CHANGES IN knitr VERSION 1.35

BUG FIXES

Fixed a regression in v1.34: now blank lines in code chunks are stripped only when collapse = FALSE but no longer stripped by default when collapse = TRUE. If you prefer blank lines to be always stripped, set strip.white = TRUE globally or on the per chunk basis (thanks, @IndrajeetPatil, rstudio/rmarkdown#2220, #2046).
In knitr::combine_words(), when words is length 2 and and = "", sep will now be used (thanks, @eitsupi, #2044).

CHANGES IN knitr VERSION 1.34

NEW FEATURES

Added a package option latex.tilde so that users can customize the tilde symbol, e.g., you can set knitr::opts_knit$set(latex.tilde = "\\hlopt{\\scriptsize{\\ensuremath{\\sim}}}"). This is because the default tilde is too high in the PDF output, and it's often more desirable to use a middle tilde instead (thanks, @brry #1992, @jaredlander #492).
For the tikz engine, the class options of the standalone document classs can be specified via the chunk option engine.opts$classoption (thanks, @XiangyunHuang, #1985). The default value is tikz, i.e., \documentclass[tikz]{standalone} is used by default.
Added the ability to pass additional arguments to dvisvgm when using the tikz engine with fig.ext = "svg" by using dvisvgm.opts in engine.opts (thanks, @andrewheiss, #2039). Recent versions of dvisvgm now allow you to embed fonts into SVG files as base64-encoded WOFF files, so tikz chunks can embed fonts using like so: ```{tikz, fig.ext="svg", engine.opts=list(dvisvgm.opts = "--font-format=woff")}.
Added a new targets engine (ropensci/targets#503, @wlandau). Details: https://books.ropensci.org/targets/markdown.html.
The chunk option cache.globals can take values TRUE and FALSE now (in addition to a character vector). When FALSE, it tries to find all symbols in the code chunk, no matter if they are local or global variables. When NULL or TRUE, it tries to find all global variables (thanks, @knokknok, #1898).

MAJOR CHANGES

An error is now thrown when an inline code result is not coercible to character. This has always been the assumed behavior but it happens to be different in certain formats with unknown automatic coercion. This is now enforced to prevent any unexpected output. An inline code expression must evaluate to a character vector or an object coercible by as.character() (#2006).
The markdown package has been moved from Imports to Suggests in knitr's DESCRIPTION, which means it is no longer a hard dependency but has become a soft dependency of knitr (#1864). One consequence for package developers is that if you use the vignette engines based on markdown (such as knitr::knitr), you will have to explicitly declare the (soft) dependency on markdown, because the implicit dependency through knitr is no longer there.
For R packages that use the knitr::rmarkdown engine to build their vignettes, rmarkdown must be declared as a dependency in the package DESCRIPTION (e.g., in Suggests). Previously, knitr would fall back to using the markdown package to build vignettes if rmarkdown is not declared (#1864, #2020).
write_bib() only uses the first URL if multiple are found in a package (#2028).

MINOR CHANGES

The attribute data-external="1" will be added to <iframe>s generated by include_url() to prevent Pandoc from embedding the URL as base64 data (thanks, @IndrajeetPatil, https://stackoverflow.com/q/67477667/559676).
The chunk option strip.white = TRUE used to work only when the chunk option collapse = FALSE. Now the two options are independent, i.e., strip.white also works when collapse = TRUE (thanks, @kbvernon, #2011).
When building R Markdown vignettes but Pandoc is not available, the vignette engine will emit a message instead of a warning before falling back to using the markdown package.
The internal function is_abs_path() has been removed. Users (if any) should use the exported function xfun::is_abs_path() instead.

BUG FIXES

Fix an issue with the RStudio IDE when using knitr::include_url() or knitr::include_app() in interactive Notebook mode. This will no more cause an error but print the list object as is (thanks, @systemnova, #2015).
Fix a regression with fig.keep chunk option used in chunks with only one figure where the figure was not showing in output (thanks, @fmichonneau, #1993).
Allow vignettes to be tangled (and output compared) in R CMD check if they have a corresponding .Rout.save (thanks, @lentinj, #2018).

CHANGES IN knitr VERSION 1.33

NEW FEATURES

Exported the internal functions sew() (previously named wrap()) and is_low_change() to make it possible for other graphics systems such as rgl to work in knitr like base and grid graphics in base R (thanks, @dmurdoch, #1892 #1853).
For sql code chunks, parameterized queries will be executed through native database api if the chunk option params is provided (thanks, @byapparov, #1987).

BUG FIXES

Reverted the fix for #1595 since it caused problems in kableExtra (thanks, @bttomio, haozhu233/kableExtra#607), and applied a different fix to the original problem (i.e., add {} before [).

MAJOR CHANGES

The internal S3 generic function wrap() has been renamed to sew().

MINOR CHANGES

For package vignettes that use the vignette engine knitr::docco_linear or knitr::docco_classic, knitr will signal an error during R CMD build if the markdown package is not declared as a soft dependency in the Suggests field in DESCRIPTION. The same treatment has been applied in the previous version of knitr for the vignette engine knitr::knitr (#1864).
R CMD build will signal an error when a package uses R Markdown vignettes to be build with the rmarkdown package but the rmarkdown package is not installed (thanks, @nsheff, #1864).
The force_v1 argument of knit2html() defaults to getOption('knitr.knit2html.force_v1', FALSE) now, which means its value can be configured via a global option.

