Riccardo (Jack) Lucchetti schrieb:
IMO, we have two colliding needs here: on one hand we'd like to ship
1.7.6 asap to get rid of the problems that have surfaced recently with
session files. On the other hand, I think the issue above is potentially
very disruptive for anyone who wishes to write a non-trivial set of
gretl functions.
Hence my proposal: let's define PROTECT_LISTS as 1 in CVS and use the
next two weeks to stress test the function apparatus by running all the
inp scripts we've got to see what this change breaks. If all goes well,
we release 1.7.6 at the end of the month. In the meantime, the rest of
the code should stay frozen (apart from obvious bugs, like the
upper()/lower() one I fixed yesterday).
Personally I think the plan "1.7.6 ASAP" implies a release sooner than
the end of this month. As I said before, I don't work with session files
(yet) so I'm not affected myself, but isn't the current crashing state
of the official release a little embarassing? IMHO the other
list-related design bug is not nearly as important. And the complicated
description suggests that a solution shouldn't be rushed and should also
be properly tested, something which is not going to happen in two weeks.
(BTW, I'm also really surprised that not more complaints are arriving on
the users list. Given that (I think) I know that a significant amount of
people are using gretl more or less regularly, that seems to mean that
either those people are not using session files, or that they are
currently being scared away from using gretl. Or what else could be the
case?)
just my 1.99 cents
-sven