Re: [Gcl-devel] Dotted lists

[Paul F. Dietz]

"Vadim V. Zhytnikov" wrote:

> The only thing which bothers me is that we
> still need to go through all GCL list processing
> functions and check their reaction on non
> proper lists.

There's no substitute for a good set of regression tests.
Not only should we check all the list processing functions
on dotted lists, we should rerun these tests (and tests of
of all other parts of the system and language spec) whenever
the system changes.

I've registered an ssh password with Savannah and will try
to create a subdirectory of gcl/ for the tests I have so far
(2000+ tests, including tests for all the functions in section
14 of the CL spec).  I estimate a comprehensive set of tests of
the CL spec might run as many as 10,000 tests, so there's
a lot of work left to do.

        Paul Dietz
        address@hidden