The generic-?? wrong-fallback was not in lowerNullCoalesce: coercing
?A -> ?B (differing payload, e.g. the ?i32->?i64 call-arg coercion when
instantiating unwrap_or(99, ?i32)) routed through .optional_wrap, which
unconditionally unwrapped the source and re-wrapped as ALWAYS-PRESENT, so
a null became present-zero everywhere (args, returns, field init,
var-decl, ??). Add a CoercionPlan.optional_to_optional (conversions.zig)
+ a presence-preserving arm in coerceMode (coerce.zig): has_value ->
present: unwrap+coerce-child+wrap-present; absent: constNull(dst); merge
via a dst_ty block param. lowerVarDecl gains a !src_is_optional guard so
an annotated x : ?B = <?A> routes through the same arm (also makes
aggregate-payload var-decl ?[3]i64->?[]i64 / ?Concrete->?Protocol work).
Alias-optional struct-literal default already works (grouping + 0166);
a 1-tuple default ?(i32,) ?? 5 now emits a clean diagnostic instead of an
LLVM PHI abort (no implicit scalar->1-tuple coercion per spec).
Regressions: optionals/0916 (generic ??), 0917 (alias struct default),
0918 (var-decl optional->optional), diagnostics/1202 (1-tuple default) +
a conversions.test.zig unit test. Verified by 3 adversarial reviews,
suite 798/0.
opt?.xs[i] typed and lowered the index over the optional CONTAINER
(?[N]T); getElementType returned .unresolved, so index_get reached LLVM
with an unresolved element type and panicked. Mirroring the 0101
!-unwrap fix: add lowerOptionalChainIndex (optional_has_value -> some:
unwrap + index (index_gep+load for ?*[N]T, else index_get) +
optional_wrap; none: const_null; merge -> ?ElemType, element-optional
flattened). The typer + dispatch guard compute the element via
ptrToArrayElem(child) orelse getElementType(child), so value-arrays,
slices, many-pointers, AND pointer-to-array (?*[N]T) children resolve.
Null receivers short-circuit (no null deref).
Regression: examples/optionals/0915-optional-chain-array-field-index.sx.
Verified by 3 adversarial reviews, suite 794/0. Filed broader pre-existing
gap 0183 (indexing a non-indexable type panics instead of diagnosing).
lowerNullCoalesce fed resolveOptionalInner's .unresolved (returned for a
non-optional lhs) into the merge-block params / optionalUnwrap / RHS
target type, reaching codegen and panicking 'unresolved type reached
LLVM emission'. Guard: when inferExprType(nc.lhs) is a resolved
non-optional type, emit a located diagnostic and bail; an .unresolved
lhs (prior error) is excluded to avoid double-report. ?? is optional-only
per specs.md (error unions use or/catch), so rejecting a failable lhs is
correct; comptime panic closed too.
Regression: examples/diagnostics/1200-diagnostics-null-coalesce-non-optional.sx.
Verified by 3 adversarial reviews, suite 790/0. Filed adjacent bug 0180
(?? lowering defects for generic/alias/tuple optional lhs).
0173: resolveArrayLiteralType gained no arm for [N]T/[]T heads, so a
([2]?i64).[...] head lost its ?i64 element type and a bare null reached
LLVM as const_null(.unresolved). Route structural heads through
resolveTypeWithBindings; validate an undefined element name in the head
via UnknownTypeChecker (semantic_diagnostics.zig) instead of a silent
empty-struct stub (no-silent-fallback).
0174: positional .{...} against a TUPLE target now coerces each element
to TupleInfo.fields[i] (was neither struct nor array, so uncoerced).
0175: a positional struct literal with a bare-variable element was
misclassified as a named shorthand (parser puns .{x} -> x=x), zeroing
the fields. has_names now consults the struct definition to reclassify a
punned non-field name as positional; positional coercion uses the
lowered value's real getRefType.
Regressions: optionals/0914, types/0199, types/0200, diagnostics/1196.
Verified by 4 adversarial reviews; suite 784/0. Filed adjacent bug 0176
(protocol-typed struct field method call aborts).
[N]?T arrays were corrupted: a positional literal .{ null, 7 } stored
bare T/null elements into {T,i1} optional slots because array elements
were never coerced (getStructFields is empty for an array, so the
i<struct_fields.len field-coercion gate never fired). A present element
then read back as absent and direct indexing segfaulted.
lowerStructLiteral's positional branch now computes array_elem_ty for
array/vector targets and coerces each element to it; lowerArrayLiteral
generalizes its slice-only coercion to coerce every element via
coerceToType (layout-aware: scalar->{T,i1}, pointer-sentinel->one-word,
array->slice, concrete->protocol). Verified by 3 adversarial reviews,
suite 780/0.
Regression: examples/optionals/0913-optionals-array-of-optionals.sx.
Filed adjacent pre-existing bugs: 0173 (typed .[null,..] element), 0174
(tuple positional-element coercion), 0175 (positional struct literal
variable element zeroed).
The RHS of a null-coalesce was lowered with no target type, so a bare
struct literal default (x ?? .{ ... }) produced a struct_init with
.ty == .unresolved that panicked in emitStructInit. lowerNullCoalesce
now saves self.target_type, sets it to the optional's resolved child
before lowering nc.rhs, and restores it (leak-free). Verified across
struct/slice/enum/tuple/protocol/nested-optional/generic child types by
3 adversarial reviews.
Regression: examples/optionals/0912-null-coalesce-struct-literal.sx.
