Optimization found by Claude in Display configuration.

[martinHkristensen-cpu]

I will let him speak for himself :-) It was done on latest update.

---

Bug report & fix: File descriptor leak + UI latency in display_connector

Environment

• Device: Elegoo Neptune 4 Max
• Firmware: OpenNept4une
• display_connector: fc84326 (latest main, March 2026)
• Python: 3.13

---

Issue 1: File descriptor leak in _try_connect_on_different_baudrates (nextion library)

Severity: High — causes service degradation and eventual failure without reboot

Root cause:

The nextion library's _try_connect_on_different_baudrates() iterates through 8 baudrates on every connect/reconnect attempt. Each failed attempt opens a serial port file descriptor that is not properly closed on failure. This leaks ~8 fd's per reconnect cycle.

Over time this produces:

[Errno 24] Too many open files

...causing all subsequent connect attempts to fail silently, the display to go unresponsive, and elevated CPU load + SoC temperature from the continuous retry loop.

Observed fd growth: ~44 open fd's after 45 minutes of idle (baseline should be ~19)

Fix:

Override _get_priority_ordered_baudrates() in TJCClient in src/tjc.py to only attempt the correct baudrate for TJC displays:

class TJCClient(Nextion):
    is_reconnecting = False

    def _get_priority_ordered_baudrates(self):
        return [115200]

Result: fd count stable at ~19 baseline. Display remains responsive indefinitely without reboot.

---

Issue 2: Unnecessary 50ms delay per display command in communicator.py

Severity: Medium — causes UI sluggishness under normal operation

Root cause:

_update_data_leaf() has a hardcoded asyncio.sleep(0.05) after every single write command. During a typical status update (temperature, position, speed, progress etc.) this introduces hundreds of milliseconds of artificial delay that accumulates in the async queue.

# communicator.py _update_data_leaf()
await self.write(command)
await asyncio.sleep(0.05)  # <-- unnecessary, asyncio event loop handles scheduling

Fix: Remove the sleep entirely. The asyncio event loop handles cooperative scheduling without it.

Also reduce navigate_to settle time from 250ms to 100ms — sufficient for TJC page transitions:

await asyncio.sleep(0.10)  # was 0.25

---

Issue 3: Unguarded self.data_mapping attribute access in update_data()

Severity: Low — defensive fix

update_data() falls back to self.data_mapping with no guard, producing an unhelpful AttributeError if the attribute is not set on the subclass.

Fix:

data_mapping = getattr(self, "data_mapping", None)
if data_mapping is None:
    self.logger.error("update_data called but no data_mapping defined on communicator")
    return

---

Files changed:

• src/tjc.py — add _get_priority_ordered_baudrates() override to TJCClient
• src/communicator.py — remove asyncio.sleep(0.05), reduce navigate_to settle time, guard data_mapping