html_paged() and other formats using paged.js now informs knitr that its JS and CSS dependencies will be loaded by setting the is.paged.js global knit option. Use isTRUE(knitr::opts_knit$get("is.paged.js")) to detect it (thanks, @sylvaine31, @davidgohel, #332).Set overflow-x: clip in default.css to fix an issue with paged.js rendering in recent browser (thanks, @jimjam-slam, #292, pagedjs/pagedjs#84).
chrome_print() now attempts to remove the output file before printing to it and throws a clear error if it cannot be removed, as this is potential sign of a locked file, e.g a pdf opened in a PDF reader on Windows (thanks, @aito123, #124, #297).
Figure inserted using markdown syntax and having a caption with (#fig:lab) are now correctly listed in the LOF (thanks, @adamvi, #283).
Fix an issue in jss_paged() with Pandoc 2.17 and above.
Fix an issue in html_paged() with LOT and LOF not showing anymore with Pandoc 2.17 and above (thanks, @adamvi, #280).
Document that the html_letter() output format modifies the default value of the fig_caption parameter and allow the user to change it (thanks, @NewGraphEnvironment, #234).
Created an option pagedown.chrome.log to ease chrome_print() debugging: when set to TRUE, the logs of Chrome are saved in the working directory.
Added support for Sass files with .scss and .sass extension (thanks, @felipecrp, #244).
loft.lua filter following changes in Pandoc 2.16 (#249).find_chrome() now searches for Microsoft Edge on Windows. That allows chrome_print() to work seamlessly with Microsoft Edge (thanks, @cderv, #160 and #225).chrome_print(), when printing a document to PDF, the default transfer mode between Chrome and R now uses a stream when this option is available in Chrome. This change ensures that PDF files of any size can be generated (#206 and #224).knitr::kables() are now vertically aligned by default in html_paged(), thesis_paged() and jss_paged() output formats (thanks, @cderv and @andrew-fuller, #214).Fixed a bug introduced in html_paged() by pagedown 0.14, the content of each link was lost if it was not textual (thanks, @mcanouil, #226).
Fixed figure and table numbering format in jss_paged() (thanks, @gglee4ai, #184).
html_paged(), added two new options lot-unlisted and lof-unlisted. If lot-unlisted (resp. lof-unlisted) is set to TRUE in the YAML options, the list of tables (resp. figures) will not be included in a table of contents (thanks, @beatrizmilz, #221).In html_paged(), links are now sanitized to avoid impossible line breaking of long URLs in a justified text (thanks, @julientaq).
In order to be compatible with the flextable package, pagedown::html_paged() sets the value of the ft.shadow chunk option to FALSE by default (thanks, @tvroylandt and @davidgohel, #216).
The fig_caption argument is no longer hard-coded to FALSE in html_resume() (thanks, @nplatonov, #208).
The multiple use of an abbreviation no longer ends up with duplicated entries in the list of abbreviations (thanks, @linogaliana, #218).
The default value of the counter-reset CSS property is correctly set to 0 instead of 1 (see https://developer.mozilla.org/en-US/docs/Web/CSS/counter-reset). To reset a page CSS counter to 1, the following declaration must be used: counter-reset: page 1 (#202).
Numbered example lists (https://pandoc.org/MANUAL.html#numbered-example-lists) are correctly numbered (thanks, @atusy, #122 and #202).
Periods are now supported in titles (thanks, @yves-amevoin and @martinschmelzer, #84, #185 and #202).
Parts titles in the table of contents no longer crash chrome_print().
chrome_print() is now compatible with the stream transfer mode which can be used to generate large PDF files (#205).
chrome_print() no longer ignores runtime exceptions in Chrome. An R warning is now raised when Chrome encounters a runtime exception (#203).
html_paged(), the title of the list of abbreviations can now be modified with the loa-title field in the YAML header (thanks, @jtrecenti, #199).The option anchor_sections is disabled internally. This option is for rmarkdown::html_document() to generate anchor links for sections and currently it does not work well for pagedown format for now (#195).
In chrome_print(), fixed a bug when the R session temporary directory and the current directory are mounted on different Linux file systems. In that case, chrome_print() failed to add an outline to the PDF and raised the warning cannot rename file ..., reason 'Invalid cross-device link' (#193).
chrome_print() no longer ignores the Chrome DevTools event Inspector.targetCrashed. An R error is now raised when Chrome crashes (#190).chrome_print() now has a new argument outline, with which the user can generate the outline bookmarks for the PDF file. Note, this feature requires Ghostscript being installed and detected by tools::find_gs_cmd() (thanks, @shrektan, #174 and #179).html_resume() template, avoid page breaks after section titles and inside subsections (thanks, @kevinrue, #170).html_resume() template, an icon inserted using inline HTML in a section title takes precedence over the default icon and the data-icon property (thanks, @Tazinho, #168).In chrome_print(), fixed some connection problems to headless Chrome: in some situations, the R session tries to connect to headless Chrome before a target is created. Now, chrome_print() controls the target creation by connecting to the Browser endpoint (thanks, @gershomtripp, #158).
