… and much, much more (#27) * Add support for software Pulse-width Modulation (PwM) on all channels summary of updates in this patch: pwm: - implement PwM interface as per spec - implement basic software PwM functionality - add logic tests for PwM interface - extend threading object hierarchy to integrate PwM thread control with existing thread control flow logic this includes remaming _PollThread as _LineThread - add _line_thread_* thread type constants - add thread_{stop,start} interface to _Line to avoid direct manipulation of _LineThread - add dutycycle and frequency fields to class _Line - update spec to reflect new internal changes (in line_* interface) - bump spec revision number to 2 - s/line_start_poll/line_poll_start/g - add dictionary mapping thread types to thread entry points examples: - fix broken examples - add flashing light example - deprecate morese code example - remove old wrap.py prototype - fix and enforce style in examples - set executable bit on all examples misc: - add a note about notes.sh to README - add header to RPi/_GPIO - mention that the project is free software in the README and header - export the PWM name in RPi/GPIO/__init__.py for access via GPIO name - s/is_all_bools/is_all_bools_or_directions/g this function now returns True when GPIO.HIGH or GPIO.LOW are the data or in the list - fix GPIO.HIGH and GPIO.LOW all being interpreted as True by output() - remove the directory named 'misc' that contained files to reproduce a random bug and should not have been in the master branch - fix test-style.sh not scanning tests and examples - fix a lot of style errors in the tests - replace some magic numbers with hopefully more descriptive constants - remove large ascii art TODO and redundant comments at the top of _GPIO.py tl;dr: Anyway, here's software PwM Signed-off-by: Joel Savitz <joelsavitz@gmail.com> * add missing constants (#24) Signed-off-by: Joel Savitz <joelsavitz@gmail.com> * add script to enable non-root access to gpio pins Signed-off-by: Joel Savitz <joelsavitz@gmail.com> * Rename SPECS directory to packaging to fix (#22) this directory is for files that pertain to distro-specific packaging Signed-off-by: Joel Savitz <joelsavitz@gmail.com> * Integrate a number of existing RPi.GPIO API examples and much more changes in this patch: .coveragerc: - report coverage information on RPi/core.py instead of RPi/_GPIO.py due to file name change Makefile: - %s/unit-cov/test-cov/g - removed an extraneous comment RPi/GPIO/__init__.py: RPi/GPIO_DEVEL/__init__.py: - Added header docstrings - formatted and sorted API imports RPi/_GPIO.py -> RPi/core.py: - File renamed to clarify purpose - convert header to docstring format - sort and organize functions and constants - fix documentation errors and gramatical mistakes - add automatic invokation of cleanup() at interpreter exit - add RPi.GPIO.{RPI_INFO, RPI_REVISION, VERSION} constants - fix various cases of mishandling of iterable channel and value lists - fix pwm thread never releasing lock if thread dies due to exception when the main thread revokes access, however I suspect that this actually be a surface fix for another bug, will look into further (#29) - small tweaks to improve modularity - add support for single value and iterable lists of values for RPi.GPIO.cleanup() - allow PwM to successfully start without explicitly setting up a channel as an output because this library has a more relaxed attitude about those things - add warn on invalid call to PWM.start() - fix board numbering mode being actually pretty broken due to inconsistent mishandling of channel inputs of various forms - fix line_get_value being actually very broken and not retuning anything and therefore always returning None - fix gpio_function() being actually entirely broken so it returns the current function representing what is going on on a GPIO channel as RPi.GPIO would, but this requires further testing (#30) - fix RPi.GPIO.PWM.start() not returning the truth of whether a PwM thread has been successfully started or not examples/callback2.py: - add shabang examples/flash18.py: - remove debuginfo output toggle examples/input_examples.py: examples/output_examples.py: examples/pwm2.py: examples/pwm_usage.py: examples/random_usage.py: examples/try_import.py: - copy example code from Ben Croston's RPi.GPIO wiki on sourceforge - see the files for the sources - ensure that the code works with this library. I found quite a few bugs integrating these files - fix a few mistakes in the code - fix style examples/morse.py: - update morse code to current API version - add basic input validation - fix style examples/pwm.py: - renamed to pwm1.py for consistency examples/pwm1.py: - possible issue: takes two Ctl+C's to stop - this may be revelvant to (#29) examples/pwm3.py: examples/pwm4.py: - Copy example code from some forum - see the files for the sources non_root_permission.sh: - expand note on lack of persistence of effects post reboot spec/spec.tex: - fix some typos - tweak style and title - add entry for channel_valid_or_die - update some of the technical spec spec/sources.bib: - replace \url with \texttt spec/spec.pdf: - regenerate document with 1 inch margins test-style.sh: - scan RPi/core.py instead of RPi/_GPIO.py due to name change - add scan of examples/morese.py since it is again working tests/test_gpio.py: - style tweaks - fix test_gpio_function by making it actually test for the feature that now is possibly implemented correctly - modify test_setdebug info to turn it off before the next call to Reset to avoid printing a lot of text to the terminal - extend test_cleanup to validate the new features of cleanup() tests/test_pwm.py: - more thoroughly test PWM.start() Signed-off-by: Joel Savitz <joelsavitz@gmail.com>

PwM thread does not stop gracefully when example exists #29

Description

Metadata

Metadata

Assignees

Labels

Type

Fields

Projects

Milestone

Relationships

Development

Issue actions