------- Additional Comments From squell at alumina dot nl 2005-07-24 16:42 ------- (In reply to comment #7)
> Yes, the standard requirements for iterators exhibit inconsistencies > at many places; for example an InputIterator is not required (by the > Standard) to be copy-constructible; consequently it cannot be used as > function parameter type. Yes, I noticed the standard isn't explicit about that, but isn't that requirement implied in the semantics for '*r++'? I read somewhere the LWG believed the requirement to be implicit. (found: http://www.open-std.org/jtc1/sc22/wg21/docs/lwg-defects.html#241) > BAck to the issue of this PR; I think there is way too vagueness in > the standard and the "obvious" thing to do is to use the natural type > of the intermediate results. However, we may actually want to submit > a DR (I'll do it). I would suggest this issue be suspended in the > meantime. I agree the paragraph should probably make it explicit what type is being used. I think there's another subtle error in the Effects clause, since dereferencing an input iterator invalidates previous copies, and order of evaluation is unspecified. I noticed this also happens in unique_copy. It works on Input Iterators but its semantics are specified in terms of '*i == *(i-1)' or 'pred(*i, *(i-1))'. Perhaps this should be part of the DR, or a seperate one? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22634
