try foo() catch (e) { } // legal
try foo() catch e { } // parse error with a migration hint
Same capture style as the for-loop. All four catch shapes keep working
with the parenthesized binding — block, bare-expression body, and the
== match sugar — and the no-binding forms are unchanged. onfail follows
the same rule (onfail (e) { }); its expression-cleanup form is
disambiguated by the paren-group-before-brace lookahead, so
onfail (f()); stays an expression cleanup.
AST unchanged; the printer renders the parens; the #run escape help
text updated. Corpus migrated (57 catch + 3 onfail bindings, in-source
parser test strings, specs incl. grammar rules, readme untouched —
no catch examples there).
Regression: examples/1157-diagnostics-catch-binding-needs-parens.sx;
re-captured stderr for 1010/1013/1037/1123 (migrated source echoed in
carets + help text).
An alloca built at its use site re-executes on every pass through that
block, and LLVM reclaims allocas only at ret — so loop-body locals,
nested-loop index slots, and emitter spill temps (ig.tmp, sret slots, ABI
coercion temps, byval materialization) grew the stack per iteration and
long loops segfaulted on stack exhaustion.
New LLVMEmitter.buildEntryAlloca inserts after existing entry-block
allocas and restores the builder position; every LLVMBuildAlloca site
reachable during instruction emission now routes through it.
Initialization stores stay at the use site (per-iteration re-init is
unchanged), and entry slots become mem2reg-promotable. The 35 .ir
snapshot diffs are pure alloca position moves (type multisets verified
identical per file).
Regression: examples/0047-basic-loop-local-stack-reuse.sx (segfaulted
pre-fix on both the 1M-iteration body-local loop and the 3M-iteration
nested loop).
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— the issue is superseded and its behavior is covered by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.
`any_to_string` runs `type := type_of(val)`; for an `.any` operand
`type_of` lowers to `struct_get(val, 0)` to read the Any's tag. At
runtime a first-class Type value is the aggregate `{ tag=.any, value=tid }`
so the read succeeds, but the comptime interpreter stores a Type as a bare
`.type_tag(tid)` and the comptime `struct_get` arm had no case for it — it
raised `CannotEvalComptime`, which `runComptimeSideEffects` swallowed into
`void_val`, truncating the `#run` while still building with exit 0.
- interp.zig: comptime `struct_get` handles a `.type_tag(tid)` base by
mirroring the runtime Any-Type layout (field 0 -> `.any` tag, field 1 ->
the type id), so `type_of` of an Any-held Type evaluates as it does at
runtime and execution continues.
- emit_llvm.zig: `runComptimeSideEffects` no longer swallows a side-effect
bail; it prints a loud diagnostic and sets `comptime_failed`
(-> error.ComptimeError, non-zero exit), matching the const-init path.
A truncated `#run` can no longer ship a successful build.
Regression: examples/0613-comptime-print-any-type.sx (all five lines print,
exit 0). Resolves issue 0096.
Resolves issue 0090. The `{}` integer formatter mis-rendered both ends of
the 64-bit range:
- `int_to_string` computed the magnitude as `0 - n`, which overflows for
`s64::MIN` (its magnitude is unrepresentable as a positive s64) — the
value stayed negative, the digit loop ran zero times, so only `-`
printed. It now extracts digits straight from `n` (per-digit
`|n % 10|`, `n` truncating toward zero), never negating MIN.
- `any_to_string`'s `case int:` formatted every integer as s64, so a u64
all-ones value printed as `-1`. There was no `uint` type-category to
distinguish signedness. Added an additive `type_is_unsigned(T)`
reflection builtin (static fold + dynamic interp/LLVM paths, mirroring
`type_name`), backed by the new `TypeTable.isUnsignedInt` predicate, and
a `uint_to_string` formatter (unsigned decimal via long-division over
four 16-bit limbs). `case int:` routes through `type_is_unsigned(type)`.
The 16-bit-limb split is factored into a shared `decompose_u16x4`, now
reused by `int_to_hex_string` (no second unsigned-math routine).
Regression: examples/0046-basic-int-formatter-extremes pins both extremes
plus a width spread; unit tests cover `isUnsignedInt`. Docs (specs.md
representation note, readme std API) updated for unsigned/extreme `{}`
behavior. IR snapshots refreshed for the two new std functions.
The global-init constant serializers in emit_llvm.zig printed a diagnostic
on an unserializable value and then RETURNED an undef/null placeholder and
CONTINUED emitting. For a comptime `#run` global that yields a function
reference (`fp :: #run pick();` where pick returns a function), the build
fell through to the JIT and segfaulted calling through the undef pointer
(exit 134) — a silent miscompile dressed up as a printed error.
Route every genuine bail in the serialization family through a new
`failGlobalInit` helper: it sets `comptime_failed` (so core.generateCode
aborts with a non-zero exit after emit()) and returns an undef placeholder
that never ships, because the halt fires before object emission / JIT. This
covers the comptime func_ref leaf, the require_resolved aggregate func_ref
leaf, the top-level + vtable func_ref globals, the comptime-init catch, and
the remaining heap-walk / aggregate-shape bails. Unresolved-function
diagnostics now name the function instead of its (stdlib-unstable) IR index.
The require_resolved=false Pass-0 placeholder is unchanged (func_map is
empty until Pass 1; the aggregate is re-emitted with require_resolved=true).
Regression: examples/1128-diagnostics-comptime-global-funcref-rejected.sx —
a `#run` global returning a function ref now exits 1 with the diagnostic
(was: exit 134 segfault). Fail-before/pass-after verified.
A module-global initialized with an enum literal silently zero-initialized
to the first tag (`chosen : Color = .green` read back as `.red`), and an
enum tag inside a global array/struct was rejected as non-constant. The
constant serializer had no enum-literal arm.
Add `Lowering.constEnumLiteral`: serialize an enum literal to a
`ConstantValue.int` holding the variant's tag value, resolved against the
destination enum type and respecting explicit variant values; the global's
type drives the backing width at emit time. Wired into `globalInitValue`
(scalar global) and `constExprValue` (array element / struct field / nested
aggregate). A non-enum destination or unknown variant is diagnosed loudly,
never silently zero-initialized. The compiler-injected OS/ARCH globals now
serialize to their real `.unknown` tag (6 / 4); runtime reads are unchanged
(they resolve through comptime_constants), so only the static initializer in
the pinned .ir snapshots changes.
Remove the silent `func_ref => orelse LLVMConstNull` fallbacks in the LLVM
constant emitters: aggregate func_ref leaves carry a `require_resolved` flag
(transient null in Pass 0, loud diagnostic if still unresolved in the
Pass-1.5 re-emit), a top-level func_ref global is resolved in
initVtableGlobals, and the comptime (#run) path bails loudly instead of
emitting a null function pointer.
Regression: examples/0139-types-global-enum-literal-init.sx (scalar, array,
struct field, explicit-value enum u16 stride, struct-array with enum field);
negative: examples/1127-diagnostics-global-enum-literal-bad-variant.sx.
Mark issue 0082 RESOLVED.
A module-global aggregate initializer rejected a `null` literal in a
pointer (or optional-pointer) field as "must be initialized by a
compile-time constant". `Lowering.constExprValue` had no `.null_literal`
arm, so the null leaf returned no constant and the whole aggregate looked
non-constant — even though `null` is the compile-time zero pointer (a
top-level scalar `p : *s64 = null;` already serialized fine).
Add `.null_literal => .null_val` to constExprValue. While here, make the
two LLVM constant emitters exhaustive: emitConstAggregate and the
top-level init_val switch in emit_llvm.zig previously ended in a silent
`else => LLVMConstNull(...)` catch-all (the silent-arm class CLAUDE.md
mandates rooting out). They now handle every ConstantValue tag explicitly
(.null_val/.zeroinit -> all-zero constant, .undef -> LLVMGetUndef,
.func_ref resolved, nested .vtable is a hard @panic tripwire). The
reject-loud path for genuinely non-constant fields is preserved.
