docs: park the M1 architecture report + review sweep; drop 0.1-to-0.2 cruft#344
docs: park the M1 architecture report + review sweep; drop 0.1-to-0.2 cruft#344mfw78 wants to merge 8 commits into
Conversation
… triage The decision-ready architecture report (target 3-layer arch, capability matrix, reth/alloy DX sketches, R1-R8 red-team, Phase 0-4 migration plan, and the seven platform decisions of 2026-07-14) and the whole-train review-comment sweep. Parked on a branch so they are durable in git ahead of the reconciliation, not yet folded into dev/m1.
Nothing is released, so there is no 0.1 to migrate from; the version strings are cruft normalizing to a single @0.1.0. The "Migration from 0.1" prose in docs/08 is left for the reconciliation pass.
mfw78
added
the
base: dev/m1
…oundary + sequencing
…nd issue bodies The videre reorganisation plan (issue-milestone-plan), the apply script and per-issue bodies used to file the tracker, and the v1 baseline, kept as the record of how the milestones and issues were determined.
Byte-exact, wasm-tools-validated target for the M1 contract fold (#370): three videre packages, the #360 opaque-status body, the R6 host change, and the rename/normalize map. Records the four locked decisions (thin to the EVM-only surface, big-endian minimal-length amounts, plain-enum status with a standard #360 detail trailer, erc20 without a chain id) and resolves the quote func/record name collision as quote returning a quotation. Notes the supersede in split-plan §7.4.
[venue] section; keeper declares one body_version, adapter declares the body_versions set, install requires membership. nexum-runtime routes [venue] opaquely; the videre-host install predicate enforces. Adapter manifest is the install-time authority, WIT export stays for introspection, macros emit both from one IntentBody so they cannot drift. Decision only; enforcement is M2.
The git-filter-repo + jj + mergiraf recipe and the byte-identical tip oracle, validated by a dry-run of the version-normalize slice. Records the package-scoped-rule hazard (WASI carries its own @0.2.0) and the base-owned-blob finding: a train-range fold cannot normalize nexum:host/shepherd:cow (L1/base packages), so base normalize lands on develop while the train fold handles only the train-owned videre reshape.
lgahdl
left a comment
lgahdl
left a comment
There was a problem hiding this comment.
Low-risk docs park, fine to merge. One thing worth calling out before it does: the PR body deletes docs/migration/0.1-to-0.2.md and acknowledges one leftover cross-reference in docs/08-platform-generalisation.md, but that's not the full list — live references to that path also remain in docs/00-overview.md (lines 20, 312, 381), docs/05-sdk-design.md (296, 436, 608, 961), docs/04-state-store.md:70, and docs/07-rpc-namespace-design.md (110, 1236). All of these will 404 once this merges. Since those files aren't touched by this diff, flagging here rather than as an inline comment — worth a quick follow-up sweep (or folding into the docs/08 cleanup already planned for the reconciliation pass) to either fix the links or explicitly note them as known-broken alongside the acknowledged one.
Everything else checked out clean: the two main parked docs (venue-platform-architecture.md, m1-review-sweep-triage.md) are internally consistent and clearly marked as decision-ready/pre-reconciliation status, and the epic/gap body files I spot-checked read as coherent planning prose.
2 participants
What
Parks two planning artifacts in git so they are durable ahead of the M1 reconciliation, and removes one piece of pre-release cruft:
docs/design/venue-platform-architecture.md— the decision-ready architecture report (target 3-layer arch, capability matrix, reth/alloy DX sketches, R1–R8 red-team, Phase 0–4 migration plan, and the seven platform decisions of 2026-07-14).docs/design/m1-review-sweep-triage.md— the whole-train review-comment sweep.docs/migration/0.1-to-0.2.md— nothing is released, so there is no 0.1 to migrate from.Why
The design docs existed only as untracked working-tree files. This makes them durable in git without folding them into
dev/m1yet — they inform the reconciliation rather than being part of it. Kept as a draft on a separate branch deliberately.Testing
Docs only — no build impact.
AI Assistance
Report + sweep authored with Claude; parked here verbatim.
Note: the "Migration from 0.1" prose in
docs/08-platform-generalisation.mdis also cruft but left for the reconciliation pass (it is a content edit, not a file removal).