rose-ash/plans/content-on-sx.md

content-on-sx: Documents, blocks & collaborative editing on Smalltalk

DRAFT outline. The CMS vertical — blog, WYSIWYG editor, Ghost sync. Depends on persist-on-sx (document history as an event log). Ghost/CMS sync stays a thin external adapter (Python/FFI) until a native replacement exists.

rose-ash's blog domain is content management: a block-based WYSIWYG editor, navigation, Ghost CMS sync. A document is a tree of live blocks; editing is a stream of operations; collaboration needs conflict-free merge. That is an object model — blocks are objects, edits are messages, and a document is the object graph responding to them. Smalltalk's "everything is an object responding to messages" maps directly to a block/WYSIWYG model, and a semilattice (CRDT) merge keeps concurrent edits conflict-free.

End-state: a Smalltalk-on-SX document model (typed blocks, structural ops), operation log + CRDT merge for collaborative editing, versioning/history via the event store, and a render boundary to HTML/SX. External CMS (Ghost) sync is an injected adapter, not core.

Status (rolling)

bash lib/content/conformance.sh → 238/238 (Phases 1–4 COMPLETE + extensions: HTML escaping)

Ground rules

• Scope: only lib/content/** and plans/content-on-sx.md. May import from lib/smalltalk/, and (once it exists) lib/persist/. Do not edit substrates.
• Architecture: a document is an ordered tree of blocks (objects); an edit is a message (insert/update/move/delete); concurrent edits merge via a commutative (CRDT/semilattice) operation so order doesn't matter. History is the persist event stream; any version is a replay.
• Determinism: merge must be commutative + idempotent (test: apply ops in any order / twice → same document).
• Commits: one feature per commit. Progress log + tick boxes.

Architecture sketch

Edit op                                 Rendered document
  (insert block after id) ...             HTML / SX tree
        │                                       ▲
        ▼                                       │
lib/content/block.sx                    lib/content/render.sx
  — typed blocks as objects               — block tree → HTML/SX
  — heading/text/image/embed              — (reuses SX render boundary)
        │                                       ▲
        ▼                                       │
lib/content/doc.sx                      lib/content/merge.sx
  — ordered block tree                    — CRDT/semilattice op merge
  — apply op, structural moves            — concurrent-edit reconciliation
        │                                       ▲
        ▼                                       │
lib/content/api.sx ── (content/edit) (content/render) (content/history) ──┐
        │                                                                  │
        ├── op log + versions → persist                                      │
        └── Ghost/CMS sync  → injected external adapter (thin, non-core) ──┘

Phase 1 — Block document model

• block.sx — typed block objects
• doc.sx — ordered tree, apply edit op, structural moves
• render.sx — block tree → HTML/SX
• api.sx + tests + scoreboard + conformance.sh

Phase 2 — Op log + versioning

• edit ops as persist events; replay to any version
• (content/history doc), diff between versions

Phase 3 — Collaborative merge (CRDT)

• commutative/idempotent op merge
• concurrent-edit tests (any order, double-apply → identical)

Phase 4 — External sync + federation

• Ghost/CMS sync via injected adapter (import/export)
• federated documents (peer-authored blocks) — trust-gated stub
• tests: round-trip import/export, conflict on concurrent external edit

Extensions (post-roadmap)

• HTML escaping at the render boundary (String>>htmlEscaped: & < > ")
• asSx wire string-escaping (" and \ in SX string literals)

Progress log

• 2026-06-07 — Extension: HTML escaping at the render boundary. Added String>>htmlEscaped (recursive char walk escaping & < > ", order-safe so & isn't double-escaped) and routed every asHTML text/attr through it — heading, text, code body + language, quote, image src/alt, embed url, list items. Render stays fully polymorphic in Smalltalk; escaping lives at the boundary. +8 render tests (incl. <script> payloads, attr breakout, ampersand-once). asSx wire-escaping deferred to next. Suite 238/238.

• 2026-06-07 — Phase 4 fed.sx (Phase 4 COMPLETE — roadmap done): trust-gated federation. Peer ops carry provenance (:author, :sig stub); none are auto-accepted. The trust gate is a pluggable predicate (acl-on-sx hook) with a trusted-actor-list convenience stub. content/merge-peer[-with] applies only accepted ops through the CvRDT and quarantines the rest ({:state :accepted :rejected}). Concurrent local/external edits reconcile deterministically: same-field LWW by (ts,actor), commutative, idempotent; untrusted ops never touch state. 20 tests; suite 230/230.

