I rebuilt the OOT from scratch - new directory, gr_modtool newmod, add, etc.
Then copied over just the edited source code, leaving all the other files
intact.
Except to edit /lib/CMakelist.txt to include my c++ (non impl) code. It
still fails in the exact
same way.
I modified /usr/lib/python3/dis
HI Josh - here is the entire file ./python/bindings/python_bindings.cc:
/*
* Copyright 2020 Free Software Foundation, Inc.
*
* This file is part of GNU Radio
*
* SPDX-License-Identifier: GPL-3.0-or-later
*
*/
#include
#define NPY_NO_DEPRECATED_API NPY_1_7_API_VERSION
#include
namespace
That looks right. And in python_bindings.cc, do you see the import of base
block methods like:
PYBIND11_MODULE(blocks_python, m)
{
// Initialize the numpy C API
// (otherwise we will see segmentation faults)
init_numpy();
// Allow access to base block methods
py::module::impo
Hi Josh - thanks for your help.
from ./python/bindings/hermesNB_python.cc:
void bind_hermesNB(py::module& m)
{
using hermesNB= ::gr::hpsdr::hermesNB;
py::class_>(m, "hermesNB", D(hermesNB))
.def(py::init(&hermesNB::make),
py::arg("RxFreq0"),
... long list of arg
Tom,
What does your hermesNB_python.cc look like?
There should be a declaration for the binding in there that looks like:
py::class_ wrote:
> I'm working on porting my OOT to gr 3.9 The 3.7 and 3.8 versions work
> fine.
> The ported code is compiled and make installed. My OOT is visible in
I'm working on porting my OOT to gr 3.9 The 3.7 and 3.8 versions work fine.
The ported code is compiled and make installed. My OOT is visible in GRC,
and
I've added to a new flowgraph.
My OOT is in category HPSDR and the module is hermesNB (of several).
When I try to execute a simple flowgraph
