Skip to content

v1.17.0

Latest

Choose a tag to compare

@github-actions github-actions released this 12 Nov 21:46
· 1 commit to main since this release
a20391e

First combined upstream release after the community fork. Includes six interface
languages and expanded device support from community 1.16.0, plus the additions
and fixes in this section.
Development and future updates use HeyOkay/HaloBattery. Existing community-fork
installations need a manual install of the upstream build to change their update
source; their settings and battery history remain compatible.

Included from community 1.16.0

  • Six interface languages and expanded device support from community 1.16.0 are
    included for users upgrading directly from upstream 1.15.0.
  • Community fixes improve shutdown, idle charging animations, history saves and
    remaining-time estimates.
  • Portable startup in unwritable folders and device discovery after Windows Fast
    Startup include the community corrections.
  • Battery reporting includes Xbox One S Bluetooth coarse levels, Logitech charging
    and shared cable/receiver icons, MCHOSE G7 smoothing and corrected VXE R1 naming.
  • PlayStation and Steam Controller polling, Corsair Void v2 writes, Barracuda X
    protocol fallback and probe-file output include the community fixes.
  • The retained 1.16.0 changelog
    contains the detailed device list and contributor credits.

Added

  • ATK V75X USB cable (320f:5055): shares the receiver's icon and is preferred
    when both transports are connected. The PR author confirmed battery and charging
    on the cable; receiver support was already integrated.

  • BEITONG KP40D (20bc:515c, 2.4 GHz receiver): reads the controller's own battery estimate and charging status through its vendor HID collection, instead of the unusable XInput / Windows battery placeholder. Requires one controller in both XInput and Windows.Gaming.Input and one matching receiver; ambiguous configurations keep the existing XInput fallback.

  • VAXEE wireless mice: the battery level (in 5 % steps) and charging state
    over the vendor feature-report protocol, on a VAXEE 4K dongle or on the cable.
    Invalid battery values are refused instead of displayed as a full charge.
    Ported from ryanlewis/mousebatt, with
    the battery command from stuffz/mouse-battery-monitor;
    confirmed on hardware on an E1 Wireless over its receiver (other models and
    the cabled path are untested).

  • Per-device Lucide Speaker, Webcam, Heart Pulse (health device), House (smart home),
    Watch (smartwatch), TV Minimal (display/TV) and Radar (tag/tracker) pictograms,
    available through the Icon menu in all six interface languages. Speaker, House
    and TV Minimal are fitted to match the existing glyph-to-ring spacing.

  • Per-device Controller logo selection: Default, Xbox, PlayStation or Nintendo
    Switch, with optional brand colors and the classic multicolor PlayStation logo.
    Choices are saved per device; automatic device types and status-file kinds are preserved.

  • ASUS ROG Harpe II Ace / Extreme on its SPEEDNOVA receiver (0b05:1ad0) or cable (0b05:1c69): integrate the sourced battery exchange from upstream #364 without replacing existing OMNI/headset support. The PR author confirmed the Ace receiver; cable and Extreme remain unverified.

  • Corsair HS80 RGB Wireless (original): battery level over direct USB
    (1b1c:0a69) or its wireless receiver (1b1c:0a6b), sharing one tray icon.
    Charging is inferred from USB presence; the powered-off charging identity
    (1b1c:0a6a) keeps an unknown-level icon when no battery reading is available.
    Protocol from ST1CKEL/HS80_Control PR #2.

  • ASUS ROG Strix Morph 96 Wireless on the ROG OMNI receiver (0b05:1ace, pair id
    1cea) - the same keyboard read as the Azoth and Strix Scope II (battery byte 6), added
    from the reporter's pair list. Confirmed accurate by @Pikldoornob in
    #340; the separate Maono request remains open.

  • URX Core75 (36b0:3002, 2.4 GHz dongle) - read through its VIA raw HID channel
    (ff60:0061, report id 0, 32-byte packets): custom-channel value 7 answers the battery,
    08 00 07 <level>. The framing comes from URX HUB's own device table and the reply was
    confirmed on hardware by @TheLastGamer18 (#245):
    the app detected the keyboard and the percentage matched URX HUB. A probe run with the
    charging cable in returned the same bytes, so this board does not expose a charger state -
    the app shows the level only.

  • Windows installer HaloBattery-<version>-setup.exe (Inno Setup, installer/HaloBattery.iss), built with every release and pre-release. It installs per user to %LOCALAPPDATA%\Programs\HaloBattery without an admin prompt, adds a Start menu entry, an optional desktop icon and an optional "Start with Windows" entry, and registers an uninstaller in Settings → Apps. Uninstalling closes the app, removes its autostart and notification registry entries, and asks before deleting the settings. Installing any version over another upgrades or repairs the one install in place (one folder, one uninstall entry, no leftover library files); an older setup asks before downgrading, and refuses when run silently unless /ALLOWDOWNGRADE is given. The in-app updater keeps working for installed copies and updates the version shown in Settings → Apps.