Regression: examples/0138 (array-of-struct null ptr fields, array of
all-null pointers, nested struct-in-struct null ptr) and the negative
examples/1126 (null ptr field beside a non-const field still errors).
Fail-before/pass-after verified.
A string `==`/`!=` used as an operand of a short-circuit `and`/`or` emitted
invalid LLVM (`PHI node entries do not match predecessors!`). String compares
expand into their own memcmp sub-CFG during LLVM emission, so the operand
finishes in a later basic block (`str.merge`) than the one the IR block
started in. `fixupPhiNodes` wired the short-circuit merge PHI's incoming edge
to `block_map[ir_block]` (the block the IR block started as), recording a
stale predecessor (`%entry`/`%and.rhs.0`).
Fix: record the builder's actual insertion block after emitting each IR
block's instructions (`term_block_map`, via `LLVMGetInsertBlock`) and use it
as the PHI predecessor. General — corrects the incoming block for any operand
that emitted intermediate basic blocks (string `==`, value `match`, …), not
just string `==`.
Regression: examples/0045-basic-string-eq-short-circuit.sx (string `==` on
both sides of `and` and of `or`, plus a match-value + enum-payload `==` shape).
Fails (LLVM abort) pre-fix, passes after.
The `type_name` / `type_eq` reflection builtins resolved their Type arg's IR
type via `getRefIRType(...) orelse TypeId.s64`, then gated `== .any`. A failed
must-succeed lookup silently became `.s64` (`!= .any`), classifying a boxed
`Any` arg as bare i64 and reading the wrong value with no diagnostic.
Add the sibling classifier `LLVMEmitter.reflectArgRepr`, which routes the
lookup through `argIRTypeOrFail` (the issue-0074 `.unresolved` resolver) and
returns `{ boxed, bare, unresolved }`. The three emit sites in ops.zig
(`type_name` + `type_eq` x2) now switch on it: `.boxed` extracts the Any value
field, `.bare` uses the value directly, `.unresolved` hits a hard `@panic`
tripwire — never silently treated as bare. Real args always resolve, so the
happy path is byte-identical (suite stays 361/0, zero snapshot churn).
Secondary `lower.zig` `null_literal`/`undef_literal => target_type orelse .void`
confirmed intentional (typeless-literal default deliberately handled by
emitConstNull/emitConstUndef as null-ptr / undef-i64) — left with an invariant
comment, not the `.unresolved` tripwire.
Regression test in emit_llvm.test.zig asserts the loud path: fail-before with
`orelse .s64` yields `.bare`; pass-after yields `.unresolved`.
Four FFI call-arg lowering sites resolved an argument's IR type via
`getRefIRType(arg_ref) orelse .void` — a silent fallback to the load-bearing
real type `.void`. A failed lookup there is a codegen invariant violation, but
`.void` is treated by downstream `toLLVMType` → `abiCoerceParamType` →
`coerceArg` as a legitimate void-typed foreign argument, corrupting the call
ABI with no diagnostic.
Add one shared resolver `LLVMEmitter.argIRTypeOrFail` that returns the
dedicated `.unresolved` sentinel on a failed lookup — never `.void`/`.s64` — so
the failure cannot masquerade as a real type and trips `toLLVMType`'s existing
hard `@panic` tripwire at the call site. Route all four sites through it:
- src/ir/emit_llvm.zig JNI constructor (NewObject) arg loop
- src/backend/llvm/ops.zig objc_msgSend arg loop
- src/backend/llvm/ops.zig JNI non-virtual call arg loop
- src/backend/llvm/ops.zig JNI Call<Type>Method arg loop
Happy path is byte-identical (every real arg already has a resolved type); FFI
examples stay green with zero snapshot churn.
Regression test (fail-before/pass-after) in src/ir/emit_llvm.test.zig asserts an
unresolvable FFI arg ref now yields `.unresolved`, not the old silent `.void`.
Move the final inline emitInst handler groups (terminators, box/unbox-Any,
reflection, switch-branch, closure-creation, vector, block-param, misc) into
the Ops facade in src/backend/llvm/ops.zig. emitInst is now pure dispatch:
every arm delegates to self.ops().*, leaving only setInstDebugLocation plus
one-line delegations.
Widen the shared infra the moved bodies reach (emitFailableMainRet, getBlock,
anyTag, isSignedTypeEx, coerceToI64/coerceToI64Signed/coerceFromI64,
emitFieldValueGet) to pub on LLVMEmitter; helper and ref-tracking sections
stay put. Pure relocation: emitted LLVM IR byte-identical, zero snapshot churn.
Relocate the struct, enum, union, array/slice, tuple, and optional
opcode handler bodies out of emitInst into the existing Ops facade.
Each moved arm now delegates via self.ops().emit<Op>(...); shared infra
stays on LLVMEmitter, with resolveAggregate/resolveGepStructType widened
to pub as the GEP handlers require. Pure relocation, behavior-preserving:
zero snapshot churn (361/0).
Relocate the Calls (objc_msg_send / jni_msg_send / call / call_indirect)
and Call-extensions (call_builtin / compiler_call / call_closure) emitInst
handler groups out of emit_llvm.zig into the existing Ops facade. Each
emitInst arm now delegates via self.ops().emit<Op>(...). Behavior-preserving
pure relocation; emitted LLVM IR is byte-identical (361/0 examples, no
snapshot churn).
Shared call infra stays on LLVMEmitter, widened pub only as the moved
bodies require: extractSlicePtr, loadJniFn, getObjcMsgSendValue, the math
F32/F64 declarators + types, getOrDeclareWrite/getWriteType, ffiCtors,
materializeByvalArg, emitCStringGlobal, emitJniConstructor, and the Jni
slot-offset constants. emitJniConstructor remains in emit_llvm.zig (A7.3
decision); the moved jni arm calls it via self.e.emitJniConstructor(...).
Relocate the `// ── Memory ──`, `// ── Globals ──`, `// ── Conversions ──`,
and `// ── Pointer ops ──` opcode handler bodies out of `emitInst` in
src/ir/emit_llvm.zig into the existing `Ops` facade in
src/backend/llvm/ops.zig. Each `emitInst` arm now delegates via
`self.ops().emit<Op>(...)`. Widen `emitConversion`, `coerceArg`, and
`getRefIRType` to `pub` (the only helpers the moved bodies call).
Pure relocation: zero snapshot churn.
Move the Constants/Arithmetic/Bitwise/Comparisons/Logical opcode handler
bodies out of emitInst into a new Ops facade in src/backend/llvm/ops.zig.
emitInst's scalar arms now delegate via self.ops().*; the shared infra they
call (mapRef/resolveRef/matchBinOpTypes/emitCmp/emitCmpOrdered/emitStrCmp/
emitStringConstant/reflection + isFloatOrVecFloat/isSignedType) stays on
LLVMEmitter, widened to pub as needed. Pure relocation: zero snapshot churn.
Move getOrCreateJniSlots (the cls/methodid slot-cache builder) out of
emit_llvm.zig into the FfiCtors backend *LLVMEmitter facade. Behavior-preserving
— self.* -> self.e.* only.
- FfiCtors gains getOrCreateJniSlots (pub). The jni_slots cache + mangleJniKey
stay on LLVMEmitter; mangleJniKey is widened to pub (the facade calls it back,
like lazyDeclareCRuntime/emitPrivateCString), and JniSlotPair is widened to pub
(the facade returns it; the call site consumes it). 1 call site routed via
ffiCtors().
