On 22 Aug 2016, at 17:30, Drew C me@drewc.ca wrote:
Hey,
How does this differ from the "Monthly-or-so" tests that Quicklisp does with cl-test-grid? Is there anything beyond `make test-lisp`, or is this a simple "try to build the ASDF master branch on Linux/x64 and report if it fails"?
For example, I notice that they used sbcl-1.0.58 in the last test[1] . What issues did you have with those earlier versions that were'fixed' in 1.1.13? Is there a report of your testing available that I could look at beyond a quick email?
Beyond that, looks good!
Thanks,
Drew Crampsie
[1] https://common-lisp.net/project/cl-test-grid/ql/quicklisp-2016-06-28-diff.ht...
Dear Drew,
I’ve only now learnt of cl-test-grid. It looks like a great idea. And it’s an actual project: A concerted, automated effort. What I summed up in my e-mail was merely a cleanup of my notes that I’ve been collecting over the course of a few days or weeks, with a limited scope (Linux-only e.g.). Hence there are no reports beyond what my e-mail provides, I’m afraid.
The goals are, as Robert already mentioned, very different, too: I cared about ASDF here, whereas cl-test-grid appears to care about quicklisp-installable libraries. cl-test-grid will also end up testing ASDF indirectly, but an old version (2.x rather than 3.x), and in a less systematic fashion than the ASDF test suite (no targeted regression testing e.g.).
If you’re a developer, you might find both tests interesting. Maybe you’re working on a project that has a few dependencies and also needs a recent version of ASDF. How portable is your project then? That’s answered by these two tests. But my e-mail was primarily targeted at the developers of ASDF.
Elias