Re: [Boost-users] [filesystem]Sorting based on equivalence information?

From: Rush Manbert (rush_at_[hidden])
Date: 2005-11-14 15:34:24

Paul Giaccone wrote:
> Rush Manbert wrote:
>>My first question: Can anyone suggest a better method of achieving the 
>>duplicates detection?
>> 
>>
> 
> Maybe you could use the graph library to build a graph of your directory 
> structure and then work with that to find duplicate paths, loops, etc.  
> I know next to nothing about the library, but this sounds like a 
> graph-theoretic problem to me.
> 
Thanks for the suggestion. The big issue with this approach is actually 
the same problem I have with my "second question" in the original post. 
The equivalence information is what really identifies a directory or 
file, not the path represented as a string. Once I start inheriting from 
or  modifying the path class to get at that information (which I will 
need in order to uniquely identify the graph nodes), I might as well 
put them all in a set and test whether a new path is already in the set.

- Rush