glpk soname bump expected?

[Michael Schwendt]

On Mon, 4 Feb 2013 09:13:33 -0700, Jerry James wrote:

> On Mon, Feb 4, 2013 at 9:07 AM, Orion Poplawski wrote:
> > Hmm, that makes it seem even more likely that upstream fat-fingered
> > something.
> >
> > Although: http://upstream-tracker.org/versions/glpk.html
> > does indicate that ABI has been broken (although it has been done so in the
> > past without bumping the soname).
> 
> Looks like this has been brought up on an upstream mailing list:
> 
> http://lists.gnu.org/archive/html/help-glpk/2013-01/msg00081.html

Yes, using the libtool versioning scheme enforces a soname change for
changed/added/removed symbols. It's weird to do that in a minor release
4.47 -> 4.48, however. And one would need to examine the releases prior
to 4.47 to understand history (i.e. whether it has arrived at -version-info
32:0:32 without bumping the soname before because of keeping cur=age).

-- 
Fedora release 19 (Rawhide) - Linux 3.8.0-0.rc6.git0.1.fc19.x86_64
loadavg: 0.32 0.38 0.24