- Mar 07, 2003
- Mar 06, 2003
-
-
pfdietz authored
-
pfdietz authored
Modified some tests so they don't make ecl crash (instead they just fail). Also, modified the RT test harness to work around other obnoxious bugs in ecl.
-
pfdietz authored
Conditionalized parts of some MAKE-STRING and MAKE-SEQUENCE tests to control whether we assume a missing :initial-element argument is assigned some single default value, or if those operations are free to leave the sequence/string filled with arbitrary legal garbage.
-
- Mar 05, 2003
-
-
pfdietz authored
-
- Mar 04, 2003
- Mar 03, 2003
- Mar 02, 2003
-
-
pfdietz authored
Added tests for (e,c)typecase where the type specifier is a class object. Tests for handler-bind; stub file for tests for handler-case. More subtypep tests on array types.
-
- Mar 01, 2003
-
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
Added test that all builtin symbol type specifiers that should represent nonempty types are not subtypes of NIL.
-
pfdietz authored
Changes to make the test suite be case agnostic, so it can work with the 'modern mode' of Allegro CL. This mode is not standard, but I'd like the suite to be able to test all the other parts of ACL's mlisp for compliance.
-
pfdietz authored
-
- Feb 27, 2003
- Feb 26, 2003
-
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
It was pointed out to me that the 'should be prepared' terminology for error checking applied to the case where the first argument of map-into is NIL. In that case, it's an extra step to check whether the other arguments are sequences, so 'lazy' checking can skip it. Modified test map-into.error.2 to accept either outcome.
-
- Feb 25, 2003
- Feb 24, 2003
- Feb 23, 2003
-
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
Moved loading of tests into separate scripts, divided approximately by major section of the ANSI CL spec.
-
pfdietz authored
Modified to run under Lispworks, which unfortunately already had a package (nick)named RT. Changed the regression test package to be REGRESSION-TEST, and make the RT nickname be available only with #-lispworks. Removed a bunch of gratuitous use-package forms. Added lispworks binary filenames to .cvsignore and make clean. Added a test to MAKE-SYMBOL that illustrates a place where Lispworks is noncomplant (this is related to the failure on MAKE-SYMBOL.ERROR.8)
-
pfdietz authored
-
- Feb 22, 2003
-
-
pfdietz authored
Toned down the error/cerror tests, losing the assumption that the :format-control or :format-arguments initializers for simple-conditions give the values that actually end up in those slots.
-