test_runner: emit diagnostic data with test points

Also updated the lexer to better detect EOL and new lines

Author: manekinekko

lib/internal/test_runner/runner.js

@@ -137,7 +137,7 @@ class FileTest extends Test {
 
     switch (kind) {
       case TokenKind.TAP_VERSION:
-        // TODO(manekinekko): handle TAP version
+        // TODO(manekinekko): handle TAP version coming from the parser.
         // this.reporter.version(node.version);
         break;
 
@@ -149,8 +149,7 @@ class FileTest extends Test {
         this.reporter.subtest(indent, node.name);
         break;
 
-      case TokenKind.TAP_TEST_OK:
-      case TokenKind.TAP_TEST_NOTOK:
+      case TokenKind.TAP_TEST_POINT:
         // eslint-disable-next-line no-case-declarations
         const { todo, skip, pass } = node.status;
         // eslint-disable-next-line no-case-declarations
@@ -169,27 +168,20 @@ class FileTest extends Test {
             indent,
             node.id,
             node.description,
-            null,
+            details(node.diagnostics),
             directive
           );
         } else {
           this.reporter.fail(
             indent,
             node.id,
             node.description,
-            null,
+            details(node.diagnostics),
             directive
           );
         }
         break;
 
-      case TokenKind.TAP_YAML_START:
-        break; // skip
-
-      case TokenKind.TAP_YAML_END:
-        this.reporter.details(indent, details(node.diagnostics));
-        break;
-
       case TokenKind.COMMENT:
         if (indent === kDefaultIndent) {
           // Ignore file top level diagnostics

lib/internal/test_runner/tap_lexer.js

@@ -1,13 +1,17 @@
 'use strict';
 
-const { SafeSet, MathMax } = primordials;
+const { SafeSet, MathMax, StringPrototypeIncludes } = primordials;
 const {
   codes: { ERR_TAP_LEXER_ERROR },
 } = require('internal/errors');
 
+const kEOL = '';
+const kEOF = '';
+
 const TokenKind = {
   EOF: 'EOF',
   EOL: 'EOL',
+  NEWLINE: 'NewLine',
   NUMERIC: 'Numeric',
   LITERAL: 'Literal',
   KEYWORD: 'Keyword',
@@ -43,7 +47,7 @@ class Token {
       line: stream.line,
       column: MathMax(stream.column - valueLength + 1, 1), // 1 based
       start: MathMax(stream.pos - valueLength, 0), // zero based
-      end: stream.pos - 1, // zero based
+      end: stream.pos - (value === '' ? 0 : 1), // zero based
     };
 
     // EOF is a special case
@@ -116,7 +120,7 @@ class TapLexer {
     this.#source = new InputStream(source);
     this.#lastScannedToken = new Token({
       kind: TokenKind.EOL,
-      value: TokenKind.EOL,
+      value: kEOL,
       stream: this.#source,
     });
   }
@@ -132,8 +136,8 @@ class TapLexer {
         this.#lastScannedToken = token;
       }
 
-      if (token.kind === TokenKind.EOL) {
-        // Store the current chunk + EOL token
+      if (token.kind === TokenKind.NEWLINE) {
+        // Store the current chunk + NEWLINE token
         tokens.push([...chunk, token]);
         chunk = [];
       } else {
@@ -142,14 +146,11 @@ class TapLexer {
     }
 
     if (chunk.length > 0) {
-      tokens.push(chunk);
+      tokens.push([...chunk, this.#scanEOL()]);
     }
 
-    if (tokens.length > 0) {
-      tokens.at(-1).push(this.#scanEOF());
-    } else {
-      tokens.push([this.#scanEOF()]);
-    }
+    // send EOF as a separate chunk
+    tokens.push([this.#scanEOF()]);
 
     return tokens;
   }
@@ -171,8 +172,8 @@ class TapLexer {
 
     if (this.#isEOFSymbol(char)) {
       return this.#scanEOF();
-    } else if (this.#isEOLSymbol(char)) {
-      return this.#scanEOL(char);
+    } else if (this.#isNewLineSymbol(char)) {
+      return this.#scanNewLine(char);
     } else if (this.#isNumericSymbol(char)) {
       return this.#scanNumeric(char);
     } else if (this.#isDashSymbol(char)) {
@@ -196,7 +197,7 @@ class TapLexer {
     );
   }
 
-  #scanEOL(char) {
+  #scanNewLine(char) {
     // In case of odd number of ESCAPE symbols, we need to clear the remaining
     // escape chars from the stack and start fresh for the next line.
     this.#escapeStack = [];
@@ -205,18 +206,26 @@ class TapLexer {
     this.#isComment = false;
 
     return new Token({
-      kind: TokenKind.EOL,
+      kind: TokenKind.NEWLINE,
       value: char,
       stream: this.#source,
     });
   }
 
+  #scanEOL() {
+    return new Token({
+      kind: TokenKind.EOL,
+      value: kEOL,
+      stream: this.#source,
+    });
+  }
+
   #scanEOF() {
     this.#isComment = false;
 
     return new Token({
       kind: TokenKind.EOF,
-      value: TokenKind.EOF,
+      value: kEOF,
       stream: this.#source,
     });
   }
@@ -292,7 +301,7 @@ class TapLexer {
 
     // If we encounter a hash symbol at the beginning of a line,
     // we consider it as a comment
-    if (!lastCharacter || this.#isEOLSymbol(lastCharacter)) {
+    if (!lastCharacter || this.#isNewLineSymbol(lastCharacter)) {
       this.#isComment = true;
       return new Token({
         kind: TokenKind.COMMENT,
@@ -371,7 +380,12 @@ class TapLexer {
   }
 
   #scanTAPKeyword(word) {
-    if (word === 'TAP' && this.#lastScannedToken.kind === TokenKind.EOL) {
+    const isLastScannedTokenEOLorNewLine = StringPrototypeIncludes(
+      [TokenKind.EOL, TokenKind.NEWLINE],
+      this.#lastScannedToken.kind
+    );
+
+    if (word === 'TAP' && isLastScannedTokenEOLorNewLine) {
       return new Token({
         kind: TokenKind.TAP,
         value: word,
@@ -395,7 +409,7 @@ class TapLexer {
       });
     }
 
-    if (word === 'not' && this.#lastScannedToken.kind === TokenKind.EOL) {
+    if (word === 'not' && isLastScannedTokenEOLorNewLine) {
       return new Token({
         kind: TokenKind.TAP_TEST_NOTOK,
         value: word,
@@ -406,7 +420,7 @@ class TapLexer {
     if (
       word === 'ok' &&
       (this.#lastScannedToken.kind === TokenKind.TAP_TEST_NOTOK ||
-        this.#lastScannedToken.kind === TokenKind.EOL)
+        isLastScannedTokenEOLorNewLine)
     ) {
       return new Token({
         kind: TokenKind.TAP_TEST_OK,
@@ -415,7 +429,7 @@ class TapLexer {
       });
     }
 
-    if (word === 'pragma' && this.#lastScannedToken.kind === TokenKind.EOL) {
+    if (word === 'pragma' && isLastScannedTokenEOLorNewLine) {
       return new Token({
         kind: TokenKind.TAP_PRAGMA,
         value: word,
@@ -476,7 +490,7 @@ class TapLexer {
     return char === undefined;
   }
 
-  #isEOLSymbol(char) {
+  #isNewLineSymbol(char) {
     return char === '\n' || char === '\r';
   }