On 9/28/23 11:26, Jason Merrill wrote:
On 9/28/23 05:55, Richard Sandiford wrote:
poly_int was written before the switch to C++11 and so couldn't
use explicit default constructors. This led to an awkward split
between poly_int_pod and poly_int. poly_int simply inherited from
poly_int_pod and added constructors, with the argumentless constructor
having an empty body. But inheritance meant that poly_int had to
repeat the assignment operators from poly_int_pod (again, no C++11,
so no "using" to inherit base-class implementations).
All that goes away if we switch to using default constructors.
The main complication is ensuring that braced initialisation still
gives a constexpr, so that static variables can be initialised without
runtime code. The two problems here are:
(1) When initialising a poly_int<N, wide_int> with fewer than N
coefficients, the other coefficients need to be a zero of
the same precision as the explicit coefficients. This was
previously done in a for loop using wi::ints_for<...>::zero,
but C++11 constexpr constructors can't have function bodies.
The patch instead uses a series of delegated initialisers to
fill in the implicit coefficients.
Perhaps it's time to update the bootstrap requirement to C++14 (i.e. GCC
5, from eight years ago). Not that this would affect this particular
patch.
IIRC the primary reason we settled on gcc-4.8.x was RHEL7/Centos7. With
RHEL 7 approaching EOL moving the baseline forward would seem to make sense.
I'd want to know if this affects folks using SuSE's enterprise distro
before actually making the change, but I'm broadly in favor of moving
forward it it's not going to have a major impact on users that are using
enterprise distros.
jeff
