-
Cleaned
sourcepos
records when they come from metadata (thanks, @dmurdoch, #111). -
The markdown package is in the maintenance-only mode now. It is feature-complete, and will receive no updates except for fixing CRAN problems. New development will continue only in litedown: https://github.com/yihui/litedown.
-
Provided three internal functions
html_document
,html_vignette
, andpdf_document
as compatibility layers to functions of the same names in the rmarkdown package (thanks, @jangorecki, #108). -
The default HTML template no longer wraps meta variables
include-before
andinclude-after
inside<div></div>
, because their values may contain incomplete HTML tags, e.g.,include-before = '<div>'
andinclude-after = '</div>'
.
- Verbatim code blocks of the form
```{lang attr1 attr2 ...}
were not correctly rendered.
- Raw blocks (
```{=lang}
) were broken in the previous version when the support for code block attributes was added.
-
Added support for attributes on fenced code blocks, e.g.,
```{.lang .class2 #id attr="value"}
(thanks, @thothal, #106). -
Fixed the bug that the option
number_sections: true
doesn't work for HTML output when then input contains certain Unicode characters (thanks, @fyuniv, #104). -
Added support for rendering HTML Widgets such as ggplotly (thanks, @fyuniv, #105).
- Fixed the superfluous warning about path lengths in
mark_html()
(thanks, @kenjisato, #103).
-
The
file
argument ofmark()
will be treated as a file path only if the file exists and the value is not wrapped inI()
. Previously, it would be treated as a file path when it has a file extension, which could lead to confusing errors like #100 (thanks, @LukasWallrich). -
When there are emojis in the text,
mark()
may fail to identify and embed web resources (thanks, @tdhock, yihui/knitr#2254).
-
Added support for footnotes, fenced
Div
s, section numbers,{}
attributes for images/headings/fencedDiv
s, and appendices. Seevignette('intro', package = 'markdown')
for details. -
A lot of enhancements to the HTML slides format. See
vignette('slides', package = 'markdown')
for details. -
Added
vignette('article', package = 'markdown')
to demonstrate how to write an HTML article. -
If the input to
mark()
is a file, the output will also be a file by default. Previously the output would be text. If you wantmark()
to return text output when the input is a file, you may specify the argumentoutput = NULL
. -
The Markdown option
base64_images
has been renamed toembed_resources
. This option can take two possible values,"local"
and"https"
, meaning whether to embed local and/or web (https) resources. You can specify none, either, or both of them. Seevignette('intro', package = 'markdown')
for details. -
Removed the option
standalone
from the list of Markdown options. Please use the argumenttemplate = TRUE/FALSE
ofmark()
instead. The optionstandalone = TRUE
was equivalent totemplate = TRUE
. -
Added the option
auto_identifiers
(enabled by default) to automatically add IDs to headings, e.g.,# Hello world!
will be converted to<h1 id="hello-world">Hello world!</h1>
. You can certainly override the automatic ID by providing an ID manually via the{#id}
attribute, e.g.,# Hello world! {#hello}
. -
Renamed the
mathjax
option tojs_math
to allow for other JS math libraries. The default library was changed from MathJax to KaTeX. To continue using MathJax, you may setjs_math: mathjax
. -
Removed the option
mathjax_embed
from the list of Markdown options. To embed the MathJax library, enable"https"
in theembed_resources
option instead. Note that only MathJax v3 can be partially embedded, and lower versions cannot. -
Renamed the option
highlight_code
tojs_highlight
, and added support for an alternative syntax highlighting JS library Prism.js, which became the default. To continue using the old defaulthighlight.js
, you may set thejs_highlight
option tohighlight
. -
The default version of MathJax has been changed from v2 to v3.
-
The default version of highlight.js has been changed from 11.6.0 to 11.7.0, and the default style has been switched from
github
toxcode
.
-
Values of meta variables
title
,author
, anddate
(if provided) will be transformed to the target output format before they are passed into templates. -
Fixed the bug that the default CSS was not added to HTML output.
-
Removed dependency on the mime package.
-
Added experimental support for HTML slides:
markdown::mark_html(..., meta = list(css = c('default', 'slides'), js = 'slides'))
. If you prefer knittingRmd
documents in RStudio, you may use the output format:output: markdown::html_format: meta: css: [default, slides] js: [slides]
See https://yihui.org/en/2023/01/minimal-r-markdown/ for a demo.
-
Empty
\title{}
in LaTeX output will be removed (along with\maketitle
). -
highlight.js is loaded from https://www.jsdelivr.com/package/gh/highlightjs/cdn-release by default now. This means more languages are supported (not only R), but also means syntax-highlighting will not work offline at the moment (it will be improved in future).
-
MathJax failed to load in the previous version. The bug has been fixed now.
-
Removed the function
markdownExtensions()
.
-
Switched the underlying Markdown rendering engine from the C library sundown (which has been deprecated for a decade) to the R package commonmark (thanks, @jeroen, yihui/knitr#1329).
-
The functions
renderMarkdown()
andmarkdownToHTML()
have been renamed tomark()
andmark_html()
, respectively. The old names are still kept in this package for backward-compatibility. -
Removed the arguments
stylesheet
andfragment.only
inmark_html()
. Forstylesheet
, please use the argumentmeta = list(css = ...)
to provide the CSS stylesheet. Forfragment.only
, please usemark_html(template = FALSE)
ormark_html(options = '-standalone')
instead offragment.only = TRUE
. Currently these old arguments are still accepted internally, but may be deprecated and dropped in the long run. -
The
file
argument ofmark()
andmark_html()
can also take a character vector of Markdown text now. -
Removed functions
rendererExists()
,rendererOutputType()
, andregisteredRenderer()
. They were primarily for internal use. -
Deprecated the function
markdownExtensions()
. All extensions should be specified via theoptions
argument of functions likemark()
, e.g.,mark(options = '+table+tasklist')
. See all options on the help page?markdown::markdown_options
. -
Renamed
markdownHTMLOptions()
tomarkdown_options()
.
