Go to file
Michael Teichgräber 6402ae5405 parseRule: treat the occurrence of a parse error as a valid condition
The previous implementation was too strict. A parse error can
occur if the input is a single line consisting of spaces and tabs only.
2012-05-30 01:29:37 +02:00
cmd/markdown rename: Options -> Extensions 2012-05-04 16:20:29 +02:00
misc misc/c2go.sed: update 2012-05-04 16:36:55 +02:00
tests add package tests, comprising tests from the MarkdownTest_1.0.3 suite 2012-05-04 12:42:16 +02:00
.hgignore initial import 2010-11-21 23:04:39 +01:00
.hgtags add tags go.r58 and go.r59 2011-12-03 00:47:57 +01:00
doc.go rename: Options -> Extensions 2012-05-04 16:20:29 +02:00
elemheap.go mk_element: turn global allocation into a per-parser, reusable allocation. 2012-04-29 23:54:10 +02:00
LICENSE initial import 2010-11-21 23:04:39 +01:00
Makefile add package tests, comprising tests from the MarkdownTest_1.0.3 suite 2012-05-04 12:42:16 +02:00
markdown.go parseRule: treat the occurrence of a parse error as a valid condition 2012-05-30 01:29:37 +02:00
output.go cosmetics 2012-04-30 00:58:06 +02:00
parser.leg rename: Options -> Extensions 2012-05-04 16:20:29 +02:00
parser.leg.go parser.leg.go: update 2012-05-04 16:36:37 +02:00
portid Same as previous patch but for the underlined variant. [jgm/peg-markdown] 2012-04-30 10:16:52 +02:00
README.md README.md: update 2012-05-04 16:36:04 +02:00

This is an implementation of John Gruber's markdown in Go. It is a translation of peg-markdown, written by John MacFarlane in C, into Go. It is using a modified version of Andrew J Snodgrass' PEG parser peg -- now supporting LEG grammars --, which itself is based on the parser used by peg-markdown.

Support for HTML output is implemented, but Groff and LaTeX output have not been ported. The output is identical to that of peg-markdown.

I try to keep the grammar in sync with the C version, by cherry-picking relevant changes. In the commit history the corresponding revisions have a suffix [jgm/peg-markdown].

A simple benchmark has been done by comparing the execution times of the Go binary (cmd/main.go) and the original C implementation's binary needed for processing a Markdown document, which had been created by concatenating ten Markdown syntax descriptions.

The C version is still around 1.3x faster than the Go version.

Installation

Provided you have a copy of Go 1, and git is available,

go get github.com/knieriem/markdown

should download and install the package according to your GOPATH settings.

See doc.go for an example how to use the package. There has been an API change recently: Where you previously wrote

buf, err := ioutil.ReadAll(os.Stdin)
...
doc := markdown.Parse(string(buf), markdown.Extensions{Smart: true})
doc.WriteHtml(w)

you would now write:

p := markdown.NewParser(&markdown.Extensions{Smart: true})

w := bufio.NewWriter(os.Stdout)
p.Markdown(os.Stdin, markdown.ToHTML(w))
w.Flush()

One purpose of the change is to have a Parser that can be reused between invocations of the converter.


To create the command line program markdown, run

go build github.com/knieriem/markdown/cmd/markdown

the binary should then be available in the current directory.

To run tests, type

go test github.com/knieriem/markdown

At the moment, tests are based on the .text files from the Markdown 1.0.3 test suite created by John Gruber, imported from peg-markdown. The output of the conversion of these .text files to html is compared to the output of peg-markdown.

Development

There is not yet a way to create a Go source file like parser.leg.go automatically from another file, parser.leg, when building packages and commands using the Go tool. To make markdown installable using go get, parser.leg.go has been added to the VCS.

Make parser will update parser.leg.go using leg which is part of knieriem/peg at github , if parser.leg has been changed, or if the Go file is missing. If a copy of peg is not yet present on your system, run

go get github.com/knieriem/peg

Then make parser should succeed.

Extensions

In addition to the extensions already present in peg-markdown, this package also supports definition lists (option -dlists) similar to the way they are described in the documentation of PHP Markdown Extra.

Definitions (<dd>...</dd>) are implemented using ListTight and ListLoose, on which bullet lists and ordered lists are based already. If there is an empty line between the definition title and the first definition, a loose list is expected, a tight list otherwise.

As definition item markers both : and ~ can be used.

Todo

Subdirectory Index

  • cmd/markdown command line program markdown