On Mon, Aug 22, 2016 at 9:18 AM, Trevor Saunders <tbsau...@tbsaunde.org> wrote: > On Sun, Aug 21, 2016 at 09:00:26PM +0100, Pedro Alves wrote: >> On 08/20/2016 03:29 AM, Mike Stump wrote: >> > On Aug 10, 2016, at 10:03 AM, Oleg Endo <oleg.e...@t-online.de> wrote: >> >> >> >> Or just wait until people have agreed to switch to C++11 or C++14. I >> >> don't think in practice anybody uses an C++11-incapable GCC to build a >> >> newer GCC these days. >> >> gdb will drop support for building with a C compiler any week >> now, and even though we're starting out with C++03, just like gcc, >> it'd be great to require C++11 (or later). Having gcc itself >> switch to C++11 too would make proposing it for gdb so much >> easier... > > huh, I would have sort of expected the oposit, if gdb was to require > C++11, but gcc didn't then you could still use gdb on antique systems > without a C++11 compiler by first building gcc. > >> So +1 from me, FWIW. :-) > > I'd mostly agree, at least requiring a compiler with rvalue references > would be pretty useful. > >> > >> > I use the system gcc 4.4.7 on RHEL to build a newer cross compiler... I >> > could bootstrap a newer native compiler, if I had too. >> > >> >> Yeah. I wonder whether the community would in general be fine with >> that too. > > I personally don't have any machines where the system compiler is that > old, but its worth noting the last C++11 features came in 4.8.1 I think > and for libstdc++ its even later.
Most of our auto-testers are still on GCC 4.1 and GCC 4.3 host compilers (all but one actually). The oldest still maintained SLES has GCC 4.3 as host compiler (though GCC 4.8 is available as well, just not as 'gcc'). The "current" SLES has GCC 4.8 as host compiler. So unless there are very compelling reasons to require C++11 I'd rather not go down that route at this very moment. Richard. (yeah, time to update some of our autotesters - but then it will most likely be simply retiring them) > Trev > >> >> Thanks, >> Pedro Alves >>