Filed adjacent pre-existing bug 0172 (?? on a non-optional lhs panics).
lowerIfExpr emitted optional_has_value only for the binding form; a bare
'if opt' passed the raw {T,i1} aggregate to condBr, where emitCondBr's
catch-all struct arm silently folded it to 'i1 true' (structs always
truthy) — a silent miscompile that took the present-branch for null
optionals. while / and / or shared the same defect.
Reduce bindingless optional conditions to optional_has_value in
lowerIfExpr/lowerWhile and via a new lowerBoolCondition helper for and/or
operands. Replace the silent-true emitCondBr arm with a lowering-time
diagnostic (checkConditionType/isValidConditionType) rejecting conditions
whose type isn't bool/integer/pointer/optional; the backend @panic is now
an unreachable tripwire.
Regressions: examples/optionals/0908..0910 + diagnostics/1194 (negative).
Verified by 3+3 adversarial reviews.
Filed adjacent bugs found during review: 0168 (array-of-optionals element
load), 0169 (optional->bool coercion), 0170 (closure-optional layout).
Five adversarial reviews of the issue-0160 fix surfaced three more bugs in the
touched optional-chain / optional-coercion code; all fixed here:
1. A COLD generic-instance getter through `?.` (`?*Vec(i64)` `.getter`, never
called directly first) panicked with "unresolved type reached LLVM emission":
a cold instance method is absent from resolveFuncByName, so the getter's
return type resolved to .unresolved → a ?unresolved merge type. lowerOptionalChain
and getterReturnTypeOnDeref now warm the monomorph (ensureGenericInstanceMethodLowered)
before querying its return type. (The 0907 test passed only by luck — List(i64)
is warmed by stdlib use; 0907 now also exercises a cold user generic.)
2. A real-field read through a `?*T` chain (`op?.field`, op: ?*T) reinterpreted
the pointer bits as the field (silent garbage) — the some-branch real-field
path didn't load through the pointer. It now derefs `?*T` before the field
access. (Pre-existing — the else-branch predates 0160 — but it's the same
function and a silent miscompile, so fixed here.)
3. `?[]T = array` skipped the array→slice promotion (corrupt .len/.ptr): the
lowerVarDecl optional arm wrapped the raw array. It now coerces the value to
the optional's child type (array→slice) before wrapping.
Regression examples 0906/0907 extended to cover all three. Distinct PRE-EXISTING
bugs the reviews surfaced in untouched subsystems are filed as issues 0161
(struct-literal vs scalar), 0162 (#run returning an optional aggregate), 0163
(untagged-union payload-binding match).
Two fixes for optional interactions surfaced by the #set/#get review. The
original issue 0160 mis-diagnosed (A) as an optional-chain bug; the chain works
fine for real fields. The actual bugs:
(A) A bare struct literal `.{ ... }` against an optional target `?T` was built
into the optional's {payload, has_value} layout instead of the inner T, then
re-wrapped — corrupting the value (a multi-field payload's first field clobbered
by the has_value flag, or a `?T` arg silently null) or failing LLVM
verification. lowerStructLiteral now builds the inner T, materializes it, and
wraps via coerceToType; lowerVarDecl's previously-UNCONDITIONAL optional wrap is
guarded so an already-`?T` value isn't double-wrapped. Fixed across var-decl,
arg, return, nested field, reassignment, and array-element contexts.
(B) `#get` accessors are now reachable through an optional chain (`obj?.getter`):
lowerOptionalChain dispatches the getter via a synthetic receiver, and
expr_typer types `obj?.getter` through a shared getterReturnTypeOnDeref helper
(handles `?T` and `?*T`, value and pointer optionals, and generic-instance
getters like List.len). The `#set` write side through `?.` is intentionally left
matching real-field behavior (optional-chain assignment unsupported).
Regression tests: examples/optionals/0906 (struct-literal → optional) and 0907
(accessor through chain). issues/0160 marked RESOLVED with the corrected root
cause.
A method `name :: (self: *T, value: V) #set { ... }` (or `=> expr;`) is the
write counterpart of a `#get` accessor: `obj.name = rhs` dispatches to it as
`obj.name(rhs)` when no real field matches. Plumbed parallel to `#get`:
- lexer/token `#set`; `FnDecl.is_set` + `Function.is_set`; parsed in the same
marker slot as `#get` (no return type, exactly self + one value param).
- get+set coexistence: a setter registers/mangles/dispatches under an effective
`name$set` name (`$` is illegal in sx identifiers, so unmistakable), keeping a
same-name `#get` under the plain `name`. Resolution is declaration-order-
independent: a plain read query picks the non-setter, a `name$set` write query
picks the setter (accessorEffName / accessorNameMatches / structMethodFn).
- write dispatch in lowerAssignment via tryLowerPropertyAssignment: plain assign
synthesizes `obj.name$set(rhs)`; compound `OP=` is get-modify-set and
evaluates the receiver EXACTLY ONCE (bound to a synthetic local); read-only
(#get-only) and write-only (#set-only + compound) emit clear diagnostics; a
real field of the same name still wins. Multi-assign property targets dispatch
the setter too (tryLowerPropertyStore, via a pre-lowered-Ref binding).
Payoff: List gains a `len` #set, so `xs.len = n` works; the `.items.len = N`
write workarounds in sched.sx + ui/* + platform/* revert to `xs.len = N`.
issues/0160 records an optional-chain interaction surfaced by the review (a
pre-existing `?T` value-optional read miscompile that blocks getter-through-`?.`).
A method declared `name :: (self: *T) -> R #get => expr;` is invoked via
no-paren field syntax (`obj.name`) instead of `obj.name()`. It is an ordinary
method (registered `Type.method`, flagged is_get); field-access lowering and
inference dispatch to it when no real field of that name exists, by synthesizing
a no-arg `obj.name()` call routed through the normal call path (so receiver
address-of and generic binding are reused).
- Lexer/token: `#get`. Parser: parsed after the return type in parseFnDecl;
hasFnBodyAfterArrow treats it as a body marker so struct-body methods parse.
- Resolution: getAccessorFor handles a generic-struct instance and a plain
struct. A REAL field of the same name wins (a getter never shadows stored
data). An explicit postfix-deref receiver (`p.*.getter`) dispatches on the
inner pointer so it takes the working auto-deref path.
- Works on plain + generic structs (incl. getters returning the type param),
in expressions/conditions/args/loop-bounds, chained, and via a pointer
receiver. Examples: types/0196 (basic) + types/0197 (stress).
Known narrow limitations (clean errors / workarounds, not silent): a getter
RESULT used directly as a method/getter receiver (`o.gi.dbl`) errors — bind it
to a local first; a getter named `len`/`ptr` returning non-i64 mis-infers
(the .len/.ptr builtin-field shortcut).
emitSubslice handled a struct (slice/string) base and an array base, but a
many-pointer [*]T base is an LLVM pointer kind — it fell through to the else arm
that mapped the result to LLVMGetUndef(slice_ty), so a slice of a many-pointer
(mp[lo..hi]) had a garbage .len/.ptr and iterating it segfaulted.
Add a LLVMPointerTypeKind branch: the base value IS the data pointer, so GEP by
lo and len = hi - lo (the caller supplies the bound; no length is read from the
unbounded pointer). An open-ended mp[lo..] has no resolvable upper bound (a [*]T
carries no length), so lowerSliceExpr now diagnoses it instead of emitting a
.length op that yields garbage.
A List (whose items is [*]T) is now iterable with for items[0..len] (e);
applied in Scheduler.deinit. Regressions: examples/types/0195 (valid slice +
List for-each) + examples/diagnostics/1192 (open-ended rejection).
A plain union initialized with a struct literal (b : Overlay = .{ f = 3.14 })
silently miscompiled — it fell through the generic struct-literal path
(getStructFields returns empty for a union), building a malformed structInit
whose overlapping zero-fill clobbered the named member, so it read back 0.0
(and a type-pun read segfaulted).
lowerStructLiteral now detects a plain-union target and dispatches to a new
lowerUnionLiteral, which writes each named member into a union-sized slot via
the same lvalue resolver the u.member = v assignment path uses, then loads the
union value back. Validity: the named members must share one arm — a single
direct member, or several promoted members of the same anonymous-struct variant.
Overlapping members, members from different arms, and positional union literals
are rejected with a diagnostic (no silent last-wins); an empty .{} yields an
undefined union (matching the --- form).
specs.md updated. Regressions: examples/types/0194 (valid forms) +
examples/diagnostics/1191 (overlap rejection).
A comparison with int-vs-float (or two float widths) operands emitted cmp on
the raw operands with no promotion, unlike the arithmetic arms -- producing a
mixed-type compare the LLVM verifier rejects / mis-evaluates. lowerBinaryOp now
coerces each operand to the promoted common type (from arithResultType) via
coerceToType (SIToFP / FPExt) for the ordering/equality arms when the promoted
type is a float, so LLVM gets a well-typed fcmp.
Regression: examples/0189-types-int-float-compare-promote.sx
A *self method called directly on arr[i] (or a deref place) fell through to an
alloca+store-of-value, so the callee mutated a throwaway copy and the live slot
was never written. fixupMethodReceiver now takes the real address of
.index_expr/.deref_expr receivers via lowerExprAsPtr (normalized to *T),
mirroring the explicit-argument path. A comptime-pack index (xs[i] where xs is
a pack) is excluded -- a pack has no runtime storage to address -- so it keeps
flowing through the general copy path.
Regression: examples/0188-types-method-array-index-receiver.sx
type_resolver "Type" -> .type_value; const_type result + emitConstType now a
bare 8-byte i64 handle (not a 16-byte Any box). Migrated every .any ref meaning
"a Type value", leaving real boxed-Any refs:
- "Any holds a Type" meta-marker tag .any -> .type_value at all 4 consumers
(reflectArgTypeId, reflectTypeId, the comptime type_tag-as-struct path,
resolveTypeCategoryTags "type").
- reflection-builtin return types (type_of/declare/define) -> .type_value;
runtime type_of(any) reads the tag as a .type_value (no re-box).
- expr_typer: a bare type-name expr is .type_value (backtick is_raw exempt).
- reflectionArgIsType accepts .type_value OR .any (a reflection arg can be a
bare Type or a boxed Any).
- comptime switch_br accepts a .type_tag discriminant (type-category match).
- a bare function name in a Type slot -> const_type(its function type), not a
func-ref (fixes a JIT crash); old string-box kept only for genuine Any params.
- field-not-found diagnostic + formatTypeName render .type_value as "Type".
Fixed 3 unit tests asserting the old .any behavior. 697/0 both gates (gate ON
bails cleanly to legacy since the VM doesn't model Type values yet) + 494 unit
tests. 24 snapshots regenerated (22 .ir const_type shape; 2 .stderr Any->Type).
The mutating compiler-API, minting types LAZILY at lowering time (single pass,
the existing runComptimeTypeFunc path — so the write side is legacy-only; the
VM isn't wired at lowering time, and the read-side readers stay dual-path):
declare_type(name) -> Type forward nominal handle (≈ declare)
pointer_to(t) -> Type build *T references
register_type(handle, kind, members) ONE kind-branching fill (≈ unified define)
register_type branches on kind IN THE COMPILER (subsuming define's per-kind
dispatch); codes match type_kind: 1 struct, 2 actual .@"enum", 3 tagged_union,
4 tuple. Members are {name: string, ty: Type}. A non-generic `-> Type` builder is
now flagged is_comptime (decl.zig) so its dead body permits the welded calls.