- emitJniConstructor intentionally NOT moved in this slice: it is emission-heavy
(resolveRef/mapRef/coerceArg/getRefIRType/extractSlicePtr/loadJniFn/
emitCStringGlobal — 100+ internal callers for the first two), so relocating it
would pub-expose the emitter's core value-emission machinery. Consistent with
A7.2 keeping emitFieldValueGet in emit_llvm.zig. Pending an explicit decision.
Gate: zig build, zig build test, bash tests/run_examples.sh -> 361/0
(JNI anchors 1402/1408/1418/1425 green, no churn).
Move the DWARF debug-info emission out of emit_llvm.zig into a DebugInfo backend
*LLVMEmitter facade (field `e`). Behavior-preserving relocation — self.* ->
self.e.* only.
- src/backend/llvm/debug.zig (DebugInfo): debugEnabled + diFileFor (private) +
initDebugInfo / beginFunctionDebug / endFunctionDebug / setInstDebugLocation /
finalizeDebugInfo (pub). The mutable DI state (di_builder/di_cu/di_files/
di_scope/current_func_file) + the shared source map (import_sources/main_file)
stay on LLVMEmitter; the facade reads/writes them via self.e.*.
- Routed the 5 pass-order call sites in LLVMEmitter.emit (init/finalize/
begin/end/setInstDebugLocation) through a new debugInfo() accessor.
- setDebugContext stays on LLVMEmitter (shared-state setter; callers in main.zig/
core.zig/test). sourceForFile stays on LLVMEmitter and is widened to pub — it is
shared with reflection's trace-frame emission (emitTraceFrame), not debug-only.
- No DI logic / module-flag / DWARF-version / scope-line change.
Gate: zig build, zig build test, bash tests/run_examples.sh -> 361/0 (no churn).
Move the LLVM type-mapping and C-ABI coercion helpers out of emit_llvm.zig into
the first src/backend/llvm/ modules. Behavior-preserving relocation — the only
rewrites are module plumbing and self.* -> self.e.* facade access.
- src/backend/llvm/types.zig (TypeLowering): toLLVMType + toLLVMTypeInfo.
- src/backend/llvm/abi.zig (AbiLowering): abiCoerceParamType / abiCoerceParamTypeEx
/ needsByval / materializeByvalArg.
- Both are backend *LLVMEmitter facades (field `e`) — the backend analogue of the
IR-side *Lowering facades, NOT a *Lowering facade. They reach the cached LLVM
handles, IR type table, module data layout, builder, and the memoizing
composite-type getters via self.e.*.
- LLVMEmitter stays the facade: toLLVMType (~97 callers) + abiCoerceParamType /
abiCoerceParamTypeEx / needsByval / materializeByvalArg kept as thin wrappers
delegating through new typeLowering()/abiLowering() accessors. Zero caller
churn. toLLVMTypeInfo deleted (sole caller moved).
- Widened getStringStructType / getAnyStructType / getClosureStructType to pub
(the moved toLLVMTypeInfo calls them back; their memoization stays on
LLVMEmitter). verifySizes stays in emit_llvm.zig (size-assertion pass, not type/
ABI lowering). No ABI/type logic, branch order, diagnostic text, or snapshot
changed. Circular import (emit_llvm <-> backend/llvm) resolves via the pointer
facade.
Gate: zig build, zig build test, bash tests/run_examples.sh -> 361/0
(1202 .ir + the 2 ABI unit tests unchanged, no churn).
Test-first scaffolding for LLVM backend modularization (Phase A7.1) before the
type/ABI helpers move into src/backend/llvm/{types,abi}.zig. Visibility-only
change to the targets — no behavior change. Closes the ARCH-SAFETY "no generic
ABI snapshot" gap.
- 2 new emit_llvm.test.zig tests:
- abiCoerceParamType across every C-ABI size bucket: <=8 -> i64, 9-16 ->
[2 x i64], >16 -> ptr, HFA (all-float/all-double, <=4 fields) -> unchanged,
string -> ptr, slice -> ptr, scalar -> unchanged. Built via a local
internStruct helper (field slice in the module arena -> no testing-allocator
leak); asserts against emitter.cached_* + LLVMArrayType2.
- needsByval: true only for >16-byte non-HFA struct; false for <=16 / HFA /
string / slice / non-struct.
- 1 new .ir snapshot: 1202-ffi-cc-c-large-aggregate (the canonical callconv(.c)
>16-byte byval example that directly documents abiCoerceParamType) — pins the
byval param path end-to-end (5 byval + entry reload + 2 sret from Arena.init).
Path-free + idempotent (verified across two captures). Suite count unchanged
(snapshot added to an existing example).
- Widened abiCoerceParamType + needsByval to pub (visibility only;
abiCoerceParamTypeEx/materializeByvalArg/verifySizes stay private — move with
callers in sub-step 2). No logic touched.
- Recorded the A7.1 coverage inventory + residual gaps (wasm32 usize->i32 branch,
fn-ptr large-aggregate 1203/1204) in ARCH-SAFETY.md.
Gate: zig build, zig build test, bash tests/run_examples.sh -> 361/0 (no churn
beyond the new 1202 .ir).
A bare failable `#run` (no catch/or) whose error escapes used to segfault (const
form `x :: #run f()`) or silently succeed (statement form `#run f();`). Now the
compiler reports the raised tag name + the resolved return trace at the #run site
and halts with a non-zero exit.
- lower.zig: a failable #run's comptime function returns the full failable tuple
(so the error slot is inspectable) while the global is typed as the success
value; failable side-effects return the tuple instead of void.
- emit_llvm.zig: read the always-on comptime trace buffer (extern sx_trace_*);
comptimeErrChannel + checkComptimeFailable split the result (non-zero tag →
reportComptimeEscape + comptime_failed flag; success → value part). Wired into
emitGlobals (const) and runComptimeSideEffects (statement, now filtered by the
__run name; buffer cleared before each eval).
- core.zig: generateCode returns error.ComptimeError when comptime_failed, so the
driver aborts before JIT/link.
catch / or / onfail compose at comptime exactly as at runtime; a successful bare
#run yields the value. Regressions: examples/1037-errors-comptime-run-escape
(diagnostic, exit 1) + 1038-errors-comptime-run-handled (exit 164). Suite: 326.
A source path with no directory component (`sx build main.sx` from the
project dir — what the chess app does) made `diFileFor` emit a `DIFile`
with an empty `directory:`, so the compile unit's `DW_AT_comp_dir` was
"". Apple's ld then silently drops the *entire* object's debug map (0
N_OSO) and the binary is undebuggable — lldb resolves no sx source.
Builds whose path had any directory (`.sx-tmp/x.sx`, `examples/x.sx`)
were unaffected, which is why small repros + the stepping smoke passed
and only the bundled chess app hit it.
Fix: diFileFor falls back to "." (and "/" for a root-level file) when
the path has no directory component, so comp_dir is never empty.
Verified: chess (`sx build --target macos --emit-obj main.sx`) now
links with OSO=1 and lldb resolves `frame at main.sx:82:8`. Regression
guard added to the DWARF unit test (asserts `DIFile(... directory: ".")`
for a bare filename). Gates: zig build, zig build test, run_examples.sh
-> 291 passed, debug-stepping smoke ok.
Each trace frame now shows the offending source line with a `^` caret
under the column — in the catch-handler formatter, the failable-main C
reporter, and the comptime path.
The source line is embedded at compile time as a 5th Frame field
(line_text), not read from disk at runtime: the file field is a
basename and a runtime read would add a filesystem dependency that
fails under the test harness and on locked-down targets.
- errors.lineAt(src, offset): shared helper for the whole source line.
- Frame gains line_text (mirrored in emit_llvm getFrameStructType,
trace.sx Frame, sx_trace.c SxFrame). emitTraceFrame embeds it; the
interp .trace_resolve extracts it from the source map.
- trace.sx (new spaces helper) and the C reporter render the line +
a col-aligned caret, guarded on a non-empty line_text.
Snapshots 243/244/247/253 regenerated. Gates: zig build, zig build
test, run_examples.sh -> 291 passed.
#run failures now print the same `func at file:line:col` trace as
runtime, resolved in-process via the interpreter's IR/source tables.
- Read-side context-split op `.trace_resolve` (mirror of .trace_frame),
lowered from a name-recognized `__trace_resolve_frame(u64) -> Frame`.
- emit_llvm: inttoptr the operand to *Frame + load (the value
.trace_frame stamped in).
- interp: unpack (func_id << 32 | span.start); resolve func/file from
module.functions and line/col via SourceLoc.compute over a new
source_map (setSourceMap wired at every production interp site).
- trace.sx: frame_at -> u64; to_string routes each frame through
__trace_resolve_frame, so one source works in both machines.
Compiled path behavior unchanged (243/244/247 identical; it now loads
via the op). New examples/253-comptime-trace.sx exercises the comptime
path. Gates: zig build, zig build test, run_examples.sh -> 291 passed.
Return-trace frames now resolve to real `func at file:line:col`
in-process — no DWARF, no symbolizer.
- New niladic, span-stamped `.trace_frame` IR op (mirrors is_comptime):
carries no operands; each backend derives the frame from context.
lower.zig's placeholderTraceFrame emits it; the existing
sx_trace_push call consumes it.
- emit_llvm: resolve the op's span + current function to
{file(basename), line, col, func}, build an interned Frame global
({string,i32,i32,string}, strings cached by content), push its
address (ptrtoint).
- interp: pack (func_id << 32 | span.start) for the comptime resolver
(slice 3b); never a pointer.
- sx_trace.c report_unhandled derefs SxFrame; trace.sx gains the Frame
struct, frame_at -> *Frame, and field-reading to_string. Layout
mirrored in 3 places with cross-ref comments.
Verified JIT + AOT. Snapshots 243/244/247 regenerated (placeholder ->
func at file:line:col). Gates: zig build, zig build test,
run_examples.sh -> 290 passed.
Attach LLVM debug metadata so a captured return-address PC resolves to
file:line:col (the runtime half E3.3 needs) and sx binaries become
debuggable in lldb/gdb.
- llvm_api.zig: bind llvm-c/DebugInfo.h (DIBuilder C API was unbound).
- emit_llvm.zig: DIBuilder + one DICompileUnit/DIFile on the main file,
a DISubprogram per function (LLVMSetSubprogram), and a DILocation per
instruction from Inst.span (errors.SourceLoc.compute, scoped to the
subprogram). Plus the "Debug Info Version"/"Dwarf Version" module
flags and LLVMDIBuilderFinalize.
- Gated on opt none/less + a wired source map (setDebugContext from
core.zig), mirroring lower.zig's tracesEnabled; release strips it.
Verified: sx ir/sx asm --opt none show correct DILocations + .loc
directives; the 290-example JIT suite (-O0 -> debug on) verifies and
runs unchanged. +2 DWARF unit tests.
The last E4 item: a comptime call-frame dump.
- New nullary `interp_print_frames` IR op (inst/print). The interpreter
maintains a `call_chain` side-stack (push/pop a FuncId around each sx-bodied
`call`, freed in deinit) and `printInterpFrames` appends the chain to its
output — most-recent-last, with the dump frame itself skipped. emit_llvm
makes the op a no-op: compiled code has no interpreter stack, and the only
caller is `process.exit`'s dead `is_comptime()` branch.
- Lowered from a name-recognized `__interp_print_frames()` builtin
(tryLowerReflectionCall + inferExprType → void).
- `trace.print_interpreter_frames()` wraps the builtin; wired into
`process.exit`'s comptime branch (process.sx now imports trace.sx).
- Frame source locations await IR-offset resolution (the comptime analog of
DWARF), so only function names print today.
examples/252-interp-frames.sx (top-level `#run` drives the dump; exit 0).
Phase E4 (entry-point + stdlib error story) is now 100% complete.
Stdlib slice of Phase E4, plus the noreturn codegen fix that enables it.
noreturn codegen (the enabling bug): E1.4c made `noreturn` type-system-only;
this is its first backend consumer and it crashed LLVM verification. Fixed:
- lower.zig: a `-> noreturn` body lowers as statements ending in `unreachable`
(ensureTerminator emits unreachable; the two body-lowering sites no longer
treat the last expr as a `ret`).
- emit_llvm.zig: a `void`/`noreturn` call result stays unnamed (direct +
foreign call sites) — LLVM rejects a named void value.
- finishCatchHandler: a `noreturn` value-carrying catch body (which is not an
IR terminator) closes the handler with `unreachable` instead of feeding a
bad value into the merge phi. Shared by lowerCatch + lowerCatchOverChain.
is_comptime(): new nullary `.is_comptime` IR op (inst/print/interp/emit_llvm) —
interp evaluates true, emit_llvm emits constant false, so `if is_comptime()`
dead-codes out of compiled binaries. Recognized by name in
tryLowerReflectionCall + inferExprType (no std.sx decl, which would emit a
spurious `declare @is_comptime` into every module).
library/modules/log.sx: warn/info/debug/err — interpolate like print, write
`LEVEL: <msg>` to stderr. (`error` is reserved → the level is `log.err`.)
process.exit(code) -> noreturn + assert(cond, msg) in process.sx. `exit` is
POSIX `_exit(2)` (immediate, no cleanup; sx print is unbuffered so nothing is
lost), bound to "_exit" which also avoids a link-level clash with the sx `exit`
function's own name.
examples 248 (exit 0), 249 (exit 42), 250 (exit 1). #caller_location, the
comptime-exit diagnostic flush, and trace.print_interpreter_frames deferred to
E4.1b.
Extends the failable-main entry-point wrapper to a value-carrying main.
`main :: () -> (int, !)` now exits the integer value on success (truncated
to u8, like a plain integer main) and reports the header + trace to stderr
+ exits 1 on an escaping error (same reporter as the pure `-> !` form).
- lower.zig validateMainSignature: accept a 2-field `{int, error_set}`
tuple return (set needs_trace_runtime) instead of rejecting it. Multi-
value `-> (T1, T2, !)` and non-integer value slots still reject — there's
no single integer exit code to map them to (sharpened diagnostic).
- emit_llvm.zig: the `.ret` arm detects a value-carrying main (tuple ending
in `.error_set`) and extracts `{value, tag}` (extractvalue 0/1) before
calling emitFailableMainRet, now generalized to take an optional `value`
(null → pure `-> !`, success exits 0; present → success exits the value).
C reporter unchanged.
All E4.2 entry-point shapes (void / int / `-> !` / `-> (int, !)`) now done.
examples/245-failable-main-value.sx (exit 64); 239 comment refreshed.
A pure-failable `main` (`-> !` / `-> !Named`) that lets an error reach the
function boundary now exits 1 and prints `error: unhandled error reached
main: error.<tag>` + the return trace to stderr, instead of returning the
raw tag id truncated as the exit code with no diagnostic. Success exits 0;
a `catch`-absorbed error exits 0 (buffer cleared).
Codegen wrapper so JIT and AOT behave identically (no host-side special-
casing):
- emit_llvm.zig: the `.ret` arm detects a failable main and routes to
new `emitFailableMainRet` — `icmp ne tag, 0` → success block `ret i32 0`
/ error block GEPs the tag name out of the always-linked tag-name table,
calls `sx_trace_report_unhandled`, `ret i32 1`. main's bare-u32 returns
(success `ret(0)` + each raise's `ret(tag)`) all funnel through it.
- sx_trace.c: new `sx_trace_report_unhandled(tag, name, name_len)` prints
the header + surviving frames to stderr (placeholder frame format mirrors
trace.sx until DWARF/E3.0). Lives next to the buffer it reads.
- lower.zig validateMainSignature: the pure-failable arm sets
needs_trace_runtime so the AOT path auto-links sx_trace.c even when the
body emits no other push/clear.
Value-carrying `-> (T, !)` main stays gate-rejected (multi-slot wrapper is
a separate slice). examples/244-failable-main.sx.
`{}` on an error-set value printed `<?>` (any_to_string had no error_set
category). Now it renders the tag name (`BadDigit`), reusing the existing
any_to_string dispatch.
Pieces:
- New `error_tag_name_get` IR op (UnaryOp): tag id -> name. Lowered from a new
`error_tag_name(e) -> string #builtin` (std.sx). Handled across inst.zig
(op def), print.zig, interp.zig (comptime: tags.getName), and emit_llvm.zig.
- emit_llvm getOrBuildTagNameArray: an always-linked `[N x {ptr,i64}]` global
of tag names indexed by global tag id (the TagRegistry namespace, slot 0 =
""). error_tag_name_get zext's the u32 tag value and GEPs into it. Built once;
not trace-gated, so it works in release too (per the spec's "tag-name table
always shipped").
- resolveTypeCategoryTags gains an `error_set` category so the
`case error_set:` arm in any_to_string matches; that arm coerces the Any to
u32 (`xx val`) and calls error_tag_name. (cast(type) didn't recover the tag
id for error-set values; the u32 coercion does.)
examples/240-error-tag-interpolation.sx: bound tags + a catch-bound tag print
their names. Regenerated ffi-objc-call-06-sret-return.ir — pure block-renumber
drift from adding one if-arm to the shared any_to_string (verified
semantically identical after collapsing block numbers).
Gates: zig build, zig build test, bash tests/run_examples.sh (277 passed; lone
failure is the user's uncommitted 213-canonical-map pack WIP).
First sema/types step. Implemented in the IR layer (ir/types.zig +
type_bridge.zig + lower.zig), NOT src/sema.zig — lowering doesn't consume
sema; the frontend Type is LSP-only. Mirrors how enums are handled.
- ir/types.zig: new `.error_set` TypeInfo kind (ErrorSetInfo {name, tags:
[]u32}; identity = name, like enum) with a u32 runtime layout (size/align
4, LLVM i32) per the locked error-slot ABI. New TagRegistry on TypeTable
(global tag pool: name -> u32, monotonic, id 0 reserved for "no error").
internTag/getTagName/errorSetType helpers; `.error_set` arms in all 7
exhaustive switches + findByName.
- emit_llvm: toLLVMTypeInfo -> i32. print: writeType -> set name.
- type_bridge: resolveInlineErrorSet (mirrors resolveInlineUnion) +
.error_set_decl arm.
- lower.zig: registerErrorSetDecl (rejects empty `error { }` with a
diagnostic) wired into both top-level decl switches + the block-local one.
- tests: ir/types.test (TagRegistry 0-reserved + identity; errorSetType u32
layout + named display + dedup; sorted storage) and ir/type_bridge.test
(decl -> type + tag interning + re-resolve dedup).
End-to-end: `Foo :: error { A, B }` + main compiles + runs (exit 0) — first
ERR syntax to survive the full pipeline; empty set rejects with a diagnostic.
Inferred bare `!`, error.X value, and == typing deferred to slice 2 / E1.2.
zig build, zig build test, and 254/254 examples green.
An unannotated param resolving to a plausible .s64 was the classic
silent-default trap (root of the 2.5 multi-param-closure bug). Replace it
with a dedicated TypeId.unresolved at slot 0, so a zero-initialised or
forgotten TypeId trips the sentinel instead of masquerading as a real type.
- types.zig: TypeId.unresolved = 0 (void moves to 17); TypeInfo.unresolved;
sizeOf/toLLVMType @panic on it (codegen tripwire); hash/eql/printer cover it.
- type_bridge: inferred_type => .unresolved (was .s64).
- resolveParamType: emit "parameter 'x' has no type annotation" for a
genuinely-unannotated value param (comptime/variadic/pack params exempt --
they resolve via per-call substitution).
- lowerLambda: resolve unannotated params from the target closure signature;
otherwise emit "cannot infer type of lambda parameter".
- CLAUDE.md: .void documented as an UNACCEPTABLE failed-type sentinel (it
conflates with a real, heavily-checked type); prescribe a distinct
.unresolved-style value + codegen tripwire.
Snapshot churn: one .ir (ffi-objc-call-06) -- the runtime type-name table and
typeof match arms renumber by the new builtin slot; program output unchanged.
Add a `pack` variant to IR `TypeInfo` — an ordered, interned sequence of
per-position element types (`PackInfo { elements: []const TypeId }`) — with
constructor (`packType`), structural equality + hashing, and a `pack(T0, …)`
printer. A pack is comptime-only: it lowers to flat positional args before
codegen and has no runtime layout, so `sizeOf` and `toLLVMType` bail loudly
rather than inventing a size. 5 unit tests (N=0/1/3, dedup, order/arity
distinctness, distinct-from-tuple, printer).
Also: give TypeTable an arena for the slices its constructors dupe (freed at
deinit), and add the missing `usize`/`isize` arms to `sizeOf` (a latent
non-exhaustive switch) so types.test.zig compiles and runs leak-free.
Previously, `t : Type = f64` stored a boxed string carrying the literal
name "f64"; comparisons and `type_of`/`type_name` round-trips lost the
underlying TypeId. This switches `Type` to a runtime-representable Any
pair: `{ tag = .any.index() (meta-marker), value = TypeId.index() }`.
Mechanism:
- `const_type` emits a 16-byte Any aggregate via insertvalue.
- `TypeId.any` advertises 16 bytes / 8-byte alignment so structs that
embed `t: Type` size correctly under verifySizes.
- `lowerBinaryOp` folds `==`/`!=` between static type-refs to a
`const_bool`, and decomposes runtime Any-vs-Any compares via
`unbox_any` so LLVM doesn't see icmp on aggregates.
- `lowerMatch`'s `is_type_match` path unboxes Any-typed subjects to
the i64 type tag before the switch, so `case type:` etc. fire.
- `lowerRuntimeDispatchCall` (used by `case T: ... cast(t) val`) does
the same unbox for the type-tag arg.
- `type_of(val: Any)` rebuilds an Any with `{.any, tag_of(val)}` so
the result is itself a `Type` value, not a bare i64.
- `buildPackSliceValue` stops re-boxing const_type — the value is
already canonical Any.
- `__sx_type_names` now indexes by TypeId across the whole table
using the new `types.formatTypeName` (structural names for `*T`,
`[]T`, `[N]T`, `?T`, `Vector(N,T)`, function/closure/tuple) so
runtime `type_name(t)` works for compound types.
- `interp.zig`'s comptime `type_name` accepts either the bare
`.type_tag` Value or the Any-boxed aggregate it now sees.
- `scanDecls` registers `Vec4 :: Vector(4, f32)` style aliases in
`type_alias_map` (before the `fn_ast_map` check; `Vector` IS a
`#builtin` fn). Lets `Vec4` in expression position lower as
`const_type(<vector tid>)`.
- `isStaticTypeArg` becomes scope-aware: a name shadowed by a runtime
local is not static. `isStaticTypeRef` is the symmetric helper for
the eq fold.
- `inferExprType` returns `.any` for bare type names (identifier and
type_expr) so pack arg types are correct.
Side effect: `print("{}", Vec4)` now prints the structural name
`Vector(4,f32)` rather than the alias literal `Vec4` — 12-meta's
expectation updated. Aliases stay pointer-equal to their target
(`Vec4 == Vector(4, f32)` is true).
