Re: Handling macro change in an exported library header
Steve Langasek <firstname.lastname@example.org> writes:
> My understanding from reading this thread is that the signature of the
> function did *not* change, but that in previous versions of the library
> the macro definition in the header was simply broken, with the result
> that anything built using that definition would not work correctly.
> So it sounds to me like there's no ABI change here, just a header
Ah, yes, in that case there's no library SONAME change, just a bug that
may have propagated into other packages in the archive.
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>