- 23 Oct, 2002 1 commit
-
-
pfdietz authored
Updated many tests to use CLASSIFY-ERROR in place of HANDLER-BIND. This makes the tests shorter and opens the possibility they can be compiled, since CLASSIFY-ERROR ensures error calls are safe calls. Added more DO tests.
-
- 22 Oct, 2002 4 commits
- 19 Oct, 2002 11 commits
- 18 Oct, 2002 7 commits
- 17 Oct, 2002 1 commit
-
-
pfdietz authored
These are the same as the original except they are guaranteed to return T when a true value is to be returned. Added NOTNOT, which coerces true to T. These changes are intended to make the test more fully standard compliant, avoiding the unwarranted assumption that a function that returns true is returning T (even though this is the case in many implementations.) Added tests for NIL, T, NOT, and NULL.
-
- 15 Oct, 2002 3 commits
- 14 Oct, 2002 6 commits
- 13 Oct, 2002 7 commits
-
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
-------------- Hello, Thank you for your tests. They have shown two new bugs in the SBCL compiler (in addition to >160 failed tests). I think the following tests are wrong: map-string.30: (SIMPLE-VECTOR size), not (SIMPLE-VECTOR type) map-array.7: ARRAY is not a subtype of SEQUENCE map-array.8: SIMPLE-ARRAY is not a subtype of SEQUENCE map-string.31: --"-- map-string.32: (SIMPLE-ARRAY CHARACTER 5) is a 5-dimensional array find-error.4, position-error.4: see CLHS 1.4.2 for 'should be prepared to signal an error' -- Regards, Alexey Dejneka
-
pfdietz authored
-
pfdietz authored
-
pfdietz authored
-