Support tags in dollar-quoted Postgres strings

FIX: Properly support tags in PostgreSQL `4073` quoted strings.
codemirror · Mar 19, 2024 · 7ca27f7 · 7ca27f7
1 parent a615a63
commit 7ca27f7
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 9 deletions.
diff --git a/src/tokens.ts b/src/tokens.ts
@@ -57,11 +57,22 @@ function readLiteral(input: InputStream, endQuote: number, backslashEscapes: boo
   }
 }
 
-function readDoubleDollarLiteral(input: InputStream) {
-  for (;;) {
-    if (input.next < 0 || input.peek(1) < 0) return
-    if (input.next == Ch.Dollar && input.peek(1) == Ch.Dollar ) { input.advance(2); return }
-    input.advance()
+function readDoubleDollarLiteral(input: InputStream, tag: string) {
+  scan: for (;;) {
+    if (input.next < 0) return console.log("exit at end", input.pos)
+    if (input.next == Ch.Dollar) {
+      input.advance()
+      for (let i = 0; i < tag.length; i++) {
+        if (input.next != tag.charCodeAt(i)) continue scan
+        input.advance()
+      }
+      if (input.next == Ch.Dollar) {
+        input.advance()
+        return
+      }
+    } else {
+      input.advance()
+    }
   }
 }
 
@@ -195,9 +206,13 @@ export function tokensFor(d: Dialect) {
     if (inString(next, Space)) {
       while (inString(input.next, Space)) input.advance()
       input.acceptToken(whitespace)
-    } else if (next == Ch.Dollar && input.next == Ch.Dollar && d.doubleDollarQuotedStrings) {
-      readDoubleDollarLiteral(input)
-      input.acceptToken(StringToken)
+    } else if (next == Ch.Dollar) {
+      let tag = readWord(input, "")
+      if (input.next == Ch.Dollar) {
+        input.advance()
+        readDoubleDollarLiteral(input, tag)
+        input.acceptToken(StringToken)
+      }
     } else if (next == Ch.SingleQuote || next == Ch.DoubleQuote && d.doubleQuotedStrings) {
       readLiteral(input, next, d.backslashEscapes)
       input.acceptToken(StringToken)

diff --git a/test/test-tokens.ts b/test/test-tokens.ts
@@ -34,6 +34,10 @@ describe("Parse PostgreSQL tokens", () => {
   it("parses double dollar quoted string literals", () => {
     ist(parser.parse("SELECT $$hello$$"), 'Script(Statement(Keyword,String))')
   })
+
+  it("parses tagged double dollar quoted string literals", () => {
+    ist(parser.parse("SELECT $body$\nhello\n$body$"), 'Script(Statement(Keyword,String))')
+  })
 })
 
 describe("Parse BigQuery tokens", () => {
@@ -86,4 +90,4 @@ describe("Parse PL/SQL tokens", () => {
   it("parses alternative quoting mechanism - unclosed", () => {
     ist(parser.parse("SELECT q'~foo'bar' FROM DUAL"), 'Script(Statement(Keyword,String))')
   })
-})
+})