fix(writer): reject NUL bytes in field names

A NUL inside a column name aborts the reference toolchain: the
Arrow FFI schema export hits a panic-cannot-unwind in arrow-rs
(ffi_stream::get_schema) and SIGABRTs the whole process (measured
against vortex-jni 0.75.0). Our pure-Java path handles NUL names
fine, so without the guard we emit files the canonical reader dies
on. Everything else is opaque UTF-8 and round-trips intact both
directions — "", whitespace-only, "$$$$$", newlines, emoji — all
measured and recorded in docs/compatibility.md.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: dfa1

docs/compatibility.md

@@ -36,6 +36,7 @@ resolves only the standalone decoders in `reader`; no encoder class is loaded.
 | `vortex.variant` arbitrary nested objects | Rust (`vortex.parquet.variant`) | ⚠️ Java encodes/decodes variant columns of **typed scalar** values (constant / chunked-of-constants core, optional shredded child); Java↔Rust round-trip verified. Arbitrary nested JSON objects and real path-based shredding need the `vortex.parquet.variant` physical encoding — deferred ([ADR 0014](../adr/0014-variant-encoding-strategy.md)). |
 | Arrow extension array import affecting Variant shape | Rust 0.74.0 (#8125) | Untested. Re-run integration fixtures against v0.74.0 once published. |
 | Duplicate struct field names | Rust writer rejects ("StructLayout must have unique field names"); Rust reader tolerates foreign files (first-match access) | ⚠️ Deliberate divergence on read: Java rejects such files with `VortexException("duplicate field name in file schema")` instead of tolerating them — the name-keyed `Chunk` API cannot represent both columns, and silent column loss is worse than a loud failure on a file the reference writer refuses to produce. Java's writer mirrors the Rust writer's rejection. Empty column name `""` is legal on both sides (pinned by `ColumnNameEdgeCasesIntegrationTest`). |
+| NUL (`U+0000`) in a field name | Aborts the Rust toolchain: Arrow FFI schema export hits a panic-cannot-unwind in `arrow-rs` (`ffi_stream::get_schema`) and SIGABRTs the process (measured against vortex-jni 0.75.0) | ⚠️ Java's writer rejects NUL in field names (`IllegalArgumentException`) so we never emit a file the canonical reader dies on; Java's reader tolerates such files (pure-Java path, no C strings). All other names are opaque UTF-8 and round-trip intact both directions — including `""`, whitespace-only, `$`-runs, newlines, and emoji (measured). |
 
 ## Encodings
 

writer/src/main/java/io/github/dfa1/vortex/writer/VortexWriter.java

@@ -140,6 +140,13 @@ private VortexWriter(
             if (!uniqueNames.add(name)) {
                 throw new IllegalArgumentException("duplicate field name: " + name);
             }
+            // A NUL byte in a field name aborts the reference toolchain's Arrow FFI export
+            // (panic-cannot-unwind in arrow-rs ffi_stream::get_schema, SIGABRT — measured
+            // 2026-07-04): never emit a file the canonical reader dies on. Everything else,
+            // including "" and whitespace-only names, is legal and round-trips.
+            if (name.indexOf('\u0000') >= 0) {
+                throw new IllegalArgumentException("field name contains NUL (U+0000)");
+            }
         }
         this.channel = channel;
         this.schema = schema;

writer/src/test/java/io/github/dfa1/vortex/writer/VortexWriterTest.java

@@ -97,6 +97,26 @@ void create_duplicateFieldNames_throwsIllegalArgumentException(@TempDir Path tmp
         }
     }
 
+    @Test
+    void create_fieldNameWithNulByte_throwsIllegalArgumentException(@TempDir Path tmp) throws IOException {
+        // Given — a NUL byte inside a field name. Legal in our pure-Java stack, but it aborts
+        // the reference toolchain's Arrow FFI export (panic-cannot-unwind in arrow-rs, SIGABRT,
+        // measured 2026-07-04) — so the writer must never emit it. Whitespace-only and empty
+        // names stay legal (see ColumnNameEdgeCasesIntegrationTest).
+        var schema = new DType.Struct(
+                List.of("col\u0000hidden"),
+                List.of(new DType.Primitive(PType.I64, false)),
+                false);
+        Path file = tmp.resolve("nul.vtx");
+
+        // When / Then
+        try (var ch = FileChannel.open(file, StandardOpenOption.CREATE, StandardOpenOption.WRITE)) {
+            assertThatThrownBy(() -> VortexWriter.create(ch, schema, WriteOptions.defaults()))
+                    .isInstanceOf(IllegalArgumentException.class)
+                    .hasMessageContaining("NUL");
+        }
+    }
+
     // ── writeChunk validation ─────────────────────────────────────────────────
 
     @Test