Den 2016-08-04 kl. 07:47, skrev Alwin Leerling:A large part of the documentation of glibmm (and gtkmm and many other *mm modules) is copied from the documentation of the wrapped C functions in glib (gtk+, etc.). The gmmproc command in glibmm modifies the documentation slightly, but not enough. If you browse the whole reference documentation of glibmm, gtkmm, etc. you will find many (probably more than 100) documentation errors in functions that return std::string or Glib::ustring, or take const std::string& or const Glib::ustring& parameters. The wrapped C functions take or return const char*.On Wed, 2016-08-03 at 22:49 -0400, Hubert Figuière wrote:On 03/08/16 07:06 PM, Alwin Leerling wrote:Of course, my bad. Forgive my momentary lapse of reasoning. Still, should you free std::string objects?Because of https://en.wikipedia.org/wiki/Resource_Acquisition_Is_Initialization you don't have to do anything. These are objects, not pointers to. HubExactly my point. Thus the documentation is wrong. I believe this is the forum to report bugs, or documentation inaccuracies or is there another way to bring this to a maintainers attention. Alwin It's possible to manually fix such documentation errors, but I doubt that anyone wants to do it. It wouldn't take too much time to fix the 4 functions that you've found, but they're not worse than many many other functions. I think this is a kind of documentation error that you have to accept, unfortunately. Unless you, or someone else, make a considerable improvement of the gmmproc command. Kjell |