Stateful allocators in libstdc++v [was: Re: Patch: stl_vector.h]

[Matt Austern]

On Oct 21, 2004, at 11:00 PM, Dhruv Matani wrote:\

>>> My favorite solution is that splicing or swapping between containers
>>> with unequal allocators should be undefined behavior.  Howard Hinnant
>>> argues that swap should be defined and should swap both the contents
>>> and the allocators themselves.  (You may notice a potential
>>> complication with that solution...)
>>
>> Yes, but for swap() it seems fine.. The problem only seems to creep in
>> when we apply the same logic to splice. I'm talking about applying
>> solution(3) to swap here.
>>
>> Then again for splice we can go with your solution about throwing, but
>> undefined behaviour seems a bit harsh to me ;-)
>>
>> Then again, since an implementation has the liberty to choose O(N) or
>> O(1) size, wherein the complexity requirements for splice change,  
>> there,
>> they can go with option(2) which you have proposed. However, it
>> introduces a new situation wherein splice can throw!
>>
>> All this probably out of the scope of this list. Maybe I'll start a
>> thread on comp.std?
>
> Ok, I did a bit of googling, and this issue seems to have been  
> discussed
> quite widely already, thus not probably mandating another discussion.
>
> The links for everyone's reference are:
> http://groups.google.com/groups? 
> q=swap+member+function+containers+swap+allocator&hl=en&lr=&selm=948vq8% 
> 24vdf%241%40nnrp1.deja.com&rnum=7
> http://groups.google.com/groups? 
> q=swap+member+function+containers+swap+allocator&hl=en&lr=&selm=9a4jtg% 
> 243lfso%241%40ID-62495.news.dfncis.de&rnum=3
> http://groups.google.com/groups? 
> q=swap+member+function+containers+swap+allocator&hl=en&lr=&selm=gridd 
> -1407011016210001%40modemcable215.92-202-24.mtl.mc.videotron.ca&rnum=9
> http://groups.google.com/groups? 
> q=swap+member+function+containers+swap+allocator&hl=en&lr=&selm=XHMbc.3 
> 0983%24jVF.981%40news04.bloor.is.net.cable.rogers.com&rnum=10
> http://groups.google.com/groups? 
> hl=en&lr=&threadm=948vq8%24vdf%241%40nnrp1.deja.com&rnum=7&prev=/ 
> groups%3Fq%3Dswap%2Bmember%2Bfunction%2Bcontainers%2Bswap%2Ballocator%2 
> 6hl%3Den%26lr%3D%26selm%3D948vq8%2524vdf%25241%2540nnrp1.deja.com%26rnu 
> m%3D7
>
> Matt, is a resolution forthcoming in the present meeting?
>

No.  It's an open issue, it's related to a number of other open issues,  
and it was not discussed at the October '04 meeting.

			--Matt