chore: drop accidental .claude/worktrees gitlink; ignore worktrees dir #10
Loading…
Reference in a new issue
No description provided.
Delete branch "feat/010-chore-drop-accidental-claude-worktrees-gitlink"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
chore: drop accidental .claude/worktrees gitlink; ignore worktrees dir
Along the way:
Touches 15 files:
.gitignore,docs/impl/runtime-testing/README.md,docs/impl/runtime/R1-executor-core.md,docs/impl/runtime/R8-the-floor.md,docs/impl/runtime/README.md,docs/impl/runtime/substrate-and-supervisor.mdand 9 more.Split. The decision in
reason_labelis a pure function now, with its own tests insrc/rust/crates/ck_driver/src/repl.rs.Nothing outside calls
events. It is pub(crate) insrc/rust/crates/ck_rt/src/lib.rsnow.A clear improvement. The change in
src/rust/crates/ck_driver/src/repl.rsis the part I would have argued for.@ -0,0 +136,4 @@}}fn reason_label(reason: DespawnReason) -> &'static str {reason_labelreads, decides and writes in one function. The deciding part is the piece I would want to test on its own. Worth splitting?@ -0,0 +96,4 @@/// Every event recorded so far, in emission order.#[must_use]pub fn events(&self) -> &[MetaEvent] {eventsis public. Does anything outside the crate call it, or is it pub because that was easier? Every public item is a promise.Invert the runtime cascade to floor-first per the "everything is an actor" decision: build the real actor substrate first, not a throwaway tree-walker. The executor's eval becomes the behavior an actor runs (order 2). Docs (docs/impl/runtime/): - README.md re-sliced with an Order column; 4 locked sub-decisions; the bootstrap-branch seam (master owns/extends ck_hir with the owned Core-Erlang tree + side-table types; runtime consumes via ck_eval; order 1 insulated, order 2 rendezvous with the lowering). - substrate-and-supervisor.md: new order-1 spec. - R1 reframed (eval in a behavior, order 2); R8 reframed (multi-actor, order 10). - Correct event vocabulary to the closed MetaEvent set: ActorSpawned / ActorDespawned (ArenaReclaimed is proposed/M7, not v1 -- reclaim observed via RuntimeState). Binary is `ck`. Code -- ck_rt (order 1, increment 1: lifecycle + reclaim): - Runtime::boot() brings up the universal root supervisor. - Runtime::spawn(kind, behavior): single carrier, run-to-completion; emits ActorSpawned -> ActorDespawned{Normal}; reclaims (safe arena stand-in, unsafe_code still forbidden). - MetaStream (order-1 subset) + RuntimeState (live set + arena high-water) -- the two post-interpret artifacts, now real. - 4 unit tests gate the lifecycle / reclaim / fresh-id invariants. - Registered ck_rt in the workspace + the architecture dependency graph. Gates: clippy --workspace (pedantic+all=deny) 0/0; cargo test --workspace all pass; cargo fmt --check clean.