refactor(napi/oxlint, napi/parser): remove source length from RawTransferMetadata (#12483)

Refactor to oxlint JS plugins and raw transfer. Remove source text length field from `RawTransferMetadata`. It's not necessary - source text is part of `Program`, so the length can be obtained from there, it doesn't need to be duplicated.

crates/oxc_ast_macros/src/generated/structs.rs

@@ -96,7 +96,7 @@ pub static STRUCTS: phf::Map<&'static str, StructDetails> = ::phf::Map {
         ("DebuggerStatement", StructDetails { field_order: None }),
         ("CommentNodeId", StructDetails { field_order: None }),
         ("Character", StructDetails { field_order: Some(&[0, 2, 1]) }),
-        ("RawTransferMetadata2", StructDetails { field_order: Some(&[0, 3, 1, 2]) }),
+        ("RawTransferMetadata2", StructDetails { field_order: Some(&[1, 2, 0]) }),
         ("StringLiteral", StructDetails { field_order: None }),
         ("TSInferType", StructDetails { field_order: None }),
         ("TSThisType", StructDetails { field_order: None }),
@@ -243,7 +243,7 @@ pub static STRUCTS: phf::Map<&'static str, StructDetails> = ::phf::Map {
         ("SpreadElement", StructDetails { field_order: None }),
         ("MetaProperty", StructDetails { field_order: None }),
         ("TSIntersectionType", StructDetails { field_order: None }),
-        ("RawTransferMetadata", StructDetails { field_order: Some(&[0, 3, 1, 2]) }),
+        ("RawTransferMetadata", StructDetails { field_order: Some(&[1, 2, 0]) }),
         ("ConditionalExpression", StructDetails { field_order: None }),
         ("PrivateInExpression", StructDetails { field_order: None }),
         ("TSStringKeyword", StructDetails { field_order: None }),

crates/oxc_linter/src/generated/assert_layouts.rs

@@ -11,22 +11,20 @@ use crate::*;
 const _: () = {
     // Padding: 3 bytes
     assert!(size_of::<RawTransferMetadata2>() == 16);
-    assert!(align_of::<RawTransferMetadata2>() == 4);
-    assert!(offset_of!(RawTransferMetadata2, data_offset) == 0);
+    assert!(align_of::<RawTransferMetadata2>() == 8);
+    assert!(offset_of!(RawTransferMetadata2, data_offset) == 8);
     assert!(offset_of!(RawTransferMetadata2, is_ts) == 12);
-    assert!(offset_of!(RawTransferMetadata2, source_len) == 4);
-    assert!(offset_of!(RawTransferMetadata2, _padding) == 8);
+    assert!(offset_of!(RawTransferMetadata2, _padding) == 0);
 };
 
 #[cfg(target_pointer_width = "32")]
 const _: () = {
     // Padding: 3 bytes
     assert!(size_of::<RawTransferMetadata2>() == 16);
-    assert!(align_of::<RawTransferMetadata2>() == 4);
-    assert!(offset_of!(RawTransferMetadata2, data_offset) == 0);
+    assert!(align_of::<RawTransferMetadata2>() == 8);
+    assert!(offset_of!(RawTransferMetadata2, data_offset) == 8);
     assert!(offset_of!(RawTransferMetadata2, is_ts) == 12);
-    assert!(offset_of!(RawTransferMetadata2, source_len) == 4);
-    assert!(offset_of!(RawTransferMetadata2, _padding) == 8);
+    assert!(offset_of!(RawTransferMetadata2, _padding) == 0);
 };
 
 #[cfg(not(any(target_pointer_width = "64", target_pointer_width = "32")))]

crates/oxc_linter/src/lib.rs

@@ -247,13 +247,11 @@ impl Linter {
         // `external_linter` always exists when `oxlint2` feature is enabled
         let external_linter = self.external_linter.as_ref().unwrap();
 
