BREAKING! Don't parse unicode escapes in PostgreSQL unicode strings a…

…nd identifiers We can't really do that at the CST parser because the escape character can be defined outside of the string itself. We could pull it off with some trickery, but better to leave this out of the scope for this parser.
nene · Feb 11, 2024 · f715a15 · f715a15
1 parent 47705bd
commit f715a15
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 22 deletions.
diff --git a/src/parser.pegjs b/src/parser.pegjs
@@ -6,7 +6,6 @@
     parseBitBlob,
     parseTextBlob,
   } from "./utils/blob";
-  import { parseUnicodeEscapes } from "./utils/unicode";
   import {
     createBinaryExprChain,
     createBinaryExpr,
@@ -7225,7 +7224,7 @@ string_literal_unicode_single_quoted_qq
     return loc({
       type: "string_literal",
       text: text(),
-      value: parseUnicodeEscapes(str.value),
+      value: str.value,
     });
   }
 
@@ -7234,7 +7233,7 @@ string_literal_unicode_double_quoted_qq
     return loc({
       type: "string_literal",
       text: text(),
-      value: parseUnicodeEscapes(str.value),
+      value: str.value,
     });
   }
 

diff --git a/src/utils/unicode.ts b/src/utils/unicode.ts
diff --git a/test/identifier.test.ts b/test/identifier.test.ts
@@ -206,7 +206,7 @@ describe("identifier", () => {
     it("parses unicode identifier", () => {
       expect(parseExpr(`U&"d\\0061t\\+000061"`)).toMatchInlineSnapshot(`
         {
-          "name": "data",
+          "name": "d\\0061t\\+000061",
           "text": "U&"d\\0061t\\+000061"",
           "type": "identifier",
         }

diff --git a/test/literal/string.test.ts b/test/literal/string.test.ts
@@ -326,7 +326,7 @@ describe("string literal", () => {
         {
           "text": "U&'d\\0061t\\+000061'",
           "type": "string_literal",
-          "value": "data",
+          "value": "d\\0061t\\+000061",
         }
       `);
     });