[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Binary-only packages

Michael Meskes wrote:
> Sven Rudolph writes:
> > While the guidelines don't mention this case I don't like to see
> > binary-only packages in the regular distribution. An important part of
> > the Debian notion of free, the availability of source code, isn't
> > given; therefore I suggest to put your package into non-free.
> I cannot agree here. I don't think a package should be name non-free just
> because it's source isn't available. I think non-free should (and does) mean
> that the package (that is the binary) is not free to be distributed

I agree.  "non-free" can mean anything.  A binary-only package can be:

  - free distribution
  - free to use
  - non-free to modify
  - non-free to include (in whole or in part) in your own packages
  - non-free of bugs

(I don't think there is a single Debian package that truely free if you
include that last one.)

> > I think you should release a source package containing only the
> > scripts necessary to build the binary package and a pointer to the
> > original binary archive.
> Good idea. But without a diff file it looks like an original Debian package.

Trust me as writer/maintainer of the Netscape package...  This is a pain!
It means two different archives that must be kept in sync.  It makes it
difficult to use the nice features of dpkg to track installed files.

> > Reason: If a new version of the original archive appears someone
> > else can package it. Or you might have some error in the
> > debian.control file, e.g. a wrong dependency. When the packaging files
> > are available everyone can fix it without having to recreate them.
> > This won't happen often, but it might be necessary.
> You got a point with this.

Yup.  Good point.  I can see no reason why the "diff" file couldn't be
uploaded, though.  That would include all the "debian.*" files.  Patches
to programs can be distributed without legal hassles, so any source fixes
would also be acceptable.  The author would just have to specify to what
the patch file had to be applied.

                               ( bcwhite@verisim.com )

    In theory, theory and practice are the same.  In practice, they're not.

Reply to: