From: Christophe R. <cs...@ca...> - 2003-05-27 10:26:51
|
Dag-Erling Smorgrav <de...@of...> writes: >> This compilation failure has been fixed in the CVS, thanks. However, >> I should note that while failure to compile one of the contributed >> systems is not ideal, neither should it hold back distribution of sbcl >> proper; the install.sh script tests for whether a contrib has >> successfully passed its regression tests, and does not install it if >> not. > > Umm, the build bombs at this point. I think the build doesn't bomb at that point; it just looks like it. (In case you haven't noticed yet, our presentation isn't too hot :-) What happens for each contrib is something like this: attempt to build; if fails, give up. attempt to test; if fails to run tests or fails a test, give up. touch test-passed file. What you're seeing at the end of your build is sb-simple-streams giving up at the first of these stages. So it gives you back a backtrace and then a shell prompt, because sb-simple-streams is the last of the contribs, so that's the end of the build, whether sb-simple-streams succeeds in building or not. > Plus, installing a partial package would really mess with the plist. > Remember, I'm not just trying to install SBCL - I'm trying to > package it for FreeBSD. Right. So if we included a Win32 interface, would you want to package that part for FreeBSD too? Well, maybe you would, I dunno, but it's possible that there would be something that didn't make sense to include from the contrib/ section. What are the criteria to deem a package complete? I can't speak for all sbcl developers, but I'd be surprised if any considered that contrib/ was a part of the core of sbcl; indeed, its raison d'etre is to house loosely-coupled extensions. Yes, we aim to have them working -- where that even makes sense -- on all X OSes and Y architectures, but a failure in one of the contribs is not a fundamental failure in a way that failing to compile ordinary lisp code would be. Maybe it would help to explain what the requirements are for your packaging system? I personally have no experience at all of it. (Maybe the other packagers on this list -- Kevin? Miles? -- can discuss the tradeoffs they've made?) Cheers, Christophe -- http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757 (set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b))) (defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge) |