mirror of
https://github.com/pionxzh/chatgpt-exporter.git
synced 2026-10-02 19:07:43 -05:00
Pionxzh
4c8fe6ea01
The markdown round trip re-serialized every text node, which escaped characters like `_` and `[` (`snake\_case`), decoded `<` into a bare `<`, and rewrote trailing-space line breaks as backslashes (#300). Copy text and line breaks from the source instead. The round trip still normalizes block syntax (ATX headings, `*` emphasis, backtick fences, `---` rules) and aligns tables, now by display width so CJK text and emoji line up in a monospace font. The text export no longer needs its workaround for an escaped leading `[`.
37 lines
1.4 KiB
TypeScript
37 lines
1.4 KiB
TypeScript
import { describe, expect, it } from 'vitest'
|
|
import { fromMarkdown, toMarkdown } from '../src/utils/markdown'
|
|
|
|
function format(input: string) {
|
|
return toMarkdown(fromMarkdown(input), input)
|
|
}
|
|
|
|
describe('markdown re-serialization', () => {
|
|
it('normalizes block syntax', () => {
|
|
expect(format('Title\n=====\n\n_a_ and __b__\n\n***\n\n~~~\ncode\n~~~'))
|
|
.toBe('# Title\n\n*a* and **b**\n\n---\n\n```\ncode\n```\n')
|
|
})
|
|
|
|
it('keeps text as written instead of escaping it', () => {
|
|
expect(format('snake_case and [brackets] and 1 * 2 and #tag'))
|
|
.toBe('snake_case and [brackets] and 1 * 2 and #tag\n')
|
|
})
|
|
|
|
it('keeps character references', () => {
|
|
expect(format('<div> & ©')).toBe('<div> & ©\n')
|
|
})
|
|
|
|
it('keeps trailing-space line breaks (#300)', () => {
|
|
expect(format('one \ntwo')).toBe('one \ntwo\n')
|
|
})
|
|
|
|
it('does not repeat container prefixes on continuation lines', () => {
|
|
expect(format('> first line\n> second line')).toBe('> first line\n> second line\n')
|
|
expect(format('- item\n continued')).toBe('- item\n continued\n')
|
|
})
|
|
|
|
it('aligns tables by display width', () => {
|
|
expect(format('| 名稱 | n |\n| --- | --- |\n| 蘋果 | 1 |\n| apple | 2 |'))
|
|
.toBe('| 名稱 | n |\n| ----- | - |\n| 蘋果 | 1 |\n| apple | 2 |\n')
|
|
})
|
|
})
|