Hello Andreas,

thank you for your feedback.

Andreas Beckmann a écrit :
> ... First
> hit was
>   swi-prolog
> You may want to take a look at their approach to combine building both
> Arch: any and *-java packages from one source. It seems to boil down to
> make the java package(s) Arch: some instead of Arch: all.
> 
> IIRC splitting source packages to create multiple source packages (or
> reintroducing copies of the same source in the archive) is not recommended.

I am releasing a new version of wims' package (4.03a-4), with the 
following changelog:

------------------8<---------------------------
wims (4.03a-4) unstable; urgency=low

  * restored the build of the package wims-java-applets. This package is now
    considered as Architecture: some. This is a bad trick, since it can
    be installed seamlessly on JVM-lacking architectures, and needs no
    JVM in localhost to be useful for other machines. However this trick
    will be used as a turnaround, just to enable wims' acceptance in
    testing. If a better solution is found, it will be applied then.
  * fixed the parameter WIMS_VERSION in Makefile
------------------8<---------------------------

> I'm pretty sure the problem of Arch: all packages being not buildable on
> all architectures (because some build-deps(-indep) are not available on
> all architectures) has been discussed before.

I thank in advance anyone who would hint a better solution to this issue.
In a few words, it should:
- allow the binary package wims-java-applets to be built once (in some
  architecture providing a JVM)
- allow all other binary packages to be built with Architecture: any
- allow this binary package to be installed on *all* architecture, since it
  has no dependency at all on the server machine.

Best regards,                   Georges.

Attachment: signature.asc
Description: Digital signature

Reply via email to