Rebase merge of AST-breaking changes

.changeset/metal-radios-swim.md

@@ -0,0 +1,5 @@
+---
+swc_ecma_ast: major
+---
+
+fix(es/ast): Fix unicode lone surrogates handling

bindings/binding_core_node/src/bundle.rs

@@ -281,6 +281,7 @@ impl swc_core::bundler::Hook for Hook {
                     span,
                     raw: None,
                     value: file_name.into(),
+                    lone_surrogates: false,
                 }))),
             },
             KeyValueProp {

bindings/binding_core_wasm/__tests__/simple.js

@@ -111,6 +111,7 @@ describe("parse", () => {
                   "decorators": [],
                   "identifier": {
                     "ctxt": 2,
+                    "nodeId": 4294967295,
                     "optional": false,
                     "span": {
                       "end": 254,
@@ -157,6 +158,7 @@ describe("parse", () => {
                   "decorators": [],
                   "identifier": {
                     "ctxt": 2,
+                    "nodeId": 4294967295,
                     "optional": false,
                     "span": {
                       "end": 267,

bindings/binding_core_wasm/src/types.rs

@@ -1725,7 +1725,7 @@ export interface JSXAttribute extends Node, HasSpan {
 export type JSXAttributeName = Identifier | JSXNamespacedName;
 
 export type JSXAttrValue =
-  | Literal
+  | StringLiteral
   | JSXExpressionContainer
   | JSXElement
   | JSXFragment;

bindings/binding_minifier_wasm/src/types.rs

@@ -1726,7 +1726,7 @@ export interface JSXAttribute extends Node, HasSpan {
 export type JSXAttributeName = Identifier | JSXNamespacedName;
 
 export type JSXAttrValue =
-  | Literal
+  | StringLiteral
   | JSXExpressionContainer
   | JSXElement
   | JSXFragment;

crates/swc/tests/fixture/issues-7xxx/7678/output/1.js

@@ -1 +1 @@
-let str="\uD83D\uDC68\\u200D\uD83D\uDE80";let obj={"\uD83D\uDC68\\u200D\uD83D\uDE80":"wrong"};
+let str="\\uD83D\\uDC68\\u200D\\uD83D\\uDE80";let obj={"\\uD83D\\uDC68\\u200D\\uD83D\\uDE80":"wrong"};

crates/swc/tests/tsc-references/unicodeExtendedEscapesInStrings10_ES5.1.normal.js

@@ -3,4 +3,4 @@
 //  2. Let cu1 be floor((cp – 65536) / 1024) + 0xD800.
 // Although we should just get back a single code point value of 0xD800,
 // this is a useful edge-case test.
-var x = "\u{D800}";
+var x = "\ud800";

crates/swc/tests/tsc-references/unicodeExtendedEscapesInStrings10_ES6.1.normal.js

@@ -3,4 +3,4 @@
 //  2. Let cu1 be floor((cp – 65536) / 1024) + 0xD800.
 // Although we should just get back a single code point value of 0xD800,
 // this is a useful edge-case test.
-var x = "\u{D800}";
+var x = "\ud800";

crates/swc/tests/tsc-references/unicodeExtendedEscapesInStrings11_ES5.1.normal.js

@@ -3,4 +3,4 @@
 //  2. Let cu2 be ((cp – 65536) modulo 1024) + 0xDC00.
 // Although we should just get back a single code point value of 0xDC00,
 // this is a useful edge-case test.
-var x = "\u{DC00}";
+var x = "\udc00";

crates/swc/tests/tsc-references/unicodeExtendedEscapesInStrings11_ES6.1.normal.js

@@ -3,4 +3,4 @@
 //  2. Let cu2 be ((cp – 65536) modulo 1024) + 0xDC00.
 // Although we should just get back a single code point value of 0xDC00,
 // this is a useful edge-case test.
-var x = "\u{DC00}";
+var x = "\udc00";

crates/swc/tests/tsc-references/unicodeExtendedEscapesInTemplates10_ES5.1.normal.js

@@ -3,4 +3,4 @@
 //  2. Let cu1 be floor((cp – 65536) / 1024) + 0xD800.
 // Although we should just get back a single code point value of 0xD800,
 // this is a useful edge-case test.
-var x = "\u{D800}";
+var x = "\ud800";

crates/swc/tests/tsc-references/unicodeExtendedEscapesInTemplates11_ES5.1.normal.js

@@ -3,4 +3,4 @@
 //  2. Let cu2 be ((cp – 65536) modulo 1024) + 0xDC00.
 // Although we should just get back a single code point value of 0xDC00,
 // this is a useful edge-case test.
-var x = "\u{DC00}";
+var x = "\udc00";

crates/swc/tests/vercel/full/utf8-1/output/index.js

@@ -5,7 +5,7 @@ import r from './on-demand-entries-client';
 import { addMessageListener as n, connectHMR as c } from './error-overlay/websocket';
 var o = JSON.parse(document.getElementById('__NEXT_DATA__').textContent);
 window.__NEXT_DATA__ = o;
-var s = o.assetPrefix, i = o.page, _ = null, u = __webpack_hash__, d = (s = s || '') + (s.endsWith('/') ? '' : '/') + '_next/static/webpack/';
+var s = o.assetPrefix, i = o.page, _ = null, d = __webpack_hash__, u = (s = s || '') + (s.endsWith('/') ? '' : '/') + '_next/static/webpack/';
 n(function(a) {
     if ('\uD83D\uDC93' !== a.data) try {
         var r = JSON.parse(a.data);
@@ -28,7 +28,7 @@ n(function(a) {
                                 5
                             ]), [
                                 4,
-                                fetch('undefined' != typeof __webpack_runtime_id__ ? "".concat(d).concat(u, ".").concat(__webpack_runtime_id__, ".hot-update.json") : "".concat(d).concat(u, ".hot-update.json"))
+                                fetch('undefined' != typeof __webpack_runtime_id__ ? "".concat(u).concat(d, ".").concat(__webpack_runtime_id__, ".hot-update.json") : "".concat(u).concat(d, ".hot-update.json"))
                             ];
                         case 2:
                             return [
@@ -38,7 +38,7 @@ n(function(a) {
                         case 3:
                             return e = t.sent(), a = '/' === i ? 'index' : i, (Array.isArray(e.c) ? e.c : Object.keys(e.c)).some(function(e) {
                                 return -1 !== e.indexOf("pages".concat(a.startsWith('/') ? a : "/".concat(a))) || -1 !== e.indexOf("pages".concat(a.startsWith('/') ? a : "/".concat(a)).replace(/\//g, '\\'));
-                            }) ? document.location.reload(!0) : u = _, [
+                            }) ? document.location.reload(!0) : d = _, [
                                 3,
                                 5
                             ];

crates/swc_bundler/examples/bundle.rs

@@ -195,6 +195,7 @@ impl swc_bundler::Hook for Hook {
                     span,
                     raw: None,
                     value: file_name.into(),
+                    lone_surrogates: false,
                 }))),
             },
             KeyValueProp {

crates/swc_bundler/src/util.rs

@@ -33,6 +33,7 @@ pub(crate) trait VarDeclaratorExt: Into<VarDeclarator> {
                         span: DUMMY_SP,
                         raw: None,
                         value: name.into(),
+                        lone_surrogates: false,
                     }
                     .assign_to(Ident::new_no_ctxt(atom!("INJECTED_FROM"), DUMMY_SP)),
                 ]
@@ -193,6 +194,7 @@ fn metadata(key: &str, value: &str) -> PropOrSpread {
             span: DUMMY_SP,
             value: value.into(),
             raw: None,
+            lone_surrogates: false,
         })
         .into(),
     })))

crates/swc_bundler/tests/deno.rs

@@ -1120,6 +1120,7 @@ impl swc_bundler::Hook for Hook {
                     span,
                     raw: None,
                     value: file_name.into(),
+                    lone_surrogates: false,
                 }))),
             },
             KeyValueProp {

crates/swc_bundler/tests/fixture.rs

@@ -153,6 +153,7 @@ impl swc_bundler::Hook for Hook {
                     span,
                     raw: None,
                     value: file_name.into(),
+                    lone_surrogates: false,
                 }))),
             },
             KeyValueProp {

crates/swc_common/src/lib.rs

@@ -39,6 +39,7 @@ pub use swc_eq_ignore_macros::{EqIgnoreSpan, TypeEq};
 pub use self::{
     eq::{EqIgnoreSpan, TypeEq},
     errors::{SourceMapper, SourceMapperDyn},
+    node_id::NodeId,
     pos::{
         hygiene, BytePos, CharPos, FileName, Globals, Loc, LocWithOpt, Mark, MultiSpan, SourceFile,
         SourceFileAndBytePos, SourceFileAndLine, Span, SpanLinesError, Spanned, SyntaxContext,
@@ -59,6 +60,7 @@ mod eq;
 pub mod errors;
 pub mod input;
 pub mod iter;
+mod node_id;
 pub mod pass;
 pub mod plugin;
 mod pos;

crates/swc_common/src/node_id.rs

@@ -0,0 +1,27 @@
+use crate::util::take::Take;
+
+#[derive(Eq, Hash, Debug, PartialEq, Clone, Copy, serde::Serialize, serde::Deserialize)]
+#[cfg_attr(
+    any(feature = "rkyv-impl"),
+    derive(rkyv::Archive, rkyv::Serialize, rkyv::Deserialize)
+)]
+#[cfg_attr(feature = "rkyv-impl", derive(bytecheck::CheckBytes))]
+#[cfg_attr(feature = "rkyv-impl", repr(C))]
+#[cfg_attr(feature = "shrink-to-fit", derive(shrink_to_fit::ShrinkToFit))]
+pub struct NodeId(#[cfg_attr(feature = "__rkyv", rkyv(omit_bounds))] u32);
+
+impl NodeId {
+    pub const DUMMY: NodeId = NodeId(u32::MAX);
+}
+
+impl Default for NodeId {
+    fn default() -> Self {
+        NodeId::DUMMY
+    }
+}
+
+impl Take for NodeId {
+    fn dummy() -> Self {
+        NodeId::DUMMY
+    }
+}

crates/swc_ecma_ast/src/expr.rs

@@ -1199,6 +1199,15 @@ pub struct TplElement {
     /// You may need to perform. `.replace("\r\n", "\n").replace('\r', "\n")` on
     /// this value.
     pub raw: Atom,
+
+    /// The string value contains lone surrogates.
+    ///
+    /// `cooked` is encoded with `\u{FFFD}` to mark the lone surrogate as an
+    /// escaped value.
+    ///
+    /// For example, a "\uD808" is a lone surrogate, and it's encoded as
+    /// `\u{FFFD}D808`.
+    pub lone_surrogates: bool,
 }
 
 impl Take for TplElement {
@@ -1208,6 +1217,7 @@ impl Take for TplElement {
             tail: Default::default(),
             cooked: None,
             raw: Default::default(),
+            lone_surrogates: false,
         }
     }
 }
@@ -1219,12 +1229,14 @@ impl<'a> arbitrary::Arbitrary<'a> for TplElement {
         let span = u.arbitrary()?;
         let cooked = Some(u.arbitrary::<String>()?.into());
         let raw = u.arbitrary::<String>()?.into();
+        let lone_surrogates = u.arbitrary::<bool>()?.into();
 
         Ok(Self {
             span,
             tail: false,
             cooked,
             raw,
+            lone_surrogates,
         })
     }
 }

crates/swc_ecma_ast/src/ident.rs

@@ -9,7 +9,8 @@ use phf::phf_set;
 use rustc_hash::FxHashSet;
 use swc_atoms::{atom, Atom, UnsafeAtom};
 use swc_common::{
-    ast_node, util::take::Take, BytePos, EqIgnoreSpan, Mark, Span, Spanned, SyntaxContext, DUMMY_SP,
+    ast_node, util::take::Take, BytePos, EqIgnoreSpan, Mark, NodeId, Span, Spanned, SyntaxContext,
+    DUMMY_SP,
 };
 
 use crate::{typescript::TsTypeAnn, Expr};
@@ -91,6 +92,7 @@ impl From<&'_ BindingIdent> for Ident {
             ctxt: bi.ctxt,
             sym: bi.sym.clone(),
             optional: bi.optional,
+            node_id: bi.node_id,
         }
     }
 }
@@ -179,6 +181,9 @@ pub struct Ident {
     #[cfg_attr(feature = "__rkyv", rkyv(omit_bounds))]
     pub ctxt: SyntaxContext,
 
+    #[cfg_attr(feature = "__rkyv", rkyv(omit_bounds))]
+    pub node_id: NodeId,
+
     #[cfg_attr(feature = "serde-impl", serde(rename = "value"))]
     pub sym: Atom,
 
@@ -565,6 +570,7 @@ impl<'a> arbitrary::Arbitrary<'a> for Ident {
             sym,
             optional,
             ctxt: Default::default(),
+            node_id: NodeId::DUMMY,
         })
     }
 }
@@ -592,6 +598,7 @@ impl Ident {
             ctxt,
             sym,
             optional: false,
+            node_id: NodeId::DUMMY,
         }
     }
 

crates/swc_ecma_ast/src/jsx.rs

@@ -5,9 +5,8 @@ use swc_common::{ast_node, util::take::Take, EqIgnoreSpan, Span, DUMMY_SP};
 use crate::{
     expr::{Expr, SpreadElement},
     ident::Ident,
-    lit::Lit,
     typescript::TsTypeParamInstantiation,
-    IdentName,
+    IdentName, Str,
 };
 
 /// Used for `obj` property of `JSXMemberExpr`.
@@ -190,12 +189,7 @@ pub enum JSXAttrName {
 #[cfg_attr(feature = "shrink-to-fit", derive(shrink_to_fit::ShrinkToFit))]
 pub enum JSXAttrValue {
     #[tag("StringLiteral")]
-    #[tag("BooleanLiteral")]
-    #[tag("NullLiteral")]
-    #[tag("NumericLiteral")]
-    #[tag("RegExpLiteral")]
-    #[tag("JSXText")]
-    Lit(Lit),
+    Str(Str),
 
     #[tag("JSXExpressionContainer")]
     JSXExprContainer(JSXExprContainer),