  • Razer Viper V4 Pro (1532:00E5 on the cable, 1532:00E6 through the receiver) added to the Razer mouse table with transaction id 0x1F, the id OpenRazer uses for it (openrazer/openrazer#2751). Confirmed on hardware by @amlidos in #333: the icon appears through the receiver, the level matches Synapse, and the charging state follows the cable.

  • Rapoo VT9 Pro V2 (24ae:14a9) - the receiver publishes the same ff00:000e
    vendor collection the VT7 is read on, so it joins that table entry. Unverified on
    hardware; a test build is out for #235.

  • Corsair M75 Wireless (1b1c:1bdc, the Slipstream receiver) - it publishes the
    bragi vendor collections (ff42:0001/0002), so it is read with the Dark Core's
    framed get-property exchange. Confirmed on hardware by @SpaceJesusCosmic in
    #301.

  • VGN Dragonfly 4K (3554:f505 receiver, 3554:f506 on the cable) - the Compx exchange already used for the VXE/Pulsar receivers, battery command 0x04. Confirmed on the receiver by @TweaknFreak (#323): the percentage agreed with the vendor tool. The cable id is claimed from the report; charging over the cable is not yet reported back.

  • Logitech G535: its dongle is long-only and has no HID++ 2.0 feature table, so the
    provider's 2.0 ping never got an answer and the headset was shown as asleep. It is now
    read through its 1.0 charge register (11 FF 05 0D) before the ping, with
    HeadsetControl's G535 voltage curve (#339).
    The register is answered on the ff43:0202 collection, not the consumer one, so both are
    tried. Confirmed on hardware by @ti-henrq: 4069 mV -> 85 % and the icon appears.

  • Compx 2.4 GHz keyboards (3554:fa09: AULA F75, LEOBOG Hi75C Pro; 3554:f523:
    Teevolution Terra Pro), Corsair Sabre v2 Pro Ultralight (1b1c:2b2a),
    ATK X1 Pro
    (373b:101a) and ATK Fire X 8K (373b:101b) - all three publish the
    Compx ff02:0002 control collection the Pulsar/VXE family is read on, so they use
    the same 17-byte power query; the keyboard keeps a keyboard pictogram. The ATK X1 Pro is
    confirmed on hardware in #230 (level
    within ~5 % of ATK HUB, charging follows the cable); the keyboard receiver and the Fire X
    8K are unverified, with a test build out for
    #247.

  • Keychron Nape Pro (3434:d026, the Link-KM receiver) - the same b3 06 exchange as the
    Ultra-Link 8K / M5, with its vendor collection on interface 1 (ffc1:0001) instead of 4, so
    the provider now picks the vendor page wherever it is. Unverified on hardware; a test build
    is out for #270.

  • Endgame Gear XM2we battery and charging through its original 2.4 GHz receiver
    (3367:1961), using documented status queries only. Battery and charging transitions verified on an XM2we receiver on Windows.

  • Plantronics Voyager 8200 UC behind the BT600 adapter (047f:02ee, 047f:02f7) - Poly's tunnelled get/set protocol on ffa2:0003, battery GET 0A 1A, level as level / (count - 1) * 100. Confirmed on hardware by @DindonOz (#328): the level matched Poly's app and the charging state was correct.

  • HyperX Pulsefire Haste 2 Wireless (03f0:0d97, 03f0:0f98) - NGENUITY's
    universal HID protocol: 50 02 asks for the battery, 51 02 answers with the level
    and charge status (plus voltage and temperature in the diagnostics). Unverified on
    hardware; a test build is out for #304.

  • i-Rocks M31R (14a5:c009 receiver, 14a5:c010 cable) - the Compx frame over a
    different transport: the request is a feature report on ff02:0002, the reply an
    input report on ff01:0000, and the receiver is asked 0x03 (online) first because
    it answers a power query from cache while the mouse sleeps. Confirmed on hardware by
    @jimchen5209 in #309: level matches
    the vendor app, charging follows the cable, the sleeping receiver greys out and returns,
    and with receiver + cable the wired mouse shows while the wireless one is greyed.

  • Alienware Pro Wireless Gaming Keyboard (413c:211d, 2.4 GHz receiver) - the level Command Center
    reads from the receiver's ff00:0001 collection (feature report 0e a1, answer
    0e a1 <level> <state>), taken from the #287 USBPcap captures. The report must be sent with
    report id 0 in front (00 | 0e a1 ...); the first test build omitted it, so the receiver
    ignored the writes and the keyboard sat at "no link". The charging capture also settles the
    charger state: byte 3 is 0x00 on battery and 0x04 with the cable in. Confirmed on
    hardware by @notisv
    (#287): the level matched Bluetooth, Windows and Command Center
    (0e a1 5c 00 = 92 %) and the charging state followed the cable.

