Apator HCA: align field decode with wmbusmeters apatoreitn driver

[mac22]

Replaces the ad-hoc Apator heat cost allocator (E-ITN) decode with a 1:1 port of the authoritative wmbusmeters apatoreitn driver (drivers/src/apatoreitn.xmq, verified against wmbusmeters 3.0.0 analyze output and against the on-meter display).

Payload map (CI=0xA0 mfct-specific, 15 bytes)

b0 season_start_date_lo (packed date raw = 0xA000 | lo)
b1-2 pad
b3-4 previous_hca uint16 LE (DIF 0x42, storagenr 1)
b5-6 esb_date_raw uint16 LE, 0 = seal intact
b7-8 current_hca uint16 LE (DIF 0x02, storagenr 0)
b9-10 current_date packed date
b11-12 temp_room_prev_avg /256 °C (matches on-meter display)
b13-14 temp_room_avg /256 °C
Packed date (year base 2000, same for all three date fields): day = raw % 32, month = (raw >> 5) % 16, year = 2000 + (raw >> 9) % 32.

All three frame wrappings from the upstream grammar are accepted: bare 15 bytes, 0xA0-prefixed, and the CI=0xB6 form 0xA0<15 bytes>.

Fixes vs the previous decode

• previous/current HCA were swapped (b3-4 is previous per DIF storagenr 1, b7-8 is current);
• temperature is two 16-bit /256 °C fields (prev-season and current-season room averages), not a single whole-degree byte with flag bytes around it;
• the packed date was rendered as a raw "Marker" — now decoded, year base 2000;
• season start date and ESB (electronic seal break) date were not decoded at all: season raw is 0xA000 | b0; a non-zero b5-6 means the seal was broken on the decoded date (shown as "Seal BROKEN" + date, otherwise "Seal OK").

Field naming follows apatoreitn (current/previous HCA, season start, ESB, temp_room_prev_avg/temp_room_avg) so output stays comparable with upstream.

Tests

tests/test_apator_hca.c now covers the two telegrams from the upstream apatoreitn.xmq test suite (direct frame with broken seal, CI=0xB6 wrapped frame), the 0xA0-prefixed variant, a field-verified live capture (APDU only, no meter ID), and the bad-length fallback. make -C tests check passes.

🤖 Generated with Claude Code