Discount is free software released under the terms of a BSD-style license.
If you find it useful, please consider making a contribution to help support onward development.
Discount 2.2.0, released 26-Feb-2016.
This is my implementation of John Gruber’s Markdown text to html language. There’s not much here that differentiates it from any of the existing Markdown implementations except that it’s written in C instead of one of the vast flock of scripting languages that are fighting it out for the Perl crown.
Markdown provides a library that gives you formatting functions suitable for marking down entire documents or lines of text, a command-line program that you can use to mark down documents interactively or from a script, and a tiny (3 programs so far) suite of example programs that show how to fully utilize the markdown library.
My markdown also does, by default, various smartypants-style substitutions.
markdown program is a trivial compiler that reads in a markdown
file and writes out a html document or – if you use the
-d flag –
an outline showing the parse tree. It does have a few options;
-dis, as previously mentioned, the flag that makes markdown produce a parse tree instead of a html document.
-F <flags>sets various flags that change how markdown works. The flags argument is a somewhat less than obvious bitmask – for example,
markdownto not do the smartypants translations on the output. (there are cases – like running the test suite – where this is a useful feature.)
-o filetells markdown to write the output to
-V tells you a markdown version number and how the package
was configured. For example
$ markdown -V markdown: discount 1.0.0 DL_TAG HEADER TAB=8
There are 17 public functions in the markdown library, broken into three categories:
MMIOT *mkd_in(FILE *f, int flags) reads a markdown input file
and returns a MMIOT containing the preprocessed document.
(which is then fed to
markdown() for final formatting.)
MMIOT *mkd_string(char *bfr, int size, int flags) reads the
markdown input file that’s been written into
returns a preprocessed blob suitable for feedin to
This function exists because annotations uses
access message files instead of traditional file i/o. (If
you’re going to port Markdown to an AS/400, this function is
the droid you’ve been looking for.)
int markdown(MMIOT *doc, FILE *out, int flags) formats a document
mkd_string()) and writes the
resulting HTML document to
int mkd_line(char *bfr, int size, char **out, int flags) allocates a
buffer, then formats the text string into that buffer.
text string, allocates a buffer,
The differences from
markdown() are it doesn’t support quoting,
footnotes (“reference links”,) multiple paragraphs, lists, code
sections, or pure html sections.
int mkd_generateline(char*bfr, int size, FILE *out, int flags)
formats the text string and writes the resulting HTML fragment to
It is exactly like
mkd_line() except that it writes the output to
int mkd_compile(MMIOT *doc, int flags) takes a document created by
mkd_string() and compiles it into a tree of block
int mkd_generatehtml(MMIOT *doc, FILE *out) generates html from
a compiled document.
int mkd_document(MMIOT *doc, char **text) returns (in
pointer to the compiled html document, and (in the return code)
the size of that document.
int mkd_css(MMIOT *doc, char **out) allocates a buffer and populates
it with any style blocks found in the document.
int mkd_generatecss(MMIOT *doc, FILE *out) prints any style blocks in
int mkd_toc(MMIOT *doc, char **out) allocates a buffer, populates it
with a table of contents, assigns it to out, and returns the length of
To get a table of contents, you must
compile() the document
MKD_TOC flag (described below)
int mkd_generatetoc(MMIOT *doc, FILE *out) writes a table of contents
to out; other than writing to a
FILE*, it operates exactly like
int mkd_dump(MMIOT *doc, FILE *f, int flags, char *title) prints
a block structure diagram of a compiled document.
void mkd_cleanup(MMIOT *doc) releases the MMIOT allocated for the
char *mkd_doc_title(MMIOT *doc)returns the
char *mkd_doc_author(MMIOT *doc)returns the
char *mkd_doc_date(MMIOT *doc)returns the
void mkd_e_url(MMIOT*, char* (callback)(char*,int,void*))sets up a callback function that is called whenever discount processes a
<link>construct. The callback function is passed a pointer to the url, the size of the url, and a data pointer (null or supplied by
void mkd_e_flags(MMIOT*, char *(callback)(char*,int,void*))sets up a callback to provide additional arguments to the tags generated by
<link>constructs. If, for instance, you wanted to add
target="_blank"to every generated url, you could just make a callback function that returned that string.
void mkd_e_free(char *, void*)is called to free any allocated memory returned by the url or flags callbacks.
void mkd_e_data(MMIOT*, void*)assigns a callback data area to the url & flags callbacks.
