[Bug]: macOS: "Capture Region" binding triggers full screenshot instead of region selector

[Taifunov]

Pre-flight checklist

• I searched existing issues and this is not a duplicate.
• I am on the latest release or a recent master build.
• I quit Logi Options+ before running OpenLogi (the two apps fight over HID++ access and only one can own a receiver at a time).

Which part of OpenLogi?

Not sure

OpenLogi version

0.8.3

Operating system

macOS

OS version & architecture

macOS Tahoe 26.6.2

Device model

MX Keys

How is the device connected?

Bluetooth (direct, no receiver)

Affected area(s)

• Device discovery / detection
• Button remapping
• DPI control
• SmartShift
• Per-application profiles
• Battery status
• Settings / configuration (TOML)
• Auto-update
• Menu bar / tray
• Other

What happened?

I bound key F17 to the "Capture Region" action (confirmed selected/checked in the binding picker under System > Capture Region). Pressing F17 fires a full-screen screenshot instead of opening the interactive region selector.

"Capture Region" was added in #296 as a distinct action from "Screenshot" (System > Screenshot), specifically to send the region-selection shortcut (Cmd+Shift+Ctrl+4 on macOS) instead of the full-screen one. On 0.8.3 / macOS Tahoe 26.6.2, binding to "Capture Region" behaves the same as binding to "Screenshot" — the full screen is captured, with no region selection.

Steps to reproduce

1. Open OpenLogi > Devices > (keyboard) > Keys
2. Select an unused key (e.g. F17) and bind it to System > Capture Region
3. Press the bound key
4. Observe: a full-screen screenshot is taken/copied, not a region selection

Expected behavior

Pressing the key should invoke the interactive region-selection capture (equivalent to Cmd+Shift+Ctrl+4), matching the behavior described in #296.

Diagnostics report

### OpenLogi Diagnostics

**App**
- OpenLogi (GUI): v0.8.3 (release)
- Agent: v0.8.3 (connected)
- IPC protocol: GUI 29 / agent 29
- Inventory: ready
- OS: macOS 26.6.2 (arm64)
- Locale: en-UA (UI: follow system)
- Accessibility: granted · Input hook: installed
- Launch at login: no · Menu bar: yes · Update check: off
- Running from: source build (dev)
- Config: schema 7 · 2 configured device(s) · thumbwheel 14

**Assets**
- Source: user cache · Index: loaded (210 models) · User cache: present
- Cache path: ~/.local/share/openlogi/assets · Bundle assets: absent

**Devices (2)**
- G604 Lightspeed — mouse (codename: G604 Wireless Gaming Mouse)
  - Connection: Bluetooth (direct) · Online: yes · Battery: 90% (discharging, full)
  - Capabilities: buttons=no, pointer=yes, lighting=yes
  - Model: 0b024 (model-ids: b024/4085/0000, ext-model: 00)
  - Transports: eQuad, BTLE
  - Render: g604 · direct
- MX Keys — keyboard (codename: MX Keys Wireless Keyboard)
  - Connection: Bluetooth (direct) · Online: yes · Battery: 100% (discharging, full)
  - Capabilities: buttons=yes, pointer=no, lighting=no
  - DPI: unsupported
  - Model: 0b35b (model-ids: b35b/408a/0000, ext-model: 00)
  - Transports: eQuad, BTLE
  - Render: mx_keys · direct

openlogi list output

inventory read from the running agent)
MX Keys (—, vid=046d pid=b35b)
  └─ slot 255 ● MX Keys Wireless Keyboard (keyboard, wpid=?, battery=100% full (discharging))
          model_ids=[b35b,408a,0000] ext=00 serial=— unit_id=ae46b4d1 transports=equad+btle

G604 LIGHTSPEED (—, vid=046d pid=b024)
  └─ slot 255 ● G604 Wireless Gaming Mouse (mouse, wpid=?, battery=90% full (discharging))
          model_ids=[b024,4085,0000] ext=00 serial=— unit_id=56adffdc transports=equad+btle

Logs

2026-09-02T10:50:54.599115Z DEBUG RPC{rpc.deadline=2026-09-02T10:51:04.599000375Z otel.kind="client" otel.name=Agent.protocol_version rpc.trace_id=00}: tarpc::client: SendRequest
2026-09-02T10:50:54.599274Z DEBUG RPC{rpc.deadline=2026-09-02T10:51:04.599000375Z otel.kind="client" otel.name=Agent.protocol_version rpc.trace_id=00}: tarpc::client: ReceiveResponse
2026-09-02T10:50:54.599316Z DEBUG RPC{rpc.deadline=2026-09-02T10:51:04.599308333Z otel.kind="client" otel.name=Agent.declare_client rpc.trace_id=00}: tarpc::client: SendRequest
2026-09-02T10:50:54.599395Z DEBUG RPC{rpc.deadline=2026-09-02T10:51:04.599308333Z otel.kind="client" otel.name=Agent.declare_client rpc.trace_id=00}: tarpc::client: ReceiveResponse
2026-09-02T10:50:54.599439Z DEBUG RPC{rpc.deadline=2026-09-02T10:51:04.599428250Z otel.kind="client" otel.name=Agent.snapshot rpc.trace_id=00}: tarpc::client: SendRequest
2026-09-02T10:50:54.599580Z DEBUG RPC{rpc.deadline=2026-09-02T10:51:04.599428250Z otel.kind="client" otel.name=Agent.snapshot rpc.trace_id=00}: tarpc::client: ReceiveResponse
(inventory read from the running agent)
MX Keys (—, vid=046d pid=b35b)
  └─ slot 255 ● MX Keys Wireless Keyboard (keyboard, wpid=?, battery=100% full (discharging))
          model_ids=[b35b,408a,0000] ext=00 serial=— unit_id=ae46b4d1 transports=equad+btle

