Skip to content

An open source design system that's fully customizable and agent ready

TypeScript

fix(Markdown, richtext): read an indented table with its own columns (#7120)

Cindy Zhang51bf85cd6c8b
diff --git a/.changeset/markdown-indented-table-columns.md b/.changeset/markdown-indented-table-columns.md
new file mode 100644
index 0000000000..87ede89234
--- /dev/null
+++ b/.changeset/markdown-indented-table-columns.md
@@ -0,0 +1,8 @@
+---
+'@astryxdesign/core': patch
+---
+
+[fix] Markdown: read an indented table with its own columns
+@cixzhang
+
+`Markdown` no longer adds an empty first column to a table whose rows are indented, such as ` | a | b |`. The indentation before a row's first pipe used to read as a cell of its own.
diff --git a/packages/core/src/Markdown/indentedTables.test.tsx b/packages/core/src/Markdown/indentedTables.test.tsx
new file mode 100644
index 0000000000..a174ff1ec2
--- /dev/null
+++ b/packages/core/src/Markdown/indentedTables.test.tsx
@@ -0,0 +1,53 @@
+// Copyright (c) Meta Platforms, Inc. and affiliates.
+
+import {describe, expect, it} from 'vitest';
+import {
+ createIncrementalState,
+ parseMarkdown,
+ parseMarkdownAst,
+ parseMarkdownIncremental,
+} from './parser';
+
+type Json = {type: string; value?: string; children?: Json[]};
+
+/** Each table row's cell texts. */
+function tableRows(markdown: string): string[][] {
+ const text = (node: Json): string =>
+ node.value ?? (node.children ?? []).map(text).join('');
+ const table = (parseMarkdownAst(markdown).children as unknown as Json[]).find(
+ node => node.type === 'table',
+ );
+ return (table?.children ?? []).map(row => (row.children ?? []).map(text));
+}
+
+describe("a table row's indentation opens no cell", () => {
+ it.each([
+ ['indented two spaces', ' | a | b |\n | - | - |\n | 1 | 2 |'],
+ ['indented three spaces', ' | a | b |\n | - | - |\n | 1 | 2 |'],
+ ['indented with a tab', '\t| a | b |\n\t| - | - |\n\t| 1 | 2 |'],
+ ['indented without outer pipes', ' a | b\n - | -\n 1 | 2'],
+ ['at the margin, unchanged', '| a | b |\n| - | - |\n| 1 | 2 |'],
+ ])('reads a table %s as two columns', (_, markdown) => {
+ expect(tableRows(markdown)).toEqual([
+ ['a', 'b'],
+ ['1', '2'],
+ ]);
+ });
+
+ it('keeps an empty first cell written inside the pipes', () => {
+ expect(tableRows(' | | b |\n | - | - |\n | | 2 |')).toEqual([
+ ['', 'b'],
+ ['', '2'],
+ ]);
+ });
+
+ it('streams an indented table to the full parse at every character', () => {
+ const markdown = 'Intro\n\n | a | b |\n | - | - |\n | 1 | 2 |\n\nAfter';
+ const state = createIncrementalState();
+ let streamed = parseMarkdownIncremental('', state);
+ for (let end = 1; end <= markdown.length; end++) {
+ streamed = parseMarkdownIncremental(markdown.slice(0, end), state);
+ }
+ expect(streamed).toEqual(parseMarkdown(markdown));
+ });
+});
diff --git a/packages/core/src/Markdown/parser.ts b/packages/core/src/Markdown/parser.ts
index 996150cc6e..57149010d0 100644
--- a/packages/core/src/Markdown/parser.ts
+++ b/packages/core/src/Markdown/parser.ts
@@ -3197,8 +3197,12 @@ function sourceEndsInParagraph(source: string, opts: ResolvedOptions): boolean {
function splitTableRow(line: string): string[] {
let start = 0;
let end = line.length;
- if (line.startsWith('|')) {
- start = 1;
+ // A row's indentation, and its leading pipe, open no cell.
+ while (start < end && (line[start] === ' ' || line[start] === '\t')) {
+ start++;
+ }
+ if (line[start] === '|') {
+ start++;
while (start < end && line[start] === ' ') {
start++;
}
diff --git a/packages/richtext/src/markdownIndentedTables.test.ts b/packages/richtext/src/markdownIndentedTables.test.ts
new file mode 100644
index 0000000000..372485a5dc
--- /dev/null
+++ b/packages/richtext/src/markdownIndentedTables.test.ts
@@ -0,0 +1,43 @@
+// Copyright (c) Meta Platforms, Inc. and affiliates.
+
+import {describe, expect, it} from 'vitest';
+import {parseMarkdownAst} from '@astryxdesign/core/Markdown/parser';
+import {markdownToEditorStateJSON} from './markdownSerializers';
+
+type Json = {type: string; text?: string; value?: string; children?: Json[]};
+
+const textOf = (node: Json): string =>
+ node.text ?? node.value ?? (node.children ?? []).map(textOf).join('');
+
+/** Each row's cell texts in the first table found. */
+function rowsOf(nodes: ReadonlyArray<Json>, rowType: string): string[][] {
+ const table = nodes.find(node => node.type === 'table');
+ return (table?.children ?? [])
+ .filter(row => row.type === rowType)
+ .map(row => (row.children ?? []).map(textOf));
+}
+
+const richRows = (markdown: string) =>
+ rowsOf(
+ (JSON.parse(markdownToEditorStateJSON(markdown)) as {root: Json}).root
+ .children ?? [],
+ 'tablerow',
+ );
+
+const coreRows = (markdown: string) =>
+ rowsOf(parseMarkdownAst(markdown).children as unknown as Json[], 'tableRow');
+
+describe('an indented table imports with its own columns, as core reads it', () => {
+ it.each([
+ ' | a | b |\n | - | - |\n | 1 | 2 |',
+ '\t| a | b |\n\t| - | - |\n\t| 1 | 2 |',
+ '| a | b |\n| - | - |\n| 1 | 2 |',
+ ])('imports %j as two columns', markdown => {
+ const expected = [
+ ['a', 'b'],
+ ['1', '2'],
+ ];
+ expect(richRows(markdown)).toEqual(expected);
+ expect(coreRows(markdown)).toEqual(expected);
+ });
+});
diff --git a/packages/richtext/src/markdownTable.ts b/packages/richtext/src/markdownTable.ts
index c0be205561..339cfdbcfc 100644
--- a/packages/richtext/src/markdownTable.ts
+++ b/packages/richtext/src/markdownTable.ts
@@ -66,15 +66,19 @@ const CELL_TRANSFORMERS: Array<Transformer> = [
type ColumnAlignment = 'left' | 'center' | 'right' | null;
/**
- * Splits a row into trimmed cell sources: one leading pipe, trailing spaces,
- * and one trailing pipe are dropped, and the rest splits on unescaped pipes.
- * An escaped pipe stays escaped for the cell's inline import.
+ * Splits a row into trimmed cell sources: the row's indentation, one leading
+ * pipe, trailing spaces, and one trailing pipe are dropped, and the rest
+ * splits on unescaped pipes. An escaped pipe stays escaped for the cell's
+ * inline import.
*/
function splitTableRow(line: string): Array<string> {
let start = 0;
let end = line.length;
- if (line.startsWith('|')) {
- start = 1;
+ while (start < end && (line[start] === ' ' || line[start] === '\t')) {
+ start++;
+ }
+ if (line[start] === '|') {
+ start++;
while (start < end && line[start] === ' ') {
start++;
}