Boost logo

Boost :

Subject: Re: [boost] [heap][formal review]
From: John Bytheway (jbytheway+boost_at_[hidden])
Date: 2011-06-13 17:49:04


On 13/06/11 17:54, Tim Blechmann wrote:
<snip>
>> I'm still not convinced that Stability is an essential requirement, or
>> that the library's approach to the solution is adequate. I would have
>> preferred the library not attempt provide such a feature. The same
>> could be said for iterators.
>
> depending on the use case it is definitely a requirement ... i have seem many
> workarounds in real-world software, where the stability of non-stable priority
> queues need some ugly workarounds ...

To weigh in briefly, I have needed a stable priority queue on at least
one occasion, and achieved it through a rather tiresome and convoluted
method involving a heap of linked lists indexed by a hash table. I
would value this being moved into a library that has thought more
carefully about the best way to solve it.

John Bytheway


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk