On 02/16/2015 11:33 PM, Kevin Kofler wrote:
Ralf Corsepius wrote:
- A GCC-5 was pushed to F22 (I thought FESCO had decided _not_ to push
GCC-5 to F22) and already is causing FTBFSes there[1]
They decided to push GCC 5 to F22, but with the GCC 4 C++ ABI as the default
C++ ABI to avoid having to rebuild everything C++ this late in the cycle.
IMO, this was a bad and non-helpful decision. They should not have
pushed any GCC-5 "this late in the cycle" and should have waited with
pushing GCC-5 to rawhide until F22 is release.
Right now, we are facing packages whose (non-gcc related) F22-bugs are
unfixable because the issues w/ GCC-5 F23 are propagating through.
Other GCC 5 changes can still cause breakage though.
Exactly.
I already am facing a couple of F22 packages which GCC 5 broke, because
- the implicit default c-standard seems to have changed. I am facing
packages to whom I now have to add -std=c89 or -std=gnu89 to let
building succeed.
[BTW: One family of packages which is victim of this change is GCC-4,x
itself. Older versions seem to require -std=gnu89]
- the GCC version numbers have changed, which is causing FTBFSes in
poorly implemented packages which somewhere depend on hard-coded GCC
version numbers.
Ralf
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