The flags argument in
mkd_generatehtml() is a mask of the following flag bits:
|MKD_NOLINKS|| Don’t do link processing, block
|MKD_NOIMAGE|| Don’t do image processing, block
|MKD_NOPANTS|| Don’t run
|MKD_NOHTML||Don’t allow raw html through AT ALL|
|MKD_TAGTEXT|| Process text inside an html tag; no
|MKD_NO_EXT||Don’t allow pseudo-protocols|
|MKD_CDATA|| Generate code for xml
|MKD_NORELAXED||Emphasis happens everywhere|
|MKD_NOTABLES||Don’t process PHP Markdown Extra tables.|
|MKD_TOC||Do table-of-contents processing|
|MKD_1_COMPAT||Compatability with MarkdownTest_1.0|
|MKD_SAFELINK||Paranoid check for link protocol|
|MKD_NOHEADER||Don’t process document headers|
|MKD_TABSTOP||Expand tabs to 4 spaces|
|MKD_NOALPHALIST||Forbid alphabetic lists|
|MKD_NODLIST||Forbid definition lists|
|MKD_EXTRA_FOOTNOTE||Enable PHP Markdown Extra-style footnotes.|
text\‘\’ is translated to “text”.
"double-quoted text"becomes “double-quoted text”
'single-quoted text'becomes ‘single-quoted text’
don'tis “don’t.” as well as anything-else’t. (But
foo'tbaris just foo'tbar.)
it'sis “it’s,” as well as anything-else’s (except not foo'sbar and the like.)
1/4th? ¼th. Ditto for
. . .also becomes …
A^Bbecomes AB. Complex superscripts can be enclosed in
My markdown was written so I could replace the fairly gross homemade
text to html prettifier that I wrote for annotations, so I’ve extended
it in a few ways; I’ve put support for paragraph centering in
so that I don’t have to hand enter the
(nowadays I generate a css-styled
<p> block, because that’s xhtml
compatible instead of the now-depreciated
<center> block element.)
I’ve added support for specifying image sizes, and I’ve written a
not-earthshatteringly-horrible markup extension for definition lists.
->this is a test<-
this is a test
xheight field to the image tag:
![dust mite](http://dust.mite =150x150)
=characters, then put the body of the list item on the next line, indented 4 spaces.
=hey!= This is a definition list
- This is a definition list
A definition list label is just a regular line of markdown code, so you can put links and images into it.
In discount 1.2.3, the definition list syntax has been
extended so that you can define sequential
<dt> blocks by doing
=tag1= =tag2= data.
<dt>tag1</dt> <dt>tag2</dt> <dd>data.</dd>
tag1 : data
Ordered lists with alphabetic labels (enabled by
during configuration) are supported in the same way that numeric ordered
a. first item b. second item
I wanted to be able to apply styles inline without having
to manually enter the
html. So I redid the
 code to support some new
“protocols” within my markdown:
Text will be written verbatim to the output. The protocol
was inspired by a short thread on the markdown mailing list
about someone wanting to embed LaTeX inside
<!-- --> and
finding, to their distress, that markdown mangled it.
Passing text through in comments seems to be a path to unreadable madness, so I didn’t want to do that. This is, to my mind, a better solution.
</style>blocks and set them aside for printing via
> %class%will become
">instead of a
PHP Markdown Extra-style tables are supported;
aaa | bbbb -----|------ hello|sailor
becomes the following table:
And much of the rest of the current table syntax (alignment, handling of orphan columns) follows the PHP Markdown Extra spec.
Pandoc-style document headers are supported; if the first three
lines in the document begin with a
% character, they are taken
to be a document header in the form of
% Document title % Document author % Document date
and can be retrieved by the library functions
Note that I implement Pandoc document headers as they were documented in 2008; any Pandoc changes since then will not be reflected in my implementation.
