|
Boost : |
Subject: Re: [boost] GSoC 2010: Heaps and Queues
From: Dan Larkin (danielhlarkin_at_[hidden])
Date: 2010-04-07 13:57:13
Andrew Sutton wrote:
> Not really "internal", just pointer-like. Typically they're used for
> iterating -- you can tell by the name :) but they're also used in
> pointer-like ways. Look at the STL list interfaces, especially insert() and
> erase().
Okay, so once again, focusing on studying for my exam tonight right now,
but just doing a bit of research to get the gears turning in the
background. Standard iterator functionality requires '++', '--', '==',
'begin()', and 'end()', correct? Did I miss anything?
Other than that, I'm still a confused about the behavior of iterators
within partially ordered data structures. It's one thing to have a
pointer abstraction, but traversal seems to be an issue. In particular,
if any changes to the structure are made during the traversal, the
structure could re-arrange itself in such a way that individual nodes
could be traversed multiple times before others are visited once,
without any changes to the node value itself. This seems like poor
behavior to me, and I'm not sure what to do with it. Perhaps someone
could shed some light on the issue?
> Don't feel too bad about being distracted. I've been at a postdoc
> interview for the last 2 days. Imagine how I feel :)
Oh man. I don't even want to think about that. D:
Dan Larkin
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk