Re: Library sonames and unstable libraries
On Wed, Jan 10, 2007 at 12:29:21AM -0500, Hubert Chan wrote:
> > Assuming that I want to publish some at least partially useful
> > packages (for a start, much of the use will be via python bindings
> > rather than linking against the .so) until upstream has committed to
> > an soname policy, and made a stable release, would it be acceptable to
> > use experimental for this (ie where the package is otherwise
> > policy-clean but doesn't have clean soname bumps during the tracking
> > of SVN snapshots) or should I stick to having this in my private
> > webspace (and losing autobuilding, bugtracking and other bits and
> > pieces of functionality)?
> I believe that packages in experimental must still follow policy, so you
> shouldn't use experimental unless you figure out how to handle the
> sonames. If use will be limited to the Python bindings, you may be able
> to get away with making the C libraries private and sticking them in
> /usr/lib/mapnik, at least for now. (And, you would probably skip out on
> the -dev package.)
> Otherwise, your options are probably to just use your own private
> webspace, or make the library into a static library.
Again, thanks for comments.
Actually I could use an soname of libmapnik.so.0d for now (idea from
Josselin Mouette's talk that I recently watched the video of :) which I
can increment to my heart's content until upstream makes a release with
an soname. As I understand it moving from libmapnik.so.0d (package name
libmapnik0d) or for that matter libmapnik.so.1d to a future official
libmapnik.so.0 shouldn't be a problem should it - the ordering of
sonames doesn't matter?
Dominic Hargreaves | http://www.larted.org.uk/~dom/
PGP key 5178E2A5 from the.earth.li (keyserver,web,email)