In html_resume() template, vertical space is removed when details are omitted (thanks, @mrajeev08, #161).
Added support for pagebreaks: in output formats which use paged.js, a pagebreak can be forced using the LaTeX commands \newpage and \pagebreak or using the CSS classes page-break-before and page-break-after.
reveal.js presentations can be printed to PDF using chrome_print().
Using RStudio, any R Markdown HTML output formats can be directly printed to PDF by adding the line "knit: pagedown::chrome_print" to the YAML header: this line modifies the behavior of the "Knit" button of RStudio and produce both HTML and PDF documents.
chrome_print() with async = FALSE, the Chrome processes and the local web server are properly closed when the function exits. This regression was introduced in pagedown 0.6.Added support for MathJax in html_resume output format (thanks, @ginolhac, #146).
The chrome_print() function internally uses a private event loop provided by later 1.0.0 when async=FALSE (thanks, @jcheng5, #127).
Added support for lines numbering in html_paged(): lines can be numbered using the top-level YAML parameter number-lines (thanks, @julientaq, #115 and #129).
Added support for covers images in html_paged(): html_paged() gains two arguments, front_cover and back_cover, to insert images in the front and back covers. Textual contents can also be added in the covers using two special divs of classes front-cover and back-cover (thanks, @atusy, #134, #136 and #137).
When chrome_print() is used with verbose >= 1, some auxiliary informations about Paged.js rendering are printed (number of pages and elapsed time).
Added a template argument to the business_card() output format for passing a custom Pandoc template (thanks, @mariakharitonova, #135).
Fixed several bugs related to MathJax: local version of MathJax is now used when the mathjax parameter is set to "local" and self contained documents are rendered by default with MathJax without throwing any warning (#130).
In html_paged, the nodes tree is sanitized before Paged.js splits the content into pages. This should avoid duplicated content observed when break-after: avoid and break-before: avoid are used (#131).
Added an async argument to chrome_print(). When async = TRUE, chrome_print() returns a promises::promise object. This allows chrome_print() to be used inside a Shiny App (thanks, @ColinFay, #91).
Added the support for chapter prefix as in bookdown. Chapter of class chapter are prefixed with the word Chapter. This prefix can be changed for internationalization purpose (thanks, @brentthorne, #101 and #107).
Added the support for lists of abbreviations. If the document contains <abbr> HTML elements, a list of abbreviations is automatically built (thanks, @brentthorne, #102 and #107).
Added the new thesis_paged template (thanks, @brentthorne, #107).
pagedjs-. For instance, --width is replaced by --pagedjs-width. Users' stylesheets that use Paged.js CSS variables need to be updated. Bleeds and crop marks are now supported. Several bugs are fixed.The default stylesheet of html_paged() is updated to support the new argument clean_highlight_tags of bookdown::html_document2() introduced in bookdown 0.10 (thanks, @atusy, #100).
In chrome_print() the connection between the R session and headless Chrome now uses the native websocket client provided by the websocket package. The previous workaround which used a websocket server, a websocket tunnel and a browser-based websocket client is removed (reverts #74). chrome_print() runs faster due to this simplification.
browser is forced to redraw the document after Paged.js finished. This will fix wrong page references observed with Chrome and RStudio 1.2.xxxx (#35 and #46, thanks, @petermeissner).
jsonlite::toJSON() is now used in chrome_print() for building all the JSON messages sent to headless Chromium/Chrome: this guarantees that the JSON strings are valid (thanks, @ColinFay, #85 and @cderv, #88).
In uri-to-fn.lua, use a shallow copy of PANDOC_VERSION. This is required for Pandoc >= 2.7.3 which changes the type of PANDOC_VERSION (thanks, @andreaphsz, #111).
Insert page numbers after page references (thanks @atusy, #120).
With Pandoc 2.7.3, page numbers wrongly appear in code blocks.
chrome_print() has been significantly enhanced. Now it prints web pages and R Markdown documents to PDF through the Chrome DevTools Protocol instead of the simple command-line call (like in v0.1). It also supports capturing screenshots.