ellap() returns wrong answer for large p.

? e=ellinit([0,0,0,0,24])
%1 = [0, 0, 0, 0, 24, 0, 0, 96, 0, 0, -20736, -248832, 0, [-2.884499140614816764643276621, 1.442249570307408382321638310 + 2.498049532966812958826359087*I, 1.442249570307408382321638310 - 2.498049532966812958826359087*I]~, 2.476797314448744297908657918, -1.238398657224372148954328959 + 0.7149897981125623456981647477*I, -2.196949286467445307753379632 - 2.45092531 E-29*I, 1.098474643233722653876689816 - 1.902613892906903694312629616*I, 1.770884811823444282448628760]
? p=557018866141
%2 = 557018866141
? isprime(p)
%3 = 1
? #
   timer = 1 (on)
? ellap(e,p)
time = 2mn, 24,587 ms.
%4 = 47745597502
? 47745597502 < 2*sqrt(p)
time = 0 ms.
%5 = 0

ellap(e,p) returns 47745597502, but this is not smaller than
2*sqrt(p), violating Hasse's theorem.

I tested gp of version 2.2.3 (build from current CVS source tree) on
my FreeBSD 4.4 box & gcc version 2.95.3 20010315 (release) [FreeBSD].
But the report I got says the situation is same for the following
environments: 
    GP/PARI CALCULATOR Version 2.1.1 (released)
i686 running linux (ix86 kernel) 32-bit version

    GP/PARI CALCULATOR Version 2.0.20 (beta)
i586 running cygwin_98-4.10 (ix86 kernel) 32-bit version

Thanks.

	;#	Iwao KIMURA			;#
	;#	Faculty. Math, Dept. of Sciece	;#
	;#	Toyama University		;#