Skip to content

CHANGELOG.md

Latest commit

 

History

History
115 lines (103 loc) · 6.41 KB

CHANGELOG.md

File metadata and controls

115 lines (103 loc) · 6.41 KB
Raw

Changelog

All notable changes to hackrfpy are documented here.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

Work on the current development branch. Entries move to a versioned section on release.

Added

  • Type annotations across the entire shipped package, and mypy promoted to a blocking CI gate (disallow_untyped_defs). The package has always shipped a py.typed marker, which tells downstream type-checkers the inline annotations are real; previously only 1 of ~114 definitions was annotated, so that marker was a false promise. It is now enforced.
  • HostOps protocol (_host.py) making the contract between HackRF and the command mixins explicit and type-checkable. The mixins call ~27 methods on self that live on the host class; that dependency was previously implicit in a comment. Runtime composition and MRO are unchanged.
  • Continuous integration: GitHub Actions workflow running the test suite across Windows, Linux, and macOS on Python 3.11-3.13, plus a ruff + mypy lint job.
  • CLI test suite covering argument parsing, the mode state file, preset resolution, info/detect/doctor, rx/tx/sweep dispatch, and the module-level exit-code mapping. Overall coverage raised from 73% to 86% (cli.py from 12% to ~98%).
  • Community health files: CONTRIBUTING.md, CODE_OF_CONDUCT.md, SECURITY.md, issue templates, and a pull request template.

Changed

  • Library diagnostics now go through the standard logging module instead of print(). Records are emitted on the hackrfpy logger: warnings at WARNING, verbose progress messages at INFO. A consumer can now route, reformat, or silence hackrfpy's output like any other library.
    • Diagnostics no longer touch stdout. print_message previously wrote to stdout, so hrf sweep -v > out.csv prepended [*] mode: rx into the CSV. stdout is now reserved for data (sweep CSV, IQ on -r -) and explicitly requested output (--print-cmd). Diagnostics go to stderr.
    • Console behavior is unchanged for scripts and the CLI: warnings still appear with no setup at all, and verbose=True / -v still prints progress. If the host application has configured logging, hackrfpy stays out of the way and simply propagates records to it.
  • Packaging classifiers: removed the contradictory Operating System :: OS Independent (the library shells out to the Windows hackrf-tools binaries and Linux/macOS operation is unverified), leaving Operating System :: Microsoft :: Windows. Development status raised from 3 - Alpha to 5 - Production/Stable to match the 1.0.0 release.
  • Ruff configuration added (line-length = 100, select = ["E", "F", "W"]), and the codebase made lint-clean so the CI lint job is meaningful.

Fixed

  • The atexit backstop never reaped an orphaned PersistentReceiver. PersistentReceiver registers itself in the live-handle registry, whose shutdown hook calls if h.is_alive(): h.stop() -- but is_alive() did not exist on the class, so the resulting AttributeError was swallowed by the hook's bare except Exception and the receiver was silently left running. Found by the typing pass; is_alive() added and pinned with a regression test.
  • Reading from a closed PersistentReceiver raised a bare TypeError: 'NoneType' object is not iterable instead of a typed error; it now raises HackRFDeviceError with an actionable message.
  • sweep_stream(..., print_cmd=True) would hand StreamCtx a None and crash; it now raises HackRFValueError pointing at sweep(..., print_cmd=True).
  • Unbounded memory growth on long-lived RX/TX handles: _Process drained child stdout/stderr into lists that were never trimmed, so an open-ended capture or repeat transmit retained every per-second stats line for the life of the process. Both drain paths now share a 64 KB cap (_DRAIN_CAP).
  • transmit() now verifies the source file exists before arming TX and spawning hackrf_transfer, raising HackRFEnvironmentError instead of failing with a generic non-zero exit from the tool. The TX-mode gate is still checked first, and print_cmd dry runs skip the check.
  • SigMF sidecars now declare the hackrf namespace in core:extensions. Previously the hackrf:* gain keys were written without declaring the extension, which strict SigMF validators reject.
  • Removed unreachable dead code in core.py (an orphaned return load_iq(...) after a return, referencing three undefined names).
  • The test suite no longer writes a stray capture.sigmf-meta into the working directory on every run.
  • CITATION.cff version and release date corrected to 1.0.0 / 2026-06-16, aligning the citation metadata with pyproject.toml and the tagged release.

1.0.0 - 2026-06-16

Initial public release.

Added

  • Python wrapper and non-GUI command-line tool (hrf) for the HackRF One, driving the hackrf-tools binaries directly (no libhackrf bindings).
  • Receive: bounded and streaming capture, with decode to complex64 and self-describing SigMF (.sigmf-meta) sidecars.
  • Spectrum sweep with streaming CSV parsing.
  • Transmit: file playback and constant-wave source, guarded by an explicit operating-mode gate (transmit refuses unless the instance is switched to TX mode, which emits a one-time safety banner) and a TX gain ceiling.
  • Device management passthroughs (clock, Opera Cake, SPI flash, debug) and preflight info / detect / doctor helpers.
  • Cross-platform process lifecycle handling with best-effort clean interrupt (SIGINT on POSIX, CTRL_BREAK on Windows) and an atexit backstop for live RX/TX handles.
  • Validation layer with hard-range checks, gain snapping to real device steps, and a parameter readback (last_params) reflecting the values actually used.