- Fixed the warnings "a function declaration without a prototype is deprecated in all versions of C" (#94).
- renderMarkdown() and markdownToHTML() will signal an error if the input file is not encoded in "UTF-8".
-
The default value of the encoding argument of renderMarkdown() and markdownToHTML() has been changed from getOption("encoding") to "UTF-8". The encoding of the input file will always be assumed to be UTF-8.
-
markdownToHTML() will return a character vector encoded in UTF-8 (instead of the system's native encoding) when not writing to an output file.
- Fixed clang-UBSAN and valgrind issues (thanks, @yixuan, #92).
- the MathJax CDN URL was replaced by https://www.bootcdn.cn/mathjax/
- fixed rstudio/htmltools#30: markdownToHTML() did not work with empty files (thanks, @VermillionAzure)
-
renderMarkdown() works now even if text = character(0) or ""
-
added an
encoding
argument to renderMarkdown() since multi-byte characters in renderMarkdown() did not work on Windows (thanks, Kohske Takahashi, #63) -
fixed #64: invalid 'n' argument in rpubsUpload() (thanks, Wouter van Atteveldt)
- if renderMarkdown() returns a character vector, it will be marked with the UTF-8 encoding if it contains multi-byte characters
- when an image is the only element of its parent node in the HTML output document, it is automatically centered on the page
-
images that have already been base64 encoded will not be encoded again (#61)
-
the URL of the MathJax CDN was updated to cdn.mathjax.org
-
fixed #60: MathJax may be included even if it is unnecessary when syntax highlighting is enabled (thanks, @aoles)
-
fixed a bug which may hang R when building R Markdown vignettes in a wrong locale (thanks, Dan Tenenbaum, yihui/knitr#782)
-
if both the 'file' and 'text' arguments are provided but file = NULL, e.g. markdownToHTML(file = NULL, text = ?), markdownToHTML() can throw an error indicating the file is invalid (thanks, Tyler Rinker, r-lib/pkgdown#66)
-
markdownToHTML(text = ?, output = ?) was broken (#54)
- added an argument 'encoding' to markdownToHTML() to specify the character encoding of the input markdown file, and the HTML output file is always encoded in UTF-8 now (thanks, Kohske Takahashi, #50)
-
added 'mathjax_embed' to HTML options for embedding the MathJax JavaScript in the HTML document rather than linking to it online. Note the JavaScript code is read from the http instead of https MathJax URL. Contributed by Henrik Bengtsson.
-
added another vignette to show the HTML output of the original vignette (see browseVignettes('markdown'))
-
the default CSS style was tweaked (major changes include: page width is at most 800px, more line height, slightly larger fonts, and a different syntax highlighting theme)
-
added a new argument 'template' to markdownToHTML() so that we can customize the HTML template (by default, it uses the template 'resources/markdown.html' in this package); thanks, Nacho Caballero
-
the options markdown.HTML.stylesheet and markdown.HTML.header used in markdownToHTML() can be character vectors (they will be processed by paste(x, collapse = '\n')
- the 'text' argument in markdownToHTML() and renderMarkdown() is treated as lines of input now, i.e. if 'text' is provided, it is passed to the markdown renderer as paste(text, collapse = '\n'); in the previous versions, it was processed by paste(text, collapse = '')
- added a package vignette; see browseVignettes(package = 'markdown')
- added a new argument 'header' to markdownToHTML() to insert code into the HTML header (e.g. custom CSS styles)
-
fixed #25 and #27: minor documentation problems
-
fixed #26: the HTML output file will be written relative to the current working directory now when it contains images that need to be base64 encoded
-
fixed #28: the image URL should be decoded before the image is based64 encoded
- Yihui Xie has taken over the maintainership for this package from Jeffrey Horner
- Both Pandoc title blocks and Jekyll front matter sections are skipped when rendering markdown documents.
- C/C++ is now a supported language for code block highlighting.
- 'hard_wrap' has been dropped while 'mathjax' and 'highlight_code' have been added to the default list of html options.
- fixed parsing of math equations when located at the end of a line.
- with the new 'latex_math' markdown extensions, users can include math equations using several syntaxes. For block level equations, use
$$latex ... $$ , $$ ... $$, or [ ... ]. For inline equations, use$latex...$ ,$...$ , or ( ... ).
-
the markdown extension 'ignore_math' was replaced with 'latex_math'.
-
users can now use the markdown.HTML.stylesheet option to override the package default stylesheet.
-
setting the fragment_only rendering option or the fragment.only parameter to markdownToHTML will base64 encode images if applicable. version 0.5.1 did not.
-
fixed a GUIDgenerator bug; for escaping math equations before markdown parsing begins.
-
image encoding was fixed for the case when there are more than one included in a markdown document.
-
added fragment.only parameter to markdownToHTML
-
added new html rendering options base64_images, fragment_only, mathjax, and highlight_code
-
added new markdown extension ignore_math
-
removed safelink from default html rendering options
-
the default html rendering options are now hard_wrap, use_xhtml, smartypants, and base64_images.
- fixed syntax errors in C exports
-
added support for post-processing HTML using smartypants filter
-
added optional support for rendering a table of contents
-
changed exported C functions to use an rmd_ prefix (eliminating potential symbol conflicts with other packages)
-
changed default html rendering options to use_xhtml, hard_wrap, safelink, and smartypants
- eliminated name collision with render_markdown function in knitr