If configured with the
--with-fenced-code options, Pandoc-style
fenced code blocks are supported; blocks of code wrapped in
lines are treated as code just as if it was indented the traditional
4 spaces. Github-flavored-markdown fenced code blocks (blocks wrapped
in backtick lines) are also supported.
Neither of these formats support the github-flavored-markdown class extension where you can put a word at the end of the opening backtick line and have the block given that class.
If configured with the
--with-latex option, text wrapped in
\) is passed unchanged (except for encoding
&) to the output for processing by a LaTeX renderer.
This collides with how Markdown escapes ‘[’, ‘]’, ‘(’, and ‘)’ – if discount is built
\[ will only map to
[ if corresponding
\]s are not found in the same paragraph.
When I run the standard test suite (version 1.0.3) from
$ MARKDOWN_FLAGS=0x20004 ./MarkdownTest.pl --tidy --script=/usr/local/bin/markdown Amps and angle encoding ... OK Auto links ... OK Backslash escapes ... OK Blockquotes with code blocks ... OK Code Blocks ... OK Code Spans ... OK Hard-wrapped paragraphs with list-like lines ... OK Horizontal rules ... OK Inline HTML (Advanced) ... OK Inline HTML (Simple) ... OK Inline HTML comments ... OK Links, inline style ... OK Links, reference style ... OK Links, shortcut references ... OK Literal quotes in titles ... OK Markdown Documentation - Basics ... OK Markdown Documentation - Syntax ... OK Nested blockquotes ... OK Ordered and unordered lists ... OK Strong and em together ... OK Tabs ... OK Tidyness ... OK 22 passed; 0 failed.
When I run the old standard test suite from daringfireball,
$ MARKDOWN_FLAGS=0x22004 ./MarkdownTest.pl --tidy --script=/usr/local/bin/markdown Amps and angle encoding ... OK Auto links ... OK Backslash escapes ... OK Blockquotes with code blocks ... OK Hard-wrapped paragraphs with list-like lines ... OK Horizontal rules ... OK Inline HTML (Advanced) ... OK Inline HTML (Simple) ... OK Inline HTML comments ... OK Links, inline style ... OK Links, reference style ... OK Literal quotes in titles ... OK Markdown Documentation - Basics ... OK Markdown Documentation - Syntax ... OK Nested blockquotes ... OK Ordered and unordered lists ... OK Strong and em together ... OK Tabs ... OK Tidyness ... OK 19 passed; 0 failed.
Most of the “how to get standards compliant” changes that went in were cleaning up corner cases and blatant misreading of the spec, but there were two places where I had to do a horrible hack to get compliant:
mkd_compile()so that it would have top-level paragraphs absorb adjacent list items, but I had to retain the old (and, IMO, correct) behavior of a new list forcing a block break within indented (quoted, inside lists) blocks..
MKD_1_COMPAT(0x2000) turns it on again for testing purposes.
By default, yes, it does. The habit of compensating for broken editors that give no way to indent except for tabbing by setting tabstops to 4 is so intertwined with this language that treating tabs properly would be the moral equivalent of dropping nuclear devices into the testsuite.
But if you use a proper tabstop (8 characters), you can configure
--with-tabstop and it will expand tabs to 8
spaces. If you’ve configured your markdown like this (
TAB=8) and you need to mark up text from other
sources, you can set the input flag
MKD_TABSTOP to revert those
documents back to the icky standard 4-space tab.
To build discount, untar your selected tarball, cd into the directory it creates, then do
configure.shto generate your Makefiles. After doing this, a
makeshould give you a functional stack of programs and libraries.
Discount builds, for me, on SLS Linux, MacOS 10.5, FreeBSD 4.8, and RHEL3. It may build on Windows with mingw, but I’m not sure about that.
version 2.2.0 Many MANY tweaks over the last year, including…
amalloc()paranoia malloc library even more paranoid by putting markers at the start and the end of each allocated block.
<p></>*was splitting into 2 lines when it should have generated
mkd_compile()is called multiple times, actually recompile the document if the flags change.
__dso that a
$__cwdwith spaces in the path won’t make
make install) puke
fwrite()error check in
mkd_xxx()output functions (inspired by a patch written by Koen Punt)
mkd_document()don’t pad the generated html with a 0 unless it’s actually generated.
configure.shmessage for the
--with-(foo) variables to reflect what it’s actually doing with them.
