[ghemical-devel] Ghemical-2.95 is now released

[Michael Banck]

On Mon, Oct 08, 2007 at 01:17:54PM +0300, Tommi Hassinen wrote:
> On Sun, 7 Oct 2007, Daniel Macks wrote:
>> I think you're seeing this backwards: it isn't a case of "breaking
>> API", but rather of keeping same API (apparently) but changing the
>> SONAME as if it *were* incompatible. So you'd have a
>
> Yes, this is more or less what I intended. In libghemical not only API 
> changes but also changes in parameter files sometime justify changes in 
> SONAME.

That is sensiible I think, but then the parameter directory should maybe
rather be versioned according to the SONAME and not LIBVERSION (which is
really PACKAGE_VERSION right now), I think.  Or rather, CURRENT.  I
actually realized yesterday that the current Debian libghemical-data
package is not versioned like the library itself, so I will rename it to
libghemical2-data or something.  

> But I now understand that the SONAMEs in libghemical and liboglappth are 
> chosen badly and should be fixed ; the SONAME for
>
> libghemical should change from 2:95:0 to 3:0:0 and SONAME for
>
> liboglappth should change from 0:95:0 to 1:0:0

That would imply another round of recompiles for people/distributions
which shipped libghemical-2.95 and liboglappth-0.95 already.  We
(Debian) have not included them yet, so I do not mind too much.

> Please reply to me and tell if this makes sense and is what you expect ; 
> I'll then make new releases and fix these.

I think it would be fine to just do this for the next SONAME bump and
carry on with the current library versioning for now (and increment as
described in the libtool manual).  But changing to above library
versions wouldn't be a problem, either; I'd just like to know so I can
hold off uploading libghemical/liboglappth for now if you plan to go
ahead with the above.

By the way, mopac7 introduced the same problem, I already uploaded it to
Debian though, but it did not arrive in the archive yet.


Michael