RE: [boost] 1.30.0-b1: filesystem::path::swap
At 06:18 AM 3/12/2003, Geurt Vos wrote: > >> >Hi, >> >Is there any reason boost::filesystem::path doesn't >> >provide a swap(path &) function? If there is, I think >> >the docs should explain why, but if there isn't, well, >> >can it still be implemented before 1.30.0 goes gold? >> >> Let me turn the question around and ask what your >> expectations would be for >> a swap member beyond what is already offered by std::swap? >> >> No throw guarantee? >> > >Yes. That is, this is the main (or actually only) >reason for asking. OK, I've added it to do-list.htm. I don't want to try to add it for 1.30.0 - it's too late at this point. If you want to send me a patch, feel free! --Beman PS: In researching this, I found that the C++ standard seemed to indicate that basic_string::swap() did not give the expected guarantee not to throw. I've raised the issue with the committee, as it looks to me like a defect. Of course, the defect might just be in my expectation for basic_string::swap(). ___ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
RE: [boost] 1.30.0-b1: filesystem::path::swap
> >Hi, > >Is there any reason boost::filesystem::path doesn't > >provide a swap(path &) function? If there is, I think > >the docs should explain why, but if there isn't, well, > >can it still be implemented before 1.30.0 goes gold? > > Let me turn the question around and ask what your > expectations would be for > a swap member beyond what is already offered by std::swap? > > No throw guarantee? > Yes. That is, this is the main (or actually only) reason for asking. > More efficient? > I guess with a swap member, you kind of get this for free. With filesystem::path I consider this a non-issue, though. I mean, the time a possible extra copy takes is negligible to retrieving a directory listing or parsing a config file (opened using filesystem::ifstream). > Or are you asking that filesystem::path satisfy more container > requirements? > No, a no throw 'swap' will suffice... I mean, except for swap, filesystem::path is complete enough for what I use it for. So I obviously also don't really care about other requirements (looking at it from a user point of view). Geurt ___ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Re: [boost] 1.30.0-b1: filesystem::path::swap
At 03:23 AM 3/10/2003, Geurt Vos wrote: >Hi, >Is there any reason boost::filesystem::path doesn't >provide a swap(path &) function? If there is, I think >the docs should explain why, but if there isn't, well, >can it still be implemented before 1.30.0 goes gold? Let me turn the question around and ask what your expectations would be for a swap member beyond what is already offered by std::swap? No throw guarantee? More efficient? Or are you asking that filesystem::path satisfy more container requirements? --Beman ___ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost