## `omarchy restart trackpad` silently does nothing on `elan_i2c` touchpads

`/usr/bin/omarchy-restart-trackpad` only iterates over devices bound to the
`i2c_hid_acpi` driver. Laptops whose touchpad is bound to any other I2C HID
driver (e.g. `elan_i2c`, common on ASUS models) are not reset at all, and the
command still exits `0` without printing anything — so it looks like it worked.

### Steps to reproduce

1. On a laptop with an ELAN touchpad bound to `elan_i2c`:
   ```
   $ ls -1 /sys/bus/i2c/drivers/elan_i2c/ | grep '^i2c-'
   i2c-ELAN1000:00
   ```
2. Note there is nothing under the driver the script looks at:
   ```
   $ ls -1 /sys/bus/i2c/drivers/i2c_hid_acpi/ | grep '^i2c-'
   (empty)
   ```
3. Run the command:
   ```
   $ omarchy restart trackpad
   $ echo $?
   0
   ```
   It prints **no output at all** (not even `Done`) and exits `0`.
4. The device is still bound, i.e. nothing was reset:
   ```
   $ ls -1 /sys/bus/i2c/drivers/elan_i2c/ | grep '^i2c-'
   i2c-ELAN1000:00
   ```

### Why it happens

```bash
for dev in /sys/bus/i2c/drivers/i2c_hid_acpi/i2c-*; do
  [[ -e $dev ]] || continue
  ...
  echo "Done"
done
```

When the glob does not match, `$dev` is the literal unexpanded pattern, `[[ -e $dev ]]`
fails and the loop `continue`s. Since the `echo "Done"` is *inside* the loop, the
command produces zero output and exits `0`, which reads as success.

### Expected behavior

Reset the touchpad regardless of which I2C HID driver it is bound to, and fail
loudly (non-zero exit) when no device is found, instead of silently succeeding.

The manual equivalent that does work on this machine:

```bash
echo "i2c-ELAN1000:00" | sudo tee /sys/bus/i2c/drivers/elan_i2c/unbind
sleep 1
echo "i2c-ELAN1000:00" | sudo tee /sys/bus/i2c/drivers/elan_i2c/bind
```

Suggested fix: discover the bound device via its sysfs path instead of assuming
`i2c_hid_acpi`, e.g. by looking the touchpad's `phys`/`uniq` up under
`/sys/class/input/*/device`, or by iterating over all of
`/sys/bus/i2c/drivers/*/i2c-*` that belong to an input device.

### Context

This came up while debugging an unrelated touchpad problem, which made the
silent success actively misleading. Worth noting the same script also has an
`intel_quicki2c` branch guarded by `lsmod | grep -q`, which is fine, but the
primary loop should not fail open.

### System

- Omarchy 4.0.4-1
- Arch, kernel 7.2.5-3-omarchy
- Hyprland 0.56.2
- ASUS GL752VL
- Touchpad: ELAN ELAN1000, bound to `elan_i2c` on `i2c-1`
- Intel HD Graphics 530 (i915), NVIDIA GTX 965M

### Diagnostics

Relevant kernel lines:

```
elan_i2c i2c-ELAN1000:00: Elan Touchpad: Module ID: 0x0005, Firmware: 0x0004
input: Elan Touchpad as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-1/i2c-ELAN1000:00/input/input8
```

`hyprctl devices` lists the device and `omarchy hw touchpad` prints
`elan-touchpad`, so the compositor enumerates it normally; the issue is purely
that the reset helper does not reach it.