This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: RFC - Another step towards dynarray?
- From: Jonathan Wakely <jwakely dot gcc at gmail dot com>
- To: Ed Smith-Rowland <3dw4rd at verizon dot net>
- Cc: "libstdc++ at gcc dot gnu dot org" <libstdc++ at gcc dot gnu dot org>, RÃdiger Sonderfeld <ruediger at c-plusplus dot de>
- Date: Mon, 28 Apr 2014 01:21:54 +0100
- Subject: Re: RFC - Another step towards dynarray?
- Authentication-results: sourceware.org; auth=none
- References: <535D7114 dot 6060205 at verizon dot net>
On 27 April 2014 22:05, Ed Smith-Rowland wrote:
> Greetings,
>
> I had a start on <dynarray> in my tree and started working on it after stage
> 1 opened on gcc-4.10.
>
> After googling around I stumbled on the always_inline attribute and decided
> to try it. It seems to work.
Are you sure this behaves correctly for cases like
std::vector<dynarray> or dynarray as a member of a class?
> Anyway, somehow I had forgotten about the conversations about this that
> Jonathan and Ruediger had on libstdc++ in the 2013-09 timeframe. So I
> stopped work so as not to scoop anyone - I'm sorry if I did.
>
> I would like to offer this incomplete patch (the allocator ctors are messed
> up, also there's a DR). I think, assuming I really am right about
> [[gny::always_inline]] offering us a way forward. It would be goo to have
> someone take that solution and merge it into one of our patches (I think
> Jonathan is ahead on the allocator thing). Or work something out.
I would forget about allocator support in dynarray for now, until the
committee decide what they want to do.