--with-(foo), so pass those settings into theme in a #define
configure.shjust state that an option is not supported instead of dying.
version 2.1.8a When I put in the patch to …
I messed up the format string and made the mangled email address into a fixed bogus string. Sigh. Fixed (thanks to a patch from Alessandro Ghedini), updated (and I really have to expand the runtime configuration flags array to be long enough to fit 64 settings, but that’s a fix for a different day) and released.
version 2.1.8 After a year or so of letting the code sit and slowly accumulate fixes, a new version which fixes a wad of bugs and adds a few new features. Some of this code is front other people, and those changes will be marked with their names:
-, which means send output to stdout instead of to a file.
SAMPLE_PFX=discount- make install.samplesto install the sample programs with a package-specific prefix.
CFLAGS=-m32 ./configure.shby using CFLAGS for all build invokations of CC. (Nathan Phillip Brink firstname.lastname@example.org)
version 2.1.6 does nothing except for some bugfixes (and ignores some particularly scary ones that I /must/ fix soon) and adds two small features.
The bugfixes are:
Markdown does not escape...commit to simply push out the backslash and back up to the start of the
<[nonwhite]sequence, so -fnohtml will continue to work.
<br/>(via two spaces) as whitespace.
The features are that I now use styles for table cell alignment instead
align=, and that I’m using the 3-clause BSD license for this
release (because there is one widely used closed-source license
that claims that you can’t dynamically link with code that uses
the 4-clause license. Fine. I’ll 3-clause this release to make
the stupid GPL happy.)
<style>blocks as regular html.
The bugfixes are
Stop tripping over tables with leading
|s; the first
implementation of tables treated leading
|s as the end
of an empty table cell. 2.1.3 corrects this to properly
be merely decoration.
As a side-effect, you now need to have all the rows of the
table either with a leading
| or not; you cannot mix and
match them, sorry.
For some mysterious reason I was treating the
<br> tag as
a block-level html tag, so my html blockifier would split paragraphs
<br>s in them.
The table of contents code was apparently generating bad html (something I never noticed, because I never use that feature, alas!) but Stefano D'Angelo contributed a patch to clean up the generated html to make it correct.
tweaks table handling so that tables with leading
won’t end up generating empty false
<td></td>s, and that
tables with trailing
|’s won’t end up getting those pipes
included in the output.
corrects a defect in smartypants processing that has been
there since the beginning of time; I’d managed to misread
the reference smartypants documentation
and thought that 1 dash made an
– and 2 dashes
— (it’s actually 2 and 3 dashes respectively.)
John Foerch read the documentation recently, noticed it was
wrong, and sent me a note pointing it out. Whoops! But it’s
fixed now (as is the this page, which is, regrettably, the only
documentation about smartypants.)
version 220.127.116.11 corrects one small defect in block handling, plus changes the format of the output for failed tests in the test suite.
The defect in block handling is that the reference implementation has text block absorbing adjacent code, so
text text text code
<p>text text text code</p>
instead of a paragraph of text(s) followed by
The change in failed test output makes it output first the source code of the failed test, and then the differences between the expected output and the generated output.
implements PHP markdown extra-style fenced code sections,
where your chunks of code are surrounded by
instead of being indented 4 spaces.
Fenced code sections are a configuration option, not a
runtime option, and are enabled by using the
--with-fenced-code (FENCED-CODE in the version
There are a few optimizations inside
markdown.c to support
fenced code detection without slowing the parser down, but
those optimizations don’t cause any of my test cases to fail.
Version 2.1.1 is still a fairly experimental release despite
this, so take some care with it.
cleans up a small collection of warts and build bugs,
plus adds some fairly small enhancements to the
configure.shso that it generates all its test scripts in the build directory.
configure.sh; check to see if
.dSYMfolders are created when a test script is compiled, and, if so, don’t forget to delete them when they’re cleaned up.
makepagenow accepts markdown option flags a'la the
-fname, or in the
opts– I can’t initialize a bitfield on plan9 cc.
themeto ignore context-sensitivity on
older versions of the code are still available.