feat: reserved UDP RTP cell: hint in the 9th cell and U in the menu

[compot-dev]

Prepares the UI for #68 (pushed RTP over UDP); the receiver itself is developed separately.

Ninth cell. The 3×3 multi-view grid has nine cells for eight slots. The ninth cell is reserved for a camera that pushes RTP over UDP. Until the receiver exists it shows a hint instead of a black square:

UDP RTP · port 5600
Waiting for a stream…

Send the camera's RTP/UDP output to
<device IP>:5600

The address is the device's first site-local IPv4 address ("this device" if none is found).

U in the menu. The menu row is now Q, 1–8, U. U stops the current slot or grid and shows the UDP cell full-screen with the same hint; U is highlighted while the view is shown, and Settings/WebUI are hidden as in multi-view. A slot number or Q leaves the view and starts that slot or the grid; tapping the view opens the menu. The view is not a saved mode: pause/resume or a shared link returns to the slot. When #68 lands, this view shows the received stream.

READMEs (EN/RU/ZH): the reserved cell and the U row in the menu table. Test: QuadGridTest.udpCellHintNamesPortAndAddress.

Built and tested by CI; the device test follows.

[compot-dev]

Device test of the CI build 4c9a3bb on the AG-216 (Android 10), 8 slots filled:

• 3×3 grid: the ninth cell shows UDP RTP · port 5600 / Waiting for a stream… / Send the camera's RTP/UDP output to 172.17.32.144:5600, i.e. the box's real LAN address.
• U: all 8 sessions stop, and the hint is shown full-screen. The menu row is Q 1–8 U with U highlighted; Settings and WebUI are hidden.
• U → slot 2: the single stream starts. U → Q: the grid comes back with 8 cells. U → Home → resume: the slot comes back. No crash.
• Remote: OK on the U view opens the menu. The first OK after a touch only leaves touch mode, which is standard Android behaviour and the same in the other views.

[compot-dev]

/review

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0) 🎨 UX issues (0) 🔗 Cross-repo conflicts (0) 📜 Skill insights (0)

1. The ninth grid cell ignores taps ✓ Resolved

Description

buildQuadGrid() adds the UDP hint without the menu-opening click listener installed on the other
grid cells. In a 3×3 grid, tapping that cell does nothing, although tapping any camera cell opens
the menu.

Code

android/app/src/main/java/com/openipc/decoder/Decoder.java[626]

+                    lr.addView(udpHintCell(), new LinearLayout.LayoutParams(

Evidence

The hint is added directly without a click listener; the adjacent camera views receive one, while
the separate full-screen UDP hint has its own listener.

android/app/src/main/java/com/openipc/decoder/Decoder.java[624-636]
android/app/src/main/java/com/openipc/decoder/Decoder.java[325-330]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The ninth grid cell does not open the menu when tapped, unlike the camera cells.
## Fix Focus Areas
- android/app/src/main/java/com/openipc/decoder/Decoder.java[624-636]
## Recommended Fix
Install the same menu-opening click listener on the grid's UDP hint cell that the camera grid cells use.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

2. The UDP hint can reopen as a blank screen ✓ Resolved

Description

onShowUdp() reuses udpFullView by changing its text and visibility without moving it above the
video-clearing overlay. If a user leaves U for an unconfigured slot, clearVideo() brings its black
overlay to the front, so reopening U places the hint behind that overlay.

Code

android/app/src/main/java/com/openipc/decoder/Decoder.java[R331-334]

+        } else {
+            udpFullView.setText(udpHintText(localIpv4(), UDP_RTP_PORT));
+        }
+        udpFullView.setVisibility(View.VISIBLE);

Evidence

The new camera-switch path calls saveSettings(), which calls clearVideo() for an unconfigured
slot. That method brings the black overlay to the front, whereas reopening U only makes the
previously created hint visible.

android/app/src/main/java/com/openipc/decoder/Decoder.java[273-277]
android/app/src/main/java/com/openipc/decoder/Decoder.java[322-346]
android/app/src/main/java/com/openipc/decoder/Decoder.java[460-466]
android/app/src/main/java/com/openipc/decoder/Decoder.java[777-789]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
Reopening U after selecting an unconfigured slot can show a black overlay instead of the UDP hint.
## Fix Focus Areas
- android/app/src/main/java/com/openipc/decoder/Decoder.java[331-335]
- android/app/src/main/java/com/openipc/decoder/Decoder.java[777-795]
## Recommended Fix
Bring the UDP hint to the front each time U is shown, including when the existing view is reused. Verify the sequence U → unconfigured slot → U.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

Tip of the day

💡 Did you know, you can commit Qodo's fix in one click with committable suggestions (GitHub & GitLab)

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

[compot-dev]

Both Qodo findings are fixed in b1ba131:

• The ninth grid cell ignores taps: udpHintCell() now installs the menu-opening click listener, so the grid cell behaves like the camera cells; the full-screen view shares it.
• The UDP hint can reopen as a blank screen: bringToFront() is called each time U is shown, so the black overlay of an unconfigured slot can no longer cover the hint. The sequence U → empty slot → U is part of the device test.

/review

[compot-dev]

Device check of the CI build b1ba131: a tap on the ninth grid cell opens the menu, U shows the hint full-screen, and a tap on it opens the menu. The U → empty slot → U path could not be exercised on the device, because all 8 slots are configured; that fix (bringToFront()) is covered by review only. An ANR seen during this session (main waiting for the RenderThread, which was blocked in eglSwapBuffers → SurfaceFlinger) also happened earlier with #64, so it is not caused by this PR; it is tracked in #76.