CHANGES IN knitr VERSION 1.32

NEW FEATURES

The chunk option ref.label can be used to reuse code from other code chunks in the current code chunk. Now it can also accept a character vector of chunk labels wrapped in I() so that the chunk options of other code chunks can also be applied to the current chunk. For example, if we have a chunk ```{r, chunk-a, echo=FALSE, results='asis'}, the chunk ```{r, chunk-b, ref.label=I('chunk-a')} will inherit both the code and chunk options from chunk-a (e.g., echo=FALSE and results='asis'), but chunk-b can still override these chunk options, e.g., the actual chunk options of ```{r, chunk-b, echo=TRUE} will be echo=TRUE and results='asis'. If you want to use one code chunk with several chunk options for multiple times, and each time you want to change a small subset of the chunk options, you may use this approach ref.label = I() to avoid typing most chunk options over and over again (thanks, @atusy, #1960).
The chunk option opts.label gained a new usage. It used to mean labels in knitr::opts_template (see the help page if you are not familiar with this object). Now it can also be used to specify the labels of other code chunks, from which the chunk options are to be inherited to the current chunk. For example, ```{r, chunk-b, opts.label='chunk-a'} means that chunk-b will inherit chunk options from chunk-a. Local chunk options in a chunk (if exist) will override inherited options. For the new features about both chunk options ref.label and opts.label, see the detailed documentation and examples at https://yihui.org/knitr/options/ (thanks, @atusy, #1960).
The internal code to process R code chunks was factored out into the language engine R, which can be obtained via knit_engines$get('R') now (thanks, @abhsarma, #1963).
Added arguments dir and envir to load_cache() to specify the working directory and environment into which the cached objects are loaded. By default, all cached objects are loaded into the global environment, which may not be desirable (thanks, @LTLA, #1905).
The internal function html_screenshot(), which takes screenshots of HTML widgets and Shiny apps in knitr documents, now prefers the webshot2 package over the webshot package. This is because the backend of the webshot package is PhantomJS, which has been archived since 2018. If webshot is still preferred, use the chunk option opts_chunk$set(webshot = "webshot") (thanks, @atusy #1918, @LouisStAmour #1858).
Added a new engine, bslib, which is a special scss/sass engine for writing Sass code that wants to leverage Bootstrap Sass. It's primarily designed to work with rmarkdown::html_document_base's theme parameter (when it represents a bslib::bs_theme()), but the engine can be used anywhere a bslib::bs_theme() is set globally with bslib::bs_global_set() (thanks, @cpsievert, #1666).
When the chunk option include = FALSE, error = TRUE used to be ignored, i.e., errors are signaled nonetheless (causing the knitting process to fail). To avoid this problem, you can now use the numeric value error = 0 so that knitr will not check the include option, i.e., knitting will not fail even if include = FALSE and an error has occurred in a code chunk, which you wouldn't notice (thanks, @rundel, #1914).
When processing package vignettes that require rmarkdown or markdown, we now check that it is declared as a package dependency in the DESCRIPTION file (thanks, @dmurdoch, #1919).
For tikz code chunks, a new option engine.opts$extra.preamble allows arbitrary LaTeX code to be inserted into the preamble of the template. This allows loading of additional LaTeX and tikz libraries without having to recreate a template from scratch (thanks, @krivit, #1886).

BUG FIXES

Fixed an issue where code source and results would not show when using a numeric vector in fig.keep to select all plots (thanks, @dongzhuoer @ajrgodfrey #1579, @cderv #1955).
Graphical parameter par(col =) is now preserved as expected in following chunks when knitr::opts_knit$set(global.par = TRUE) (thanks, @ArnonL, @RolandASc, #1603).
Fixed an issue with kable(format = "latex", booktabs = TRUE) when the first cell of the table stats with [ (thanks, @jonnypenarth, #1595).
For child documents, include_graphics(error = FALSE) is the default now (#1957).
For Rnw documents, the commented %\begin{document} will no longer cause trouble (thanks, @NewbieKnitter @shrektan, #1819).
Fixed an issue with the chunk option fig.alt causing figures to disappear in rmarkdown::word_document() output. This option is currently supported for HTML output only. If provided for office output in rmarkdown, it will emit a warning and be ignored (#1966).
Spaces in messages were not properly escaped in .Rnw documents (thanks, @elbersb, #1978).
The hook hook_pdfcrop() did not crop figures generated by the tikz or dot engine (thanks, @giabaio yihui/xaringan#285, @cderv #1923).

MINOR CHANGES

For Rnw documents, if a chunk's output ends with \n, knitr will no longer add another \n to it (thanks, @krivit #1958, @jpritikin 1092).
For kable(x, format = 'simple'), it no longer generates a pipe table but a simple table instead when x has only one column (thanks, @alusiani, #1968). For kable(x, format = 'pipe'), it no longer warns when x has no column names, but just generate an empty table header.
For tikz code chunks, the default LaTeX template uses \documentclass[tikz]{standalone} instead of \documentclass{article} and \usepackage{preview} now (thanks, @XiangyunHuang, #1985).

CHANGES IN knitr VERSION 1.31

NEW FEATURES

Exported the functions pandoc_to() and pandoc_from() to get the output and input Pandoc format respectively when knitting R Markdown documents (thanks, @cderv, #1951).
Added a new chunk option fig.alt for users to specify the alternative text in the alt attribute of the <img> tag (images). Previously, the alt attribute takes value from the chunk option fig.cap (it still does so if fig.alt is NULL). If there are multiple plots/images in a chunk, you can pass a character vector to fig.alt, and it will be applied to individual images (thanks, @cderv, #1900).
include_url() and include_app() can now take a vector of URLs (thanks, @cderv, #1948).
The sql engine now correctly captures error with the chunk option error = TRUE (thanks, @colearendt, rstudio/rmarkdown#1208).
The chunk option collapse = TRUE now works as expected when the chunk option attr.* or class.* is provided. By this change, The chunk option collapse = TRUE forces attr.* and class.* be NULL except for the chunk options attr.source and class.source (thanks, @aosavi @cderv @atusy, #1902 #1906).
New links added in ?knitr::kable() help page to the Rmarkdown Cookbook relevant pages.
The table label can be disabled via kable(label = NA) now (thanks, @NickCH-K, #1937).
The table.attr argument in kable(.., format = 'html') can take the value from the global option knitr.table.html.attr now (thanks, @cderv, #1922).
Added a new argument oxford_comma to the function combine_words() (thanks, @thompsonsed, #1946).
The label prefix of tables generated by kable() can be customized via the package option label.prefix in opts_knit now. This option takes a named character vector, and the default is knitr::opts_knit$set(label.prefix = c(table = 'tab:')). If you want to use a different prefix such as tbl, you may set the option to c(table = 'tbl:') (thanks, @luispfonseca, #1890).
The global option options(knitr.device.fallback = TRUE) can be used to allow the graphical device specified in the chunk option dev to fall back to other usable devices if the specified device is not operational. Users can provide a list of fallback devices via the global option, e.g., options(knitr.device.choices = list(png = c('jpeg', 'svg'), jpeg = c('tiff'))), which means the png device can fall back to jpeg and svg (the first operational device in the list is used) and jpeg can fall back to tiff.

BUG FIXES

Line numbers in error messages are incorrect when the error occurs in a child Rmd document (thanks, @cderv @khughitt, #1945).

MAJOR CHANGES

Previously knit_exit() in a child document would not terminate the knitting process of its parent document. Now it will terminate the whole process by default. You can use knit_exit(fully = FALSE) to exit the child document only (thanks, @davidchall, #1810).

MINOR CHANGES

Vignette engines for R Markdown vignettes no longer check for availability of pandoc-citeproc since it has gone since Pandoc 2.11.

CHANGES IN knitr VERSION 1.30

NEW FEATURES

Added knitr::hooks_*() functions to get a list of output hooks for a specific format. Previously, these hooks only exist inside the knitr::render_*() functions, and users do not have direct access to them. Now they can be accessed directly, e.g., via knitr::hooks_markdown() to get a list of output hooks for R Markdown documents. You can also set the output hooks individually, e.g., knitr::knit_hooks$set(knitr::hooks_markdown()['source']) only sets the source ouput hook. See more on output hooks at https://yihui.org/knitr/hooks/#output-hooks and https://bookdown.org/yihui/rmarkdown-cookbook/output-hooks.html (thanks, @cderv, #1889).
Added an argument lib.loc to knitr::write_bib().