• 2026-06-07 — Phase 4 sync.sx (cb1): external CMS sync via an injected adapter. Core defines the shape — {:import :export} — and delegates; content/import / content/export / content/round-trip know nothing about Ghost. A Ghost-flavoured adapter confines all format translation (post :sections ↔ content blocks, all 8 kinds). Swapping in a stub raw-adapter works identically. Round-trip (export∘import and import∘export) preserves ids, types, fields, order. 14 tests; suite 210/210. Next: trust-gated federation + concurrent-external-edit conflict (via CRDT).

• 2026-06-07 — Phase 3 crdt.sx (Phase 3 complete): collaborative merge as a state-based CvRDT. Merge is a join (lub) on a semilattice → commutative, associative, idempotent by construction. Ordering = unique dense Logoot position keys (cell = (digit actor), lexicographic); presence = OR-tombstones (remove-wins); each field = an LWW-Register keyed by logical (ts, actor). Every op contributes a PARTIAL element and per-id state is their join, so update-/delete-before-insert are not lost. crdt-materialize bridges back to a Phase-1 CtDoc (sort live elements by pos → blocks). Tests prove: ops in any order converge, double-apply is a no-op, merge commutes/associates/is idempotent, concurrent inserts order deterministically, same-field LWW by (ts,actor), disjoint fields both survive, two divergent replicas converge both ways. 34 tests; suite 196/196.

• 2026-06-07 — Phase 2 store.sx (Phase 2 complete): op log + versioning over the persist event stream. content/commit! appends an edit op as a persist event to the doc's stream (content:<id>); the log is the source of truth. content/head / content/at b id seq replay the op stream to the latest / any version (materialised doc is a cache, never primary state). content/history returns per-version metadata; content/diff / content/diff-versions report added/removed/changed block ids. Backend is injected via (persist/open) — content knows nothing about which backend. Minimal persist load (event/backend/log/kv/api). 29 tests; suite 162/162.

• 2026-06-07 — Phase 1 api.sx (Phase 1 complete): content/* facade over block + doc + render. content/bootstrap! registers the hierarchy; content/edit applies one op or an op stream; content/render picks the boundary format ("html"/"sx" or keyword). Re-exports content/new, content/append, content/insert|update|move|delete, content/find, etc. content/op? distinguishes a single op from a list/block. 26 tests; suite 133/133. content/history deferred to Phase 2 (needs the persist op log).

• 2026-06-07 — Phase 1 render.sx: render boundary as polymorphic message dispatch. Every block and CtDoc answers asHTML / asSx; the document folds children via Smalltalk inject:into: (works on raw SX lists), so (asHTML doc) / (asSx doc) are pure sends with zero type-switching in SX. Lists/headings render in Smalltalk source. No HTML escaping yet (noted in render.sx — boundary concern before untrusted content). 29 tests; suite 107/107.

• 2026-06-06 — Phase 1 doc.sx: ordered block document (CtDoc) as a Smalltalk object holding an ordered block sequence. Edit ops are data dicts (insert/update/move/delete) with op-* constructors; doc-apply / doc-apply-all interpret an op stream, each returning a NEW document (input never mutated → replay-safe). Structural moves, insert-after/at, find/index, immutability all tested. 40 tests; suite 78/78.

• 2026-06-06 — Phase 1 block.sx: typed block objects as Smalltalk instances (CtBlock hierarchy: text/heading/code/quote/image/embed/divider/list). Type tag + accessors are message sends (polymorphic dispatch); fields are immutable copy-on-write via functional st-iv-set! (history-safe). Added mk-* constructors, block? predicate, lib/content/conformance.sh + scoreboard. 38/38.

Blockers

• Smalltalk-only load chain (tokenizer/parser/runtime/eval) does not load lib/r7rs.sx/spec/stdlib.sx, so r7rs aliases (car/cdr/null?) are absent. Use base SX primitives (first/rest/(= (len x) 0)) in lib/content/**. Not a substrate bug — just the load surface.