Open MLsmd opened 6 years ago
Other things to improve:
there's this goddamn (what a design fail on the side of gr-newmod) try
: / catch: pass
thing around the from _swig import *
clause. That is cancer and must go; if the C++ components aren't loadable, there should be a clear and loud exception.
The python file is generated from the .ui file using pyuic4
. Hence, the .py file should not be tracked within git at all. Instead, add_custom_command
to generate the file on the fly when building in the python/GUI/CMakeLists.txt; something like
add_custom_command(OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/user_frontend.py"
COMMAND pyuic4 -o "${CMAKE_CURRENT_BINARY_DIR}/user_frontend.py" "${CMAKE_CURRENT_SOURCE_DIR}/user_frontend.ui"
MAIN_DEPENDENCY "${CMAKE_CURRENT_SOURCE_DIR}/user_frontend.ui")
and of course, then replacing user_frontend.py in the GR_PYTHON_INSTALL by "${CMAKE_CURRENT_BINARY_DIR}/user_frontend.py"
Thanks for the review.! I did the requested changes and some more wrap up, docu, ... I think we are ready for merge in master now.
So, formally OK, now we just have to fix the build failures:
../lib/fib_sink_vb_impl.cc:84:23: error: return type of out-of-line definition of 'gr::dab::fib_sink_vb_impl::process_fig' differs from that in the declaration
fib_sink_vb_impl::process_fig(uint8_t type, const char *data, uint8_t length) {
^
../lib/fib_sink_vb_impl.h:51:11: note: previous declaration is here
int process_fig(uint8_t type, const char *data, uint8_t length);
~~~ ^
I think you meant to use void
here, as the actual implementation never returns a value?
(also, hint, maintainers love code that is going to merge smoothly into master, so a git rebase -s ours
might be a good idea)
I am not yet finished with reformatting and documenting the whole code. I will give you a hint here when I finished the work. And thanks for the tip with rebase -s ours, I am going to try it out :)
I am now finally finished with reformatting and documenting the rest of the code and ready for the merge.
Juche! Will review tomorrow :)