  • DJI Mic Mini 2 (2CA3:4011): both transmitters over the receiver's vendor interface 6,
    read passively - a seven-step gauge shown as an estimate, the charging state, and one icon
    per transmitter. It needs WinUSB bound to that interface with
    Zadig; without it the provider stays silent and says so in the
    diagnostics, and no other device is affected. Protocol from
    ShadowBitBasher/DJI-Mic-Control
    (Unlicense), re-derived on a receiver here; pyusb and libusb-package are new
    dependencies, and the release workflow bundles libusb.

  • Corsair Void Elite Wireless (1b1c:0a55) - the older Void exchange: output report c9 64, five-byte reply 64 00 <level> b1 <status>. Unverified on hardware; a test build is out for #330.

  • Turtle Beach Atlas Air (10f5:225e) - the receiver battery read is now
    confirmed on hardware by @Vetrices in #310:
    the percentage matches Swarm II, charging follows the cable and updates keep working.

  • Glorious Model O Wireless (258a:2022) - the Sinowealth framed battery query on the
    receiver's vendor channel (feature report 00 00 02 02 00 83, level at byte 8, charging
    at byte 7), from the reverse-engineering work behind GloriousBatteryMonitor-Go. Level
    confirmed on hardware by @gumocska
    (#281): the percentage matched Glorious Core. While
    the mouse charges over USB it leaves the 2.4 GHz link, so the wired id (258a:2011) now
    tells the app it is on the cable and the icon shows charging with the last level; a wall
    adapter gives no signal, because the mouse is not talking to the receiver.

  • Razer Barracuda X (2022) (1532:0552, 2.4 GHz dongle) - the dongle bridges to an
    Airoha chip and speaks RACE (50 41), not the Razer PA protocol the other Barracudas use;
    the PA writes were refused with "Incorrect function". Level comes from the headset's
    battery voltage through remote mode + GET_BATTERY (group 0x06), converted on a Li-ion
    curve, with charging inferred from the voltage spread between samples. Mirrored from
    mehdibouchami/razer-barracuda-x-battery. Confirmed on hardware in the test-build run:
    3822 mV -> 38 %, samples within 9 mV, reported as not charging; the charging inference
    still needs a run with the cable in.

  • Razer Barracuda X (2022) (1532:0536) - its receiver publishes ff00:0001, the
    collection the PA headset path reads, and the id was simply not routed there. Unverified
    on hardware; a test build is out for #256.

Not enabled in this release

  • Endgame Gear 4K (3367:1970), AJAZZ AJ159 PRO (3151:402d), Kysona Uranus
    (3151:5031), Turtle Beach Stealth 600X Gen 3 (10f5:2232), Sony INZONE H9 II
    (054c:0fa8), Lamzu Tachi (37b0:000c) and Logitech PRO X 3 LIGHTSPEED
    (046d:0b45) remain on hold. Their current exchanges
    failed on hardware or still lack evidence of protocol compatibility. The candidate
    does not open these targets; their protocol code and simulated tests are retained
    for follow-up. Confirmed XM2we, AJ179 V2 MAX, Atlas Air and original INZONE H9
    support remains enabled. See docs/maintenance.md for the evidence and next steps.

Changed

  • Show charging with an unknown battery level as a steady neutral ring and a
    smoothly rotating green arc, keeping the device glyph stationary.

  • Use local Lucide device pictograms inside a thin circular battery ring, with
    artwork rendered at the taskbar's physical icon size and refreshed on DPI changes.

  • Refresh the icon and charging previews with explicit controller brand variants.

  • Replace the remaining microphone drawing with the Lucide mic glyph, fitted and
    centered by its visible bounds with the same ring clearance as the other icons.

  • Bring the community integration history back to HeyOkay/HaloBattery, retaining contributor credits. Updates, installer links and contribution links use the original repository again; cached offers from the fork are cleared without changing other settings.

  • Document a weekly manual pre-release review and the remaining hardware follow-ups in docs/maintenance.md. Stable publication still requires a reviewed release PR.

Fixed

  • ATK V75X receiver polling: reuse successful readings for five minutes,
    retry failed requests after 30 seconds, and rate-limit typing-triggered wakes
    so bursts of input do not repeatedly poll every device.

  • Softer neutral tray ink and area-sampled ring edges, avoiding excess dark/light
    coverage at small DPI sizes while preserving glyph placement and brand colors.

  • Use a matching .in/.txt build requirements pair so Dependabot can update the hashed dependencies installed by CI and release builds.

  • Keep unrelated devices updating and retry failed disconnected icons when one
    native tray image update fails.

  • Correct colored Xbox logo alignment for light and dark taskbar themes.

  • Preserve each device's stable Windows tray identity, serialize native icon-handle
    swaps during taskbar refreshes, and retry failed icon updates.

  • Keep packaged modules ahead of loose source folders beside the EXE, preventing
    startup import errors when running a new build from an older checkout.

  • Include the libusb backend and bundled library in pre-release builds, matching stable builds for USB devices.

  • Keep Alienware and Barracuda protocol documentation in separate sections after integration; record the newer Morph 96, URX and Plantronics confirmations. A disconnected ASUS receiver no longer raises from the output-report read path.