issue-0037 fixed: ptr↔int conversion in coerceToType / bitcast emit

109/109 regression tests pass; chess Android + iOS-sim still build clean. Root cause: sx's `xx <ptr>` cast targeting an integer type (common pattern: `xx u64 = xx @some_global`) lowered to a no-op because `coerceToType` had branches for int↔float and same-kind widen/narrow, but nothing for pointer↔integer. The cast left the value as a pointer Ref, and `emitInst`'s `.ret` arm tried to coerce a `ptr` value to an `i64` slot — coerceArg had no ptr↔int branch either, fell through to undef. Why it worked in main but failed in helpers: an `alloca u64`+`store ptr @g, alloca`+`load i64, alloca` sequence preserves the address bits as raw memory, so the "store-then-load through an alloca" workaround happened to do the right thing without a real cast. A `ret i64 <ptr>` has no such intermediate slot and triggers an LLVM type mismatch. Fix layered into two existing IR opcodes: lower.zig (coerceToType): new branch — when src and dst types are ptr↔int, emit a `bitcast` IR opcode with the right from/to. Mirrors how int↔float emits `.int_to_float` / `.float_to_int`. emit_llvm.zig (.bitcast arm): dispatch ptr→int to `LLVMBuildPtrToInt` (+ trunc/zext if the target int width != 64), int→ptr to `LLVMBuildIntToPtr`. The "real bitcast" path stays for same-kind type punning. Modern LLVM's BuildBitCast rejects ptr↔int directly, hence the dispatch. The fix also closes a quiet behavior gap that affected non-`#foreign` globals (any `xx @<global>` from a helper fn). Surfaced while investigating issue-0037; verified independently with a non-`#foreign` sx-side global of type `s64`. File mechanics: issue-0037 promoted to a focused feature example per CLAUDE.md's resolution flow: examples/issue-0037.sx -> examples/102-foreign-global-from-helper.sx tests/expected/issue-0037.{txt,exit} -> tests/expected/102-foreign-global-from-helper.{txt,exit} ffi-objc-call-03 + ffi-objc-call-06 IR snapshots updated to reflect the ptr→int store-via-ptrtoint shape that's now correct at the LLVM-IR level (same bits in memory, but properly typed).
2026-05-19 19:18:31 +03:00
parent 5fad92785e
commit 0bb7b8cc27
9 changed files with 96 additions and 69 deletions
--- a/examples/102-foreign-global-from-helper.sx
+++ b/examples/102-foreign-global-from-helper.sx
@@ -0,0 +1,28 @@
+// `xx @<foreign_global>` round-trips through a non-main helper
+// function: the helper's `xx @__stdinp` cast lowers to a `bitcast`
+// IR opcode that emit_llvm.zig dispatches to `LLVMBuildPtrToInt`
+// (BitCast doesn't accept ptr↔int on modern LLVM with opaque
+// pointers), and a `u64`-returning function correctly returns
+// the address.
+//
+// Was issue-0037 — the helper used to emit `ret i64 undef` because
+// `coerceToType` had no pointer↔integer case, so the explicit
+// `xx ptr` cast produced an unchanged pointer value handed to a
+// `ret i64` slot.
+
+#import "modules/std.sx";
+
+__stdinp : *void #foreign;
+
+stdinp_addr_via_helper :: () -> u64 {
+    xx @__stdinp;
+}
+
+main :: () -> s32 {
+    direct  : u64 = xx @__stdinp;
+    via_helper := stdinp_addr_via_helper();
+    print("direct non-null = {}\n", direct != 0);
+    print("helper non-null = {}\n", via_helper != 0);
+    print("eq              = {}\n", direct == via_helper);
+    0;
+}
--- a/examples/issue-0037.sx
+++ b/examples/issue-0037.sx
@@ -1,47 +0,0 @@
-// `@__stdinp` (address-of a `#foreign` global) lowers to `undef` /
-// 0 when accessed from inside a non-main function body. From `main`
-// directly it works:
-//
-//   single file:
-//     __stdinp : *void #foreign;
-//     stdinp_addr :: () -> u64 { xx @__stdinp; }
-//     main :: () -> s32 {
-//         a : u64 = xx @__stdinp;     // a = <real symbol address>
-//         b := stdinp_addr();         // b = 0
-//         ...
-//     }
-//
-// The emitted IR for the helper is just `ret i64 undef`, suggesting
-// the `address_of(identifier=__stdinp)` branch in `lower.zig`
-// (~line 1719) doesn't see `__stdinp` in `global_names` at the time
-// the helper's body is being lowered — even though the same lookup
-// succeeds inside main.
-//
-// Likely cause: lazy/lazy-deferred body lowering ordering vs. the
-// pass that registers extern global decls into `global_names`. Or
-// global_names is scoped per-function (?). Worth verifying which
-// before fixing.
-//
-// Expected when fixed: the helper returns the same address as the
-// direct read, prints `eq = true`. Today: prints `eq = false`.
-//
-// Filed during PLAN-FFI step 0.10 (cross-file `#foreign` global —
-// the cross-file dimension surfaced the bug, but it reproduces in
-// a single file too).
-
-#import "modules/std.sx";
-
-__stdinp : *void #foreign;
-
-stdinp_addr_via_helper :: () -> u64 {
-    xx @__stdinp;
-}
-
-main :: () -> s32 {
-    direct  : u64 = xx @__stdinp;
-    via_helper := stdinp_addr_via_helper();
-    print("direct non-null = {}\n", direct != 0);
-    print("helper non-null = {}\n", via_helper != 0);
-    print("eq              = {}\n", direct == via_helper);
-    0;
-}
--- a/src/ir/emit_llvm.zig
+++ b/src/ir/emit_llvm.zig
@@ -1006,7 +1006,26 @@ pub const LLVMEmitter = struct {
            .bitcast => |conv| {
                const operand = self.resolveRef(conv.operand);
                const to_ty = self.toLLVMType(conv.to);
-                self.mapRef(c.LLVMBuildBitCast(self.builder, operand, to_ty, "bitcast"));
+                // LLVMBuildBitCast doesn't accept ptr↔int on modern
+                // LLVM. Dispatch to PtrToInt / IntToPtr when needed —
+                // lower.zig emits a `bitcast` IR op for both shapes.
+                const from_kind = c.LLVMGetTypeKind(c.LLVMTypeOf(operand));
+                const to_kind = c.LLVMGetTypeKind(to_ty);
+                if (from_kind == c.LLVMPointerTypeKind and to_kind == c.LLVMIntegerTypeKind) {
+                    const i64_val = c.LLVMBuildPtrToInt(self.builder, operand, self.cached_i64, "pti");
+                    const w = c.LLVMGetIntTypeWidth(to_ty);
+                    if (w == 64) {
+                        self.mapRef(i64_val);
+                    } else if (w < 64) {
+                        self.mapRef(c.LLVMBuildTrunc(self.builder, i64_val, to_ty, "pti.tr"));
+                    } else {
+                        self.mapRef(c.LLVMBuildZExt(self.builder, i64_val, to_ty, "pti.ext"));
+                    }
+                } else if (from_kind == c.LLVMIntegerTypeKind and to_kind == c.LLVMPointerTypeKind) {
+                    self.mapRef(c.LLVMBuildIntToPtr(self.builder, operand, to_ty, "itp"));
+                } else {
+                    self.mapRef(c.LLVMBuildBitCast(self.builder, operand, to_ty, "bitcast"));
+                }
            },
            .int_to_float => |conv| {
                const operand = self.resolveRef(conv.operand);
--- a/src/ir/lower.zig
+++ b/src/ir/lower.zig
@@ -9290,6 +9290,8 @@ pub const Lowering = struct {
        const dst_float = isFloat(dst_ty);
        const src_int = self.isIntEx(src_ty);
        const dst_int = self.isIntEx(dst_ty);
+        const src_ptr = !src_ty.isBuiltin() and self.module.types.get(src_ty) == .pointer;
+        const dst_ptr = !dst_ty.isBuiltin() and self.module.types.get(dst_ty) == .pointer;

        // Int → Float
        if (src_int and dst_float) {
@@ -9299,6 +9301,13 @@ pub const Lowering = struct {
        if (src_float and dst_int) {
            return self.builder.emit(.{ .float_to_int = .{ .operand = val, .from = src_ty, .to = dst_ty } }, dst_ty);
        }
+        // Ptr ↔ Int — explicit `xx ptr` to/from an integer-typed slot.
