Re: [boost] [block_ptr] Update

Subject: Re: [boost] [block_ptr] Update
From: Phil Bouchard (philippe_at_[hidden])
Date: 2012-07-30 18:58:39

On 7/30/2012 2:45 PM, Stewart, Robert wrote:
> Phil Bouchard wrote:
>>
>> It's been more than a year since a completed the block_ptr
>> and it seems as if more recent library are being reviewed
>> first.  This would suggest the FIFO of the schedule is not
>> followed.
>
> We often refer to the review "queue", but that is using the vernacular rather than the computer science meaning of that word.  It is not a FIFO.  The order in which reviews are conducted is dependent upon many factors, but a library will not be scheduled for review without a review manager.
>
> I just looked at http://www.boost.org/development/submissions.html and http://www.boost.org/community/reviews.html and don't see any mention of who finds the review manager, so it's understandable that you misunderstood.  It is your responsibility to secure a review manager.  One is not assigned.  Once you've found someone, you post a message to the Review Wizards asking them to accept your review manager.  If accepted, your review manager and you must then agree on some dates for your review -- you'll want to avoid conflicts with currently scheduled reviews -- and submit them to the Review Wizards who will then schedule your review.

Your explanation is appreciated.  I thought I read somewhere it was a 
first come first served basis so I apologize for the misunderstanding.

If there is any potential review manager reading this then I can say 
that it features:
- Constant complexity
- Deterministic (predictable)

We could also:
- Optimize the allocator it uses
- We could detect cyclic pointers within the same set explicitly when 
the program is on idle (for example) if memory is really needed but this 
is "non a prioritate".

Once again it can be used for any problems such as feedback neural 
networks and all heap allocations are guaranteed to be deleted on the 
termination of the program.

-Phil