Re: [boost] Re: [filesystem] new functions proposals

From: Beman Dawes (bdawes_at_[hidden])
Date: 2003-05-05 09:23:10

At 04:14 AM 5/5/2003, Vladimir Prus wrote:

 >Beman Dawes wrote:
 >
 >...
 >> * I think the preconditions should be removed. In each case, the 
behavior
 >> as specified is quite sensible even if ph.leaf().empty(), and may be
 >useful
 >> in some applications. (Assuming you agree, the assert()'s should be
 >> removed, and test cases should be added.)
 >
 >Let me ponder a little. For 'extension', the idea that empty path has 
empty
 >extension seems ok. For 'basename' this looks even more resonable. 
However,
 >for "change_extension", this is not so. If you call
 >
 >    change_extension("", ".txt")
 >
 >you'll get ".txt" as result, which has no sense at all. But I'm not sure
 >the assertion, here, is so big help. So I'll removed precondition here
 >as well, but add a bit of rationale.

Good. I agree with your reasoning.

 >However, I've a question: if is always true that ph.leaf().empty() ==
 >ph.empty.

Yes, that's both intuitively correct and follows from the specification 
path()'s return.

 > IOW, is it ever possible to have empty element in the sequence of
 >strings that compose the path?

No, not in a path which follows the portable grammar. It is a little less 
clear for a native path on a system other than POSIX or Windows, but I 
don't think we should agonize over that corner case.

Thanks,

--Beman