BUG FIXES

The option fig_caption = FALSE for rmarkdown::html_document() was unable to suppress the figure captions in some cases.
The internal function fix_options() should be called after option hooks are executed (thanks, @atusy, #1876 and #1877).
When the option options(OutDec = ) is set to a value other than ".", percentages in the chunk options out.width and out.height do not work (thanks, @omgitsmoe, #1887).

MINOR CHANGES

knitr::write_bib() removes pairs of single quotes in the titles of citation entries now.
knitr::write_bib() uses the URL in the package DESCRIPTION if it is provided, instead of the canonical CRAN URL for the package.
hook_pdfcrop() and plot_crop() will work only when both programs pdfcrop and ghostscript have been installed. Previously only pdfcrop was checked (thanks, @dalupus, #954).

CHANGES IN knitr VERSION 1.29

NEW FEATURES

Added a new function kables() (the plural form of kable()) to create a table containing multiple tables, with each table created by kable(). Previously you can only create such a table from kable(list(data1, data2)), and now you can also do kables(list(kable(data1), kable(data2))). The latter way gives you more freedom to control each table individually via the arguments of each individual kable() call.
The graphics device chunk option, dev, now supports a value of 'ragg_png' which calls the agg_png() function from the ragg package (thanks, @cpsievert, #1834).
The argument error of include_graphics() takes value from the global R option knitr.graphics.error by default, e.g., you may set options(knitr.graphics.error = FALSE) so include_graphics() won't signal an error if the graphics file to be included doesn't exist.
Added a new engine cc to compile C++ code via R CMD SHILIB like the c and fortran engines (thanks, @kingaa, #1832).

BUG FIXES

For non-R code chunks that generate plots and use the chunk option fig.cap, the plot hook in knitr stopped working in v1.27 and v1.28 (thanks, @XiangyunHuang, https://d.cosx.org/d/421249).
The sql engine fails when the first column of the database is of the type bit64::integer64 (thanks, @randy3k, #1837).
The chunk option dpi is correctly passed to showtext::showtext_opts() now (thanks, @cpsievert, yixuan/showtext#33).
Theorem environments in bookdown fail to work with Pandoc >= 2.7.3 because of an issue in the block2 engine of knitr (rstudio/bookdown#883).

MAJOR CHANGES

For R Markdown documents, the default table format generated by kable() was changed from Pandoc's simple tables to pipe tables. This fixes bugs #1788 (kable() fails to align columns containing wide characters) and #1817 (broken column headers) (thanks, @atusy, #1830).
When a code chunk generates multiple plots to be placed in figure environments, their figure labels will be of the form label-i where label is the chunk label, and i is the plot number. Previously, the labels were of the form labeli (i.e., there was not a dash before the number). If you cross-reference the figures, you need to change the reference keys from labeli to label-i.
For a non-matrix object passed to kable(), its columns are formatted individually via format(). Previously, the whole object is coerced via as.matrix() (thanks, @thebioengineer, #1827).
For packages that contain R Markdown vignettes using the vignette engines based on the markdown package (e.g., knitr::knitr), the markdown package should be declared in Suggests in the package DESCRIPTION unless it has already been declared as a dependency in other fields. For this version of knitr, you will see a warning in R CMD check in the CRAN incoming check. In the future, you may see a warning in a regular R CMD check, so please add markdown to Suggests at your earliest convenience. If it is not clear how to do it, please see #1864.

MINOR CHANGES

knitr::image_uri() calls xfun::base64_uri() instead of markdown:::.b64EncodeFile() now.
knitr::write_bib() now adds an empty line after each BibTeX entry (#1862).

CHANGES IN knitr VERSION 1.28

BUG FIXES

hook_pdfcrop() no longer crops images included via include_graphics() now (thanks, @hpages #1797, @salim-b #1800).

MAJOR CHANGES

If the hook_pdfcrop() is enabled, the non-PDF image will be cropped only if the magick package is available, otherwise it will signal a message. In the previous version, it will signal an error (thanks, @trannhatanh89, #1796).
By default, include_graphics(files) will signal an error if any files do not exist and are not web resources. To avoid the error (e.g., if you think it is a false positive), use include_graphics(..., error = FALSE) (thanks, @hadley, #1717).

CHANGES IN knitr VERSION 1.27

NEW FEATURES

The chunk option message = FALSE can now be used to suppress the message "running: command ..." when the language engine is a command such as bash, awk, perl, ruby, sas, sed, and stata, etc. (thanks, @splaisan, #1782).

BUG FIXES

The figure caption was placed incorrectly when there are multiple include_graphics() in a code chunk (thanks, @hadley #1771, @cderv #1776).
The chunk option fig.keep was buggy when taking a numeric vector. The bug was originally introduced in #1265 and discovered in the SO post https://stackoverflow.com/q/59180351/559676.
When the chunk option dev = 'tikz' and a plot in the code chunk is generated by ggplot2 with a legend on a continuous scale, tikzDevice will create a raster image for the legend. Now the path to the raster image in the .tex file is tweaked to include the dir name of the .tex file, so that the .tex file can be correctly compiled to PDF via LaTeX (thanks to @rstub for the debugging at https://stackoverflow.com/a/58410965/559676).
hook_pdfcrop() does not crop all plots when multiple graphical devices are used in a code chunk (i.e., the chunk option dev takes a vector of devices) due to a bug in the internal function all_figs() identified by @bastistician in #1766, who also proposed a fix.

MAJOR CHANGES

options('width') no longer affects caching. With this change, your previous cache will be invalidated if you update knitr from 1.26 to 1.27. If you prefer that changes in the R global option width invalidate cache (as in previous versions of knitr), you may associate it with a chunk option, e.g., cache.extra = getOption('width'). The reason for this change is that it is too costly to invalidate the cache when the width option changes---the effect of this option is only cosmetic and the code chunk output may not really rely on this option (thanks, @jaburgoyne, #1781).

MINOR CHANGES

is_html_output() returns TRUE for the Pandoc output format gfm now (thanks, @ttimbers @cderv, rstudio/rmarkdown#1756).
plot_crop() no longer calls the convert command in ImageMagick to trim non-PDF plots. Instead, the function magick::image_trim() is called. This means you no longer need to install ImageMagick, but should install the R package magick (thanks, @hpages, #1785).
When the chunk option fig.scap is used in R Markdown and the output format is LaTeX/PDF, the plot hook knitr::hook_plot_tex() will be used to create the appropriate figure environment. In previous versions, fig.scap is silently ignored (thanks, @billdenney, #1793).

CHANGES IN knitr VERSION 1.26

NEW FEATURES

write_bib() also includes all citation entries in the package's CITATION file if provided.

MINOR CHANGES

knit() no longer adjusts the global R option width; previously it sets options(width = 75) by default.

