README.md
1# pulldown-cmark
2
3[](https://dev.azure.com/raphlinus/pulldown-cmark/_build/latest?definitionId=2&branchName=master)
4[](https://docs.rs/pulldown-cmark)
5[](https://crates.io/crates/pulldown-cmark)
6
7[Documentation](https://docs.rs/pulldown-cmark/)
8
9This library is a pull parser for [CommonMark](http://commonmark.org/), written
10in [Rust](http://www.rust-lang.org/). It comes with a simple command-line tool,
11useful for rendering to HTML, and is also designed to be easy to use from as
12a library.
13
14It is designed to be:
15
16* Fast; a bare minimum of allocation and copying
17* Safe; written in pure Rust with no unsafe blocks
18* Versatile; in particular source-maps are supported
19* Correct; the goal is 100% compliance with the [CommonMark spec](http://spec.commonmark.org/)
20
21Further, it optionally supports parsing footnotes,
22[Github flavored tables](https://github.github.com/gfm/#tables-extension-),
23[Github flavored task lists](https://github.github.com/gfm/#task-list-items-extension-) and
24[strikethrough](https://github.github.com/gfm/#strikethrough-extension-).
25
26Rustc 1.34 or newer is required to build the crate.
27
28## Why a pull parser?
29
30There are many parsers for Markdown and its variants, but to my knowledge none
31use pull parsing. Pull parsing has become popular for XML, especially for
32memory-conscious applications, because it uses dramatically less memory than
33constructing a document tree, but is much easier to use than push parsers. Push
34parsers are notoriously difficult to use, and also often error-prone because of
35the need for user to delicately juggle state in a series of callbacks.
36
37In a clean design, the parsing and rendering stages are neatly separated, but
38this is often sacrificed in the name of performance and expedience. Many Markdown
39implementations mix parsing and rendering together, and even designs that try
40to separate them (such as the popular [hoedown](https://github.com/hoedown/hoedown)),
41make the assumption that the rendering process can be fully represented as a
42serialized string.
43
44Pull parsing is in some sense the most versatile architecture. It's possible to
45drive a push interface, also with minimal memory, and quite straightforward to
46construct an AST. Another advantage is that source-map information (the mapping
47between parsed blocks and offsets within the source text) is readily available;
48you can call `into_offset_iter()` to create an iterator that yields `(Event, Range)`
49pairs, where the second element is the event's corresponding range in the source
50document.
51
52While manipulating ASTs is the most flexible way to transform documents,
53operating on iterators is surprisingly easy, and quite efficient. Here, for
54example, is the code to transform soft line breaks into hard breaks:
55
56```rust
57let parser = parser.map(|event| match event {
58 Event::SoftBreak => Event::HardBreak,
59 _ => event
60});
61```
62
63Or expanding an abbreviation in text:
64
65```rust
66let parser = parser.map(|event| match event {
67 Event::Text(text) => Event::Text(text.replace("abbr", "abbreviation").into()),
68 _ => event
69});
70```
71
72Another simple example is code to determine the max nesting level:
73
74```rust
75let mut max_nesting = 0;
76let mut level = 0;
77for event in parser {
78 match event {
79 Event::Start(_) => {
80 level += 1;
81 max_nesting = std::cmp::max(max_nesting, level);
82 }
83 Event::End(_) => level -= 1,
84 _ => ()
85 }
86}
87```
88
89There are some basic but fully functional examples of the usage of the crate in the
90`examples` directory of this repository.
91
92## Using Rust idiomatically
93
94A lot of the internal scanning code is written at a pretty low level (it
95pretty much scans byte patterns for the bits of syntax), but the external
96interface is designed to be idiomatic Rust.
97
98Pull parsers are at heart an iterator of events (start and end tags, text,
99and other bits and pieces). The parser data structure implements the
100Rust Iterator trait directly, and Event is an enum. Thus, you can use the
101full power and expressivity of Rust's iterator infrastructure, including
102for loops and `map` (as in the examples above), collecting the events into
103a vector (for recording, playback, and manipulation), and more.
104
105Further, the `Text` event (representing text) is a small copy-on-write string.
106The vast majority of text fragments are just
107slices of the source document. For these, copy-on-write gives a convenient
108representation that requires no allocation or copying, but allocated
109strings are available when they're needed. Thus, when rendering text to
110HTML, most text is copied just once, from the source document to the
111HTML buffer.
112
113When using the pulldown-cmark's own HTML renderer, make sure to write to a buffered
114target like a `Vec<u8>` or `String`. Since it performs many (very) small writes, writing
115directly to stdout, files, or sockets is detrimental to performance. Such writers can
116be wrapped in a [`BufWriter`](https://doc.rust-lang.org/std/io/struct.BufWriter.html).
117
118## Build options
119
120By default, the binary is built as well. If you don't want/need it, then build like this:
121
122```bash
123> cargo build --no-default-features
124```
125
126Or put in your `Cargo.toml` file:
127
128```toml
129pulldown-cmark = { version = "0.7", default-features = false }
130```
131
132SIMD accelerated scanners are available for the x64 platform from version 0.5 onwards. To
133enable them, build with simd feature:
134
135```bash
136> cargo build --release --features simd
137```
138
139Or add the feature to your project's `Cargo.toml`:
140
141```toml
142pulldown-cmark = { version = "0.7", default-features = false, features = ["simd"] }
143```
144
145## Authors
146
147The main author is Raph Levien. The implementation of the new design (v0.3+) was completed by Marcus Klaas de Vries.
148
149## Contributions
150
151We gladly accept contributions via GitHub pull requests. Please see
152[CONTRIBUTING.md](CONTRIBUTING.md) for more details.
153
){kind=link}
){kind=link}