Fixed

• Internal CC1101 receiving nothing in 1.1.0. Routing the on-board
  radio through the subghz_devices_* plugin API left the chip in a
  state where the FIFO never accumulated bytes, so no telegrams were
  decoded. Reverted the internal path to direct furi_hal_subghz_*
  calls (the same code that worked in 1.0). The plugin abstraction is
  retained for the external module, which it was actually designed for.

Changed

• Trimmed README and GitHub Pages

Added

• External CC1101 module support. Settings → Module → External routes
  the radio through the cc1101_ext plugin on the GPIO header, with auto-
  fallback to the on-board chip when no external module is detected. Lifted
  from ProtoPirate's radio_device_loader pattern. OTG power is enabled
  only while scanning.
• New manufacturer drivers:
  • bmeters/hydrodigit — BMeters Hydrodigit water (firmware 0x13)
  • engelmann/hydroclima — Engelmann Hydroclima room-temperature HCA
  • sontex/rfmtx1 — Sontex RFM-TX1 water (legacy XOR + new OMS firmwares)
  • zenner/zenner0b — Zenner B.One water (1/256000 m³ resolution)
  • apator/apator_na1 — Apator NA-1 water
  • techem/mkradio3a — Techem MK-Radio 3a water (compact frame v0x70)
  • misc/gwfwater — GWF water (CI 0x8C ELL frames)
  • misc/bfw240radio — BFW 240-Radio HCA
• Driver engine refactor. Introduced drivers/engine/ with a uniform
  WmbusDriver interface (mvt, decode_ex, optional legacy decode) and
  a static registry. New drivers drop into drivers/europe/<vendor>/ and
  ship as one C file each — see drivers/CONTRIBUTING.md.
• Host-side regression tests. tests/test_izar and tests/test_ports
  exercise every new driver against the wmbusmeters reference telegrams.
  Wired into make -C tests check and the GitHub Actions workflow.

Fixed

• OMS short/long header was being walked as DIF/VIF. oms_split_emit
  now skips the post-CI header (4 bytes for CI=0x7A/0x5A, 12 bytes for
  CI=0x72/0x53/0x8B) before invoking the application-layer walker. This
  removes a class of phantom records (e.g. an "ACC" byte being decoded as
  a 4-byte volume) that affected every unencrypted CI=7A/72 frame.
• rfmtx1 index off-by-one. The driver was treating apdu[0] as the
  CI byte instead of post-CI, so the legacy XOR de-obfuscation reached past
  the volume field and the BCD datetime decoded as garbage. Re-anchored
  every offset against frame[11] → apdu[0].
• Added regression tests for both bugs above.

Changed

• Worker thread now talks to the radio through the subghz_devices_* API
  instead of furi_hal_subghz_* directly. The chip-FIFO drain still uses
  raw SPI, but now selects the right bus handle (furi_hal_spi_bus_handle_subghz
  for internal, _external for external) based on the active device.
• License changed from MIT to GNU GPL v3.0 or later. Aligns with the
  upstream reference implementations (wmbusmeters, rtl_433) we
  cross-checked against and ensures any redistributed derivative ships
  its complete corresponding source. Re-using earlier 1.0.x sources is
  still possible under MIT; the GPLv3 terms apply from 1.1.0 onwards.