Release Notes – v2.5.5

  • Status: Released — May 2026
  • Notes: Bug fix release. Resolves two scenarios where a Build and Repair system could go silently idle — once after the welder briefly hit capacity, and once after downstream cargo briefly filled up. Also halves the default target-scan interval (10 s → 5 s) so new work is picked up twice as fast, and lightens the Show Area overlay box so you can see through it.

Bug Fixes

Build and Repair Stuck “Inventory Full” After Welder Drains

A Build and Repair system that filled its onboard inventory to capacity could remain stuck thinking it was full even after the items had been pushed out to a connected cargo container and the inventory was actually empty. The block sat idle with no welding, grinding, or collecting — and a world reload didn’t help, because the reload path also only knew how to set the flag, never clear it.

Root cause: the “inventory full” flag had a one-way state machine. It was set to true when the welder hit max capacity, but the only path that cleared it required an active transport cycle to also be running. When auto-push (Push Components / Push Ore / Push Items Immediately) drained the welder while the BaR was otherwise idle, no transport cycle existed, so the flag stayed armed forever — and all work paths (weld, grind, collect, idle fast-path) were gated off behind it.

Fix: the flag is now bidirectional with hysteresis. It sets at ≥100 % capacity (unchanged) and clears once the welder drains below 90 %. The 10 % hysteresis prevents flicker when a partial weld nudges the inventory just under max.

Symptoms this resolves:

  • A BaR that was working fine, briefly filled up, the player added a cargo container, and the BaR never resumed.
  • A BaR that says “Block inventory is full!” in the terminal info panel while its inventory is visibly empty.
  • An entire fleet going silently idle after a busy grinding session, even after the cargo network is drained.

If you were working around this by toggling the BaR off and on, you no longer need to.

Push Backoff Stuck for Up to a Minute After Freeing Cargo Space

When all downstream cargo containers were full and a Build and Repair system’s push attempt moved zero items, the BaR set an internal “push targets are full” flag and stopped retrying. The flag only cleared on the next 30-second source rescan, and only if either a container had been added/removed or a 60-second safety backoff had expired.

If you instead made room inside an existing container (manually moved items, an Assembler consumed components, a refinery output was hauled away), neither condition fired and the BaR sat with a full welder for up to a full minute before trying to push again.

Fix: the safety backoff is shortened from 60 seconds to 15 seconds. The signature-based fast path (container added/removed) is unchanged — that still triggers an immediate retry. Combined with the inventory-full fix above, a BaR now recovers in seconds instead of minutes after you clear downstream cargo.

Faster Target-Scan Cadence

The default target-scan interval — how often a Build and Repair system rebuilds its list of weld / grind targets — has been halved from 10 seconds to 5 seconds. The idle-backoff interval (used after several consecutive empty scans) has been halved from 20 seconds to 10 seconds.

What changes for players:

  • A newly-pasted, newly-projected, or newly-damaged grid is picked up in roughly half the time.
  • The lag between toggling a terminal setting (priority, work mode, area size) and the next scan-driven rebuild is shorter — though most of those toggles already trigger an immediate rescan via FEAT-080 (v2.5.4), so the visible improvement here is mostly for cases where the rescan trigger doesn’t apply.
  • Idle BaRs cost slightly more CPU than before — but only marginally, because the saturated-skip (FEAT-075), idle-backoff (FEAT-071), and dirty-cluster-key (FEAT-072) optimisations short-circuit most of the scan work when nothing has changed.

What stays the same:

  • Source/push-target rescans still run on their own 30 s schedule (SourcesUpdateInterval).
  • The immediate-rescan path (member sees a fresher coordinator result, settings change, weld/grind completes) is unchanged.
  • The interval is internally tunable via Mod.Settings.TargetsUpdateInterval if a server admin wants to revert to the old behaviour — it’s not exposed in ModSettings.xml because it has rarely needed adjustment in practice.

Show Area Overlay Lighter and More Transparent

The terminal Show Area toggle (the box that visualises a BaR’s working area in world) used a fully-opaque black colour. On smaller working areas this completely obscured the blocks underneath, making the toggle counter-productive — you turned it on to see where the BaR would work, and instead got a black box hiding everything.

Fix: the box is now a translucent grey. You can clearly see the working volume and the blocks inside it at the same time.

Behind the Scenes

  • CheckAndUpdateInventoryFull is now profiler-instrumented with wasFull / nowFull / fillPct so future “stuck full” reports are diagnosable from a single profiling session.
  • BuildId convention continues — diagnostics from this release show 260526.x in the debug HUD, /nanobars version, and profiler manifest header.

Compatibility

  • No save format changes.
  • No settings changes — existing ModSettings.xml files load unchanged.
  • No new chat commands.

Safe to drop in over v2.5.4. No world migration needed.

This site uses Just the Docs, a documentation theme for Jekyll.