CHANGES IN knitr VERSION 1.25

NEW FEATURES

Office outputs (e.g., rmarkdown::word_document) supports the following chunk options out.width, out.height, out.extra. Their behavior follows the behavior of Pandoc's link_attributes extension (thanks, @atusy, #1746).
The chunk option fig.process can be specified by a function with the options argument to read chunk options (thanks, @atusy, #1749).
For kable(format = 'latex'), a single string consisting of l, c, and r for the align argument also works now, e.g., knitr::kable(head(iris), 'latex', align = 'llcrr'). Previously it has to be a character vector of individual letters.

MAJOR CHANGES

Reverted #1519 (which intended to fix #1487): the valign argument of kable(format = 'latex') meant to control the vertical alignment of the inner tabular environment, instead of the floating position of the outer table environment. A separate argument named position has been added to control the floating position (e.g., position = "!b").

MINOR CHANGES

The returned value of combine_words() is wrapped in xfun::raw_string() for pretty printing.

BUG FIXES

The output hook for render_sweave() failed to respect the chunk option results='asis' (thanks, Achim Zeileis, https://stackoverflow.com/q/57571790/559676).
Added the unit px to the chunk option out.width if its value is numeric (thanks, @chendaniely, #1761).
The chunk option dependson did not work for non-R engines (thanks, @nielsrhansen, #1601).

CHANGES IN knitr VERSION 1.24

MAJOR CHANGES

The input argument of knitr::knit() must be a file path (a character string). It can no longer be connections.
The encoding argument of knitr::knit() is ignored. The encoding is always assumed to be UTF-8. Please see https://yihui.org/en/2018/11/biggest-regret-knitr/ for more info.

BUG FIXES

Inline code chunk references are parsed as character vectors instead of single character strings separated by \n now. This will avoid errors when a python code chunk contains a reference to a code chunk that contains multiple statements (thanks, @Atrebas, rstudio/rmarkdown#1589).

CHANGES IN knitr VERSION 1.23

NEW FEATURES

Added mogrify support in hook_png() and hook_mogrify() in order to obtain trimmed .png files (thanks, @mhofert, #1697).
It is possible to customize the sign \times used for the scientific notation of inline numeric output via a global option, e.g., options(knitr.inline.times = '\\cdot ') (thanks, @wuffi @trentks, #1563).
Exported knit_code and documented it at ?knitr::knit_code. Here be dragons, but I know you are brave and creative (thanks, @r-cheologist @ProQuestionAsker, #1545).
Added support for specifying fenced code block attributes in Markdown output via chunk options attr.*, which are similar but a generalization to the existing chunk options class.*. For example, class.source = 'numberLines' is equivalent to attr.source = '.numberLines', but attr.source can take arbitrary attribute values, e.g., attr.source = c('.numberLines', 'startFrom="11"') (thanks, @atusy, #1710).

BUG FIXES

For R Markdown documents, inline numeric output that is formatted with the scientific notation will be wrapped in \ensuremath{}. This will fix the LaTeX error ! Missing $ inserted.. However, you are still recommended to add a pair of dollar signs around the inline output when it is formatted with the scientific notation (e.g., $`r 1.234e12`$), otherwise it will not work with other R Markdown output formats such as HTML or Word (thanks, @billdenney and many who have suffered from this issue, rstudio/rmarkdown#1527).
For an input file that contains no code chunks or inline R code expressions, purl() should return an empty string instead of the original text from the input (thanks, @jrnold, #1660).
purl() might trigger a warning for non-R code chunks (thanks, @adamcagle, rstudio/rmarkdown#1528).
When building package vignettes via R CMD build, the R scripts generated from knitr's vignette (tangling) engines are not corrected encoded in UTF-8 (thanks, Kurt Hornik).
Fixed a bug introduced when fixing #1648: include_url() no longer works at all unless the chunk option out.extra is not empty (thanks, @gabrielvazdemelo, #1691).
When evaluating perl code chunks the interpreter is spawned with the -E option, enabling all optional features such as say, rather than -e (thanks, @kiwiroy, #1699)
When the chunk label of a tikz chunk contains periods, the tikz output cannot be converted to the expected .svg (thanks, @ucpresearch, #1706).
The ... argument for raw_latex() and raw_html() is actually passed to asis_output() now (thanks, @GregorDeCillia, #1716).

MAJOR CHANGES

knitr::knit() will try to read the input file with the UTF-8 encoding first. If UTF-8 doesn't work, it will try to read with the system native encoding instead (with a warning). The encoding argument of knitr::knit() is completely ignored. In the future, only UTF-8 will be supported, and we will stop retrying with the system native encoding. The output file from knitr::knit() is always encoded in UTF-8.
spin_child() will also assume the input file is encoded in UTF-8 (thanks, Henrik, https://stackoverflow.com/q/55395422/559676).
When the chunk option dev = 'svg', grDevices::svg() is used to record plots, instead of the default PDF null device (thanks, @trevorld, #729).
Partially reverted the solution for #1403 in favor of fixing the more surprising issue #1708, i.e., avoid false positive cache dependencies among chunks more than false negative dependencies when the chunk option autodep = TRUE (thanks, @fountainer).
The LaTeX package framed.sty is no longer bundled with the knitr package. This should only affect users who write .Rnw documents. If framed.sty is missing in your LaTeX distribution, you will have to install it by yourself. You may also consider TinyTeX (https://yihui.org/tinytex/), which has included this LaTeX package.

CHANGES IN knitr VERSION 1.22

NEW FEATURES

Added chunk options class.error, class.warning, and class.message to customize the CSS classes for errors, warnings, and messages in R Markdown output, respectively (thanks, @gadenbuie, #1676).
Added a new engine sass/scss to convert Sass/SCSS to CSS using either the sass R package (LibSass) or Dart Sass executable (when R package not found, the engine option engine.opts = list(package = FALSE), or engine.path to executable is provided). After conversion, resulting CSS is treated as in the CSS engine (thanks, @emilyriederer, #1666).
The cat engine supports the chunk option eval = FALSE now (thanks, @HanOostdijk, #1618).
The chunk option out.extra can be used to include extra attributes for the <iframe> generated from knitr::include_url() or knitr::include_app() (thanks, @jvcasillas, #1648).

BUG FIXES

knit_meta_add() could exhaust the system's memory because of a wrong method of vectorization (thanks, @kevinushey @nikkoc, rstudio/rmarkdown#1538).
The output path should be quoted in pandoc() (thanks, @antoine-sachet, #1644).
When there are multiple figures with multiple captions in a code chunk, bookdown's figure numbering is incorrect (thanks, Catherine Peng, https://stackoverflow.com/q/53880195/559676).
Added an argument label to kable() so that users can manually specify a label (thanks, @Inferrator, #1655).
The chunk option fig.show='hide' doesn't work for knitr::include_graphics() (thanks, @vincentarelbundock, #1675).
The tikz engine doesn't work on Windows (thanks, Andry, https://stackoverflow.com/q/54839403/559676).
kable() now generates a table for R Markdown documents even when the data has 0 rows (thanks, @yutannihilation, #1677).
The chunk options dev = 'tikz' and external = FALSE didn't work in R Markdown (thanks, @martinschmelzer, #1649).

