Re: is the whole unstable still broken by gcc-5?
Jakub Wilk <email@example.com> writes:
> * peter green <firstname.lastname@example.org>, 2015-09-19, 06:29:
>> Another big one is the descision was taken NOT to change the sonames
>> when making sub-transitions. I presume this was done to avoid
>> deviating from upstream sonmaes
> I don't think this is the reason. We don't change SONAMEs because it's
> a huge pain for maintainers.
For upstream, it is probably wrong to change the soname: the ABI changed
due to libstdc++ changing, however upstream cannot know which C++
implementation is used when building their packages.
> (Has anyone figured a straight-forward
> way to append arbitrary suffix to SONAME for packages that use
Though not a suffix: maybe libtool's -release option? It can be used
together with -version.