links-notation 0.20.0 → 0.21.1

package/dist/index.js

@@ -275,7 +275,8 @@ var ELLIPSIS = "...";
 
 class ParseError extends Error {
   constructor(input, error) {
-    const start = error?.location?.start ?? { offset: 0, line: 1, column: 1 };
+    const start = positionAt(input, error?.location?.start?.offset ?? 0);
+    const end = positionAt(input, error?.location?.end?.offset ?? start.offset);
     const lineText = lineAt(input, start.offset);
     const summary = `line ${start.line}, column ${start.column}: ${error.message}`;
     const snippet = quote(start.line, lineText, start.column);
@@ -283,7 +284,7 @@ class ParseError extends Error {
 ${snippet}`);
     this.name = "ParseError";
     this.cause = error;
-    this.location = error.location;
+    this.location = { ...error.location, start, end };
     this.offset = start.offset;
     this.line = start.line;
     this.column = start.column;
@@ -294,12 +295,32 @@ ${snippet}`);
 }
 function lineAt(input, offset) {
   const at = Math.max(0, Math.min(offset, input.length));
-  const start = input.lastIndexOf(`
-`, at - 1) + 1;
-  const end = input.indexOf(`
-`, start);
-  const line = input.slice(start, end === -1 ? input.length : end);
-  return line.endsWith("\r") ? line.slice(0, -1) : line;
+  const before = input.slice(0, at);
+  const start = Math.max(before.lastIndexOf(`
+`), before.lastIndexOf("\r")) + 1;
+  const end = input.slice(start).search(/[\r\n]/);
+  return input.slice(start, end === -1 ? input.length : start + end);
+}
+function positionAt(input, offset) {
+  const at = Math.max(0, Math.min(offset, input.length));
+  let line = 1;
+  let column = 1;
+  for (let index = 0;index < at; index++) {
+    if (input[index] === "\r") {
+      line++;
+      column = 1;
+      if (input[index + 1] === `
+` && index + 1 < at)
+        index++;
+    } else if (input[index] === `
+`) {
+      line++;
+      column = 1;
+    } else {
+      column++;
+    }
+  }
+  return { offset: at, line, column };
 }
 function quote(number, lineText, column) {
   const [quoted, at] = windowAround(lineText, column);
@@ -400,7 +421,7 @@ function isSubstantiveBody(content) {
         return false;
       }
     }
-    if (!/\s/.test(character)) {
+    if (!/[ \t\n\r]/.test(character)) {
       hasVisible = true;
     }
   }
@@ -2220,7 +2241,7 @@ function peg$parse(input, options) {
           return false;
         }
       }
-      if (!/\s/.test(c)) {
+      if (!/[ \t\n\r]/.test(c)) {
         hasVisible = true;
       }
     }
@@ -2334,12 +2355,7 @@ class Parser {
       return;
     if (item.children && item.children.length > 0) {
       if (item.id !== undefined && item.id !== null && (!item.values || item.values.length === 0)) {
-        const childValues = item.children.map((child) => {
-          if (child.values && child.values.length === 1) {
-            return this.transformLink(child.values[0]);
-          }
-          return this.transformLink(child);
-        });
+        const childValues = item.children.map((child) => this.transformIndentedValue(child));
         const linkWithChildren = {
           id: item.id,
           values: childValues
@@ -2371,6 +2387,23 @@ class Parser {
       }
     }
   }
+  transformIndentedValue(item) {
+    const children = item.children || [];
+    if (children.length && item.id != null && !item.values?.length) {
+      return new Link(item.id, children.map((child) => this.transformIndentedValue(child)));
+    }
+    const current = this.transformLink(item);
+    if (children.length) {
+      return new Link(current.id, [
+        ...current.values,
+        ...children.map((child) => this.transformIndentedValue(child))
+      ]);
+    }
+    if (item.id == null && item.nested === undefined && current.values.length === 1) {
+      return current.values[0];
+    }
+    return current;
+  }
   combinePathElements(pathElements, current) {
     if (pathElements.length === 0)
       return current;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "links-notation",
-  "version": "0.20.0",
+  "version": "0.21.1",
   "description": "Links Notation parser for JavaScript",
   "main": "dist/index.js",
   "types": "index.d.ts",