MAJOR CHANGES

knitr::knit() starts to warn against non-UTF8 encodings (Why?). In the future, we will only support UTF-8 input. If your input file is not encoded in UTF-8, we strongly recommend that you re-save it with UTF-8.
Removed the encoding arguments in knitr::pandoc(), knitr::knit2pdf(), knitr2wp(), and knitr::Sweave2knitr(). The input files must be encoded in UTF-8.
knitr::knit2html() still has the encoding argument, but it only supports UTF-8 internally.

MINOR CHANGES

Changed tools::texi2dvi() to tinytex::latexmk() for the tikz engine to compile TikZ graphics to other formats (e.g., svg or png). This requires the tinytex package >= v0.10.10: https://github.com/yihui/tinytex.
Empty lines are no longer removed in the output of purl() (thanks, Marius Hofert).

CHANGES IN knitr VERSION 1.21

NEW FEATURES

Added styler support: the chunk option tidy = 'styler', the code will be reformatted using the styler package. The chunk option tidy = TRUE means tidy = 'formatR' and still uses formatR::tidy_source() to reformat the code. The chunk option tidy can also take a custom function to reformat the code. See the documentation for details: https://yihui.org/knitr/options/ (thanks, @lorenzwalthert, #1581).
Added a new hook function hook_gifski() to create a GIF animation from plots in a code chunk. To enable this hook, you may install.packages('gifski') and set the chunk option animation.hook='gifski'. See https://yihui.org/en/2018/08/gifski-knitr/ for examples.
Added a new object cache_engines for other language engines to handle caching. See ?knitr::cache_engines for details (thanks, @tmastny, #1518).
Can now pass additional arguments to knitr vignette engines if needed (thanks, @jimhester, #1577).
When options(knitr.include_graphics.ext = TRUE) is set, the full filename will be used in \includegraphics{} (e.g., \includegraphics{foo.pdf}) instead of the (default) base filename (e.g., \includegraphics{foo}) if a plot is embedded in the LaTeX output. This will avoid the ambiguity for LaTeX to choose the correct file (#1587).
The chunk option engine.opts can also take a list now, so that users can specify different options for different language engines (thanks, @kiwiroy, #1632).
Added functions raw_block(), raw_latex(), and raw_html() to wrap content in raw attribute blocks for Pandoc (thanks, @hughjonesd, #1633).

BUG FIXES

valign in kable_latex() does not put the float alignment to the correct location (thanks, @haozhu233, #1487, #1519).
engine.path does not work for engine = 'dot' (thanks, @billy34, #1534).
The sql engine now caches the result properly when the chunk option output.var is specified (thanks, @yutannihilation, #1544).
knit_params() mangles UTF-8 text not representable in current locale (thanks, @kevinushey, #1557).
The linesep argument in kable(format = 'latex') does not work as expected (thanks, @tinu-schneider, #1558).
MP4 generated by hook_ffmpeg() does not work in Safari (thanks, @brianzhang01, #1586).
The chunk option fig.env does not work for PDF output from R Markdown (thanks, @RSchwinn, #1620).
spin() fails to generate a proper R Markdown document when the R script contains roxygen comments as string literals or more than three backticks (thanks, @yutannihilation, #1605, #1611).
Stan chunks cannot be properly cached: the bug #1064 appeared again (thanks, Mark, https://stackoverflow.com/q/53660143/559676).
Roman numerals should not be converted to arabic numerals in inline R expressions (thanks, @rgaiacs, #1625).

MAJOR CHANGES

For R scripts passed to spin(), only the UTF-8 encoding is accepted (previously it assumes the native encoding of the system).
Due to changes in the S3 method dispatch mechanism since R 3.5.0, you will have to call registerS3method() to register knit_print methods defined in code chunks. For package authors, it no longer suffices to export a method without importing knitr::knit_print, but if you don't want to import it, there is still a workaround. See the package vignette vignette('knit_print', package = 'knitr') for details (thanks, @wch @kevinushey @skranz #1580).

MINOR CHANGES

spin() will close each individual code chunk if multiple code chunks appear consecutively (previously it only closes the last code chunk in this case) (thanks, @avehtari, #1590).

CHANGES IN knitr VERSION 1.20

NEW FEATURES

Added a roxygen_comments parameter to read_chunk() in order to control whether trailing commentary is stripped from code chunks; default behaviour is to keep roxygen comments with roxygen_comments = TRUE (thanks, @ruaridhw, #1484).

BUG FIXES

knit_filter() does not work for R Markdown input files (thanks, @omahdi, #1498).

CHANGES IN knitr VERSION 1.19

NEW FEATURES

spin() now supports code chunks of the form -- ---- label ---- allowing external SQL and Haskell files to be spun (thanks, @ruaridhw, #1492).

MAJOR CHANGES

tinytex::latexmk() is used to compile LaTeX plots generated by the chunk option dev = 'tikz' to PDF, which means you have to install the tinytex package if you use this device.
Similarly, tinytex::latexmk() is used in knit2pdf() and stitch().

BUG FIXES

Fixed rstudio/bookdown#501: the chunk option out.height can also take a percentage value when the output format is LaTeX (similar to out.width), e.g., out.height = '30%' means .3\textheight (thanks, @dataopt, #1482).
imgur_upload() starts to fail due to changes in Imgur's API (thanks, @jennybc, #1495).

CHANGES IN knitr VERSION 1.18

NEW FEATURES

added a new engine julia based on JuliaCall::eng_juliacall() to support Julia code chunks (thanks, @Non-Contradiction, #1458)
added a new chunk option fig.dim to specify fig.width and fig.height using one option, e.g., fig.dim = c(5, 7) is a shorthand of fig.width = 5, fig.height = 7 (thanks, @MichaelChirico, #1449)
added two chunk options fig.ncol and fig.sep to make it possible to arrange multiple subfigures in columns (thanks, @eliocamp, #1444)
exported function is_latex_output() and is_html_output() (thanks, @zeehio, rstudio/rmarkdown#649)
for read_chunk(), the code chunks in the external R script can be written under the lines of the form -- ---- label ---- now (# ---- label ---- is still supported); this makes it possible to read an external Haskell script (thanks, @brittAnderson, #1455)
added a $append() method to the internal function new_defaults() (thanks, @yonicd, #1417)
added another usage of engine_output() for expert users: if you are familiar with the list structure of the evaluate::evaluate() output, you can let your custom engine return a similar list and call engine_output(options, out = YOUR_LIST) (thanks, @kevinushey, #1442)
added an evaluate.inline hook to knit_hooks (thanks, @kevinushey, #1443)