Tests:
- examples/189-type-all-interactions.sx: 12-section comprehensive
coverage — literal `==`, `type_of(value) == T`, `Type` var storage,
`type_name` (static + runtime), printing Type values, generic
dispatch via `$T: Type`, `identity($T, val)`, `Wrap($T)`, reflection
builtins (`size_of`, `align_of`, `field_count`, `type_eq`),
`..$args` pack walking, `Type` in struct field, compound type
literals (`*Point`, `[4]s32`, `[]bool`, `?f64`).
- examples/12-meta.sx: expected output updated to reflect structural
name for the Vec4 alias path.
- ffi-objc-call-06-sret-return.ir: regenerated to absorb the new
type-name strings now emitted globally.
223/223 examples pass.
`abiCoerceParamType` had a libc-friendly heuristic: sx `string` /
`[]T` slice → `ptr` (drop the len, just pass the start pointer).
The heuristic is right for `#foreign` decls that mirror libc
signatures (`puts(const char *)`, `strlen(const char *)`); it's
wrong for sx-internal `callconv(.c)` (e.g. block trampolines) where
both sides see and exchange the full slice.
Split via a new `abiCoerceParamTypeEx(ir_ty, llvm_ty,
is_foreign_c_api)`. The old single-arg form forwards with
`is_foreign_c_api = true` so every call site that already collapses
keeps doing so. The function-decl emit at lines 1442 / 1454 now
passes `func.is_extern` — sx-internal `callconv(.c)` declarations
take the false path and preserve the slice as `{ptr, i64}` →
`[2 x i64]` via the general struct-coerce branch (true C ABI for
a 16-byte aggregate: passed in x0+x1 on AArch64).
`examples/188-block-string-arg.sx` flips green ("got: <hello>");
suite stays at 222/222. Foreign-decl call sites
(objc msg_send / JNI / direct extern calls) keep the libc
collapse — they pass `is_foreign_c_api = true` via the legacy
`abiCoerceParamType` shim.
`#run` / post-link callback `print` output was reaching stderr via
`std.debug.print` flushes from three sites. The runtime JIT path
already writes to fd 1 (stdout) directly. Anyone redirecting one
stream saw the two halves disappear in different places.
Switches all three flush sites + the `--- build done ---` delimiter
in main.zig to `std.c.write(1, ...)` so build-time and runtime
prints share the stream the user wrote them against (they typed
the same `print(...)` at both call sites — there's no reason for
them to land on different streams). Test runner uses `2>&1` so
snapshots are unaffected; suite stays at 218/218.
Closes issue-0047.
Fix for the silent .s64 fall-through in `type_name(<dynamic-arg>)`.
`tryLowerReflectionCall` now splits on `isStaticTypeArg(node)`:
- Static (type_expr / identifier / pack_index_type_expr / pointer
/ array / slice / optional / many_pointer / function_type_expr
/ tuple_literal / call) → fold to const_string at lower time
(today's fast path).
- Dynamic (index_expr, field_access, runtime locals, anything
else) → emit `callBuiltin(.type_name, [arg_ref])`. The interp's
arm (commit 9600ba5) reads the runtime `.type_tag` Value and
returns the per-position name.
`isStaticTypeArg(node)` is a new helper mirroring the explicit
arms of `resolveTypeArg`. Lives alongside resolveTypeArg in
lower.zig; documented to track shape changes together.
emit_llvm: the comptime reflection builtins (`type_name`,
`type_eq`, `has_impl`) now emit a silent undef-i64 placeholder.
Same reasoning as 4A.bare.1.B's relaxation of const_type's
emit_llvm arm: the JIT compiles the containing fn module-wide
even if main never calls it, so emit-time noise here is just
dead-from-main's-perspective code. Real misuse — passing a non-
Type value to one of these — is caught by the interp arm's
`asTypeId orelse bailDetail`.
`examples/171-pack-dynamic-type-name.sx` flips from "s64s64"
(silent .s64 fold per element) to "s64string" (per-position
correct via interp arm). Test runs `walk(42, "hi")` at `#run`
time so the dynamic path executes in the interp.
211/211 example tests + zig build test green.
Step 4A final-slice fix. Bare `$<pack_name>` (no `[<int>]`)
in expression position now parses + lowers to a comptime
`[]Type` slice value carrying one `const_type(TypeId)` per
pack element.
Plumbing:
- src/ast.zig: new `ComptimePackRef { pack_name }` node +
`comptime_pack_ref` variant in Data.
- src/parser.zig: `parsePrimary`'s `$` arm makes `[` optional
after the pack name. With `[<int>]` → existing
`pack_index_type_expr` (single Type value). Without → new
`comptime_pack_ref` (whole pack as []Type).
- src/sema.zig: adds the no-op switch arms for the new node
in `analyzeNode` and `findNodeAtOffset`.
- src/ir/lower.zig: `lowerExpr` arm reads `pack_arg_types[name]`
and calls `buildPackSliceValue(arg_tys)`. The helper allocas
a `[N x Any]` array, emits one `const_type(arg_tys[i])` per
slot, then a slice `{data_ptr, len}` aggregate. No active
binding → focused diagnostic + null slice placeholder. The
IR slice element type is `Any` (matches the today's
`Type → .any` mapping in type_bridge); the interp stores
raw `.type_tag` Values directly (NOT Any-boxed) so
`args[i]` at interp time reads a Type value.
- src/ir/emit_llvm.zig: relaxed `const_type` to silently emit
undef-i64 instead of the previous stderr-noisy bail. Storage
of Type values in runtime aggregates is harmless (undef in,
undef out). Use-site misuse is caught by the bails on
type_name/type_eq/has_impl and the bitcast guard.
`examples/170-pack-bare-value.sx` flips from the parse-error
lock-in to "0/1/3/4" — four call shapes of `len_of(..$args) ->
s64 { list := $args; return list.len; }`. The slice's `.len`
field carries the per-mono pack arity.
210/210 example tests + `zig build test` green.
The remaining 4A.bare slices (4 and 5) — resolveTypeArg
silent-arm fix for index_expr + smoke test of a real builder
walking $args — are separate commits per the cadence rule.
Second slice of the .type_tag activation. The reflection
intrinsics (`type_name`, `type_eq`, `has_impl`) now have
interp-time implementations that read `.type_tag` Values
directly. Today's lower-time fast path (folding to
`const_string`/`const_bool` when the type arg is statically
resolvable) stays — these interp arms are the fallback path
for when lowering emits a real `builtin_call` because the
arg is interp-time-only (e.g. `args[i]` inside a builder body
where the pack element is bound at interp execution).
Plumbing:
- New BuiltinId entries: `type_name`, `type_eq`, `has_impl`.
- Interp arms in `execBuiltinInner`:
- `type_name(t)`: reads `.type_tag` via `asTypeId`, looks up
via `module.types.typeName`, dupes the slice into the
interp allocator, returns `.string`. Non-`.type_tag` arg
→ `bailDetail` ("argument is not a Type value").
- `type_eq(a, b)`: both args must be `.type_tag`; compares
TypeIds. Either side missing → `bailDetail`.
- `has_impl(P, T)`: bails with a "not yet wired" message —
interp-time has_impl needs a queryable snapshot of the
host's `protocol_thunk_map` + `param_impl_map`, which is
its own follow-up slice. Static-arg has_impl still works
via the lower-time `tryConstBoolCondition` fast path.
- emit_llvm: explicit arms for the three new builtins that
log + map to undef-i64 (Type values are comptime-only; if
one of these reaches LLVM emit, lowering produced wrong
IR — the LLVM verifier downstream surfaces the offending
site).
Three new Zig unit tests in interp.test.zig:
- `type_name builtin on type_tag` — emits a `builtin_call`
to `type_name` with a `const_type(s64)` operand, asserts
the result is the string "s64".
