Repository navigation
chore: retire the ReScript migration subsystem (FOT-R-22) - #96
Merged
Merged
Conversation
Nine files, 1698 lines, zero tests, and no reachable entry point. MigrationObserve was never registered: `doors.ex` lists exactly three tools in @tools and that module is not among them, so no MCP client could call the 392-line tool at all. The rest of the subsystem was reachable only from an `--migration-observer` argv scan and a FEEDBACK_A_TRON_MIGRATION_MODE env read, both performed inside `Application.start/2` — the same read-at-boot shape removed from the submitter in 6e91317. ReScript is banned estate-wide, so the migration this observed cannot recur. Deleted: migration_observer, batch_reviewer, report_generator, verisim_writer, the four-module pipeline/ tree, and mcp/tools/migration_observe. Dropped {:gen_stage, "~> 1.2"} and unlocked it. Unwired both optional-child clauses and the argv/env gate from application.ex. Corrected the pre-ledger comment in submitter.ex, which listed BatchReviewer as a caller that lands in the drafted clause. The consent path is untouched: nothing here changes what the CLI prints, that it refuses to send without a typed y, or that a non-tty answers no. Verified: 318 tests, 0 failures (unchanged — the subsystem had none); `mix escript.build` clean; ./feedback-o-tron --version prints v1.0.0. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HGbKA9B3XfzRSdsjLg3j33
The code retire in 7ca576c left two tracked documents naming the nine deleted modules. A repo-wide sweep (not just lib/test/mix.exs) found no hits in config/, README.adoc, ARCHITECTURE.adoc, the CLI help text, CI, humans.txt or STATE.a2ml — no user-facing flag was ever advertised — but two records still pointed at the subsystem: - TEST-NEEDS.adoc names the nine deleted modules in its file list. The document is also superseded for an unrelated reason: it predates #93 and still claims "Unit tests: NONE (0 test files found)" against an engine with 318 passing tests. Banner rather than curation — surgically editing nine filenames out of a document whose headline claim is false would polish a lie. Re-running the audit is separate work. - The 2026-09-07 design spec listed the pipeline as "Real and tested" and "the natural seed for burst clustering at L5". Both are now wrong: it carried zero tests, and it is gone. Left unmarked, a later session plans L5 on a deleted subsystem. VeriSimDB is explicitly NOT retired. It remains a live estate integration (.machine_readable/integrations/verisimdb.a2ml, gated by dogfood-gate.yml); only the two FoT-internal writers went with the pipeline. Recorded in both documents so a future sweep does not mistake the name for dead code. The plan document's "Noticed, not in this plan" mention is historical and is deliberately left as written. asciidoctor -o /dev/null is silent on both files. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HGbKA9B3XfzRSdsjLg3j33
Contributor
|
Warning Review limit reachedNext included review available in 2 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Advanced Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (14)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
4 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Retires the ReScript migration subsystem (owner ruling FOT-R-22).
The estate has migrated fully to bun; ReScript is banned estate-wide. The
pipeline built to observe that migration is therefore moot — and it was never
load-bearing to begin with.
Why this is a delete, not a fix
Three independent reasons, each verified against the tree:
application.exdid not start the pipeline in thesupervision tree. The only reachable entry was an argv gate on
--migration-observer, a flag the CLI does not parse — so after SP1b's CLIlanded, no documented invocation could reach it.
tested".
maybe_add_mcp_serverregisters exactly threetools (
SubmitFeedback,ResearchFeedback,SynthesizeFeedback); an MCP hostnever sees
migration_observeintools/listand cannot call it.What goes
Nine modules — the migration observer, the batch reviewer, the report generator,
the GenStage producer/consumer pipeline and the two FoT-internal VeriSim writers —
plus the
gen_stagedependency frommix.exsandmix.lock.VeriSimDB itself is NOT retired. It remains a live estate integration
(
.machine_readable/integrations/verisimdb.a2ml, gated bydogfood-gate.yml).Only the two FoT-internal writers went with the pipeline.
Net effect
15 files changed, +19 / −1740.
Verification
mix test— 318 tests, 0 failuresmix escript.buildsucceeds;./feedback-o-tron --version→feedback-o-tron v1.0.0grep -c gen_stagein live code → 0 (the 5 remaining mentions are prose inthree docs, and two of them are annotated by the second commit — see below)
This branch was rebuilt from patches onto current
main(30a4a6b) rather thanrebased over #93's 43-commit ancestry. Its tree hash is byte-identical
(
fa9ed5fe…) to the tree those verification figures were measured on, so themeasurements above transfer exactly rather than being re-asserted.
Documentation honesty
The second commit does not silently rewrite history. It annotates the two docs
that are still live references and would otherwise mislead a reader:
docs/superpowers/specs/2026-09-07-fot-everywhere-design.adoc— the bulletdescribing the pipeline now records the retirement and warns against seeding L5
burst clustering on it.
TEST-NEEDS.adoc— marked SUPERSEDED, since that audit predates both SP1b — Runnable: product #93(which is why its "Unit tests: NONE" claim is false) and this retirement.
🤖 Generated with Claude Code
https://claude.ai/code/session_01HGbKA9B3XfzRSdsjLg3j33