Re: [Gcl-devel] ansi-test

[Paul F. Dietz]

Sam Steingold wrote:

> Generally (not universally) the tests are grouped by page in the
> CLHS.  These tests, for example, are testing the DEFSTRUCT page.
>
>
> sure.  but what aspect?  (I am probably asking too much here :-)

Most of the tests in structures-02.lsp use the DEFSTRUCT-WITH-TEST
macro.  This macro acts just like DEFSTRUCT, except that it sets
up a number of tests (whose names are symbols whose names have
the structure name as a prefix, followed by "/" and a number.)

In each case we're trying a different set of parameters for DEFSTRUCT.
Just which parameters are being used should be entirely obvious.

The individual tests generated by the macro are mostly explained by
comments in the structure-00.lsp, where DEFSTRUCT-WITH-TESTS is
defined.


> > There are something like 6500 tests now.  Since few or very few of
> > them fail on good implementations, it's a lot easier on me if I can
> > just document on demand when tests fail.
>  
> that's the whole point!!!
> you should document tests _as you write them_.
> every DEFTEST should have a documentation slot.

Sure, it would be slightly easier for *you* if all that documentation
were present in the tests, but there's an adequate workaround that's
a lot easier for *me*.  Why exactly should I spend a lot of my time to
save you a very small amount of your time?  I do add comments to some
of the more obscure tests.


> > As Christophe noted, this has to be printed with *print-circle* bound
> > to t to be reread properly.
>
> then you should bind *print-circle* when printing such tests.

I'll see if that's acceptable (I don't want it to make the output too
ugly, although it should only have that effect in tests that actually
have shared structure).

        Paul