package/src/ParseError.js

@@ -19,7 +19,8 @@ export class ParseError extends Error {
    * @param {Error} error - The error the generated parser threw
    */
   constructor(input, error) {
-    const start = error?.location?.start ?? { offset: 0, line: 1, column: 1 };
+    const start = positionAt(input, error?.location?.start?.offset ?? 0);
+    const end = positionAt(input, error?.location?.end?.offset ?? start.offset);
     const lineText = lineAt(input, start.offset);
     const summary = `line ${start.line}, column ${start.column}: ${error.message}`;
     const snippet = quote(start.line, lineText, start.column);
@@ -29,7 +30,7 @@ export class ParseError extends Error {
     /** @type {Error} The error the generated parser threw */
     this.cause = error;
     /** @type {{start: {offset: number, line: number, column: number}}} */
-    this.location = error.location;
+    this.location = { ...error.location, start, end };
     /** @type {number} Offset of the offending position from the start of the document */
     this.offset = start.offset;
     /** @type {number} Line the offending position is on, counted from 1 */
@@ -53,10 +54,37 @@ export class ParseError extends Error {
  */
 function lineAt(input, offset) {
   const at = Math.max(0, Math.min(offset, input.length));
-  const start = input.lastIndexOf('\n', at - 1) + 1;
-  const end = input.indexOf('\n', start);
-  const line = input.slice(start, end === -1 ? input.length : end);
-  return line.endsWith('\r') ? line.slice(0, -1) : line;
+  const before = input.slice(0, at);
+  const start =
+    Math.max(before.lastIndexOf('\n'), before.lastIndexOf('\r')) + 1;
+  const end = input.slice(start).search(/[\r\n]/);
+  return input.slice(start, end === -1 ? input.length : start + end);
+}
+
+/**
+ * Count CR, LF and CRLF as line breaks, as the grammar does. Peggy's error
+ * location counts a lone CR as an ordinary character.
+ * @param {string} input - The document being parsed
+ * @param {number} offset - Offset of the position
+ * @returns {{offset: number, line: number, column: number}} The position
+ */
+function positionAt(input, offset) {
+  const at = Math.max(0, Math.min(offset, input.length));
+  let line = 1;
+  let column = 1;
+  for (let index = 0; index < at; index++) {
+    if (input[index] === '\r') {
+      line++;
+      column = 1;
+      if (input[index + 1] === '\n' && index + 1 < at) index++;
+    } else if (input[index] === '\n') {
+      line++;
+      column = 1;
+    } else {
+      column++;
+    }
+  }
+  return { offset: at, line, column };
 }
 
 /**

package/src/Parser.js

@@ -80,13 +80,9 @@ export class Parser {
         item.id !== null &&
         (!item.values || item.values.length === 0)
       ) {
-        const childValues = item.children.map((child) => {
-          // For indented children, extract the actual reference from the child's values
-          if (child.values && child.values.length === 1) {
-            return this.transformLink(child.values[0]);
-          }
-          return this.transformLink(child);
-        });
+        const childValues = item.children.map((child) =>
+          this.transformIndentedValue(child)
+        );
         const linkWithChildren = {
           id: item.id,
           values: childValues,
@@ -128,6 +124,34 @@ export class Parser {
     }
   }
 
+  // A line below an indented ID is a value in its own right. Bare single
+  // references unwrap, but named links and descendants keep their structure.
+  transformIndentedValue(item) {
+    const children = item.children || [];
+    if (children.length && item.id != null && !item.values?.length) {
+      return new Link(
+        item.id,
+        children.map((child) => this.transformIndentedValue(child))
+      );
+    }
+
+    const current = this.transformLink(item);
+    if (children.length) {
+      return new Link(current.id, [
+        ...current.values,
+        ...children.map((child) => this.transformIndentedValue(child)),
+      ]);
+    }
+    if (
+      item.id == null &&
+      item.nested === undefined &&
+      current.values.length === 1
+    ) {
+      return current.values[0];
+    }
+    return current;
+  }
+
   combinePathElements(pathElements, current) {
     if (pathElements.length === 0) return current;
     if (pathElements.length === 1) {

package/src/comments.js

@@ -155,7 +155,7 @@ function isSubstantiveBody(content) {
         return false;
       }
     }
-    if (!/\s/.test(character)) {
+    if (!/[ \t\n\r]/.test(character)) {
       hasVisible = true;
     }
   }

package/src/grammar.pegjs

@@ -83,7 +83,7 @@
           return false;
         }
       }
-      if (!/\s/.test(c)) {
+      if (!/[ \t\n\r]/.test(c)) {
         hasVisible = true;
       }
     }

package/src/parser-generated.js

@@ -1881,7 +1881,7 @@ function peg$parse(input, options) {
           return false;
         }
       }
-      if (!/\s/.test(c)) {
+      if (!/[ \t\n\r]/.test(c)) {
         hasVisible = true;
       }
     }

