Troubleshooting
Start with the Log tab in the dock and the Last run block in the cue's own inspector. Between them they usually say it outright.
Cues
| Symptom | Cause and fix |
|---|---|
| GO skips a cue | It is disarmed — the row says so. Arm it from the right-click menu or the Basics tab. |
| An audio cue is silent after moving files | Its file has moved. The row says missing; use the relink bar at the top of the window, and see the workspace file. |
| The cue list order differs from the file you saved | The file's cue grouping was damaged — children scattered, a cue naming a group that is not in the file, or groups inside each other — and was repaired as the show opened. A bar at the top of the window says how many repairs were made. The file on disk is unchanged until you save, which writes the repaired grouping. |
| Everything goes quiet at once and the rows say audio device lost | The sound device went away mid-show — unplugged, taken by another program in exclusive mode, or a driver that stopped calling. The engine released the cues it was playing and said so in the alarm strip, the transport bar's alarm chip and the log. Settings → Audio shows the device and the engine's own readout of it, and is where another interface is chosen. When the device is back, Reconnect device sends the show's settings again and the engine reopens it: the readout updates at once, and the failure text clears. The alarm the loss raised is not the engine's to take down — it stands in the alarm strip until it is dismissed there. |
| A MIDI cue does nothing | No port: the cue names none and the show has no default output, or the interface is gone. Settings → MIDI, and the Monitor reports an interface that disappeared. |
| A show control cue does nothing | The show's desk does not take MIDI — the cue's panel says so — or the device ID does not match the console. 127 reaches everything. |
| A network cue fails with "Failed to fetch" | Nothing answered at that address: wrong host or port, or the target is not running. The Last run block has the exact error. |
| A network cue succeeds but the show stalls | The cue holds the show until the server answers. Lower its timeout, or put it in a group set to fire everything at once. |
| A script cue never finishes | Its timeout is 0, which waits for ever. Set one, or check what the script is waiting on in the log. |
| A trigger does not fire the cue | The input is off (Settings → Network or MIDI), the cue is disarmed, the address does not match exactly, or the repeat guard is swallowing a second message. The Monitor shows what actually arrived. |
The window
| Symptom | Cause and fix |
|---|---|
| A banner says the interface stopped unexpectedly | The window crashed and was recovered. The show itself kept playing — the engine is in another process. The banner carries the error; the log has the rest. |
| The cue list looks empty after opening a show | Every group is collapsed. Expand in the toolbar opens the top level. |
| The wireless panel shows no readings | No source, or the source is unreachable — Settings → Wireless reports which. Demo proves the panel works. |
| Where are the error messages? | The Log tab, and the Last run block on the cue. Alarms also appear in the transport bar, whatever panel is open. |
Known limitations
- No video or image cues. Video is driven by talking to something that plays it, through a network, OSC or MIDI cue.
- No cue-level audio effects beyond level, pan, rate and fades.
- ASIO runs a single driver at a time and at most 32 channels: a rig that needs two interfaces on two ASIO drivers at once is not covered.
- OSC input is UDP only.
- The channel ceiling a desk model states is the number the console publishes rather than a measurement: a show with more channels than the selected model takes is named under the desk connection, with the number, and every channel is sent as it always was. Only the count control stops at it, and a show already larger than the ceiling keeps its own whole range on the dial rather than being trimmed back to it.
- An SQ's membership is written over the desk's own control service — TCP 51326 outbound, with a keepalive datagram to the desk's UDP 51324 — a protocol Allen & Heath publish nothing about, spoken here from frames measured off an SQ-6 on firmware 1.6.1. The desk's own answer to a membership has never been read, so nothing here says a desk has accepted one: what a recall leaves out is worked out against what Cuerunner last sent, not against the desk. The channel mutes ride the MIDI service beside it, on 51325. The session leaves by the desk's own door when the show stops, and one the desk loses without that goodbye leaves it mute until it is reset — said in the log, on the device row and in an alarm, which is why a lost session is not dialled again in a loop.
- The tablet view is read-only by design.
- Cuerunner is distributed outside the app stores, and the release notes say whether a build is signed: Windows warns when an unsigned installer is run, and macOS warns about an unidentified developer at first launch — Gatekeeper refuses such an app on a machine that has not seen it. Each release publishes its digests twice: a
SHA256SUMSbeside the assets on its release page, naming the files that page carries, andlatest/SHA256SUMSon the download host, naming the files the cards link. - The Linux AppImage is experimental: MIDI needs ALSA development headers present when dependencies are installed.