Skip to content

v1.2.0

Choose a tag to compare

@github-actions github-actions released this 29 Sep 05:49
· 17 commits to master since this release
  • The README is rewritten for the test author (quick start, connecting a client, the scenario API
    grouped, recipes, the fixtures, what is measured and what is the simulator's model), and the
    changelog moved into this file.
  • data/fixtures/lab-2026-09.json: the devices of four test systems with the paramset
    descriptions of their exact firmware, their direct links and radio modules - the CCU centrals,
    RPI-RF-MOD, HmIP-RFUSB, HmIP-HAP, HmIP-BBL, HMIP-WRC2, HmIP-PDT, HmIPW-DRAP/-DRI16/-DRS8,
    HM-CC-TC, HM-Sec-SC, HM-LC-Sw1-Pl-2 - anonymised. tools/dump-ccu.js makes such a fixture from
    any CCU (read only, anonymised unless told otherwise). A device file given to --devices brings
    its links and bidcosInterfaces along.
  • The calls openccu-lite's daemon makes: logLevel (rfd and hs485d answer the level, 5 until set;
    hmipserver an empty string), changeKey (rfd, hmipserver; recorded in sim.keyChanges) and
    refreshDeployedDeviceFirmwareList (rfd, hmipserver; recorded in
    sim.firmwareListRefreshes), with the interfaces measured on a CCU with firmware 3.89.x.
  • init keeps the path of an XML-RPC callback url (http://host:port/cb/BidCos-RF; the path used
    to end up in the port), takes https://, and identifies a registration by its url exactly, scheme
    included, as the interface processes do - an init(url, '') with another scheme no longer removes
    it. A url that is none is a fault.
  • An XML-RPC request declared ISO-8859-1 is read as ISO-8859-1 instead of UTF-8.
  • Device and radio health: setReachable (UNREACH, and STICKY_UNREACH on BidCos until a client
    clears it; unreachWrites: 'fault' answers writes with notReachable), setLowBattery,
    setDutyCycle and setCarrierSense (in listBidcosInterfaces and, on hmip, as the radio
    module's DUTY_CYCLE_LEVEL/CARRIER_SENSE_LEVEL), offerFirmware with the update states
    updateFirmware/installFirmware then walk through, and schedule for a timeline of scenario
    calls.
  • Interface processes that misbehave: stopInterface, startInterface and restartInterface
    (forgetting the registered clients, or calling them back as rfd does after a restart),
    interfaces.<iface>.startDelay, pong and pingDelay, injectFault (delay, hang, fault,
    closed connection, for the next calls of a method) and clearFaults, and
    listenerModel: 'measured': what hmipserver and rfd do with a callback server that hangs
    (hmipserver holds every client's events back, rfd stops answering). getCallbackLog() records
    every call to a client with its timing, fireEvents() sends bursts as system.multicall batches.
  • The command line is usable out of process: every port may be 0, --ports-json <file|->
    reports the ports once every server listens, --config and --devices load options and device
    files, --wired-port, --tls with --tls-cert-out, --auth, --behavior-path,
    --no-behaviors, and --control-port serves the scenario API over HTTP on the loopback. A taken
    port exits 1 with the address in the message. behaviorPath: false starts no behaviour scripts.
  • An ENUM whose description gives DEFAULT as the index (1088 of the 5624 ENUMs in the bundled
    descriptions) started at -1 instead of that index, so a thermostat's FAULT_REPORTING was a
    service message from the start.
  • rfd and BidCos-Wired answer BIN-RPC and XML-RPC on the same port, as on a CCU, so a client
    that talks XML-RPC to rfd can be tested too; the callbacks follow the URL a client registered
    with, whichever protocol it registered over. With tls the XML-RPC half is HTTPS, BIN-RPC stays
    plain. interfaces.<iface>.protocols narrows a port to one of them.
  • A device whose firmware has no paramset description uses the description of the nearest
    firmware of the same type and VERSION instead of answering -2 Invalid device for every
    channel listDevices reports; the bundled CCU virtual remote (HM-RCV-50 2.27.8,
    BidCoS-RF:1..50) works again. paramsetFallback: false switches it off,
    sim.getMissingParamsetDescriptions() lists what is not matched exactly (reported in #1 by
    @Hypnos3, cause found by @foxriver76).

Commits since v1.1.0

  • Release 1.2.0 (1fec214)
  • rewrite the README for the test author, and move the changelog out of it (45ba7c0)
  • add a fixture dumped from real systems, and the tool that dumps one (a76e0e2)
  • answer the calls openccu-lite's daemon makes, with its callback url (3687ef1)
  • report device and radio health over time (adc5498)
  • simulate interface processes that restart, start late, hang or fault (422caec)
  • make the command line usable out of process (9e2c38c)
  • start an ENUM whose DEFAULT is an index at that index (5f35bda)
  • answer XML-RPC and BIN-RPC on the same port for rfd and BidCos-Wired (35463cd)
  • fall back to the nearest firmware for a missing paramset description (e433eb5)