- `type_eq builtin on type_tag values` — two equal Type
operands compare equal.
- (Pre-existing) `const_type yields type_tag` + `type_tag
comparison` from 4.0 still pass.
208/208 example tests + `zig build test` green. No source-
language path constructs `.type_tag` yet — the foundation is
ready for the `$args`-in-expression-position slice that
turns it on for users.
Wires the dormant `Value.type_tag(TypeId)` variant in interp.zig
so Type values flow through the comptime interpreter as
first-class kind-distinguished entities. No source-language
construction path yet — that's a follow-up. This commit is the
infrastructure foundation.
Audit findings (from interp.zig switch-walk):
- Every `else =>` arm over Value is either already loud
(`bailDetail` / `error.TypeError`) or a pass-through helper
(`materializeCtxArg`, `materializeForCall`, `resolveSlotChain`)
where transit-unchanged is semantically correct for type_tag.
No new silent paths introduced by activating the variant.
- The three pre-existing `.type_tag => return bailDetail(...)`
arms (store-at-raw-ptr, deref-non-pointer, unbox-non-aggregate)
already cover the disallowed paths cleanly.
New plumbing:
- `Op.const_type: TypeId` — dedicated opcode. Never piggybacks
on `const_int`. Result IR-type is `.any` to signal "untyped
at runtime" so downstream coercions fail loudly.
- `Builder.constType(tid)` constructor.
- Interp arm emits `Value{ .type_tag = tid }` for the op.
- emit_llvm arm bails loudly + emits an undef-i64 placeholder
(Type is comptime-only — if a Type ever reached LLVM emit,
some upstream builder leaked through; the diagnostic + LLVM
verifier downstream surface the offending site).
- `print.zig` arm prints `const type(<typeName>)`.
- `Value.asTypeId() ?TypeId` helper — the kind-honest accessor
for Type values. asInt/asFloat/asBool/asString continue to
return null for `.type_tag` (no silent coercion).
- `evalCmp` arm for `.type_tag, .type_tag` — TypeId equality.
Mixed `.type_tag` vs `.int` deliberately falls through to
the typeErrorDetail bail (a Type is not an int).
Tests (src/ir/interp.test.zig):
- `const_type yields type_tag` — confirms the variant is
produced and that asTypeId/asInt distinguish correctly.
- `type_tag comparison` — exercises cmp_eq on equal and
unequal pairs, asserts the right bool comes back.
208/208 example tests + `zig build test` green. No user-visible
behaviour change yet — `.type_tag` is constructible from Zig-
side IR builders but no sx-level syntax produces it. Next slice
wires `$args` lowering (or `$args[i]` in expression position)
to emit `const_type` per pack element.
Migrates SxSceneDelegate from the hand-rolled
objc_allocateClassPair + class_addMethod + class_addProtocol
sequence to the declarative form:
SxSceneDelegate :: #objc_class("SxSceneDelegate") {
#extends UIResponder;
#implements UISceneDelegate;
#implements UIWindowSceneDelegate;
scene_willConnectToSession_options :: (self, scene, session, options) { ... }
window :: (self) -> *void { ... }
setWindow :: (self, w) { ... }
}
emit_llvm now honors '#implements' in the class-pair init
constructor — for each #implements ProtocolAlias on the cache
entry's AST, emit before objc_registerClassPair:
proto = objc_getProtocol("ProtocolName")
class_addProtocol(cls, proto)
iOS checks 'class_conformsToProtocol' when instantiating scene
delegates; without the conformance the runtime silently rejects
the class and a default scene with no delegate gets created
instead. The protocol-getter returns null on dead-strip /
runtime mismatch (rare but possible) — the runtime treats
class_addProtocol(cls, null) as a no-op, so no explicit null
check needed.
Method bodies forward to the existing legacy free IMP functions
(uikit_scene_will_connect, uikit_window_getter,
uikit_window_setter) so we don't have to inline the scene-
connect setup logic (~80 lines).
uikit_register_classes is now tiny — just the two remaining
view-class helpers (M3.3 SxGLView + M3.4 SxMetalView). M3.5
deletes the function entirely once those land.
Chess on iOS-sim: board renders, scene delegate fires, touch
events route correctly. 183 example tests + zig build test
green.
Two coupled changes that unblock the uikit_register_classes
migration:
1) M1.2 A.3 — body's 'self' is the Obj-C id (opaque), NOT the
state struct. Matches Apple's ObjC semantics where 'self' IS
the object. Cocoa idiom 'xx self → id' works at runtime calls
(addObserver:, etc.); previously the trampoline replaced
'self' with the state-struct pointer, breaking any runtime
call that expected an id.
'*Self' substitution in resolveTypeWithBindings now points at
foreignClassStructType(fcd) — the opaque class stub — instead
of objcDefinedStateStructType(fcd).
'self.field' access on a sx-defined class instance field is
rewritten by lowerFieldAccess to go through the __sx_state
ivar:
state = object_getIvar(self, load(__<Cls>_state_ivar))
val = struct_gep(state, field_idx) → load
Both read (lowerFieldAccess) and write (lowerAssignment) take
this path. Compound ops (+=, -=, etc.) are supported via
storeOrCompound. The lookup is filtered: skip property fields
(those still go through the M2.2 msgSend getter/setter
dispatch) and foreign classes (no state).
New helpers in lower.zig:
- lookupObjcDefinedStateFieldOnPointer — match check.
- lowerObjcDefinedStateForObj — emit the object_getIvar +
ivar-global-load idiom (shared between read + write paths).
- lowerObjcDefinedStateFieldRead — the load path.
Also moved the @llvm.global_ctors registration out of the
sx-defined class-pair init constructor — global_ctors fires
DURING dyld's framework load, before UIKit registers its Obj-C
classes. objc_getClass("UIResponder") returned null, super
was null, objc_registerClassPair crashed. main's entry block
is post-framework-load but pre-user-code — exactly the right
window. New helper injectCtorIntoMain.
2) M3.1 — SxAppDelegate migrated to declarative #objc_class.
uikit_register_classes' hand-rolled objc_allocateClassPair +
class_addMethod for SxAppDelegate is gone; the compiler
synthesises the class at module init. The method bodies
forward to the existing legacy IMP free functions
(uikit_did_finish_launching, uikit_keyboard_will_change_frame)
so we don't have to inline 70+ lines of keyboard-frame logic
right now.
Also adds UIResponder foreign-class declaration and chains
UIView / UITextField to it via #extends UIResponder so the
methods that previously lived on UITextField directly
(becomeFirstResponder etc.) move to their proper home.
Chess on iOS-sim: board renders, full state intact. 183 example
tests + zig build test green.
When 'obj.method()' is called on a foreign-class pointer and the
method isn't declared on the receiver's class, the compiler walks
the '#extends' chain to find an ancestor that declared it.
Property lookup (M2.2) flows through the same chain walker.
ParentX :: #foreign #objc_class("...") { foo :: ... }
ChildX :: #foreign #objc_class("...") { #extends ParentX; }
child.foo() // now resolves — was 'no method foo on ChildX'
Two new helpers in lower.zig:
- findForeignMethodInChain(fcd, name) walks the cache via
fcd.members[i].extends → foreign_class_map[parent] → ...
Depth-capped at 16 to break accidental cycles.
- findForeignPropertyInChain(fcd, name) — same shape for fields.
ALSO fixes a latent class-hierarchy bug uncovered while testing
M2.3: emit_llvm was passing the sx alias name to
objc_allocateClassPair(super, ...) rather than the actual Obj-C
runtime class name. For 'SxThing :: #objc_class(...) { #extends
NSObjectBase; }' where 'NSObjectBase' is aliased to "NSObject",
emit_llvm produced 'objc_getClass("NSObjectBase")' → NULL →
'objc_allocateClassPair(NULL, ...)' → SxThing's super-class link
was broken → '[sx_thing hash]' bypassed NSObject and crashed in
the forwarding machinery.