MAJOR CHANGES

the python engine is based on reticulate::eng_python() now; this means all Python code chunks are evaluated in the same Python session; if you want the old behaviour (new session for each Python code chunk), you can set the chunk option python.reticulate = FALSE (thanks, @kevinushey, #1440)
the auto_pdf argument of include_graphics() defaults to FALSE now; you can set options(knitr.graphics.auto_pdf = TRUE) to change the default to TRUE (thanks, @mebden, #1456)

BUG FIXES

fixed #1446: knitr may fail to parse code chunks in blockquotes in Markdown (thanks, @rgaiacs)
fixed #1460: kable() does not generate the table caption when the data only has one column (thanks, @Henrik-P)
fixed #1464: subcaptions are not rendered if the chunk option out.extra = NULL and the output format is LaTeX (thanks, @eliocamp)
fixed #1462: kable() may mis-order column names of data generated by sparklyr (thanks, @JohnMount)

CHANGES IN knitr VERSION 1.17

MAJOR CHANGES

the automatic detection of chunk dependencies (via the chunk option autodep = TRUE) is more conservative by default now; chunk B depends on chunk A if any variables in B are created in A, no matter if these variables are local or global in B; you can use the chunk option cache.globals to manually provide a vector of variable names that should be considered "global" to avoid the dependency when local variables in B are also found in A (thanks, @knokknok, #1403)
for engines css and js, the source code will be displayed by default just like R code chunks, and you can hide the source code using the chunk option echo = FALSE (thanks, @skadauke, #1408)
for kable(), the digits argument will be applied to columns of the class difftime (thanks, @RoyalTS, #1396)

BUG FIXES

when a table only has one column, kable() does not work in R Markdown documents (thanks, @expersso, yihui/printr#31)
message(..., appendLF = FALSE) does not work (thanks, Anastasius, http://stackoverflow.com/q/44059508/559676)
hook_optipng() does not work correctly when a code chunk does not contain any R-generated plots (thanks, @@zevross-spatial, #1404)
markdown lines starting with three backticks won't be considered inline code chunks (thanks, @gadenbuie, #1416)

CHANGES IN knitr VERSION 1.16

NEW FEATURES

adds two new options, class.output and class.source to apply additional HTML classes to output and source chunks, which can be useful for R Markdown documents with HTML output formats since you can define custom CSS styles for these classes (thanks, @ijlyttle #1335 and @AmeliaMN #1333)
added a go engine (thanks, @hodgesds, #1330)
added a sql.show_interpolated chunk option for sql engine. For {sql, sql.show_interpolated=TRUE}, interpolated SQL query will be displayed in the rendered document instead (thanks, @saurfang, #1357)
the chunk option results=FALSE is an alias of results='hide' now (thanks, @hadley, #1360)
added a wrapper function knit2pandoc(), and reStructuredText vignettes will be compiled through Pandoc instead of rst2pdf (thanks, @trevorld, #1386)
longest common whitespace prefix in code chunks is now stripped off along with the eventual prefix (see #1391)

MAJOR CHANGES

a warning will be emitted when the chunk option fig.show='hold' and the output format is Word (rstudio/bookdown#249); fig.show='hold' will be changed to 'asis' at the same time
purl() no longer writes verbatim non-R code chunks to the output R script; these code chunks will be commented out in the output
purl() can now also retrieve the code from inline expressions; it is disabled by default for compatibility, and can be enabled via options(knitr.purl.inline = TRUE) (thanks, @kforner, #1345)
the chunk option fig.show also applies to plots included via include_graphics(), e.g. fig.show='hold' will put all plots to the end of a chunk (thanks, @statnmap, #1382)

BUG FIXES

kable() should not ignore the table caption if provided when the input is a list, the table format is pandoc, and the document output format is not HTML or LaTeX, e.g. Word output (thanks, @ugroempi, rstudio/bookdown#368)
opts_chunk$set() supports referencing a sql connection by variable name, e.g. knitr::opts_chunk$set(connection = "db") (thanks, @javierluraschi, #1362)
the chunk option dev = 'CairoPS' will lead to double filename extensions .eps.ps (thanks, @qifei9, #1364)
the output file of write_bib() is now encoded in UTF-8 (thanks, @arnyeinstein, #1374)
the <embed> tag should be used instead of <img> for PDF images in HTML output (thanks, Daijiang Li, http://stackoverflow.com/q/43549930/559676)

CHANGES IN knitr VERSION 1.15.1

NEW FEATURES

added a new hook function hook_pngquant() that can call pngquant to optimize PNG images (thanks, @slowkow, #1320)

BUG FIXES

not really a knitr bug, but knit_params() should be better at dealing with multibyte characters now due to the bug fix in the yaml package vubiostat/r-yaml#6

CHANGES IN knitr VERSION 1.15

NEW FEATURES

NA values can be displayed using different characters (including empty strings) in kable(); you can set the option knitr.kable.NA, e.g. options(knitr.kable.NA = '') to hide NA values (#1283)
added a fortran95 engine (thanks, @stefanedwards, #1282)
added a block2 engine for R Markdown documents as an alternative to the block engine; it should be faster and supports arbitrary Pandoc's Markdown syntax, but it is essentially a hack; note when the output format is LaTeX/PDF, you have to define \let\BeginKnitrBlock\begin \let\EndKnitrBlock\end in the LaTeX preamble
figure captions specified in the chunk option fig.cap are also applied to HTML widgets (thanks, @byzheng, rstudio/bookdown#118)
when the chunk option fig.show = 'animate' and ffmpeg.format = 'gif', a GIF animation of the plots in the chunk will be generated for HTML output (https://twitter.com/thomasp85/status/785800003436421120)
added a width argument to write_bib() so long lines in bib entries can be wrapped
the inline syntax `r#code` is also supported besides `r code`; this can make sure the inline expression is not split when the line is wrapped (thanks, Dave Jarvis)
provided a global R option knitr.use.cwd so users can choose to evaluate the R code chunks in the current working directory after setting options(knitr.use.cwd = TRUE); the default is to evaluate code in the directory of the input document, unless the knitr option opts_knit$set(root.dir = ...) has been set
if options(knitr.digits.signif = TRUE), numbers from inline expressions will be formatted using getOption('digits') as the number of significant digits, otherwise (the default behavior) getOption('digits') is treated as the number of decimal places (thanks, @numatt, #1053)
the chunk option engine.path can also be a list of paths to the engine executables now, e.g., you can set knitr::opts_chunk$set(engine.path = list(python = '/anaconda/bin/python', perl = '/usr/local/bin/perl')), then when a python code chunk is executed, /anaconda/bin/python will be called instead of the system default (rstudio/rmarkdown#812)
introduced a mechanism to protect text output in the sense that it will not be touched by Pandoc during the conversion from R Markdown to another format; this is primarily for package developers to extend R Markdown; see ?raw_output for details (which also shows new functions extract_raw_output() and restore_raw_output())

