On Sun, 8 May 2022 at 15:09, Paul Smith <paul@xxxxxxxxxxxxxxxxx> wrote:
>
> On Sun, 2022-05-08 at 09:16 +0100, Jonathan Wakely wrote:
> > > > xmethods don't help me (IIUC) because I'm in the middle of some
> > > > Python function and the value I want to retrieve is in a Python
> > > > variable, not in a GDB variable, so I can't easily access it with
> > > > parse_and_eval().
> > >
> > > gdb.set_convenience_variable('mgr', val['mgr'])
> > > init = gdb.parse_and_eval('$mgr->initialized')
> > >
> > > This will use the xmethod to evaluate the expression.
> >
> > And then:
> >
> > if init:
> > return gdb.parse_and_eval('*$mgr')
>
> Yes... this is why I qualified my statement with "easily" :).
>
> Also if I'm doing this as I search through a 700,000 element container,
> which is already not super-speedy, the performance is likely not going
> to be great.
>
> I imagine the cleanest solution is to have a Python API for STL that
> works with and returns gdb.Values, then have the pretty-printers call
> that API and format the result. Then others can use that API as well.
>
> I get that the current suite of Python functions are geared
> specifically towards using STL objects directly from the GDB UI: both
> pretty-printing and accessing them. As someone who's writing my own
> Python functions for a complex C++ program, however, an API as above
> would be more helpful; we have lots of methods that grovel through
> complex structures looking for specific elements, etc.
Then you are the perfect person to scratch this itch :-)
>
> And, I understand the answer "patches welcome" :). I'm just putting
> this out there for people to consider.
