feat: process embedded languages in text blocks with language identifier comments

Author: jtkiesel

packages/prettier-plugin-java/src/comments.ts

@@ -42,6 +42,9 @@ export function determineFormatterOffOnRanges(cst: JavaNonTerminal) {
 
 export function isFullyBetweenFormatterOffOn(path: AstPath<JavaNode>) {
   const { node, root } = path;
+  if (isNonTerminal(node) && node.location === undefined) {
+    return false;
+  }
   const start = parser.locStart(node);
   const end = parser.locEnd(node);
   return (

packages/prettier-plugin-java/src/printer.ts

@@ -6,19 +6,29 @@ import {
   isFullyBetweenFormatterOffOn
 } from "./comments.js";
 import {
+  embedTextBlock,
   isNonTerminal,
   isTerminal,
   printComment,
+  printTextBlock,
   type JavaNode,
   type JavaTerminal
 } from "./printers/helpers.js";
 import { printerForNodeType } from "./printers/index.js";
 
 export default {
   print(path: DistributedAstPath<JavaNode>, options, print, args) {
-    return hasTerminal(path)
-      ? path.node.image
-      : printerForNodeType(path.node.name)(path, print, options, args);
+    if (hasTerminal(path)) {
+      return path.node.tokenType.name === "TextBlock"
+        ? printTextBlock(path)
+        : path.node.image;
+    }
+    return printerForNodeType(path.node.name)(path, print, options, args);
+  },
+  embed(path: DistributedAstPath<JavaNode>) {
+    return hasTerminal(path) && path.node.tokenType.name === "TextBlock"
+      ? embedTextBlock(path)
+      : null;
   },
   hasPrettierIgnore(path) {
     const { node } = path;

packages/prettier-plugin-java/src/printers/helpers.ts

@@ -9,7 +9,7 @@ import type {
   IToken,
   StatementCstNode
 } from "java-parser";
-import type { AstPath, Doc, ParserOptions } from "prettier";
+import type { AstPath, Doc, Options, ParserOptions } from "prettier";
 import { builders } from "prettier/doc";
 import type { JavaComment } from "../comments.js";
 import parser from "../parser.js";
@@ -328,6 +328,94 @@ export function printClassType(
     });
 }
 
+export function printTextBlock(path: AstPath<JavaTerminal>) {
+  const [open, ...lines] = path.node.image.split("\n");
+  const baseIndent = findBaseIndent(lines);
+  const textBlock = join(hardline, [
+    open,
+    ...lines.map(line => line.slice(baseIndent))
+  ]);
+  const ancestor = path.getNode(16) as JavaNonTerminal | null;
+  return ancestor?.name === "variableInitializer" ||
+    (ancestor?.name === "binaryExpression" &&
+      ancestor.children.AssignmentOperator)
+    ? indent(textBlock)
+    : textBlock;
+}
+
+export function embedTextBlock(path: AstPath<JavaTerminal>) {
+  const language = findEmbeddedLanguage(path);
+  if (!language) {
+    return null;
+  }
+  const text = path.node.image
+    .replace(/^"""\n/, "")
+    .replace(/"""$/, "")
+    .replace(/\\u+([0-9a-fA-F]{4})/g, (_, hex) =>
+      String.fromCharCode(parseInt(hex, 16))
+    );
+  const unindentedText = stripIndent(text);
+  const decodedText = translateEscapes(unindentedText);
+
+  return async (
+    textToDoc: (text: string, options: Options) => Promise<Doc>
+  ) => {
+    const doc = await textToDoc(decodedText, { parser: language });
+    return group(indent(['"""', hardline, doc, hardline, '"""']));
+  };
+}
+
+function findEmbeddedLanguage(path: AstPath<JavaNode>) {
+  return path.ancestors
+    .find(
+      node =>
+        (isNonTerminal(node) && node.name === "blockStatement") ||
+        node.comments?.some(({ leading }) => leading)
+    )
+    ?.comments?.filter(({ leading }) => leading)
+    .reverse()
+    .map(
+      ({ image }) =>
+        image.match(/^(?:\/\/|\/\*)\s*language\s*=\s*([^\s]+)/)?.[1]
+    )
+    .find(language => language)
+    ?.toLowerCase();
+}
+
+function stripIndent(text: string) {
+  const lines = text.split("\n");
+  const indent = findBaseIndent(lines);
+  return lines.map(line => line.slice(indent)).join("\n");
+}
+
+function translateEscapes(text: string) {
+  return text.replace(
+    /\\(?:([bfntr"'\\])|([0-7]{1,3})|\n)/g,
+    (_, single, octal) => {
+      if (single) {
+        switch (single) {
+          case "b":
+            return "\b";
+          case "f":
+            return "\f";
+          case "n":
+            return "\n";
+          case "t":
+            return "\t";
+          case "r":
+            return "\r";
+          default:
+            return single;
+        }
+      } else if (octal) {
+        return String.fromCharCode(parseInt(octal, 8));
+      } else {
+        return "";
+      }
+    }
+  );
+}
+
 export function isBinaryExpression(expression: ExpressionCstNode) {
   const conditionalExpression =
     expression.children.conditionalExpression?.[0].children;

packages/prettier-plugin-java/src/printers/lexical-structure.ts

@@ -1,35 +1,12 @@
-import { builders } from "prettier/doc";
 import {
-  findBaseIndent,
   map,
   onlyDefinedKey,
   printSingle,
-  type JavaNodePrinters,
-  type JavaNonTerminal
+  type JavaNodePrinters
 } from "./helpers.js";
 
-const { hardline, indent, join } = builders;
-
 export default {
-  literal(path, print) {
-    const { TextBlock } = path.node.children;
-    if (!TextBlock) {
-      return printSingle(path, print);
-    }
-    const [open, ...lines] = TextBlock[0].image.split("\n");
-    const baseIndent = findBaseIndent(lines);
-    const textBlock = join(hardline, [
-      open,
-      ...lines.map(line => line.slice(baseIndent))
-    ]);
-    const ancestor = path.getNode(14) as JavaNonTerminal | null;
-    return ancestor?.name === "variableInitializer" ||
-      (ancestor?.name === "binaryExpression" &&
-        ancestor.children.AssignmentOperator)
-      ? indent(textBlock)
-      : textBlock;
-  },
-
+  literal: printSingle,
   integerLiteral: printSingle,
   floatingPointLiteral: printSingle,
   booleanLiteral: printSingle,

packages/prettier-plugin-java/test/unit-test/text-blocks/_input.java

@@ -55,4 +55,21 @@ public void print(%s object) {
     );
   }
 
+  void json() {
+    // language=json
+    String config = """
+          {   "name":"example",
+      "enabled"   :true,
+            "timeout":30}
+    """;
+
+    /* language = JSON */
+    String query = """
+         {
+       "sql":"SELECT * FROM users \
+    WHERE active=1 \
+    AND deleted=0",
+       "limit":10}
+    """;
+  }
 }

packages/prettier-plugin-java/test/unit-test/text-blocks/_output.java

@@ -52,4 +52,19 @@ public void print(%s object) {
             abc"""
     );
   }
+
+  void json() {
+    // language=json
+    String config = """
+      { "name": "example", "enabled": true, "timeout": 30 }
+      """;
+
+    /* language = JSON */
+    String query = """
+      {
+        "sql": "SELECT * FROM users WHERE active=1 AND deleted=0",
+        "limit": 10
+      }
+      """;
+  }
 }