The following happens with the latest version 1.2 of PFGW:-
pfgw_ver_12_win32_pfgw_winpfgw.zip

as downloaded from:-
http://groups.yahoo.com/group/primeform/files/

(and with any earlier versions I have).

1.
PFGW -f -lconf.out conf.inp
where the 1-line input file conf.inp is:-

107020..[59994 more digits]..
07205183735507325932393414052340741874514018423003356731387578805508634512164617\
85510327727621884768762773773378337629775651378147023054942973488017241823439073\
73296157932561643158554069586538600525873472764334149779859430874537673700472125\
40331013726163121755037949679928082433

(the input number has 60254 digits),

gives the correct result:-
>107020864767591639004 .. 55037949679928082433 has factors: 464587^2

2.
PFGW -f -lerror.out error.inp
where the 2-line input file error.inp is:-

ABC $b // -f{$a}
-77431 107020..[59994 more digits]..
07205183735507325932393414052340741874514018423003356731387578805508634512164617\
85510327727621884768762773773378337629775651378147023054942973488017241823439073\
73296157932561643158554069586538600525873472764334149779859430874537673700472125\
40331013726163121755037949679928082433

gives on the screen:-

>trial factoring to 7063881632927445
>PRP: 1070208647675916....0720518373550732 nnn/199285
>1070208647675916....0720518373550732 is composite
>// - Evaluator failed
>Unknown error in PFGW runtime
>Possibly a number too big for alloca version of GMP

The "....0720518373550732" means it has cut off the last 251 digits (see the end
of the actual number given above).

Does this mean there is a limit of 60K digits somewhere in the // -f{$a}
processing, perhaps?

Can this bug be corrected?
Meanwhile, is there anything I can do to get round it?

TIA

-Mike Oakes