gr-qtgui: refactoring old sinks #6790
Conversation
|
Just took a quick look at the includes. It looks like the API should be backward compatible, but removing some of the include files will break build backward compatibility. What would people think about putting back these includes and making them wrappers around the common includes? The goal is to make it so anyone compiling against the current code can compile without changes against the new code. |
|
@R-ohit-B-isht this PR picked up a bunch of unrelated commits. |
R-ohit-B-isht
force-pushed
the
qt-gui-refactoring
branch
from
c62a99e
to
bca3698
Compare
R-ohit-B-isht
force-pushed
the
qt-gui-refactoring
branch
6 times, most recently
from
ac257c5
to
45b8328
Compare
willcode
added
Backport-3.10
Signed-off-by: R-ohit-B-isht <rbtunes0@gmail.com> Signed-off-by: Jeff Long <willcode4@gmail.com>
willcode
force-pushed
the
qt-gui-refactoring
branch
from
45b8328
to
b424472
Compare
|
@R-ohit-B-isht I removed the include cmake changes and pushed to your branch. I'll work on removing the old code from the specialized include files. |
Signed-off-by: Jeff Long <willcode4@gmail.com>
|
@marcusmueller could you take a quick look. The only goal here is to reduce duplicate code, no improvements. The old include files are left as wrappers for the new ones, for API backward compatibility. We don't have any test code that explicitly tests the old includes, as far as I know. |
| @@ -466,7 +466,7 @@ templates: | |||
| ${ gui_hint() % win} | |||
|
|
|||
| cpp_templates: | |||
| includes: ['#include <gnuradio/qtgui/${type.fcn}.h>'] | |||
| includes: ['#include <gnuradio/qtgui/freq_sink.h>','#include <gnuradio/qtgui/${type.fcn}.h>'] | |||
There was a problem hiding this comment.
The second #include isn't needed here. Same in the other yml files.
| // void set_time_domain_axis(double min, double max); | ||
| // void set_constellation_axis(double xmin, double xmax, | ||
| // double ymin, double ymax); | ||
| // void set_constellation_pen_size(int size); |
There was a problem hiding this comment.
good time to also get rid of these
| # spectrumUpdateEvents_python.cc | ||
| # spectrumdisplayform_python.cc | ||
| # timeRasterGlobalData_python.cc |
There was a problem hiding this comment.
I don't think we'll ever expose these to python. Might be a good time to remove.
| # plot_raster_python.cc | ||
| # plot_waterfall_python.cc |
| /* | ||
| void | ||
| sink_impl<T>::set_time_domain_axis(double min, double max) | ||
| { | ||
| d_main_gui.setTimeDomainAxis(min, max); | ||
| } | ||
|
|
||
| void | ||
| sink_impl<T>::set_constellation_axis(double xmin, double xmax, | ||
| double ymin, double ymax) | ||
| { | ||
| d_main_gui.setConstellationAxis(xmin, xmax, ymin, ymax); | ||
| } | ||
|
|
||
| void | ||
| sink_impl<T>::set_constellation_pen_size(int size) | ||
| { | ||
| d_main_gui.setConstellationPenSize(size); | ||
| } | ||
| */ |
|
@marcusmueller I'll come back and clean up the things you mentioned another time ... not hurting anything for the moment and I want to get the RC out. |
Signed-off-by: Jeff Long <willcode4@gmail.com>
Signed-off-by: Jeff Long <willcode4@gmail.com>
|
Tested cpp gen. Time Sink fails with a topology error and GUI sink seems to lock up. We'll have to iron this out later - not going in to 3.10.8.0. |
Description
Refactoring the float and complex implementations of qtgui sink block into a single block using generics and templates.
Related Issue
Which blocks/areas does this affect?
Frequency Sink
Waterfall Sink
Sink
Time Raster Sink
Time Sink
Eye Sink
Testing Done
yes
Checklist