core: add SWIG macro for namespaced blocks
Removed whitespace and added dtools/bin/remove-whitespace as a tool to do this in the future.
The sed script was provided by Moritz Fischer.
Removes all references to guile and scheme, which are no longer needed.
The remaining pmt_serial_tags.scm is parsed by Python already. Future modifications could change this format to not confuse the point.
Merge branch 'master' into next
Conflicts: gruel/src/swig/Makefile.am
gruel: created common swig include file
This work taken from gruel_work_squashed
Removed autotools, gr-waveform, some cleanup
Nick Foster owes Nick Corgan a six-pack of beer!
docs: fixes some warnings.
There is way too much institutional knowledge in these makefiles now. Another reason I'm glad we're going to cmake, which is being kept more sane as all of the problems are already known.
docs: more fixes for generated swig files. This worked before the merge, but this forces the build system to generate these before processing the other swig files.
docs: better install of swig docs .i files since we actually have a variable to take care of it already.
docs: installing generated _doc.i files under autotools.
docs: autotools Makefile.am generates empty swig_docs .i files; builds in autotools and cmake.
The autotools build does not auto-generate the Python documentation for each block like we do with the cmake build, but the empty .i files are required to build things regardless. The swig_doc.py and doxyxml files are distributed with the autotools build in case someone wants to generate this on their own (see README.doxyxml for how to do this)....
work on swig docs, added to core and digital
cmake: fixed up swig deps properly for make -j*
Also took the time to fix up the swig index file generation.This may have not-been auto-regenerating before as well.
tags: added swigging for gr_tags header
swig: fix rules for generated index files
gr: the CMakeLists.txt took a chill pill
core: make the lib/swig depend on the generated targets
swig: added python script to auto-extract dependencies
Merge branch 'next' of http://gnuradio.org/git/gnuradio into next
swig: language independent exception handler
core: added source deps to swig files (regen on .h,.i)
gr: squashed cmakelists.txt into one commit
core: API declaration macros for core class and function symbols
Minor bug fixes (from last two commits)
Removing docstring generation from this branch.
Fixed Makefiles to generate swig_doc.i
Merged upstream from next
Have swig/guile wrap enums and constants as scheme variables, not functions.
Use load-extension-global instead of load-extension
Merge branch 'next' into guile-next
Regenerated Makefile.swig.gen's
Rename libguile- to libguile-gnuradio- to match guile library naming convention.
The stuff in gnuradio-core ends up with funky names, but that couldbe fixed by renaming gnuradio_core_filter.i -> core_filter.i etc.
Merge branch 'grorg-master'
Major Makefile.am housecleaning. Passes distcheck.
Move all occurrences of swig_built_sources out of Makefile.am's.Move all SWIG related use of BUILT_SOURCES out of Makefile.am's.Clean up 'if PYTHON' conditionalization in gr-*
Still left to do: fix Makefile.swig CLEANFILES and no_dist_files...
Adding typedef for uint64_t and int64_t so we can use them through SWIG.
Clean up lib/swig/Makefile.am, Makefile.common and Makefile.swig
Confirmed that it builds and make checks on all four combintationsof --{enable,disable}-{python,guile}.
Have not tested make dist, but expect that there may be some problemswith it. I'm pretty sure that not all files that need to be removed...
go back to the generating the lists here, so make check works again
remove -/gnuradio_swig_bug_workaround.h, it's not generated anymore.
don't generate this file anymore
regenerated
don't generate the list of scripts here, use the version in Makefile.swig
Add support for guile std::vector< std::complex<float> >.
Modify GR_SWIG_BLOCK_MAGIC to not strip package prefix when usingGUILE outside of gnuradio-core.
This avoids the problem with the guile bindings where multiple blocksend up with the name gr:sink (e.g., audio_alsa_sink). With thischange, it ends us as gr:audio-alsa-sink....
update .gitignore
Remove .gitignore from empty directory
Create guile QA framework.
Also moves hand coded files out of gnuradio-core/src/lib/swig/gnuradiothat were getting nuked by make clean.
Reduce guile load time from 4.5 to 1.2 seconds.
Rewrite re-export-all to only export symbols from the module name supplied.
add script names to swig_built_sources, not BUILT_SOURCES. Clean the generated files than now live in subdirectories.
Rename coerce.scm to runtime-shim.scm. Modify core.scm to use runtime-shim.
Export a few more items
Cleanup gr:connect and gr:disconnect for Guile.
Rename {dis,}connect to {dis,}primitive_connect in .i file.Update python code to reflect change.
Routines to coerce blocks and connect them
move good comment to here where it belongs
Modules now load cleanly!
gnuradio-core modules are almost loading cleanly.
Move sizeof_* from gnuradio.i to runtime.i to avoid multiple definitions
Version that builds guile goops bindings, but dependencies are hosed.
Doesn't compile cleanly, but does enough to allow experimentation withgoops wrappers. We're currently seeing the nasty interaction betweenthe package system, the "export" syntax, and generic-functions....
Add %typecheck for complex.
Fixes problem with moving-average-cc.
Move std_complex.i to gnuradio-core/src/lib/swig/guile.
Seems to basically work. There's some simple test code inlinein gnuradio_core_general.i.
rm gnuradio_swig.scm. Got turned into gnuradio_core.scm
Missed a few: s/_swig_py_/_/
rename s/gnuradio_swig_python/gnuradio_core/g
regenerate
s/_swig_py_/_core_/g
Rename guile libs to libguile-* and install them in $(libdir).
This follows the Guile recommendations and should remove the need foradding an additional directory to LD_LIBRARY_PATH.
Remove #warning from gr_swig_block_magic.i
Rengenerate Makefile.swig.gen's
Remove leading _ from guile library names coded into .i files
Set new install locations for guile scm code and libraries.
I think we may want to rename the guile .so's and install them intolibdir instead...
Remove dead and broken code from GR_SWIG_BLOCK_MAGIC.
Looks like this hasn't been used since we converted the hier_blockstuff to C++, and it's been broken since then.
Move #ifdef SWIGPYTHON outside of %define
There's apparently a time-of-evaluation issue here.The docs say they reevaluate the contents of %define at expansiontime. Apparently SWIGPYTHON et al. aren't defined in whatever scopethat may be.
Move common stuff outside of if PYTHON
Move TOP_SWIG_IFILES outside of if PYTHON
Add stub for gnuradio_swig.scm to get make to work
correct typo in script name
move new suffix rules to common Makefile so everything can use them.
add comments
use suffixes to build scm and py files
correct which lines are under comment
load the general and gengen files, now that they compile.
put the ifdef in the right place
gitignore more swig generated files
gitignore swig generated files
correct copyright date
improve comment
install the scm files for guile too
load what'll load for now
add top level guile file
use foreach...subst to make a list of generated files instead of cut & paste for BUILT_SOURCES