Graph support: forward declare_type handles + pointer_to express a mutually-
recursive A<->B graph (*A, *B, B-by-value) before bodies are filled. register_type
is idempotent — re-filling a nominal slot (a minting module reached via two import
edges) re-mints identically rather than erroring (nominalIdent reads identity from
any nominal kind).
Fixes (issue 0142):
- A fully payloadless comptime-minted enum was minted as an all-void tagged_union,
whose IR size disagrees with its LLVM size -> verifySizes panic. Now mints a real
.@"enum" (register_type kind 2 AND the metatype defineEnum).
- Bare `EnumType.variant` qualified construction of a payloadless variant wasn't
supported (failed for hand-written enums too — the type name lowered to a Type
value). Added in lowerFieldAccess via isPayloadlessVariant; payload-carrying
variants keep their call form.
Examples: 0631 (graph + actual enum + reflection), 0632 (make_enum all-void),
0633/0634/0635 (namespaced / bare / multi-edge import of a minted type), 0187
(qualified variant construction). Unit tests added.
Parity 697/697 (gate OFF and -Dcomptime-flat).
Introduce the welded comptime `compiler` library (`#library "compiler"` +
`abi(.zig) extern compiler`), per design/comptime-compiler-api.md, and unify
`callconv(...)` into the new `abi(...)` annotation.
abi(...) replaces callconv(...):
- New ABI enum { default, c, zig, pure }; `abi(.c|.zig|.pure)` parses in the
postfix slot before extern/export (and standalone). `kw_callconv` -> `kw_abi`.
- Migrated 52 sx files, the call-convention-mismatch diagnostic, and docs
(readme/specs) from `callconv(.c)` to `abi(.c)`.
Phase 1 — welded compiler library (parse -> registry -> validation -> bridge):
- `abi(.zig) extern compiler` parses on fn decls (carries abi/extern_lib) and
struct decls (StructDecl.abi/extern_lib).
- `#library "compiler"` is the comptime-only internal surface — never dlopen'd.
- src/ir/compiler_lib.zig: the binding registry (the safety boundary). `Field`
welded to StructInfo.Field with layout baked from the real Zig type
(@offsetOf/@sizeOf); `findType`/`findFn`. Welded structs are layout-validated
at registration (field set + total size) as a header checked against the impl.
- Host-call bridge: a `fn abi(.zig) extern compiler` dispatches under the
comptime interp to its registered Zig handler (intern/text_of round-trip),
never dlsym. IR Function.compiler_welded; validated in declareFunction.
- Comptime-only enforcement: a runtime call to a welded fn is a clean
build-gating error (emitCall), not an undefined-symbol link failure.
Phase 2.1 — byte-layout weld foundation:
- Decision: full byte-layout weld (sx struct laid out byte-identically to the
bound Zig type). Registered StructInfo (first non-natural / Zig-reordered
layout). `computeWeldPlan` — pure offset-ordered element plan + padding +
sx-field->LLVM-element remap; unit-tested. Emit/interp wiring is the next
sub-step (2.2+, see current/CHECKPOINT-COMPILER-API.md).
Examples: 0625/0626 (welded struct + fn round-trip), 1183/1184/1185
(layout-mismatch, unexported-fn, runtime-call diagnostics).
`arr[lo..hi]` at comptime bailed for any non-string base — the interp's
.subslice op only handled string-backed values. Worse, the open-ended
`hi` came from a .length op that misread a 2-element array as a {ptr,len}
fat pointer (returning the 2nd element, not the count), so even lo/hi
weren't valid ints.
Fix, interp-only (runtime already handles arrays via LLVMTypeOf):
- Thread the base operand's IR type onto the Subslice op (base_ty); the
interp uses it to tell a bare array (elements = aggregate fields) from a
{data,len} slice (elements in the data field) — indistinguishable by
Value shape alone.
- Fold an open-ended slice's hi to the array's static length for fixed
arrays at lower time (runtime emitLength folds the same constant, so the
IR result is unchanged — no snapshot churn — but the comptime interp no
longer hits the ambiguous .length op).
- subsliceElements() resolves the element list (array/slice, inline or
slot_ptr-backed) and subslice returns a proper {data,len} slice value.
Suite green (678), no .ir changes.
A `"s"` input operand feeds a function/global symbol; the template's
%[name] emits the platform-mangled name, so `bl %[fn]` / `call %[fn]`
branches DIRECTLY to it (PC-relative, no register load — one fewer
indirection than register-indirect `blr`).
Lowering: an `"s"` input lowers its RHS normally (a function name →
`ptr @fn`); the rejection added last commit is removed. Emit: a symbol
operand is passed with its OWN llvm type (LLVMTypeOf) and no coercion —
the function value is a `ptr`, and the old coerce-to-register-int path
mistyped it and failed the verifier. New asmIsSymbol helper.
Verified on aarch64: examples/1656 (sx → asm → bl _cb → sx → 42); the
emitted asm is a direct `bl <_cb>` (objdump-confirmed), IR constraint
`...,s,...`(ptr @cb). Flipped 1656 from the rejection lock to a runnable
aarch64 example. zig build test green (665 corpus, 446 unit).
A symbol operand (constraint "s") feeds a function/global symbol whose
mangled name the template emits — enabling a DIRECT `bl %[fn]` (one
fewer indirection than register-indirect `blr`). Until now `"s" = fn`
fell through to emit and produced an LLVM-verifier crash (param type
mismatch). Reject it at lowering with a clear diagnostic instead, and
lock that with examples/1656-platform-asm-symbol-operand.sx. The next
commit implements it and flips the example to run (→ 42).
