- Jun 19, 2024
-
-
Raymond Toy authored
-
Raymond Toy authored
-
- Aug 04, 2023
-
-
-
This allows testing an extrinsic system that implements part of the standard without replacing the implementation's systems.
-
If we don't do that, we only get somewhat mysterious errors when loading the compiled file fails.
-
Instead of multiple parameters for each type of message, there is only one parameter (c:*suppress-compiler-messages*) which takes a type of messages to suppress. By setting that to t we were suppressing all messages.
-
-
-
-
- Jun 07, 2023
-
-
Tarn Burton authored
-
- May 24, 2023
-
-
Tarn Burton authored
Add ability to specify expected failures See merge request !55
-
Tarn Burton authored
- Add EXIT keyword argument to do-tests. This will exit with a non-zero status code on test failures and a zero status code on success. - Add EXPECTED-FAILURES keyword argument to DO-TESTs. If this argument is provided then it will override the contents of *EXPECTED-FAILURES*. If it is a pathname then the test names will be read from the file with the CL-TEST package as *PACKAGE*. Keywords in this file or in a supplied list will be interpreted as the names of notes to disable. - For some implementations load the expected failures from a file in the expected-failures folder.
-
- May 11, 2023
-
-
Daniel Kochmański authored
The license file was missing. I've added it based on the project's author repository: https://github.com/pfdietz/ansi-test/blob/master/LICENSE Closes #26. C.f #40.
-
Daniel Kochmański authored
define-method-combination: add tests for the :arguments option See merge request !54
-
- Jan 21, 2023
-
-
Marius Gerbershagen authored
-
- Nov 10, 2022
-
-
Daniel Kochmański authored
Use copy of dispatch table in test See merge request !52
-
Tarn Burton authored
-
- Oct 17, 2022
-
-
Daniel Kochmański authored
Include char-code in read-symbol.13 and read-symbol.14 See merge request !50
-
-
Daniel Kochmański authored
Fix #35: Disable test shiftf.7 Closes #35 See merge request !51
-
-
- Sep 08, 2022
-
-
Daniel Kochmański authored
Fix #34: parse-namestring.5 should test with logical host "CLTEST" Closes #34 See merge request !49
-
- Sep 07, 2022
-
-
Raymond Toy authored
Currently `parse-namestring.5` creates a pathname with a logical host of "CLTEST". It then tests that the corresponding namestring has a host of "CLSTEST". That should be "CLTEST" instead.
-
- Mar 24, 2022
-
-
Daniel Kochmański authored
Fixes to FORMAT.E tests Closes #32 and #33 See merge request !50
-
Christophe Rhodes authored
It invokes undefined behaviour. Fixes #33
-
Christophe Rhodes authored
Fixes #32
-
Daniel Kochmański authored
Fix non-standard LOOP construct See merge request !49
-
Christophe Rhodes authored
-
- Dec 27, 2021
-
-
Daniel Kochmański authored
Compile tests in batches See merge request !38
-
Daniel Kochmański authored
objects: make-load-form: add tests for correct ordering of load time forms See merge request !43
-
Daniel Kochmański authored
Add test for upcasing in logical pathname parsing See merge request !48
-
- Oct 03, 2021
-
-
Marius Gerbershagen authored
Check for symbol length in character.1.body See merge request !46
-
Marius Gerbershagen authored
fix bogus error in ensure-directories-exist.8 See merge request !47
-
- Oct 01, 2021
-
-
Marius Gerbershagen authored
The test would succeed only the first time it was run, since it was creating a temporary directory each run which wasn't cleaned up.
-
- Jul 31, 2021
-
-
Tarn Burton authored
-
- Mar 26, 2021
-
-
Marius Gerbershagen authored
Nowhere in the specification does it say that an invalid proclamation has to result in a type-error. All other proclaim.error tests only check for an error.
-
- Feb 15, 2021
-
-
Marius Gerbershagen authored
-
- Jan 26, 2021
-
-
Raymond Toy authored
- Ignore sse2f fasls created by cmucl/x86 - Add comment to makefile on how to run with cmucl
-
- Jan 24, 2021
-
-
Raymond Toy authored
Cmucl converts a string to Unicode NFC form before creating a symbol. This means some symbols have names that longer than 1 character. The test assumes the name has one character which causes an error in cmucl. This just adds a check that the length is 1 before extracting the character from the symbol name. This doesn't change behavior of the test (which was already assuming a length of 1), but allows cmucl not to signal an error but to finish the test with a meaningful list of characters whose corresponding symbol name is more than one character long. A couple of minor tweaks: - Ignore sse2f fasl created by cmucl/x86. - Add comment to makefile on how to run with cmucl.
-
- Jan 17, 2021
-
-
Raymond Toy authored
Use char-equal in position as indicted by phoe See merge request ansi-test/ansi-test!45
-