package/tests/IndentedIdNestedValues.test.js

@@ -0,0 +1,38 @@
+import { test, expect } from 'bun:test';
+import { Parser } from '../src/Parser.js';
+import { formatLinks } from '../src/Link.js';
+
+const cases = [
+  [
+    'outer:\n  inner:\n    value1\n    value2\n  value3',
+    '(outer: (inner: value1 value2) value3)',
+  ],
+  [
+    'statement:\n  subject:\n    I\n  verb:\n    love\n  object:\n    you:\n      very\n      much',
+    '(statement: (subject: I) (verb: love) (object: (you: very much)))',
+  ],
+  [
+    'document:\n  (metadata: title author date)\n  content:\n    paragraph1\n    (paragraph2: text (with: nested structure))\n    paragraph3',
+    '(document: (metadata: title author date) (content: paragraph1 (paragraph2: text (with: nested structure)) paragraph3))',
+  ],
+  [
+    'image:\n    file: .gitpod.Dockerfile\n    context: ./docker-content',
+    '(image: (file: .gitpod.Dockerfile) (context: ./docker-content))',
+  ],
+  [
+    'level1:\n  level2:\n    level3a\n    level3b\n  level2b',
+    '(level1: (level2: level3a level3b) level2b)',
+  ],
+  [
+    'root:\n  child1\n  child2\n    grandchild',
+    '(root: child1 (child2 grandchild))',
+  ],
+];
+
+test('indented ID retains nested values', () => {
+  for (const [source, expected] of cases) {
+    const parser = new Parser();
+    expect(formatLinks(parser.parse(source))).toBe(expected);
+    expect(parser.parse(source)).toEqual(parser.parse(expected));
+  }
+});

package/tests/IndentedIdSyntax.test.js

@@ -123,7 +123,6 @@ test('Indented ID with deeper nesting', () => {
   child2
     grandchild`;
 
-  // This should work but the grandchild will be processed as a separate nested structure
   const result = parser.parse(input);
   expect(result.length).toBeGreaterThan(0);
 
@@ -131,6 +130,7 @@ test('Indented ID with deeper nesting', () => {
   const rootLink = result[0];
   expect(rootLink.id).toBe('root');
   expect(rootLink.values.length).toBe(2);
+  expect(formatLinks(result)).toBe('(root: child1 (child2 grandchild))');
 });
 
 test('Empty indented ID should work', () => {

package/tests/Issue312Conformance.test.js

@@ -0,0 +1,46 @@
+import { test, expect } from 'bun:test';
+import { Parser } from '../src/Parser.js';
+
+const parser = new Parser();
+
+function render(node) {
+  if (!node.values || node.values.length === 0) {
+    return `<${node.id ?? ''}>`;
+  }
+  const head =
+    node.id === null || node.id === undefined ? '' : `<${node.id}>: `;
+  return `(${head}${node.values.map(render).join(' ')})`;
+}
+
+function parsesAs(source) {
+  return parser.parse(source).map(render).join('\n');
+}
+
+test('CR is a line break', () => {
+  expect(parsesAs('a\rb')).toBe('(<a>)\n(<b>)');
+  expect(parsesAs('a\r\nb')).toBe('(<a>)\n(<b>)');
+});
+
+test('only grammar whitespace makes an empty document', () => {
+  expect(parsesAs(' \t\n\r')).toBe('');
+  for (const character of ['\u0085', '\u00a0', '\u2028', '\u3000', '\ufeff']) {
+    expect(parsesAs(character)).toBe(`(<${character}>)`);
+  }
+});
+
+test('Unicode spaces are substantive even quote bodies', () => {
+  for (const character of ['\u0085', '\u00a0', '\u2028', '\u3000', '\ufeff']) {
+    expect(parsesAs(`''${character}''`)).toBe(`(<${character}>)`);
+  }
+});
+
+test('trailing indentation is document whitespace', () => {
+  expect(parsesAs('a\n  ')).toBe('(<a>)');
+  expect(parsesAs('a\n  \t')).toBe('(<a>)');
+});
+
+test('a single reference is a link value at every depth', () => {
+  expect(parsesAs('1')).toBe('(<1>)');
+  expect(parsesAs('(1)')).toBe('(<1>)');
+  expect(parsesAs('((1))')).toBe('((<1>))');
+});

package/tests/ParseErrorPosition.test.js

@@ -50,6 +50,14 @@ test('reports the line a late defect is on', () => {
   expect(error.lineText).toBe('e: f: g');
 });
 
+test('lone CR counts as one line break in diagnostics', () => {
+  const error = syntaxError('a\rb: c: d');
+
+  expect(error.line).toBe(2);
+  expect(error.column).toBe(5);
+  expect(error.lineText).toBe('b: c: d');
+});
+
 test('reports the end of the document when a group is never closed', () => {
   const error = syntaxError('a (b\n');