Brummel
6fdb45d2f2
Single iter shipping the post-milestone-24 follow-up named in
docs/specs/2026-05-14-retire-per-type-print-effects.md. After this
iter the only surviving direct-output effect-op is `io/print_str`;
all per-type print primitives are replaced by the polymorphic
`print` helper (prelude, iter 24.3).
Components:
- 92 examples/*.ail fixtures migrated (do io/print_<T> x) →
(app print x); 6 .prose.txt snapshots regenerated via `ail prose`.
- Four-site lockstep compiler deletion: crates/ailang-check/src/builtins.rs
(3 effect_ops.insert blocks + 3 list() rows + the
install_io_print_float_signature test + module + EffectOpSig
doc-comments); crates/ailang-codegen/src/lib.rs lower_app
(3 arms + lowers_io_print_float test); crates/ailang-codegen/src/synth.rs
builtin_effect_op_ret match-arm pattern. Dead `intern_string`
helper removed as a follow-up.
- Five incidental test-body migrations (ailang-check x2, ailang-core
spec_drift + design_schema_drift, ailang-surface/src/lex.rs,
ailang-prose/src/lib.rs round-trip test).
- Cat B test-harness patch: six IR-shape tests in
crates/ail/tests/e2e.rs gained a monomorphise_workspace call
before lower_workspace_with_alloc so they follow the same
pipeline as `ail build` (the home-rolled desugar+lift loop
stayed because mono's precondition is "already lifted"; mono
inserts after lift).
- Six doc-comment touch-ups (lex.rs module doc, parse.rs
diagnostic example, ail/src/main.rs x2, runtime/str.c %g anchor,
crates/ailang-core/specs/form_a.md surface-spec example).
- DESIGN.md seven-site sweep (Decision 11 example, Polymorphic
print past-tense, Heap-Str output sentence, effect-op
invocation comment, Float NaN paragraph re-anchored on
float_to_str, two "What is supported" lists).
- Three E2E test-comment polish + four IR-snapshot refresh + one
canonical-hash pin update (plan-unanticipated downstream
consequences of the corpus migration).
- bench/{check,compile_check,cross_lang}.py: all exit 0; no
ratification needed.
- Roadmap entry struck through; per-iter journal at
docs/journals/2026-05-14-iter-rpe.1.md.
Tests 564/0/3. cargo clippy and cargo doc: zero warnings.
Two upstream codegen bugs surfaced during the first BLOCKED
attempt and were fixed in separate iters before this retry:
- 1fb225e bugfix: mono cursor misalignment at poly-free-fn Var
with class-constrained Forall.
- feb9413 bugfix: print leak — propagate ret_mode through rigid
substitution + prelude Show.show ret_mode.
Known debt (carried forward for next /audit):
- Emitter.strings field is functionally dead post-iter (orphan
after intern_string removal); cycles over empty map harmlessly.
95 lines
3.0 KiB
Plaintext
95 lines
3.0 KiB
Plaintext
; Bench fixture: explicit-mode pair of bench_list_sum.
|
|
;
|
|
; Same algorithm and same sizes as bench_list_sum.ail (build a list of
|
|
; 0..N-1, sum it, print) but with full `(borrow)` / `(own)` annotations
|
|
; on every fn-param signature in the hot path. Under --alloc=rc the
|
|
; codegen now emits `inc`/`dec` instructions: each cell allocated by
|
|
; `cons_n_acc` is dec'd as `sum_list_acc` walks the chain.
|
|
;
|
|
; The IntList is `(drop-iterative)` so the param drop on the LCons arm
|
|
; stays O(1) stack regardless of length — without this, dec-ing a
|
|
; 3M-cell chain at scope close would recursive-overflow.
|
|
;
|
|
; This is the "RC-fair" arm of the cross-language bench (21'f);
|
|
; pairs with bench/reference/list_sum_explicit_free.c which adds
|
|
; matching free() calls. Together they answer "what does RC's full
|
|
; alloc+dec cost look like vs C's full malloc+free cost?".
|
|
;
|
|
; Sizes match bench_list_sum exactly:
|
|
; 100_000 * 24 B = 2.4 MB
|
|
; 1_000_000 * 24 B = 24 MB
|
|
; 3_000_000 * 24 B = 72 MB
|
|
;
|
|
; Expected stdout (one int per line):
|
|
; 100_000 -> 4999950000
|
|
; 1_000_000 -> 499999500000
|
|
; 3_000_000 -> 4499998500000
|
|
|
|
(module bench_list_sum_explicit
|
|
|
|
(data IntList
|
|
(doc "Singly-linked Int list. drop-iterative so Own-param drops are O(1) stack.")
|
|
(ctor INil)
|
|
(ctor ICons (con Int) (con IntList))
|
|
(drop-iterative))
|
|
|
|
(fn cons_n_acc
|
|
(doc "Tail-recursive: prepend (n-1, n-2, ..., 0) onto acc. Returns owned chain.")
|
|
(type
|
|
(fn-type
|
|
(params (con Int) (own (con IntList)))
|
|
(ret (own (con IntList)))))
|
|
(params n acc)
|
|
(body
|
|
(if (app == n 0)
|
|
acc
|
|
(tail-app cons_n_acc
|
|
(app - n 1)
|
|
(term-ctor IntList ICons (app - n 1) acc)))))
|
|
|
|
(fn cons_n
|
|
(doc "Build [0..n-1] :: IntList. Returns owned chain; caller owns and consumes.")
|
|
(type
|
|
(fn-type
|
|
(params (con Int))
|
|
(ret (own (con IntList)))))
|
|
(params n)
|
|
(body (app cons_n_acc n (term-ctor IntList INil))))
|
|
|
|
(fn sum_acc
|
|
(doc "Tail-recursive sum. Owns xs; consumes via LCons-arm move-into-tail-call.")
|
|
(type
|
|
(fn-type
|
|
(params (own (con IntList)) (con Int))
|
|
(ret (con Int))))
|
|
(params xs acc)
|
|
(body
|
|
(match xs
|
|
(case (pat-ctor INil) acc)
|
|
(case (pat-ctor ICons h t)
|
|
(tail-app sum_acc t (app + acc h))))))
|
|
|
|
(fn sum_list
|
|
(doc "Sum every element. Owns xs, hands it to sum_acc which consumes it.")
|
|
(type
|
|
(fn-type
|
|
(params (own (con IntList)))
|
|
(ret (con Int))))
|
|
(params xs)
|
|
(body (app sum_acc xs 0)))
|
|
|
|
(fn run_one
|
|
(doc "Build, sum, print. The owned list flows from cons_n into sum_list and is fully consumed.")
|
|
(type (fn-type (params (con Int)) (ret (con Unit)) (effects IO)))
|
|
(params n)
|
|
(body
|
|
(app print (app sum_list (app cons_n n)))))
|
|
|
|
(fn main
|
|
(type (fn-type (params) (ret (con Unit)) (effects IO)))
|
|
(params)
|
|
(body
|
|
(seq (app run_one 100000)
|
|
(seq (app run_one 1000000)
|
|
(app run_one 3000000))))))
|