Fix: ObjcDefinedClassEntry gains a 'parent_objc_name' field
pre-resolved by lower.zig's 'resolveObjcParentName' through
foreign_class_map (which has the alias → foreign_path mapping).
emit_llvm just reads the resolved name from the entry.
153-objc-extends-chain.sx exercises both fixes:
1-level: SxThing → NSObject — t.hash() walks one #extends.
2-level: SxLeaf → SxMiddle → NSObject — chained #extends.
Both return real NSObject.hash values from libobjc.
183 example tests pass (+1). zig build test green.
Bodied methods without a '*Self' first param (parser marks
is_static=true) are now registered as Obj-C CLASS methods on
the metaclass.
Each such method gets:
- A synthesized FnDecl + body lowering through the existing
M1.2 A.2 path.
- A C-ABI trampoline 'emitObjcDefinedClassStaticImp' — same
shape as the instance trampoline but skips the __sx_state
ivar read (no instance state) and passes only
'__sx_default_context' (plus user args) to the sx body.
- An entry in ObjcDefinedMethodEntry with 'is_class=true'.
emit_llvm's class-pair init constructor now computes the
metaclass once up-front (via object_getClass(cls)) and shares
it between the +alloc IMP registration (M1.2 A.5) and the
M2.1(b) class-method registrations. The per-method registration
loop picks the target via 'method.is_class ? metaclass : cls'.
149-objc-class-method-static-imp.sx end-to-end on macOS:
SxFoo :: #objc_class("SxFoo") {
answer :: () -> s32 { return 42; }
}
// [SxFoo answer] via objc_msgSend → 42
// class_getClassMethod(SxFoo, sel_answer) → non-null
Still TODO for M2.1: the (a) class-LEVEL constant form
'layerClass :: Class = CAEAGLLayer.class();' — needs parser
extension to recognize 'name :: Type = expr;' inside #objc_class
blocks, plus lazy-init-slot synthesis.
179 example tests pass (+1). zig build test green.
Delete the bail at lower.zig:4407 that diagnosed sx-defined Obj-C
class dispatch as 'not yet supported'. Both foreign and
sx-defined '#objc_class' decls now flow through the same
'lowerObjcMethodCall' path — instance methods on sx-defined
classes dispatch via objc_msgSend, and the registered IMP
trampolines (M1.2 A.4b.iii) route to the sx bodies.
The runtime non-Obj-C branch (.swift_class / .swift_struct /
.swift_protocol) keeps its 'not yet supported' diagnostic;
M1.2 only addresses the Obj-C runtimes.
Constructor reorder in emit_llvm: emitObjcDefinedClassInit
runs BEFORE emitObjcClassInit. Otherwise the Phase 3.1
class-cache populator calls objc_getClass("SxFoo") before our
constructor registers the class — cache slot stored null and
'SxFoo.method()' dispatched against a null class pointer.
ffi-objc-defined-class-01-instance.sx (the integration test
from the plan) now runs the full lifecycle on macOS:
f := SxFoo.alloc() // synthesized +alloc IMP fires
f.bump() // dispatch → IMP trampoline → sx body
f.bump() // state persists across calls
f.bump()
f.get() // → 3
release_fn(f, sel_release) // synthesized -dealloc fires
The user declares 'alloc :: () -> *SxFoo;' bodyless to give the
synthesized +alloc IMP a typed contract at sx call sites —
same convention as foreign classes today.
M1.2 complete: A.0 A.1 A.2 A.3 A.4 A.4b.i A.4b.ii A.4b.iii
A.5 A.6 A.7. End-to-end class-synthesis foundation works.
177 example tests pass (+1 from the integration test). zig
build test green.
For every sx-defined #objc_class, emit a C-callconv -dealloc IMP
that runs at refcount-zero. Frees the sx state struct, nils the
ivar, then chains to [super dealloc] so NSObject's runtime
cleanup (object_dispose, associated-object teardown, KVO, etc.)
runs as usual.
-dealloc IMP (self: id, _cmd: SEL) -> void
state = object_getIvar(self, load @__<Cls>_state_ivar)
free(state) // free(NULL) is safe
object_setIvar(self, ivar, NULL)
sup = alloca { receiver: *void, super_class: *void }
sup.receiver = self
sup.super_class = load @__<Cls>_class
sel_dealloc = sel_registerName("dealloc")
objc_msgSendSuper2(&sup, sel_dealloc)
return
Two new per-class globals:
- '__<Cls>_class' : *void — populated by emit_llvm's
class-pair init constructor with the freshly-allocated Class
pointer (after objc_registerClassPair).
- The existing '__<Cls>_state_ivar' is also consulted to find
the state struct.
The -dealloc IMP is registered on the class itself (instance
method) via class_addMethod with encoding 'v@:'. emit_llvm
ALSO stores cls_val into '__<Cls>_class' so the trampoline
can build the objc_super struct.
internStringConstantGlobal helper added to lower.zig — interns
C strings as [N:0]u8 globals with byte-level aggregate inits.
Used here for the 'dealloc' selector string.
147-objc-class-dealloc-roundtrip.sx verifies end-to-end on
macOS: alloc + release fires the IMP, and a second alloc/release
cycle proves runtime state isn't corrupted. class_getMethod-
Implementation confirms the IMP is registered.
176 example tests pass (+1). zig build test green.
Still gated: sx-side 'obj.method()' calls bail at lower.zig:4407
with the existing diagnostic. A.7 opens the gate — last sub-step
of M1.2.
For every sx-defined #objc_class, emit a C-callconv +alloc IMP
that the Obj-C runtime calls when '[Cls alloc]' fires (from sx
code, UIKit instantiation, Info.plist principal class, etc.):
+alloc IMP (cls: Class, _cmd: SEL) -> id
instance = class_createInstance(cls, 0)
state = malloc(STATE_SIZE)
memset(state, 0, STATE_SIZE)
object_setIvar(instance, load(@__<Cls>_state_ivar), state)
return instance
STATE_SIZE = max(typeSizeBytes(state struct), 1) — always at
least one byte so the ivar is never null after +alloc returns.
The IMP is registered on the METACLASS (class methods live there
— every Class object's isa points to the metaclass) in emit_llvm's
class-pair init constructor:
metaclass = object_getClass(cls)
sel_alloc = sel_registerName("alloc")
class_addMethod(metaclass, sel_alloc, alloc_imp, "@@:")
That override wins over NSObject's default +alloc; runtime
instantiations get the __sx_state ivar bound automatically.
Per-instance allocator binding (the plan's full design — store
the Allocator value in the state struct so -dealloc frees through
the same one) is deferred. libc malloc/free is fine for v1; we'll
upgrade once Month 4's autoreleasepool + ARC ops shake out.
REFACTOR: collapsed five duplicate 'get<Name>Fid' helpers and
their cache fields (object_getIvar, object_setIvar,
class_createInstance, malloc, memset) into a single
'ensureCRuntimeDecl(name, params, ret) -> FuncId'. The helper
checks for an existing decl by name first (avoids the
'class_createInstance.1' duplicate-symbol crash when stdlib's
'#foreign' decl is already in the module). One helper instead
of one-per-function = ~150 lines deleted.
object_getIvar / object_setIvar added to stdlib std/objc.sx
so user code can use them too (146 exercises object_getIvar
to verify __sx_state was bound to a non-null state pointer
after +alloc).
146-objc-class-alloc-roundtrip.sx end-to-end against macOS:
'[SxFoo alloc]' returns non-null AND object_getIvar(instance,
__sx_state) returns the state ptr. Real Obj-C runtime, no
mocks.
175 example tests pass (+1). zig build test green.
