Thursday, August 27, 2015

Regressions and improvements in quicklisp 2015-08-04

The difference between quicklisp 2015-07-10 and quicklisp 2015-08-04 is shown in these reports:
grouped by lisp implementation
grouped by library

(Both reports show the same data, just arranged differently)

As we see, most of the new failures are caused by drakma failing to load on CCL 1.8, CLISP and SBCL 1.0.58, because drakma.asd now specifies test-op, but in a way supported only by new ASDF, and not supported by the ASDF shipped with these old lisp implementations. Maybe the drakma-caused failures are not critical for everyone, on the other hand it is trivial to fix.

There are other regressions as well as some improvements. If you care about some of the results, but don't understand the cause of particular failure, or how to read the report, just ask in the comments or on the mailing list, we will be glad to help, investigate and reproduce the problem.

BTW, you can click test result links in the reports to open the corresponding log.

A little bit more about the cl-test-grid project. We regularly collect test results of all the Quicklisp libraries on various lisp implementations, and compute difference between Quicklisp versions to detect changes (regressions or improvements). That way we can monitor the CL ecosystem as a whole.