RunicGateway/docs
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs(rust): protocol 13 — the reload as the rigs showed it (W1-W5, D180)

Browse Source 
PROTOCOL.md §19.1: how the reload is asked for (Carbon: c.reload on the
server console), which log is read (Carbon.Core.log on Carbon), the pinned
failure lines on both frameworks, a load confirmed by a log read, and the
compiler's give-up retried once. PLAYER_WALK.md: the Carbon reload path and
failure words as walked. PLAN_FIXES.md: D180.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E14m6SuuY6i1vASFeGDBeY
This commit is contained in:
claude
2026-09-26 20:21:26 -05:00
parent cfa37c44e3
commit b8d9740ed0
3 changed files with 48 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
modules/rust/PLAN_FIXES.md
View File

@@ -1,7 +1,7 @@
# `module-rust` — plan fixes
**Status:** plan, approved 2026-09-26; its open questions were answered the same day (D169–D175), and step 1's
shape was settled as work began (D176–D179). **Everything in it — fixes and redesigns alike — lands before Module-Rust's cutover
shape was settled as work began (D176–D179), and its rig walk's fixes by D180. **Everything in it — fixes and redesigns alike — lands before Module-Rust's cutover
(phase 19, [`PLAN.md`](PLAN.md) §34, D145; D169).** Everything here comes from the first walk of
[`PLAYER_WALK.md`](../../rust-link/PLAYER_WALK.md) with a real player in the game — the Oxide pass, walked by
the org lead on the `rust-oxide` rig with every frame checked on the console, the sidecar and the site's
@@ -63,6 +63,7 @@ Taken by the org lead during and straight after the walk.
| **D177** | **F9 and F10 open protocol 13.** They are built first, on `edge` in Rust-Plugins, Rust-Link and Module-Rust; the rest of §6 step 2 joins the same bump, and all of it releases together. Rejected: F9 alone as protocol 13 with the rest as 14; and a wider window now as a stop-gap. |
| **D178** | **A reload fails on evidence, not on a clock.** While a configuration reload is pending the plugin reads the framework's new log lines each second and rolls back the moment they show the target plugin failing to compile or initialise. The ceiling is 30 s; when it passes, the plugin checks whether the target is loaded (a hook it missed) before restoring anything, and reports what actually loaded (F10). A framework hook for a failed load, if the rig shows one, replaces the log read. Rejected: the clock alone with a longer ceiling, which leaves a broken plugin down for the whole ceiling. |
| **D179** | **The configuration page polls the write.** The save is recorded as `reloading`; ingest settles that row from the outcome frame, and the page polls it every couple of seconds until it does. Rejected: pushing the outcome over the admin event stream. |
| **D180** | **The step-1 rig walk's fixes stay inside the bridge, and the console is a fair way in.** The walk (2026-09-26) found Carbon's Oxide-compatible `ReloadPlugin` does not reload (W3), Carbon's log is `Carbon.Core.log` (W4), Carbon loads a plugin whose config threw and raises `Loaded` anyway (W5), Oxide's "Could not initialize plugin" line was missed (W2), and Oxide's idle compiler drops a reload with "Timed out waiting for plugin to be compiled" (W1). All are fixed in `RunicGateway.cs` (PROTOCOL.md §19.1), never by changing a third-party plugin. Running server console commands (`c.reload`) is approved, and so is a helper plugin that hooks and runs console commands where one is needed (D168). |
## 2. Fixes