Implements indirect-memory (`=*m`) `-> @place` outputs — the last
substantive asm feature. Unlike a write-through `=` output (which
returns a value that is then stored), an indirect output passes the
place ADDRESS to the asm and the asm writes through it; there is no
return slot.
emitInlineAsm:
- indirect outputs are excluded from the LLVM return type;
- their pointer is passed as an opaque `ptr` call arg, placed FIRST
(the arg-consuming constraint order is: output-section indirect
pointers, then inputs, then read-write tied seeds);
- each indirect arg gets an `elementtype(T)` call-site attribute
(required in the opaque-pointer era), T = the pointee type;
- the store-back loop skips indirect outputs (already written).
New asmIsIndirect helper. Lowering stops rejecting `*` (constraint kept
verbatim; `=*m` reaches the constraint string as-is). asmOperandIndex
is unchanged — indirect outputs still count as operands, so `%[name]`
${N} numbering holds.
Verified by running on aarch64: store-through-pointer (str x9, %[out]
→ 42, IR `=*m,~{x9}` with `ptr elementtype(i64)`) and a mixed case
(indirect + value output + input → `=*m,=r,r`, indirect ptr arg first,
${0}/${1}/${2} correct). 1652 flipped from the rejection lock to a
runnable aarch64 example (ir-only elsewhere). zig build test green
(661 corpus, 446 unit).
A scalar `::` constant folds to its value and has no storage. The
unary `.address_of` lowering (src/ir/lower/expr.zig) skipped the
alloca path (is_alloca == false) and resolveGlobalRef (scalar consts
get no storage global), falling through to the generic addr_of arm,
which reinterpreted the folded value as a pointer:
`inttoptr (i64 <value> to ptr)`. That wild pointer segfaulted on
deref and emitted invalid stores for inline-asm `-> @const`.
Diagnose instead, in the address_of(identifier) path: a non-alloca,
non-ref-capture, non-pack-elem scope binding (local scalar const) and
a module_const_map name not backed by storage (module scalar const)
both report "cannot take the address of constant '<name>' — a scalar
'::' constant has no storage …" and return a placeholder Ref. Chose
diagnose over materializing read-only storage (consistent with the
fold-only scalar model). Array/struct consts keep real storage and
stay addressable (@K/@LIT unchanged).
Also gives the ASM stream's planned output-to-const rejection for
free — asm `-> @const` lowers through the same path. Regression:
examples/1177-diagnostics-addr-of-const-rejected.sx. Resolves 0138.
Implements read-write (`+r` / `+{reg}`) `-> @place` outputs. LLVM has
no `+` constraint, so a read-write place lowers to:
- an output `=` constraint (return slot, stored back through the
place after the call), with the leading `+` rewritten to `=`; plus
- a TIED input constraint (the decimal index of that output) appended
after the regular inputs, seeded with the place's loaded value
passed as a call arg.
Tied inputs are appended last so existing operand indices (%[name] ->
${N}) are undisturbed; asmOperandIndex stays correct. Lowering no longer
rejects `+` (indirect `*` still rejected). emitInlineAsm grows the
arg/param arrays by the rw count, loads each seed, and emits the tied
constraint.
Verified by running: increment-in-place (41 -> 42) and a mixed case
(rw place + regular input + value output) producing the textbook
"=r,=r,r,0" constraint with correct ${N} indices. 1650 flipped from
the rejection lock to a runnable aarch64-pinned example (ir-only
elsewhere). zig build test green (658 corpus, 446 unit).
An asm result can be STORED through a place (a local / struct field) instead of
returned; the place output does not join the result tuple.
- parser.zig: `-> @place` parses `@place` as an ordinary address-of expression
→ an out_place operand (the in-function form; reuses the existing `@` prefix).
- inst.zig: AsmOperand gains out_ty (the output slot's value type) so emit can
build the combined return struct without re-deriving from Inst.ty.
- lower/expr.zig: out_place operand = the lowered @place address, out_ty = the
pointee. Read-write (`+`) and indirect-memory (`*`) constraints rejected loudly
(not yet implemented) rather than miscompiled.
- ops.zig emitInlineAsm: the LLVM return type is built from ALL outputs
(out_value + out_place); after the call, out_place slots are stored through
their address and out_value slots rebuild the sx result. Fast path when there
are no place outputs (the struct return IS the result — pure-value asm IR
unchanged).
Verified: write-to-local (42), struct field, mixed value+place (v=10 b=20), `+`
rejected. Locked with 1649-platform-asm-place-output (mixed, runs on aarch64).
zig build test green (657 corpus, 446 unit).
Replaces the N>1 "Phase E" bail with a shared asmResultType helper (lowering +
inferType) that derives the result type from the out_value operands: 0→void,
1→T, N→a named tuple (fields named via the §II.5 effective-name rule).
Key realization: toLLVMType(tuple) already produces a literal struct {T1,…,Tn} —
exactly what LLVM's multi-output inline asm returns — so emit needs NO change.
Building the op with a tuple result type makes the asm call return the struct,
which IS sx's tuple value (destructured by the normal tuple_get path).
inferType's .asm_expr arm now also delegates to asmResultType (single owner), so
`return asm`, `x := asm`, and `q, r := asm` all agree on the type.
Verified end-to-end on aarch64: split(0x1234)→(lo=52,hi=18), a udiv/msub
divmod→(3,2). IR: `call { i64, i64 } asm "divq ${4}",
"={rax},={rdx},{rax},{rdx},r,~{cc}"(…)` → extractvalue → tuple.