MAJOR CHANGES

the minimal version of R required for knitr is 3.1.0 now (#1269)
the formatR package is an optional package since the default chunk option tidy = FALSE has been there for a long time; if you use tidy = TRUE, you need to install formatR separately if it is not installed
:set +m is no longer automatically added to haskell code chunks (#1274)

MINOR CHANGES

the package option opts_knit$get('stop_on_error') has been removed
the confusing warning message about knitr::knit2html() when building package vignettes using the knitr::rmarkdown engine without pandoc/pandoc-citeproc has been removed (#1286)
the default value of the quiet argument of plot_crop() was changed from !opts_knit$get('progress') to TRUE, i.e., by default the messages from cropping images are suppressed

BUG FIXES

the chunk option cache.vars did not really behave like what was documented (thanks, @simonkth, #1280)
asis_output() should not be merged with normal character output when results='hold' (thanks, @kevinushey, #1310)

CHANGES IN knitr VERSION 1.14

NEW FEATURES

improved caching for Rcpp code chunks: the shared library built from the C++ code will be preserved on disk and reloaded the next time if caching is enabled (chunk option cache = TRUE), so that the exported R functions are still usable in later R code chunks; note this feature requires Rcpp >= 0.12.5.6 (thanks, @jjallaire, #1239)
added a helper function all_rcpp_labels(), which is simply all_labels(engine == 'Rcpp') and can be used to extract all chunk labels of Rcpp chunks
added a new engine named sql that uses the DBI package to execute SQL queries, and optionally assign the result to a variable in the knitr session; see https://rmarkdown.rstudio.com/authoring_knitr_engines.html for details (#1241)
fig.keep now accepts numeric values to index low-level plots to keep (#1265)

BUG FIXES

fixed #1211: pandoc('foo.md') generates foo_utf8.html instead of foo.html by default
fixed #1236: include = FALSE for code chunks inside blockquotes did not work (should return > instead of a blank line) (thanks, @fmichonneau)
fixed #1217: define the command \hlipl for syntax highlighting for Rnw documents (thanks, @conjugateprior)
fixed #1215: restoring par() settings might fail when the plot window is partitioned, e.g. par(mfrow = c(1, 2)) (thanks, @jrwishart @jmichaelgilbert)
fixed #1250: in the quiet mode, knit() should not emit the message "processing file ..." when processing child documents (thanks, @KZARCA)

MAJOR CHANGES

knitr will no longer generate screenshots automatically for HTML widgets if the webshot package or PhantomJS is not installed

MINOR CHANGES

if dev = 'cairo_pdf', the cairo_pdf device will be used to record plots (previously the pdf device was used) (#1235)
LaTeX short captions now go up to the first ., : or ; character followed by a space or newline (thanks, @knokknok, #1249)

CHANGES IN knitr VERSION 1.13

NEW FEATURES

kable now accepts a single multi-character string for alignment, so kable(..., align=c('c','l')) can be shortened to kable(..., align='cl')
code chunks that generate metadata may be cached now; it requires htmlwidgets >= v0.6 and htmltools >= 0.3.3 if you cache code chunks that contain HTML widgets or Shiny inputs/outputs (#1158)
when the output format is not HTML, HTML widgets used to fail to render; now knitr will try to generate static screenshots for HTML widgets automatically using the webshot package; you can specify alternative screenshots via the chunk option screenshot.alt (which takes a character vector of image paths), and pass more options to webshot::webshot() via the chunk option screenshot.opts, e.g. list(delay = 3, cliprect = 'viewport')
added two functions include_url() and include_app() to embed URLs in the output (the latter is for Shiny app URLs); when the output format is HTML, iframe will be used; otherwise screenshots of the URLs will be used
screenshotting for HTML widgets and URLs can be forced for all output formats via the chunk option screenshot.force = TRUE; if you set the chunk option screenshot.force = FALSE, knitr will just render these content normally and not take screenshots
added a new chunk option fig.link, which can be used to attach hyperlinks on figures, e.g. you can add a link to a screenshot of a Shiny app so that readers can check out the actual live app after clicking on the static screenshot (this chunk option currently only works for Markdown and LaTeX output)
syntactical errors in code chunks will be allowed when the chunk option error = TRUE and the package version of evaluate is at least 0.8.4; previously knitr would just stop on parsing errors (r-lib/evaluate#65)
PNG/JPEG images included via include_graphics() also respects the chunk option dpi now; if it is numeric and the chunk option out.width is not set, the output width (in inches) of an image will be automatically calculated from the actual width (in pixels) divided by dpi; note this feature requires the packages png and/or jpeg to be installed, and you can disable the feature using dpi = NA (thanks, @hadley, rstudio/bookdown#38)
added a new hook function named evaluate in knit_hooks so that users can redefine an evaluator to evaluate the code chunk; the default is evaluate::evaluate(), and your custom evaluator must be compatible with evaluate::evaluate() in terms of the argument names and the data structure of the returned value (a list of values with special classes)
added a new function combine_words() to combine multiple words / phrases into a single string, which may be useful in inline R expressions, e.g. combine_words(c('a', 'b', 'c')) returns a, b, and c
render_markdown() gained a new argument fence_char to customize the character to be used as the code blocks fence, e.g. it can be a backtick, or a tilde, depending on the Markdown rendering engine (thanks, @tinyheero, #1161)
the pandoc() function no longer assumes Markdown input (thanks, @scls19fr, #1170)
added a new function knit_meta_add() so that users can manually inject metadata into the current knitr session
for the tikz engine, if fig.ext = 'svg', dvisvgm will be called to convert the DVI output of TikZ to SVG; you need to install dvisvgm, and Windows users have to install GhostScript as well (thanks, @dkilfoyle, #1177)
new js and css engines which surround their content with <script> and <style> tags respectively, and print no output when not in an HTML document
for LaTeX tables, kable() supports short captions now via the caption.short argument, e.g. kable(..., caption = 'A long caption', caption.short = 'A short caption') (thanks, @ismayc, #1199)
added three global R options knitr.sanitize.errors, knitr.sanitize.warnings, and knitr.sanitize.messages to mask or change the messages, e.g. if options(knitr.sanitize.errors = TRUE) and the chunk option error = TRUE, the actual error message will be replaced by a character string like "An error occurred"; these options can also accept character values so you can customize the messages to be displayed, e.g. options(knitr.sanitize.warnings = 'You had a warning from the code'); see rstudio/shiny#1123 for the motivation of these options

BUG FIXES

when the chunk option cache.rebuild = TRUE, the cache database should be rewritten (thanks, Oleg Mayba)
include_graphics() did not work in inline R expressions (thanks, @WastlM, #1166)
the cex parameter was not correctly restored in the case of opts_knit$set(global.par = TRUE) (http://stackoverflow.com/q/35606445/559676)
for Rnw documents, when there are two instances of \documentclass{}, knitr might mistakenly treat the second instance as the actual command to declare the document class (thanks, #1180, @ekstroem)
corrected the environment for evaluating R scripts in stitch_rhtml() and stitch_rmd() (thanks, @Hughan, #1207)

