Converts between Markdown (MDX) and Textile, for those of you unfortunate enough to have to maintain a legacy system that relies on Textile.

Pure Ruby, no runtime dependencies.

Add to your Gemfile:

gem 'mdx-tex'

bundle install

Or install directly:

gem install mdx-tex

require 'mdx_tex'

MdxTex.to_textile(markdown: '# Hello **world**')
# => "h3. Hello *world*"

# Override options per call
MdxTex.to_textile(markdown: '# Hello', header_level: 'h1')
# => "h1. Hello"

# Or set them globally
MdxTex.configure do |config|
  config.header_level = 'h2'
  config.list_depth = 1
end

# Per-call options take precedence over global config
MdxTex.to_textile(markdown: '- item', list_depth: 2)

MdxTex.to_markdown(textile: 'h3. Hello *world*')
# => "### Hello **world**"

Unlike to_textile, to_markdown takes no options. Heading level maps directly (h1 -> #, h2 -> ##, ..., h6 -> ######), the list base depth is auto-detected per document, and ordered list numbers are computed from per-depth counters that reset on blank or interrupting lines.

Adds to_textile and to_markdown directly to String. Off by default.

Any Ruby app:

require 'mdx_tex/core_ext/string'

'# Hello **world**'.to_textile
# => "h3. Hello *world*"

'- item'.to_textile(list_depth: 1)
# => "* item"

'h3. Hello *world*'.to_markdown
# => "### Hello **world**"

Or load it through config:

MdxTex.configure do |config|
  config.enable_string_extension = true
end
MdxTex.load_string_extension!

Rails (config/initializers/mdx_tex.rb):

MdxTex.configure do |config|
  config.enable_string_extension = true
end

The Railtie loads this automatically after initialization.

Markdown -> Textile converts all headings (# through ######) to the same Textile tag. The number of # doesn't matter; header_level controls which tag to use (default: h3).

Requires a space after #. #NoSpace won't convert. Heading hierarchy is not preserved (## and ### both produce the same tag) — we plan to fix this in a future version.

Textile -> Markdown maps the Textile tag directly to the matching number of #s.

Requires a space after hN.. h3.NoSpace won't convert.

Markdown -> Textile turns **text** and __text__ into *text*. Multiple spans and mixed delimiters work on the same line.

Textile -> Markdown turns *text* into **text**. Whitespace-padded asterisks (* text *) are not bold per Textile rules and are left unchanged.

Markdown -> Textile converts - item to * item. Nesting uses 2-space indentation. list_depth controls the base asterisk count (default: 3).

Textile -> Markdown auto-detects the depth-1 base per document (the smallest leading-asterisk run anywhere in the input) and emits Markdown with 2-space indentation per nesting level. This makes the conversion robust to input authored at different base depths.

Markdown -> Textile converts 1. item to # item. The actual number is ignored. Nesting uses 2-space indentation.

Textile -> Markdown auto-detects the depth-1 base (smallest leading-# run) and emits incrementing numbers per depth. Counters reset on any non-ordered-list line (blank line, plain text, header, unordered item), so each separate list starts at 1. again.

The following syntax passes through as-is in both directions (no conversion):

• Italic (*text* / _text_ in Markdown; _text_ in Textile)
• Strikethrough (~~text~~ / -text-)
• Links ([text](url) / "text":url)
• Images (![alt](url) / !url!)
• Inline code (`code`) and code blocks
• Blockquotes (> / bq.)
• Tables
• Horizontal rules (---)
• Task lists (- [ ], - [x])
• Textile-specific syntax: attributes (h3{color:red}.), block modifiers other than hN., footnotes, acronyms

Configuration applies to to_textile only. to_markdown auto-detects everything from the input and takes no options.

Bad values raise InvalidHeaderLevelError or InvalidListDepthError.

Markdown input:

### **Title**

- item one
- **item two**
  - nested with **bold**

1. **first**
2. second

After MdxTex.to_textile (default config):

h3. *Title*

*** item one
*** *item two*
**** nested with *bold*

# *first*
# second

After feeding that back through MdxTex.to_markdown:

### **Title**

- item one
- **item two**
  - nested with **bold**

1. **first**
2. second

The round-trip is stable for canonical inputs: headers at the configured header_level, bold using ** (not __), unordered lists at depth 1 matching list_depth, and ordered list numbers already incrementing per depth.

git clone https://github.com/gbudiman/mdx-tex.git
cd mdx-tex
bundle install

Uses overcommit for pre-commit RuboCop. After cloning:

bundle exec overcommit --install
bundle exec overcommit --sign

bundle exec rspec

bundle exec rubocop

bin/release patch   # 0.1.10 → 0.1.11
bin/release minor   # 0.1.10 → 0.2.0
bin/release major   # 0.1.10 → 1.0.0

Runs specs, bumps the version, commits, tags, and pushes. GitHub Actions handles the RubyGems publish.

MIT