1640 → the x86_64 multi-output IR lock (ir-only); 1647 → a multi-output example
that runs on aarch64.
zig build test green (655 corpus, 446 unit).
lowerAsmExpr stops bailing and builds the inline_asm op: resolves each operand's
effective name (§II.5 — explicit [name] else the {reg} pin), interns
template/constraints/clobbers, lowers input Refs, derives the result TypeId
(0→void, 1→T). Adds the last deferred validation (every %[name] must name an
operand). Multi-output (N>1) bails with a named "Phase E" diagnostic.
emitInlineAsm (backend/llvm/ops.zig) ports Zig's airAssembly: assembles the LLVM
constraint string (outputs → inputs → ~{clobber}, ',' → '|'), rewrites the
template (%[name]→${N}, %%→%, $→$$, %=→${:uid}), then LLVMGetInlineAsm +
LLVMBuildCall2 (AT&T dialect). Dispatch wired in emit_llvm.zig (replacing the C.0
@panic tripwire).
inferType gains an .asm_expr arm (expr_typer.zig) so a bare `x := asm {…-> T}`
binding types correctly — without it the binding inferred .unresolved and
silently produced 0.
llvm_shim.c: LLVMInitializeNativeAsmParser() — the JIT must assemble inline asm
at run time.
Verified end-to-end on the aarch64 host: `mov`/`add` with register-class inputs
and a value output run (exit 42/99), `nop volatile` runs (exit 0). IR is
textbook: `call i64 asm "add ${0},${1},${2}", "=r,r,r"(…)`.
Locked with 1645 (aarch64 add, runs; ir-only on non-aarch64) + 1646 (:= binding).
Updated 1640 (now Phase-E bail) + 1642 (now runs).
zig build test green (654 corpus, 446 unit).
Extends lowerAsmExpr with a pinnedRegister(constraint) helper and two §II.5
operand-naming checks, in the compile path before the codegen bail:
- reject the echo form `[eax] "={eax}"` — a label identical to the register its
own constraint pins is redundant (the operand is already auto-named after the
register); the useful form is a label that differs (`[quot] "={rax}"`);
- reject duplicate operand names (ambiguous %[name] / result field).
Locked with 1643-platform-asm-echo-name and 1644-platform-asm-duplicate-name.
zig build test green (652 corpus, 445 unit).
Restructures the .asm_expr lowering arm into lowerAsmExpr, which validates the
asm shape with specific named diagnostics BEFORE the not-yet-implemented codegen
bail, so the user sees the real problem first. Two checklist items enforced:
- template must be a compile-time-known string ("..." or #string), not a
runtime expression;
- an asm with no value outputs must be `volatile` (else its effects could be
deleted) — mirrors Zig's rule.
Valid shapes still bail with the "codegen not yet implemented" message. Result-
type derivation + the operand auto-naming rule stay deferred to Phase C, where a
real IR op makes the result type observable/testable.
Locked with 1641-platform-asm-missing-volatile (the volatile error) and
1642-platform-asm-nop-volatile (no-output + volatile accepted → codegen bail).
zig build test green (650 corpus, 445 unit).
`asm volatile? { "tmpl", [name]? "constraint" (-> Type | = expr), …,
clobbers(.…) }` now parses into a flat-operand AsmExpr/AsmOperand (ast.zig +
parser.zig parseAsmExpr, dispatched from parsePrimary on .kw_asm). `volatile`
and `clobbers` are recognized contextually (not reserved). `-> @place`
write-through is rejected with a clear "Phase 2" parse error.
Codegen is not implemented yet (IR op + LLVM emit are Phases C–E), so lowering
bails LOUD + named via an explicit .asm_expr arm in lower/expr.zig (not the
generic unknown_expr else) — emitPlaceholder makes hasErrors() abort the build
on the message.
The new asm_expr tag forced (and got) arms in three exhaustive Node.Data
switches: sema.zig analyzeNode + findNodeAtOffset, semantic_diagnostics.zig
checkBindingNames — each recurses into template + operand payloads.
Design: adopted the operand auto-naming rule (design §II.5) — name auto-derived
from a {reg} pin, explicit [name] only when it differs or for register-class
operands, echo form rejected. Typing-stage rule; parser stores name: ?[]const u8.
Locked with examples/1640-platform-asm-parse.sx (multi-output divmod: named
operands, register pins, clobbers — parses then bails, called from main).
Also files issue 0137 (pre-existing, orthogonal: `sx run` with no `main`
segfaults via an unguarded JIT entry lookup in target.zig — not an asm bug).
zig build test green (648 corpus, 445 unit).
Reword every 'foreign' comment to the extern/runtime-class vocabulary matching the
renamed identifiers (foreign call→extern call, foreign class→runtime class, foreign
path→runtime path, the #foreign-literal comment mentions → extern, etc.). Also fixes
two USER-FACING issues: the 'expected … #foreign … after type annotation' parse error
no longer advertises the removed keyword, and the Android 'no #jni_main' help
diagnostic now shows '#jni_class(…) extern' instead of the rejected '#foreign
#jni_class'. Removed the now-dead prefix-#foreign-vs-postfix conflict branch in
parseRuntimeClassDecl (the caller rejects #foreign before it runs).
src/ now contains 'foreign' ONLY in the hash_foreign token machinery + its 4
rejection messages — the deprecation mechanism (kept per the 9.0 recommendation; the
message MUST name #foreign to guide migration). Snapshot-neutral; suite green
(646 corpus / 444 unit, 0 failed).
