[bug#51903] cmn license etc

[Guillaume Le Vaillant <glv@posteo.net>]

[-- Attachment #1: Type: text/plain, Size: 2023 bytes --]

bil@ccrma.Stanford.EDU skribis:

> I did not remove it; for some reason it wasn't in the
> directory I used to make the tarball.

In order for the compilation of cmn to put the files in the standard
locations used in GNU Guix, I changed the asd file to replace

  :components ((:file "cmn-all\" ))

by

  :serial t
  :components ((:file "cmn-init")
               (:file "cmn-utils")
               (:file "cmn-objects")
               (:file "cmn0")
               (:file "cmn-grfx")
               (:file "cmn-glyphs")
               (:file "cmn1")
               (:file "cmn2")
               (:file "cmn3")
               (:file "cmn4")
               (:file "rqq")
               (:file "wedge")
               (:file "accent")
               (:file "pedal")
               (:file "percussion")
               (:file "ring")
               (:file "rests")
               (:file "lyrics")
               (:file "transpose")
               (:file "pmn")
               (:file "quarter"))

and it almost works.
The only issue making the compilation fail is a type error warning by
sbcl, which I think is due to a bug in the program.

In the IDENTIFY-SCORE function of the "cmn2.lisp" file, the TIED-TAG
variable is defined by a "(find :tied-to (store-data object))" form,
which means that TIED-TAG will be either :TIED-TO or NIL.
However a few lines lower, the "(setf (second tied-tag) ...)" form
indicates that TIED-TAG must be a non-empty list.

From the surrounding code, I got the impression that
"(store-data object)" probably contains lists, and that the definition
of TIED-TAG should be something like

  (find :tied-to (store-data object) :key #'car)

With this, the compilation succeeds with sbcl and ecl, and the little
"(cmn treble c4 q)" example from the README file works.

I don't know if you're interested in the modified asd file, we can keep
it as a Guix-only thing anyway. However the TIED-TAG thing looks like an
actual bug.
What do you think?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]