[Bug c++/70735] [5/6/7 Regression] problem combining std::function, generic lambdas and static variables

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70735

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Allan McRae from comment #6)
> (In reply to Jonathan Wakely from comment #5)
> > (In reply to Jonathan Wakely from comment #4)
> 
> > > This shows that Arch is using a snapshot from the gcc-5-branch on 20160209,
> > > which means it's 5.3.1, but they lie and call it 5.3.0 for some annoying
> > > reason.
> > 
> > Allan, why does Arch do this?
> > 
> > The upstream numbering scheme used since gcc 5 means that we can distinguish
> > official releases from snapshots between releases, but Arch breaks that. Why?
> 
> gcc 5.3.1 built from a snapshot tarball is calling itself 5.3.0 unless you
> use "--with-pkgversion".  I do not consider this a problem with Arch if gcc
> does not report the correct version with the default configure options.

I don't see that:

$ /tmp/gcc5/bin/gcc -v
Using built-in specs.
COLLECT_GCC=/tmp/gcc5/bin/gcc
COLLECT_LTO_WRAPPER=/tmp/gcc5/libexec/gcc/x86_64-unknown-linux-gnu/5.3.1/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-5-20160209/configure --disable-bootstrap
--prefix=/tmp/gcc5 --enable-languages=c,c++ --disable-libsanitizer
--disable-libgomp --disable-libmpx --disable-liboffloadmic
--disable-libquadmath --disable-libcilkrts --disable-libitm --disable-libvtv
--disable-libssp
Thread model: posix
gcc version 5.3.1 20160209 (GCC) 

The version should be 5.3.1 YYYYMMDD , and the paths should include 5.3.1

That was built just now from
ftp://ftp.mirrorservice.org/sites/sourceware.org/pub/gcc/snapshots/5-20160209/gcc-5-20160209.tar.bz2