The unary .not arm emitted bool_not (LLVM bitwise Not) for every
operand. Correct on i1; on an error binding — an error-set value, u32
tag at the LLVM level — a bitwise not of a nonzero tag stays nonzero,
so 'if !e' held even on a SET error and its branch read the
uninitialized success value (real segfault in the distribution repo's
sqlite tests). Plain integers had the same hole ('!7' was '~7').
Now: bool keeps bool_not; integers and error-set operands lower as the
truthiness complement (cmp_eq against a typed zero); anything else is
diagnosed instead of silently bit-flipped.
Regression: examples/1057 (set error: !e must not hold; success: !e
holds with a real value; integer truthiness) + examples/1171 (!"text"
diagnosed); both FAIL pre-fix. zig build test 426/426;
tests/run_examples.sh 600/600.
lowerEnumLiteral resolved the variant against the raw destination type,
so any non-enum destination fell into resolveVariantValue's silent
return-0 tail with the enum_init stamped as the wrong type:
- ?E destinations produced variant 0 mis-typed as the optional
(observed as variant 0 OR null, layout-dependent);
- builtin destinations (i64) silently became 0;
- unknown variants of real enums silently became variant 0;
- a destination-less literal panicked LLVM emission (unresolved
type reached codegen).
Now: optional destinations unwrap to the child enum (the coercion
layer's .optional_wrap handles E -> ?E), and the remaining shapes are
diagnosed — unknown variant (with the variant list, via the new
emitBadEnumVariant twin of emitBadVariant), non-enum destination, and
no destination (cascade-guarded: silent when the destination's type
already failed to resolve and was reported).
Regression tests: examples/0183 (return/assign/reassign into ?Enum,
non-zero variants, null path) + examples/1169/1170 (each diagnostic);
all three FAIL on pre-fix master. zig build test 426/426;
tests/run_examples.sh 598/598.
Surface rename of the signed integer family: s1..s64 become i1..i64
(u1..u64, usize, isize unchanged). 'string' keeps the s-prefix arm in
name classification; width parsing moves to the i-prefix arm next to
isize.
Internal TypeId tags follow the surface (.s8/.s16/.s32/.s64 ->
.i8/.i16/.i32/.i64), as do mono-key mangle fragments (ptr_i64,
tu_i64_bool) and all display/diagnostic formatting (i{d}).
Migrated in the same sweep: stdlib + examples + issue repros + FFI C
companions (shared symbol names like ffi_id_i64), expected
stdout/stderr/ir snapshots, specs.md, readme.md, CLAUDE.md/AGENTS.md,
implementation_plan.md, docs/, issue writeups. Vendored stb_image and
historical flow state left untouched.
zig build test: 426/426; examples suite: 595/595.
Two lowering sites materialized a local array as a whole LLVM value;
the legalizer scalarizes each such op into one SelectionDAG node per
element, and at ~64K elements the DAG combiner segfaults
(DAGCombiner::visitMERGE_VALUES → ReplaceAllUsesWith).
- lowerVarDecl: an array-typed `---` initializer emits NO store — the
slot stays uninitialized instead of receiving a whole-array undef
store. The tuple zero-init carve-out stays; non-array `---` keeps
the undef store. The interp is unchanged either way (slots start
.undef).
- lowerIndexExpr: element reads on an array with addressable storage
GEP the storage and load one element — the general-expression
sibling of 0110's lowerFor fix — without value-lowering the object
(a dead whole-array load would still reach the DAG). Storage-less
arrays keep the index_get fallback.
Sibling shape filed as 0125: any_to_string's per-array-type arms still
pass the array by value, so a 64K+ array type + any {} print crashes.
Regression: examples/0055-basic-large-stack-array.sx (sx build
segfaulted pre-fix). 22 .ir snapshots re-pinned: removed undef stores
and ig.tmp spills, in-place gep+load (instruction-shape-only churn,
reviewed).
A '*[N]T' receiver in an index expression reached LLVM emission with an
unresolved element type and tripped the panic sentinel — no read or
write spelling worked. ptrToArrayElem on Lowering recognises the shape;
the index READ path GEPs the pointee array through the pointer value
and loads the element; the write / compound-assign / lvalue /
addr-of-element paths and the expression typer resolve the element type
through the same helper (their GEP machinery already handled a pointer
base). Kept out of getElementType so slice paths don't half-accept a
raw pointer base.
Regression: examples/0176 (read, write, compound, element ptr + deref).
The globals registry (global_names) was last-wins across modules with no
per-importer gate: any module's bare K could read/write/type against an
unrelated module's same-named global (hash.sx's K table hijacked every
user K once std's namespace tail pulled hash into the program), and an
own const of an unsupported shape borrowed another module's const and
panicked at the unresolved-type tripwire.
- var_decl joins RawDeclRef: module globals are selectable raw authors.
- selectGlobalAuthor (the globals analogue of F2's selectModuleConst):
own author wins, one flat-visible author resolves, >=2 distinct flat
authors diagnose loudly, authored-but-not-visible diagnoses, and a
compiler-synthesized global (no raw author) emits untracked. A var_decl
author whose per-source registration was deduped at flat-merge (two
modules declaring the same extern symbol) serves the symbol's
registration.
- All bare-identifier global sites route through it: value read, addr-of,
assignment (store + compound), lvalue address, fn-ptr call, call param
typing, and expression type inference.
- selectModuleConst gains .own_opaque: an own const author with no
materialized per-source value (e.g. an array '::' const) blocks
borrowing another module's same-named const — the read diagnoses
cleanly instead of panicking.
- The fn-as-VALUE arm admits raw-facts-only authors: an own fn whose name
a flat-merge collision dropped from the global decl list (first-wins)
now resolves via author selection for func_ref/closure/Any shapes too.
