Hi Jonathan, Thanks for looking into this! It’s unfortunate that we can't know if the pointer in the iterator is legal to dereference. Nevertheless, inspecting smart pointers is a far more common use case, it's good that it has finally been sorted out.
Thanks, Juraj On Tue, 9 Jan 2018 at 23:10, Jonathan Wakely <jwak...@redhat.com> wrote: > On 09/01/18 18:50 +0000, Jonathan Wakely wrote: > >On 09/01/18 15:02 +0000, Jonathan Wakely wrote: > >>On 09/01/18 14:59 +0000, Jonathan Wakely wrote: > >>>On 04/01/18 11:22 +0100, Juraj Oršulić wrote: > >>>>Hi Jonathan (and the libstdc++ list). Can we revive this? I sent the > >>>>patches for improving the smart pointer pretty printers in March. They > >>>>haven't been reviewed. > >>> > >>>Thanks for the reminder. I'm testing the attached patch, which has > >>>been rebased on trunk, but I'm getting test failures for the > >>>shared_ptr printers. > >>> > >>>I can't see any difference between the expected output and what GDB > >>>prints, so I think it's some DejaGnu or Tcl/Tk oddity. The actual > >>>printer seems to work OK. > >> > >>No, the problem is that I'm just misreading the results. > >> > >>I'll finish testing and commit this soon. > > > >I changed the code slightly, so it still works for GDB versions older > >than 7.5, which don't have the fix to allow children() to return a > >list. There's no easy way to tell if the GDB interpreting the code has > >the fix or not. > > > >Thanks very much for the patches, and sorry for the delay reviewing > >them. > > > >Tested x86_64-linux, committed to trunk. > > I've just realised I didn't put your name in the ChangeLog, sorry! > (I semi-automate my changelog entries, so have to remember to do an > extra step when I'm not the author, or not the sole author). > > Fixed with this patch, committed to trunk. > > > >