+        // Emits a `bitcast` IR op; emit_llvm.zig's bitcast arm dispatches
+        // to LLVMBuildPtrToInt / LLVMBuildIntToPtr at the LLVM level
+        // since LLVMBuildBitCast itself doesn't accept ptr↔int.
+        if ((src_ptr and dst_int) or (src_int and dst_ptr)) {
+            return self.builder.emit(.{ .bitcast = .{ .operand = val, .from = src_ty, .to = dst_ty } }, dst_ty);
+        }
        // Same kind — widen/narrow based on bit width
        const src_bits = self.typeBitsEx(src_ty);
        const dst_bits = self.typeBitsEx(dst_ty);
--- a/tests/expected/102-foreign-global-from-helper.exit
+++ b/tests/expected/102-foreign-global-from-helper.exit
--- a/tests/expected/102-foreign-global-from-helper.txt
+++ b/tests/expected/102-foreign-global-from-helper.txt
@@ -0,0 +1,3 @@
+direct non-null = true
+helper non-null = true
+eq              = true
--- a/tests/expected/ffi-objc-call-03-selector-sharing.ir
+++ b/tests/expected/ffi-objc-call-03-selector-sharing.ir
@@ -107,7 +107,8 @@ entry:
  store { ptr, i64 } undef, ptr %allocaN, align 8
  %loadN = load ptr, ptr %allocaN, align 8
  %gep = getelementptr inbounds { ptr, i64 }, ptr %allocaN, i32 0, i32 0
-  store ptr %loadN, ptr %gep, align 8
+  %pti = ptrtoint ptr %loadN to i64
+  store i64 %pti, ptr %gep, align 8
  %loadN = load i64, ptr %alloca, align 8
  %gepN = getelementptr inbounds { ptr, i64 }, ptr %allocaN, i32 0, i32 1
  store i64 %loadN, ptr %gepN, align 8
--- a/tests/expected/ffi-objc-call-06-sret-return.ir
+++ b/tests/expected/ffi-objc-call-06-sret-return.ir
@@ -267,7 +267,8 @@ entry:
  store { ptr, i64 } undef, ptr %allocaN, align 8
  %loadN = load ptr, ptr %allocaN, align 8
  %gep = getelementptr inbounds { ptr, i64 }, ptr %allocaN, i32 0, i32 0
-  store ptr %loadN, ptr %gep, align 8
+  %pti = ptrtoint ptr %loadN to i64
+  store i64 %pti, ptr %gep, align 8
  %loadN = load i64, ptr %alloca, align 8
  %gepN = getelementptr inbounds { ptr, i64 }, ptr %allocaN, i32 0, i32 1
  store i64 %loadN, ptr %gepN, align 8
@@ -2371,7 +2372,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.139, label %if.else.140
@@ -2398,7 +2400,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.163, label %if.else.164
@@ -2425,7 +2428,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.166, label %if.else.167
@@ -2452,7 +2456,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.169, label %if.else.170
@@ -2506,7 +2511,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.175, label %if.else.176
@@ -2533,7 +2539,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.178, label %if.else.179
@@ -2560,7 +2567,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.181, label %if.else.182
@@ -2587,7 +2595,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.184, label %if.else.185
@@ -2614,7 +2623,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.187, label %if.else.188
@@ -2641,7 +2651,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.190, label %if.else.191
@@ -2668,7 +2679,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.193, label %if.else.194
@@ -2695,7 +2707,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.196, label %if.else.197
@@ -2722,7 +2735,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.199, label %if.else.200
@@ -2749,7 +2763,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.202, label %if.else.203
@@ -2776,7 +2791,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.205, label %if.else.206
@@ -2803,7 +2819,8 @@ entry:
  store ptr %0, ptr %alloca, align 8
  %allocaN = alloca i64, align 8
  %load = load ptr, ptr %alloca, align 8
-  store ptr %load, ptr %allocaN, align 8
+  %pti = ptrtoint ptr %load to i64
+  store i64 %pti, ptr %allocaN, align 8
  %loadN = load i64, ptr %allocaN, align 8
  %icmp = icmp eq i64 %loadN, 0
  br i1 %icmp, label %if.then.208, label %if.else.209
--- a/tests/expected/issue-0037.txt
+++ b/tests/expected/issue-0037.txt
@@ -1,3 +0,0 @@
-direct non-null = true
-helper non-null = false
-eq              = false