Regressions: examples 0835 (own const vs flat array global), 0836 (main
const vs namespaced array global, incl. inference), 0837 (own array
const never borrows cross-module — clean unresolved).
Two coupled capabilities on the road to the std restructure
(current/PLAN-STDLIB.md, issue 0114):
1. alias.Type.method() / alias.Type as a call head, alias.CONST, and
alias.Enum.variant now resolve — previously only alias.fn() and
type-position alias.Type worked. objectIsValue treats an
alias-rooted field_access as a type head; the call path strips the
alias to the existing Type.method machinery; lowerFieldAccess
resolves alias.CONST pinned to the target module and alias.Enum.x
as a typed enum literal; resolveTypeWithBindings resolves qualified
type_exprs pinned to the target.
2. The carry rule: namespaceAliasTarget resolves an alias from the
file's own edges first, then from DIRECT flat imports (one level),
diagnosing two distinct carried targets as ambiguous. All qualified
shapes work through a carried alias — the std.sx namespace tail
(mem.GPA.init() etc.) is now expressible.
Regression: examples/0831-modules-namespace-alias-carry.sx (direct +
carried, all seven shapes).
checkIntLiteralFits range-checks a literal against its integer target
(builtins + custom widths via intLiteralRange; width-64 types skip —
every representable literal is a legal bit pattern there) and diagnoses
with the type's range and an xx/cast hint. Wired into the .int_literal
arm (covers decls, assignments, call args, struct-literal fields),
lowerStructConstant, and globalInitValue.
A negated literal now folds to a single constant so -128 range-checks
as -128 rather than as an out-of-range +128 intermediate. An explicit
xx operand skips the check — truncation stays available on request
(cast(T) was already exempt: its value arg lowers without the target).
examples/0300-closures-lambda.sx pinned 133 wrapping to -3 through an
s3 param — the exact class this outlaws; updated to a fitting value.
Found during the fix and filed separately: issue 0113 (negated-literal
global initializers rejected as non-constant; pre-existing).
Regressions: examples/1156-diagnostics-int-literal-out-of-range.sx,
examples/0174-types-int-literal-boundaries.sx.
xs[1..=3] (end inclusive), xs[0<..<4] (both exclusive), xs[..=2]
(prefix form with markers, implicit 0 start), xs[2<..] (open end,
exclusive start), and xs[..] (whole collection) — lowered as lo+1 /
hi+1 on the existing subslice op. Strings slice through the same path.
An explicit end marker requires an end expression, matching the
for-header rule.
Regression: examples/0052-basic-slice-range-bounds.sx.
lowerBreak/lowerContinue emitted a bare br, and the enclosing block's
emitBlockDefers — seeing the terminator — discarded the pending entries
on the assumption a return had already drained them. The breaking
iteration's defers were silently skipped, leaking whatever the cleanup
released.
Lowering.loop_defer_base records the defer-stack height at each loop's
body start (while / for / range-for, saved and restored alongside
break_target); break/continue drain non-onfail entries down to it in
LIFO order via the non-truncating emitLoopExitDefers before branching.
Truncation stays with the lexical block exits — the same entries still
belong to the fall-through path after the branch containing the break.
break/continue outside a loop now diagnose instead of no-op'ing.
Regression: examples/0049-basic-defer-break-continue.sx (for and while,
break and continue, nested-block LIFO drain).
Sweep all src/**.zig comments that cite resolved issues (issue NNNN /
fix-NNNN / KB-N): the invariant or mechanism each comment states is
kept; the historical citation is dropped, per the no-conclusion-comments
rule. Pure-history parentheticals are removed outright. References to
the 16 still-open issues (0030, 0041-0056) are untouched, as are test
NAMES carrying regression provenance (matching the sanctioned
"Regression (issue NNNN)" example-header convention).
Also removes the issues/0019-import-non-transitive-c-scope/ fixture dir
— the issue is superseded and its behavior is covered by
examples/0706-modules-import-non-transitive.sx (the .md writeup stays).
issues/0030's repro .sx stays: that issue is an open feature request.
Gate: zig build OK; zig build test 426/426; run_examples 541/0; zero
expected/ snapshot churn.
Review follow-up to the ARCH-B split (comment/import hygiene only, no
code changes):
- Section banners that travelled to the wrong file with the B1-B8 cuts
are reworded to describe the section that actually follows (e.g.
stmt.zig's trailing "Expression lowering", expr.zig's "Control flow"
before lowerChainedComparison) or deleted where nothing follows
(4 trailing-at-EOF banners). ffi.zig's facade note no longer claims
the IMP builders "stay here" (they live in lower/objc_class.zig);
protocol.zig's namespace-lookup banner now points at
pack.zig:resolvePackProjection for the orchestrator.
- lower.zig's two lower/expr.zig alias blocks (B8.1 + B8.2 appends)
merged into one.
- 448 unused header decls pruned from the 15 lower/*.zig files (each
had inherited lower.zig's full import block; pruned to fixpoint so
cascading type-extraction consts went too).
Gate: zig build OK; zig build test 426/426; run_examples 541/0; zero
expected/ snapshot churn.
Verbatim relocation of the 30-method expression cluster (struct/array/
tuple/enum/tagged-enum literals, init blocks, field access on values and
types, optional chains, numeric limits, indexing, slicing, deref, force
unwrap, null coalesce) into src/ir/lower/expr.zig — one contiguous
1,372-line cut. 30 aliases on Lowering keep all call sites unchanged.
Nested StructConstInfo stays on Lowering (field type of
struct_const_map), flipped pub and reached via an alias const, alongside
headNameOfCallee.
Gate: zig build OK; zig build test 426/426; run_examples 541/0; zero
expected/ snapshot churn.
