Oops, I meant to reply to the lists, not just Geoff ...
On 16/09/15 18:35 +0100, Jonathan Wakely wrote:
On 16/09/15 10:08 -0700, Geoff Pike wrote:
Also, to clarify, I am primarily seeking high-level comments; I am new
here and don't want to waste anybody's time.
Hi Geoff,
This looks very interesting indeed, but we've just gone through one
big ABI break earlier this year and my nerves and our users probably
can't take another one in the near future!
That means *replacing* the current containers is unlikely in the short
term, but we could certainly offer alternatives (as e.g.
__gnu_cxx::unordered_map instead of std::unordered_map), and could
maybe have a configure option that would make them the default.
So I don't want to discourage you from working on this, but would have
been a lot more positive about the proposed changes 18 months ago :-)
P.S. While I have your attention, could I ask you to comment on
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55815 ? I think I tried
to reach you about it a while back, via someone else at Google, but
failed. That's another case where we have probably missed our chance
to change the default (i.e. std::hash) but we could still offer a
__gnu_cxx::__sip_hash functor as an extension.