G604 LIGHTSPEED (—, vid=046d pid=b024)
  └─ slot 255 ● G604 Wireless Gaming Mouse (mouse, wpid=?, battery=90% full (discharging))
          model_ids=[b024,4085,0000] ext=00 serial=— unit_id=56adffdc transports=equad+btle

macOS permissions (if applicable)

• OpenLogi has Accessibility permission (needed to remap buttons via the event tap).
• OpenLogi has Input Monitoring permission (needed for Bluetooth-direct devices and capture).

[4ni1ak]

Checked `crates/openlogi-inject/src/inject/macos.rs` on current master: `NativeAction::CaptureRegion` and `NativeAction::Screenshot` are already distinct, and match what #296 intended:

```rust
// Lock screen = Cmd+Ctrl+Q (kVK_ANSI_Q = 0x0C)
NativeAction::LockScreen => post_key(0x0C, cmd | ctrl),
// Screenshot = Cmd+Shift+3 (kVK_ANSI_3 = 0x14)
NativeAction::Screenshot => post_key(0x14, cmd | shift),
// Capture region to clipboard = Cmd+Shift+Ctrl+4 (kVK_ANSI_4 = 0x15)
NativeAction::CaptureRegion => post_key(0x15, cmd | shift | ctrl),
```

These are different virtual keycodes (0x14 vs 0x15) with different modifier sets (Cmd+Shift vs Cmd+Shift+Ctrl), posted via the same `post_key` down/up helper used for every other native action — nothing in this code path treats the two actions identically, and this hasn't changed since #296 shipped (0.8.3 already had it, confirmed via `git merge-base`).

I don't have a Mac to reproduce this on, so I can't tell whether:

• this is a real runtime/timing issue specific to posting a 3-modifier chord via a synthetic `CGEvent` (e.g. macOS's screenshot hotkey daemon behaving differently for a scripted keydown+keyup with no interval than the 2-modifier Cmd+Shift+3 case), or
• something in your specific setup (System Settings > Keyboard > Shortcuts > Screenshots) has "Screen shot selected portion to clipboard" unbound or reassigned, since OpenLogi's shortcut only works if the corresponding System Settings shortcut is enabled.

Could you confirm on a recent build:

1. That System Settings > Keyboard > Shortcuts > Screenshots > "Copy picture of selected area to clipboard" is enabled and still bound to ⌃⇧⌘4.
2. Whether pressing that exact combo manually on your keyboard opens the interactive selector (to rule out a system-level shortcut conflict vs an OpenLogi injection bug).

Leaving this open rather than guessing at a code change, since the code already does what the fix should look like and I can't reproduce or verify the failure without macOS hardware.

[quentinhayot]

I have the same issue.

• ⌃⇧⌘4 correctly triggers a region-to-clipboard screenshot. I'd like my screenshot key to do exactly that too.
• Right now, "Capture region" creates a full screenshot and saves it to my screenshot folder

If I remember correctly, on LogiOptions, I had that key set to custom key presses and manually set ⌃⇧⌘4. And it worked just fine. Could we have that in OpenLogi ?

Keyboard is an AZERTY Ergo k860 and my input language is French.

[4ni1ak]

Same root cause as #1430 (Lock Screen on AZERTY) — the Capture Region/Screenshot shortcuts were posted by hardcoded ANSI-QWERTY physical keycode instead of the resolved character, so on AZERTY's shifted digit row they don't type '3'/'4' and macOS's system shortcut never fires, hence the fallback-to-full-screenshot behavior. Fixed in #1464 (now covers all three actions). Not hardware-verified yet — would appreciate a test once it's in a build.

[4ni1ak]

Correction to my earlier comment: I'm walking back the proposed fix in #1464. On review, Screenshot/Capture Region aren't the same shape of bug as Lock Screen (#1430) — both shortcuts already hold Shift, and a digit row's physical key position produces the same digit under Shift on AZERTY as on ANSI/QWERTY, so a character-based fix here risked breaking these shortcuts for everyone rather than fixing them for AZERTY.

To actually diagnose this I'd need what #1430 had: a CGEventTap trace (or similar) of what OpenLogi is actually posting when you trigger Capture Region, captured on your hardware. If you're able to build from source or run a debug build, I can point you at how — otherwise this stays open until someone can gather that.