Closed 2bndy5 closed 1 year ago
:exclamation: No coverage uploaded for pull request base (
master@9ce1c68
). Click here to learn what that means. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #44 +/- ##
=========================================
Coverage ? 86.90%
=========================================
Files ? 16
Lines ? 2466
Branches ? 0
=========================================
Hits ? 2143
Misses ? 323
Partials ? 0
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more
There's likely a better way to monkeypatch classes with pytest's monkeypatch
fixture, but that would be way more involved as it would require replacing the SPI bus usage with a logging-like function... I'm ok with 40% test code coverage for now.
I couldn't help myself. With a miniature strate machine instead of a physical radio (and a thin SPI IO shim), I was able to test most of the code base. Only certain things that require external physical radio could not be tested (like RF24Mesh handshaking).
I was happy with 40% coverage; now I'm thrilled for > 80% coverage.
__version__
attribute is still adjusted by circuitpython-build-bundles script according togit describe
(for CirPy community bundle releases)Bug fixes
RF24Mesh.*_dhcp()
function, I found a bug and fixed it inRF24Mesh.save_dhcp()
. It was crashing when writing thejson.dump()
output to a binary file handler.UrlServiceData
class, I found a problem where the byte codes for URL suffices (.com/
,.org
) were not used. With the new fix, the byte code prefixes (http://
,https://www.
) are now only used at the start of a given URL.write_readinto()
requires the MISO and MOSI buffer to be the same length. This mandates more memory allocated for each radio object instantiated.Other enhancments
RF24.write()
will assert the length is in acceptable range only ifdynamic_payloads
is enabled. This might speed up pre-processing time for transmissions.