Re: ASDF 2.32 released
Raymond Toy <[email protected]> Tue, 5 Mar 2013 19:51:40 -0800
| Newsgroups | gmane.lisp.cmucl.devel |
|---|---|
| Message-ID | <CAG14z1FmqefgoOfX+Fj-fZbDkWjpi+t6dB3XLHVXF3hShUKmNQ@mail.gmail.com> |
On Tue, Mar 5, 2013 at 5:27 PM, Faré <[email protected]> wrote: > Dear CMUCL hackers, > > ASDF 2.32 was released in time for inclusion in the upcoming CCL 1.9. > Bummer. You just missed the 2013-03 snapshot. But I've updated cmucl's version to 2.32 (or later!) and it will be available in the next snapshot. BTW, I ran the asdf testsuite and all tests pass, of course. Ray > Once again, we invite all implementation vendors and distributors > to upgrade the version of ASDF included in their releases. > > Yes, I know, that's a lot of releases in a bit more than a month. > But it was also a lot of development, and a lot of adjustment to reality > after that development was first released. > We will keep fixing bugs and improve ASDF based on > our growing internal test suite and results from Anton Vodonosov's > cl-test-grid. > > This is still a pre-release for ASDF3. Here are the changes, > since previous release 2.31, and since previous public announcement for > 2.30. > > From 2.31 to 2.32: > * deferred-warning checks are now disabled by default, > to enhance backward compatibility with ASDF 2.26 and earlier. > In the future, we will enable them by default, > but not until we implement a smoother transition plan. > Also, CCL-specific fixes to deferred-warning support. > > * UIOP is the new official name for ASDF-DRIVER. > "Utilities for Implementation- and OS- Portability". > Old names remain as package and system nicknames: > ASDF/DRIVER ASDF-DRIVER ASDF-UTILS > > * require-system so systems may depend on implementation-provided > modules. > > * auto-upgrade attempt in operate *before* the implicit find-system. > > * class specification for components can now be a string to be read > as a symbol. > > * feature conditional dependencies are now debugged and tested: > :depends-on ("some-system" (:feature :foo "other-system)) > we only recommend them for dependencies between systems; > within systems, we recommend you use feature-conditional components > with > :if-feature :foo > and we only provide limited support for the deprecated legacy way of > using > feature-conditionals with a module having :if-component-dep-fails > :ignore > while its components specify things like: > :in-order-to ((compile-op (feature :foo))) > > * read-from-file for :version specification is made more useful by using > the ASDF-USER package rather than CL (into which interning is > forbidden) > to read the files. > > Changes from 2.30 to 2.31: > > * *default-encoding* is now UTF-8. As tested with cl-test-grid, > This is a boon for most programs, > that will compile in a more predictable way in a wider range of > settings. > A handful of unmaintained outliers still need to be fixed. > > * Downgrade is now explicitly avoided. ASDF will issue a warning > if your system is misconfigured in a way that specifies downgrade, > but will then ignore the specified ASDF downgrade. > > * version specifications, when invalid, will lead only to a warning for > now; > a future ASDF3 may restore the behavior of ASDF 2.27, > that raises an errors when there is a bad :version specification, > but for now, backward compatibility prevails. > > * ECL support received several bug fixes. > > * precompiled-system support was notably fixed, enabling single-fasl > deployment of SBCL contribs. > > —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• > http://fare.tunes.org > He who refuses to do arithmetic is doomed to talk nonsense. > — John McCarthy, in his webpage on Progress and Sustainability > _______________________________________________ > cmucl-imp mailing list > [email protected] > http://lists.zs64.net/mailman/listinfo/cmucl-imp > _______________________________________________ cmucl-imp mailing list [email protected] http://lists.zs64.net/mailman/listinfo/cmucl-imp