Den søn. 20. apr. 2025 kl. 10.41 skrev Ladislav Michl <[email protected]>:
>
> On Sat, Apr 19, 2025 at 11:10:47PM +0200, Ladislav Michl wrote:
> > On Fri, Apr 18, 2025 at 01:33:57PM +0200, Bruno Thomsen wrote:
> > > +ptx/repology = \
> > > +   $(call ptx/env) \
> > > +   ptx_dgen_rulesfiles_make="$(PTX_DGEN_RULESFILES_MAKE)" \
> > > +   ptxd_make_world_repology
> >
> > Here would be probably better to pass $(PTX_PACKAGES_INSTALL) and possibly 
> > sort it.
>
> Actually tried that with $(PTX_PACKAGES_ALL) and it does not seem to be viable
> solution. Too much has to be set up just to generate JSON report. Now trying 
> to
> model it similar way ptxdist-lint target is made.

Hi

I also started with ptxdist-lint but might have removed a little too
much of the makefile
and relied too much on the bash script part :)
It would be nice if we could run more makefile to get dynamic versions
and licenses resolved
before handoff to bash script. Or maybe we should do it more like spdx
sbom in Python.

The sorting and indentation is currently done externally by jq, but
that could be moved inside
ptxd_make_world_repology script function.

Also I am a bit unsure if this cross rule variable access is actually
more of a bug then a feature.

$ rg XORG_LIB_X11_LICENSE
rules/libx11-locale.make
24:LIBX11_LOCALE_LICENSE    = $(XORG_LIB_X11_LICENSE)

rules/xorg-lib-X11.make
24:XORG_LIB_X11_LICENSE    := MIT


Ladislav you can find latest development version of this repology support here:
https://github.com/baxeno/ptxdist/tree/repology_support
(you can send PR directly if you wish)

Output from RFC PATCH have at least 19 issues with "$(":
https://github.com/baxeno/ptxdist-repology/blob/main/repology.json

/Bruno

>
>         l.

Reply via email to