-        // Write offset of `Program` and source text length in metadata at end of buffer
+        // Write offset of `Program` in metadata at end of buffer
         let program = semantic.nodes().program().unwrap();
         let program_offset = ptr::from_ref(program) as u32;
-        #[expect(clippy::cast_possible_truncation)]
-        let source_len = program.source_text.len() as u32;
 
-        let metadata = RawTransferMetadata::new(program_offset, source_len);
+        let metadata = RawTransferMetadata::new(program_offset);
         let metadata_ptr = allocator.end_ptr().cast::<RawTransferMetadata>();
         // SAFETY: `Allocator` was created by `FixedSizeAllocator` which reserved space after `end_ptr`
         // for a `RawTransferMetadata`. `end_ptr` is aligned for `FixedSizeAllocator`.
@@ -318,10 +316,8 @@ struct RawTransferMetadata2 {
     pub data_offset: u32,
     /// `true` if AST is TypeScript.
     pub is_ts: bool,
-    /// Offset of `u32` containing source text length within buffer.
-    pub(crate) source_len: u32,
     /// Padding to pad struct to size 16.
-    pub(crate) _padding: u32,
+    pub(crate) _padding: u64,
 }
 
 #[cfg(all(feature = "oxlint2", not(feature = "disable_oxlint2")))]
@@ -329,8 +325,8 @@ use RawTransferMetadata2 as RawTransferMetadata;
 
 #[cfg(all(feature = "oxlint2", not(feature = "disable_oxlint2")))]
 impl RawTransferMetadata {
-    pub fn new(data_offset: u32, source_len: u32) -> Self {
-        Self { data_offset, is_ts: false, source_len, _padding: 0 }
+    pub fn new(data_offset: u32) -> Self {
+        Self { data_offset, is_ts: false, _padding: 0 }
     }
 }
 

napi/oxlint2/src-js/generated/constants.cjs

@@ -3,16 +3,16 @@
 
 const BUFFER_SIZE = 2147483616,
   BUFFER_ALIGN = 4294967296,
-  DATA_POINTER_POS_32 = 536870900,
+  DATA_POINTER_POS_32 = 536870902,
   IS_TS_FLAG_POS = 2147483612,
-  SOURCE_LEN_POS_32 = 536870901,
-  PROGRAM_OFFSET = 0;
+  PROGRAM_OFFSET = 0,
+  SOURCE_LEN_OFFSET = 16;
 
 module.exports = {
   BUFFER_SIZE,
   BUFFER_ALIGN,
   DATA_POINTER_POS_32,
   IS_TS_FLAG_POS,
-  SOURCE_LEN_POS_32,
   PROGRAM_OFFSET,
+  SOURCE_LEN_OFFSET,
 };

napi/oxlint2/src-js/index.js

@@ -1,6 +1,6 @@
 import { createRequire } from 'node:module';
 import { lint } from './bindings.js';
-import { DATA_POINTER_POS_32, SOURCE_LEN_POS_32 } from './generated/constants.cjs';
+import { DATA_POINTER_POS_32, SOURCE_LEN_OFFSET } from './generated/constants.cjs';
 import { getErrorMessage } from './utils.js';
 import { addVisitorToCompiled, compiledVisitor, finalizeCompiledVisitor, initCompiledVisitor } from './visitor.js';
 
@@ -167,8 +167,9 @@ function lintFile([filePath, bufferId, buffer, ruleIds]) {
   // Some rules seen in the wild return an empty visitor object from `create` if some initial check fails
   // e.g. file extension is not one the rule acts on.
   if (needsVisit) {
-    const programPos = buffer.uint32[DATA_POINTER_POS_32],
-      sourceByteLen = buffer.uint32[SOURCE_LEN_POS_32];
+    const { uint32 } = buffer,
+      programPos = uint32[DATA_POINTER_POS_32],
+      sourceByteLen = uint32[(programPos + SOURCE_LEN_OFFSET) >> 2];
 
     const sourceText = textDecoder.decode(buffer.subarray(0, sourceByteLen));
     const sourceIsAscii = sourceText.length === sourceByteLen;

napi/parser/generated/constants.js

@@ -3,16 +3,16 @@
 
 const BUFFER_SIZE = 2147483616,
   BUFFER_ALIGN = 4294967296,
-  DATA_POINTER_POS_32 = 536870900,
+  DATA_POINTER_POS_32 = 536870902,
   IS_TS_FLAG_POS = 2147483612,
-  SOURCE_LEN_POS_32 = 536870901,
-  PROGRAM_OFFSET = 0;
+  PROGRAM_OFFSET = 0,
+  SOURCE_LEN_OFFSET = 16;
 
 module.exports = {
   BUFFER_SIZE,
   BUFFER_ALIGN,
   DATA_POINTER_POS_32,
   IS_TS_FLAG_POS,
-  SOURCE_LEN_POS_32,
   PROGRAM_OFFSET,
+  SOURCE_LEN_OFFSET,
 };

napi/parser/generated/deserialize/js.js

@@ -20,7 +20,7 @@ function deserialize(buffer, sourceTextInput, sourceByteLenInput) {
   sourceByteLen = sourceByteLenInput;
   sourceIsAscii = sourceText.length === sourceByteLen;
 
-  const data = deserializeRawTransferData(uint32[536870900]);
+  const data = deserializeRawTransferData(uint32[536870902]);
 
   uint8 =
     uint32 =
@@ -5238,16 +5238,16 @@ function deserializeBoxTSExternalModuleReference(pos) {
   return deserializeTSExternalModuleReference(uint32[pos >> 2]);
 }
 
-function deserializeOptionNameSpan(pos) {
-  if (uint32[(pos + 8) >> 2] === 0 && uint32[(pos + 12) >> 2] === 0) return null;
-  return deserializeNameSpan(pos);
-}
-
 function deserializeU64(pos) {
   const pos32 = pos >> 2;
   return uint32[pos32] + uint32[pos32 + 1] * 4294967296;
 }
 
+function deserializeOptionNameSpan(pos) {
+  if (uint32[(pos + 8) >> 2] === 0 && uint32[(pos + 12) >> 2] === 0) return null;
+  return deserializeNameSpan(pos);
+}
+
 function deserializeOptionU64(pos) {
   if (uint8[pos] === 0) return null;
   return deserializeU64(pos + 8);

napi/parser/generated/deserialize/ts.js

@@ -20,7 +20,7 @@ function deserialize(buffer, sourceTextInput, sourceByteLenInput) {
   sourceByteLen = sourceByteLenInput;
   sourceIsAscii = sourceText.length === sourceByteLen;
 
-  const data = deserializeRawTransferData(uint32[536870900]);
+  const data = deserializeRawTransferData(uint32[536870902]);
 
   uint8 =
     uint32 =
@@ -5369,16 +5369,16 @@ function deserializeBoxTSExternalModuleReference(pos) {
   return deserializeTSExternalModuleReference(uint32[pos >> 2]);
 }
 
-function deserializeOptionNameSpan(pos) {
-  if (uint32[(pos + 8) >> 2] === 0 && uint32[(pos + 12) >> 2] === 0) return null;
-  return deserializeNameSpan(pos);
-}
-
 function deserializeU64(pos) {
   const pos32 = pos >> 2;
   return uint32[pos32] + uint32[pos32 + 1] * 4294967296;
 }
 
+function deserializeOptionNameSpan(pos) {
+  if (uint32[(pos + 8) >> 2] === 0 && uint32[(pos + 12) >> 2] === 0) return null;
+  return deserializeNameSpan(pos);
+}
+
 function deserializeOptionU64(pos) {
   if (uint8[pos] === 0) return null;
   return deserializeU64(pos + 8);

napi/parser/generated/lazy/constructors.js

@@ -13752,17 +13752,17 @@ function constructBoxTSExternalModuleReference(pos, ast) {
   return new TSExternalModuleReference(ast.buffer.uint32[pos >> 2], ast);
 }
 
-function constructOptionNameSpan(pos, ast) {
-  if (ast.buffer.uint32[(pos + 8) >> 2] === 0 && ast.buffer.uint32[(pos + 12) >> 2] === 0) return null;
-  return new NameSpan(pos, ast);
-}
-
 function constructU64(pos, ast) {
   const { uint32 } = ast.buffer,
     pos32 = pos >> 2;
   return uint32[pos32] + uint32[pos32 + 1] * 4294967296;
 }
 
+function constructOptionNameSpan(pos, ast) {
+  if (ast.buffer.uint32[(pos + 8) >> 2] === 0 && ast.buffer.uint32[(pos + 12) >> 2] === 0) return null;
+  return new NameSpan(pos, ast);
+}
+
 function constructOptionU64(pos, ast) {
   if (ast.buffer[pos] === 0) return null;
   return constructU64(pos + 8, ast);

napi/parser/src/generated/assert_layouts.rs

@@ -19,11 +19,10 @@ const _: () = {
 
     // Padding: 3 bytes
     assert!(size_of::<RawTransferMetadata>() == 16);
-    assert!(align_of::<RawTransferMetadata>() == 4);
-    assert!(offset_of!(RawTransferMetadata, data_offset) == 0);
+    assert!(align_of::<RawTransferMetadata>() == 8);
+    assert!(offset_of!(RawTransferMetadata, data_offset) == 8);
     assert!(offset_of!(RawTransferMetadata, is_ts) == 12);
-    assert!(offset_of!(RawTransferMetadata, source_len) == 4);
-    assert!(offset_of!(RawTransferMetadata, _padding) == 8);
+    assert!(offset_of!(RawTransferMetadata, _padding) == 0);
 
     // Padding: 7 bytes
     assert!(size_of::<Error>() == 80);
@@ -78,11 +77,10 @@ const _: () = {
 
     // Padding: 3 bytes
     assert!(size_of::<RawTransferMetadata>() == 16);
-    assert!(align_of::<RawTransferMetadata>() == 4);
-    assert!(offset_of!(RawTransferMetadata, data_offset) == 0);
+    assert!(align_of::<RawTransferMetadata>() == 8);
+    assert!(offset_of!(RawTransferMetadata, data_offset) == 8);
     assert!(offset_of!(RawTransferMetadata, is_ts) == 12);
-    assert!(offset_of!(RawTransferMetadata, source_len) == 4);
-    assert!(offset_of!(RawTransferMetadata, _padding) == 8);
+    assert!(offset_of!(RawTransferMetadata, _padding) == 0);
 
     // Padding: 3 bytes
     assert!(size_of::<Error>() == 44);

napi/parser/src/raw_transfer_types.rs

@@ -36,16 +36,13 @@ pub struct RawTransferMetadata {
     pub data_offset: u32,
     /// `true` if AST is TypeScript.
     pub is_ts: bool,
-    /// Offset of `u32` containing source text length within buffer.
-    /// Not used in this implementation. Used by `napi/oxlint2`.
-    pub(crate) source_len: u32,
     /// Padding to pad struct to size 16.
-    pub(crate) _padding: u32,
+    pub(crate) _padding: u64,
 }
 
 impl RawTransferMetadata {
     pub fn new(data_offset: u32, is_ts: bool) -> Self {
-        Self { data_offset, is_ts, source_len: 0, _padding: 0 }
+        Self { data_offset, is_ts, _padding: 0 }
     }
 }
 

tasks/ast_tools/src/generators/raw_transfer.rs

@@ -26,6 +26,9 @@ use crate::{
 
 use super::define_generator;
 
+/// Offset of length field in `&str`
+const STR_LEN_OFFSET: u32 = 8;
+
 /// Bytes reserved for `malloc`'s metadata
 const MALLOC_RESERVED_SIZE: u32 = 16;
 
@@ -1015,10 +1018,10 @@ struct Constants {
     data_pointer_pos: u32,
     /// Offset within buffer of `bool` indicating if AST is TS or JS
     is_ts_pos: u32,
-    /// Offset within buffer of `u32` containing position of `u32` containing source text length
-    source_len_pos: u32,
     /// Offset of `Program` in buffer, relative to position of `RawTransferData`
     program_offset: u32,
+    /// Offset of `u32` source text length, relative to position of `Program`
+    source_len_offset: u32,
     /// Size of `RawTransferData` in bytes
     raw_metadata_size: u32,
 }
@@ -1029,13 +1032,12 @@ fn generate_constants(consts: Constants) -> (String, TokenStream) {
         buffer_size,
         data_pointer_pos,
         is_ts_pos,
-        source_len_pos,
         program_offset,
+        source_len_offset,
         raw_metadata_size,
     } = consts;
 
     let data_pointer_pos_32 = data_pointer_pos / 4;
-    let source_len_pos_32 = source_len_pos / 4;
 
     #[rustfmt::skip]
     let js_output = format!("
@@ -1043,16 +1045,16 @@ fn generate_constants(consts: Constants) -> (String, TokenStream) {
             BUFFER_ALIGN = {BLOCK_ALIGN},
             DATA_POINTER_POS_32 = {data_pointer_pos_32},
             IS_TS_FLAG_POS = {is_ts_pos},
-            SOURCE_LEN_POS_32 = {source_len_pos_32},
-            PROGRAM_OFFSET = {program_offset};
+            PROGRAM_OFFSET = {program_offset},
+            SOURCE_LEN_OFFSET = {source_len_offset};
 
         module.exports = {{
             BUFFER_SIZE,
             BUFFER_ALIGN,
             DATA_POINTER_POS_32,
             IS_TS_FLAG_POS,
-            SOURCE_LEN_POS_32,
             PROGRAM_OFFSET,
+            SOURCE_LEN_OFFSET,
         }};
     ");
 
@@ -1085,7 +1087,6 @@ fn get_constants(schema: &Schema) -> Constants {
 
     let mut data_offset_field = None;
     let mut is_ts_field = None;
-    let mut source_len_field = None;
     for (field1, field2) in raw_metadata_struct.fields.iter().zip(&raw_metadata2_struct.fields) {
         assert_eq!(field1.name(), field2.name());
         assert_eq!(field1.type_id, field2.type_id);
@@ -1093,13 +1094,11 @@ fn get_constants(schema: &Schema) -> Constants {
         match field1.name() {
             "data_offset" => data_offset_field = Some(field1),
             "is_ts" => is_ts_field = Some(field1),
-            "source_len" => source_len_field = Some(field1),
             _ => {}
         }
     }
     let data_offset_field = data_offset_field.unwrap();
     let is_ts_field = is_ts_field.unwrap();
-    let source_len_field = source_len_field.unwrap();
 
     let raw_metadata_size = raw_metadata_struct.layout_64().size;
 
@@ -1115,7 +1114,6 @@ fn get_constants(schema: &Schema) -> Constants {
     let raw_metadata_pos = buffer_size - raw_metadata_size;
     let data_pointer_pos = raw_metadata_pos + data_offset_field.offset_64();
     let is_ts_pos = raw_metadata_pos + is_ts_field.offset_64();
-    let source_len_pos = raw_metadata_pos + source_len_field.offset_64();
 
     let program_offset = schema
         .type_by_name("RawTransferData")
@@ -1124,12 +1122,20 @@ fn get_constants(schema: &Schema) -> Constants {
         .field_by_name("program")
         .offset_64();
 
+    let source_len_offset = schema
+        .type_by_name("Program")
+        .as_struct()
+        .unwrap()
+        .field_by_name("source_text")
+        .offset_64()
+        + STR_LEN_OFFSET;
+
     Constants {
         buffer_size,
         data_pointer_pos,
         is_ts_pos,
-        source_len_pos,
         program_offset,
+        source_len_offset,
         raw_metadata_size,
     }
 }