This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Patch: stl_vector.h and vector.tcc


Gabriel Dos Reis wrote:

Mark Mitchell <mark@codesourcery.com> writes:

| Matt Austern wrote:
| | >>
| >> I believe Matt's suggestion is the right long term approach, assuming
| >> we have an implementation of #nospam. For the mid-term, I think that
| >> using bits/cpp_type_traits.h is the most viable option -- with proper
| >> uglification.
| >
| >
| > I wonder if this does have to be "long term"? If we think that we
| > have a good specification of preprocessor scoping, I bet it won't
| > be that hard to implement.
| >
| > I don't attend EWG sessions, so I don't have a good sense of how it
| > was received. Did it look to you as if the feature is ready to
| > implement as Bjarne proposed it?
| | I know that Nathan Sidwell had some pretty serious objections to
| Bjarne's proposal, but I'm not sure whether they've gotten through the
| UK committee yet, or not.


I've answered (and I think Bjarne did too) Nathan's questions on the
Evolution reflector. Nathan did not follow up, so I assumed he had no
further concerns. If there still are issues, it would be rather
helpful if he could post them as soon as possible instead of waiting
for the "slower" path to transmit them.


We'll have to let Nathan speak for himself; I have no further information. All I know is that at some point in the past Nathan was not happy with the proposal; to what extent that's been resolved, I'm not sure.

I do think that we should be very hesitant, as a matter of policy, about putting any features into G++ that have not been approved by the committee. I'm not sure exactly what *level* of approval to require; maybe Bjarne's proposal already meets a reasonable threshold, in which case subsequent comments do not apply.

Implementing these features for experimental purposes and putting them on a branch is great; that will save us time when the final version is ratified. However, putting these kinds of features on the mainline would be risky, in my opinion, for the same reason that we hesitate to implement DRs for which it is not obvious what the outcome in the committee will be.

--
Mark Mitchell
CodeSourcery, LLC
(916) 791-8304
mark@codesourcery.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]