ftExtra 0.3.1

as_paragraph_md supports empty string, "", as an input (#68).
colformat_md applies on actual content of flextable rather than input dataset (#72).
Improved support for the underline syntax in Pandoc’s Markdown (#73).

ftExtra 0.3.0

colformat_md and as_paragraph_md gains replace_na = "" as a new default parameter. Previously, NA are printed as "NA" (#63).

Drop supports for Pandoc versions < 2.0.6 in order to reduce maintenance costs. (#59).

colformat_md supports multiple paragraphs by collapsing them with a separator given to the .sep argument (default: `“”) (#43).
colformat_md can now automatically add citations to reference on R Markdown (#48).
colformat_md can now contain cells with multiple footnotes. In addition, footnote keys are placed exactly the same place of source. The previous implementation moved the keys to the end of corresponding cells (#51).
as_paragraph_md supports variety of input formats by checking enabled extensions (#54).

Fix colformat_md(pandoc_args = c("--bibliography", "example.bib")) should work even if bibliography field is missing from the YAML front matter (#41).

Drop jpeg package from suggests, which is formerly used in a vignette (#42).
colformat_md converts cells to Pandoc’s AST by a single call of Pandoc. Previously it called pandoc for each cell. This change improves performance around 15X faster (#46). Note that #46 broken support for some input formats (e.g., commonmark) and old Pandoc (< 2.0.6), but the regression is already fixed (#53, #57).
as_paragraph_md() internally uses flextable::chunk_dataframe() in order to avoid potential problems if the flextable package changes the chunk structure in the future (#56, thanks @davidgohel).

Fix math not rendered on Windows with Pandoc < 2.10 (#33)
math.lua requires Pandoc >= 2.7.3 (#35)
Fix Pandoc < 2.7.3 raising error message “option –lua-filter requires an argument SCRIPTPATH”. This is a regression from #35 (#37)
Fix CRAN warning, “file ‘https://www.r-project.org/logo/Rlogo.png’ can not be found.” (#38)

Support markdown footnote with colformat_md. Currently, one footnote per a cell is allowed, and it must be located at the end of the cell content (#22).
Add footnote_options() to configure options for footnotes (#23).
Support formatting markdown texts on header with colformat_md(part = "header") (#23).
Support single- and double-quotes by a lua filter (dfc82e0).
Support attributes with Span, Link, and Code. Useful attributes include the .underline class and the color, shading.color, and font.family attributes (#24).
Support inline code (#25).
Add the md_extensions and pandoc_args arguments to colformat_md() and as_paragraph_md() (#26, #29).
Support citation. Bibliography can be inherited from YAML front matter of an Rmd file (#27, #29).
Support math on macOS and Linux (#30), and on Windows (#31).
Fix character corruptions on Windows by forcing UTF8 (#31).
Use pandoc citeproc when YAML frontmatter contains the bibliography field (#31).

Patched to pass “CRAN Package Check Results for Package” where r-patched-solaris-x86 fails because of missing pandoc.

Markdown parser
- supports more syntax
  - Links
  - Inline images
  - Soft line breaks
  - Hard line breaks (e.g., colformat_md(as_flextable(data.frame('a\\\nb', stringsAsFactors = FALSE)))).
- stops with error if there are multiple paragraphs or block elements other than the paragraph.
colformat_md()
- controls colors of links for example by colformat_md(auto_color_link = 'red').
- gains .from argument to specify Pandoc’s --from argument.
- gains j argument so that users can choose which columns to be treated as markdown based on dplyr::select()’s semantics.
Add as_paragraph_md which parses markdown texts on selected cells rather than selected columns. Use this function within flextable::compose.

Add colformat_md() which parses markdown text in the body of the flextable object.
Add separate_header() which separates header into multiple rows based on regular expression.
Add span_header() which separates header into multiple rows based on regular expression, and spans them if the adjacent values share the same value.
Add with_blanks() which inserts blank columns based on the semantics of dplyr::select.
Add as_flextable.data.frame to convert data frames to flextable.
Add as_flextable.grouped_df to convert grouped data frames to flextable.