I noticed that this bug can be reproduced the way frans suggested iff
the installer is run at priorities high or critical (low or med don't do).
Another way to reproduce this bug is entering disagreeing passwords,
pressing <Ok> and then <Back>.
Anyway, i pot some test prints in the gtk_go() body and i could clearly
see that gtk_go() completes correctly after last <BACK> pressure and has
not been started again yet when the frontend crashes.
So, i wonder if the crash is related to the specific question asked
after the user has pressed <BACK> the last time at high pri instead of mid.
I'll keep on investigating on this, but if someone could help i'd really
apreciate it.
Atilio
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]