[Gcl-devel] Re: ANSI

[Paul F. Dietz]

Camm Maguire wrote:
> Thank you, and my apologies for the question.  I'm trying to hurry
> this summer, and am using your wonderful tet suite largely as a
> replacement for me having to read and parse the spec in detail
> myself.  I should have known that you'd get it right, but I just had
> the nagging question that this might have been designed to fit other
> complying implementations.

One way I dropped the ball in the test suite was traceability
back to the specification.  I'm going to try to add comments
explaining some of the more obscure tests, but it will take
a while.

> Is it true that your suite checks the integrity of the compiled or
> interpreted functions as supplied in the image, and not the result of
> any compiler inlining or replacement of such functions, except in the
> random tester?  I'd like to deduce from the ansi test results if
> possible that I haven't messed up compiler inlining of #'member
> et.al. for example.

You want to run the random type prop tests.  Start lisp, then:

(load "gclload1.lsp")
(load "random-type-prop-tests.lsp")

These tests are explicitly designed to test type-driven compiler
specializations of built-in functions.  The member tests are
in random-type-prop-tests-05.lsp.  The def-type-prop-test macro
prepends "RANDOM-TYPE-PROP." to the specified name to get the
name of the test, so the tests you want are (in the CL-TEST
package) named random-type-prop.member.1 through
random-type-prop.member.6.

Note: the tests of destructive functions here may be wrong,
because of issues with modification of literal constants
in compiled code.

        Paul