[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Can we just finish the C++ transition for crying out loud?

Matthias Klose wrote:
> Nathanael Nerode writes:
>> And now, a new gcc-4.0 bumped the shlibdeps for libstdc++ -- and worse,
>> depends on new binutils and new glibc.  This will undoubtedly mean that
>> either forced package breakages, significant numbers of package removals,
>> or months more of waiting will be needed.
> the dependency on glibc and binutils was already there before. now
> it's just made explicitely in the dependencies.

That change would have been fine by itself.  It's the *combination* of that
with the libstdc++ dependency bump which is going to clog things up really
badly.  If you're going to depend on new versions of binutils and glibc, and
those new versions are seriously buggy (which they are), it's not a good
idea to propagate that dependency to every C++ package which gets rebuilt
-- in the middle of a difficult ABI transition.  :-P

I will be recommending the removal from testing of every package which
depends on new gcc-4.0 and blocks the KDE cluster from getting in.  This is
because it does not seem reasonable to force in the current versions of
gcc-4.0, binutils, and glibc, given the rather serious problems with
current binutils.   It also is certainly not reasonable to wait for them to
be fixed before letting the KDE C++ transition into testing; this is such a
large group of packages that keeping them out of testing for several more
months severely diminishes the utility of the testing distribution.

ksig --random|

Reply to: