* Matěj Týč wrote on Tue, Feb 03, 2009 at 11:15:10AM CET: > > This is not possible, in general. It has nothing much to do with > > libtool either, because typically it's just the system semantics that > > allow for only one unversioned soname symlink. > > OK, but an easy check whether I can safely link with the lib would be > nice, too.
Well, a feature check would be more in line with Autoconf "philosophy" than a version-based one, I guess. True, this is less applicable for the large software stacks that desktop apps typically use, than e.g., for the numerical libraries where often, several different implementations exist that provide (nearly) the same API. > > Of course, if you don't have to be able to have multiple versions > > installed concurrently, then all you do is use a configure test like > > AC_CHECK_LIB or sisters with the latest-added symbol that you need, > > or a self-written link test if the latest needed API isn't > > distinguishable by function symbol alone. > > Another point of having a AC_CHECK_LTLIB - like macro would be a > possibility to check whether the version of the library is OK. > Sometimes it is not possible to use AC_CHECK_LIB to check whether a > function is in the library (due to calling convention). Are you hinting at C++ here? Somebody should write a decent check for C++ libraries. Or maybe one exists already out there, dunno. > Well, I just wonder what is the sense of the libtool's versioning > system since it can't be used by library users :-) It can be used. However, library versioning is primarily a concept to allow support of existing binaries. Recompiling is intended to always work against the newest versions of things. A corollary is that, when you are building software in order to distribute it in binary form, then build it against the oldest versions of the libraries that you want to be compatible with. At least that is the way it was originally intended, I think. I should admit that, in principle, libtool could go beyond that and support some sort of linking against older library versions; on some systems, this could work. But we try not to promote semantics that are unportable. And anyway there is no analogous functionality in the preprocessor: for included headers, there exists no agreed-upon versioning concept, and you typically cannot mix those from different versions either. Hope that helps. Cheers, Ralf _______________________________________________ http://lists.gnu.org/mailman/listinfo/libtool
