June 2021 - asdf-devel - mailman3.common-lisp.net

Last call before 3.3.5 bug fix release
by Robert Goldman 29 Jun '21

29 Jun '21

I'm going to release 3.3.5, which I *hope* will be the last bug fix release in the 3.3 series, either this afternoon or tomorrow. So this is the last call for bug fixes that should get in under the wire. I have reviewed the issues and the merge requests, and I don't see anything that can't be postponed. There are still issues with ABCL failing test-around-compile.script and test-run-program-unix.script, but to the best of my knowledge those are waiting for bug-fixes in ABCL, rather than in ASDF. If a new ABCL comes out and these are still broken, we can have a 3.3.6 ... So, speak now or forever hold your peace! Thanks, R

1 0

(Lack of) Inheritance of in-order-to dependencies
by akater 21 Jun '21

21 Jun '21

I tried to specify a build-time dependency for ASDF system, and failed. The defsystem option :depends-on systems makes “systems” both build-time and runtime dependencies of a system being defined. I'm interested in specifying a dependency that is required at build time only. I attempted (defsystem "test-toplevel-in-order-deps" :in-order-to ((compile-op (load-op "some-macros"))) :components ((:file "a") (:file "b"))) but this way, some-macros is loaded too late: (let ((system (find-system "test-toplevel-in-order-deps"))) (plan-actions (make-plan 'sequential-plan 'compile-op system))) ;; => ((#<PREPARE-OP > . #<SYSTEM "test-toplevel-in-order-deps">) (#<PREPARE-OP > . #<CL-SOURCE-FILE "test-toplevel-in-order-deps" "a">) (#<COMPILE-OP > . #<CL-SOURCE-FILE "test-toplevel-in-order-deps" "a">) (#<PREPARE-OP > . #<CL-SOURCE-FILE "test-toplevel-in-order-deps" "b">) (#<COMPILE-OP > . #<CL-SOURCE-FILE "test-toplevel-in-order-deps" "b">) (#<PREPARE-OP > . #<SYSTEM "some-macros">) ... (#<LOAD-OP > . #<SYSTEM "some-macros">) (#<COMPILE-OP > . #<SYSTEM "test-toplevel-in-order-deps">)) I'm not requesting a feature but still, it's unfortunate that a seemingly basic task of specifying build-time dependency is unachievable in ASDF. My hypothesis is, if (compile-op . parent) depends on “actions”, then compiling children of parent should depend on “actions” as well, and similarly for all instances of downward-operation. I'm not sure what's the best way to implement this but I wonder if you find this reasonable in the first place.

2 1

Upcoming bug fix release
by Robert Goldman 06 Jun '21

06 Jun '21

I have just updated the ASDF repository to version 3.3.4.19 I plan for this to turn into the bug fix release 3.3.5 very soon, so please, if you have the opportunity, test it with your libraries, your hardware, your lisp implementations, etc. I am particularly excited to get this out there, because we have a number of enhancements on deck that could become version 3.4, and I would dearly like 3.3.5 to be the last bug fix release in the 3.3 series. Best, Robert

5 5

Very stupid question about :module.
by Marco Antoniotti 02 Jun '21

02 Jun '21

Hi I know I should RTFM, but the wisdom of real people is better. Plus, I am lazy... I would like to have a file depending on a "collection" of other files. I know I can use :module to do so, but how do I prevent the module from being cast into a directory? I.e., I have bar.lisp zot.lisp foobar.lisp I would like to say something like (:module "barzot" :components ((:file "bar") (:file "zot")) (:file "foobar" :depends-on ("barzot") How can I tell ASDF that there isn't any "barzot" folder? I know this must be a stupid question... Thanks Marco -- Marco Antoniotti, Associate Professor tel. +39 - 02 64 48 79 01 DISCo, Università Milano Bicocca U14 2043 http://dcb.disco.unimib.it Viale Sarca 336 http://cdac2021.lakecomoschool.org I-20126 Milan (MI) ITALY

2 2