RE: Re: ASDF/INTERFACE::TRAVERSE-SUB-ACTIONS

[Marco Antoniotti]

Thank you Faré,   Let me grok your email and then I will see which works best for my needs.   cheers   Marco Antoniotti DISCo, Università degli Studi di Milano-Bicocca +39 02 6448 7901 bimib.disco.unimib.it     From: Faré Sent: Tuesday, May 5, 2020 19:44 To: Robert Goldman Cc: Marco Antoniotti; ASDF-devel Subject: Re: ASDF/INTERFACE::TRAVERSE-SUB-ACTIONS   git grep system-input-files doesn't return anything for HEAD, 3.2.0, 3.1.7, 3.1.2. But the name vaguely rings a bell, and I might have offered a function that does that long ago. A better replacement today might be    (input-files :concatenate-source-op system)   Problem: this "solution" will do what you want in case of package-inferred-systems, or systems where all or most of the functionality is in secondary systems. In these cases, you might have to do something more sophisticated, some variant of:   (defun system-lisp-files (x) (remove-if-not (lambda (p) (subpathp p (system-source-directory x))) (input-files :monolithic-concatenate-source-op x)))   or if you want to be more subtle in case of non-lisp input files, something that generalizes the below to somehow work with multiple phases of operation (see notably how it behaves for iolib):   (defun system-files (system) (mapcar 'component-pathname (remove-duplicates (remove-if-not (lambda (x) (and (typep x 'source-file) (equal (primary-system-name x) (primary-system-name system)))) (mapcar 'cdr (plan-actions (make-plan () :load-op system)))))))   If you get things to work across phases of execution, you'll be in a good position to also fix POIU.   Good luck.   —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org Poetry is what gets lost in translation. — Robert Frost     On Tue, May 5, 2020 at 12:33 PM Robert Goldman <rpgold...@sift.info> wrote: >  > Something is really wrong. Look at this transcript: >  > rpg@rpg-mbp-2: ~/lisp/asdf $ git checkout 3.3.4 > Note: switching to '3.3.4'. >  > You are in 'detached HEAD' state. You can look around, make experimental > changes and commit them, and you can discard any commits you make in this > state without impacting any branches by switching back to a branch. >  > If you want to create a new branch to retain commits you create, you may > do so (now or later) by using -c with the switch command. Example: >  >   git switch -c <new-branch-name> >  > Or undo this operation with: >  >   git switch - >  > Turn off this advice by setting config variable advice.detachedHead to false >  > HEAD is now at 082f4ed4 Bump version to 3.3.4 > rpg@rpg-mbp-2: ~/lisp/asdf $ find . -name '*.lisp' -exec fgrep -iq SYSTEM-INPUT-FILES {} \; -print > rpg@rpg-mbp-2: ~/lisp/asdf $ >  > Similarly, after loading 3.3.4 into SBCL, I see this: >  > CL-USER(1): (apropos '#:system-input-files) >  > CL-USER(2): >  > Similarly on Lispworks: > ``` > CL-USER 3 > (load "/Users/rpg/lisp/asdf/build/asdf.lisp") > ; Loading text file /Users/rpg/lisp/asdf/build/asdf.lisp >  > P"/Users/rpg/lisp/asdf/build/asdf.lisp" >  > CL-USER 4 > (apropos '#:system-input-files) >  > CL-USER 5 > > ``` >  > I just tried (require :asdf) on Lispworks, and still no system-input-files. >  > I simply have no idea where this is coming from. Please send a minimum working example for Lispworks.