|
# On some machines, ACE 3.000 would die with a segmentation fault when trying
# to construct coset representatives. Since the code seemed OK, and the
# problem only manifested itself on some Linux/gcc combinations, it was put
# down to a compiler bug (emails to/from Greg Gamble, Aug 2000). In fact, it
# seems that this problem may have been due to an uninitialised variable
# which only manifested itself in some circumstances (emails to/from Volker
# Gebhardt, Dec 2000, and the Changes file entry for 22 Dec).
# The examples below illustrate this fault, if it's present. Note that they
# all contain a CT row > 1 whose first entry <= row is row itself. It is the
# building of a coset rep for this row, either via a call to `pr' with -ve
# argument or a call to `oo', which causes the problem. The offending
# entries are CT(3,b), CT(11,a) & CT(7,a) respectively.
mess: 10000;
default;
Gr: abc;
Rel: b, aab, (c)^4, ACac;
Gen: ;
text; start; pr:-8;
Gr: ab;
Rel: a^-2*b*a*b^-1*a*b,a^3*b^-1*a^-2*b^-1;
Gen: a;
text; start; print:-12;
Gr: ab;
Rel: a^3, a*b*a/(b*a*b);
Gen: a;
text; start; oo:0;
text;
[ Dauer der Verarbeitung: 0.13 Sekunden
(vorverarbeitet)
]
|