MAJOR CHANGES

the default value of the package option eval.after is changed from NULL to fig.cap, i.e. the figure caption will always be evaluated after a code chunk is evaluated (thanks, @JoshOBrien, #1165)
the function eclipse_theme() has been removed since the website eclipsecolorthemes.org has been down for a long time

CHANGES IN knitr VERSION 1.12.3

BUG FIXES

figures without captions may fail to render correctly in R Markdown v2
fixed #1149: the sanitize and external chunk options were not passed to tikzDevice::tikz() when the tikz() device was opened to record plots in a chunk, so plots that contain special LaTeX characters would not work (thanks, @O-T)

CHANGES IN knitr VERSION 1.12

NEW FEATURES

added a chunk option fig.asp to specify the aspect ratio (i.e. the ratio height/width) of figures: when fig.asp is specified, the height of a figure (the chunk option fig.height) is calculated from fig.width * fig.asp
added a new function knit_watch() to knit input files continuously when any of them are updated, so you do not have to call a knitting function like knit(), knit2pdf(), or rmarkdown::render() repeatedly as you update an input document
kable() supports generating multiple tables from a list of data objects, and the tables will be placed side by side when the output format is HTML or LaTeX
the format argument of kable() can be a function that returns a character string of the table format (thanks, @haozhu233, #1132)
added a new function include_graphics() to embed existing image files into knitr documents; see ?include_graphics for details
added a new object opts_hooks that can be used to set up chunk option hooks to modify chunk options; see ?opts_hooks for details (thanks, @hadley, #1096)
added a chunk option ffmpeg.bitrate (default: 1M) to control the quality of WebM videos created from FFmpeg (thanks, @JsonPunyon, #1090)
added a chunk option ffmpeg.format (default: webm) to control the video format of FFmpeg; this option specifies the filename extension of the video (#712)
plots drawn by tikzDevice are also measured and recorded with that device (thanks, @krlmlr, #1066)
for the chunk option out.width, a value of the form xx% will be converted to xx/100\\linewidth when the output format is LaTeX, so we can specify figure widths as percentages for both LaTeX and HTML output
for the chunk option background (to define the background color of code chunks in Rnw documents), the value NA means knitr will no longer write \definecolor{shadecolor} to the output of an individual code chunk (you can also enable this behavior by setting it as a global chunk option) (thanks, @friendly, #1122)
added a new engine block to wrap the chunk in a LaTeX environment or an HTML div, depending on the output format; the name of the environment or the class of the div can be specified via the chunk option type
you can use the svglite device in the svglite package (https://github.com/hadley/svglite) via the chunk option dev = 'svglite'
cache = TRUE works better for the stan engine now: the Stan model object will be saved and can be loaded the next time for the cached code chunk (thanks, @dalewsteele, #1064)
the chunk option engine.env can be used to set environment variables when running engines like bash; this option is passed to the env argument of system2() (thanks, @mbojan, #1032)

BUG FIXES

the asis engine did not preserve line breaks in the code chunk
kable() could not display chron objects correctly (thanks, @liesb, #1118)
kable(format = 'latex', longtable = TRUE) did not write the table caption in the correct place (should be inside the longtable environment)
fixed #1055: the tikz device may fail if there are code chunks before \documentclass{} in Rnw documents (thanks, @ClaudiusL)
fixed #1119: when the chunk option engine != 'R', the size option does not work (thanks, @cysouw)

MAJOR CHANGES

local chunk options can override options set in opts_template now, e.g. for opts_template$set(quiet = list(echo = FALSE)), and a local chunk option echo = TRUE will override echo = FALSE in opts_template for the current code chunk (thanks, @aaronwolen, #1140)

MINOR CHANGES

added an argument force_v1 to knit2html(); when it is TRUE, knit2html() renders the input file as an R Markdown v1 document even if it is for R Markdown v2 (#1130)
the trailing space after ![]() (for plots) in Markdown output has been removed (#1049)

CHANGES IN knitr VERSION 1.11

NEW FEATURES

added an argument format.args to kable() to pass arguments to format() mainly to format numeric values, e.g. you can use , as the thousands separator via kable(..., format.args = list(big.mark = ',')) (thanks, @neuwirthe, #1086)
added an argument quiet to plot_crop() (thanks, @WastlM, #1034)
added !expr and !r support in YAML for knit_params() (#1069)
added two language engines mysql and psql (thanks, @beanumber, #1076)

MAJOR CHANGES

the two hook functions hook_rgl() and hook_webgl() have been moved from knitr to the rgl package (v0.95.1247) (thanks, @dmurdoch)
inline R expressions will always be evaluated; previously, they were not be evaluated if the global chunk option eval = FALSE, and the inline output was ?? (thanks, @Bozh, #1046)

BUG FIXES

the HTML code generated for FFmpeg animations did not work through Pandoc conversion (thanks, @s-fiebig, #1067)

CHANGES IN knitr VERSION 1.10.5

MAJOR CHANGES

this version is a patch release due to a few changes in handling the params field of the YAML metadata, including
- the character input passed to yaml::yaml.load() will always be converted to UTF-8 to make sure multibyte characters work in YAML
- the character output from yaml::yaml.load() will always be marked as UTF-8 (vubiostat/r-yaml#6)
- y, Y, n, and N will not be treated as booleans (TRUE/FALSE), although they do stand for booleans in the YAML 1.1 spec
- for purl(), the params field will be correctly written to the R script

CHANGES IN knitr VERSION 1.10

NEW FEATURES

added a new chunk option cache.rebuild (default: FALSE) to force knitr to invalidate and rebuild the cache for a chunk when cache.rebuild = TRUE (thanks, @zachary-foster, #995)
added a stata engine and improved the existing sas engine (thanks, @Hemken #979, @muschellij2 #984)
the Markdown table generated from kable() will use the width of characters to decide how many spaces are needed for padding, which can be useful for wide characters such as CJK characters (normally of width 2 per char) (thanks, @yutannihilation, #985)
for language engines, the syntax of R Markdown code chunk headers can be ```{lang, option=value} now (e.g. ```{python} and ```{Rcpp}), which is equivalent to ```{r, engine='lang', option=value} (thanks, @JanSchulz, #963)
added the ... argument to all_labels() to make it possible to filter chunk labels according to local chunk options, e.g. you can use all_labels(engine == 'Rcpp') to find the labels of all code chunks that have the chunk option engine = 'Rcpp', and pass the character vector of labels to ref.label to gather all Rcpp code chunks into a single chunk
added a new function knit_params() to extract report parameters from the YAML metadata of R Markdown documents; see ?knit_params for details (thanks, @jjallaire, #997)
added support for code chunks inside Markdown blockquotes; see the example below (thanks, @r-gaia-cs, #1022)
```
  > Here is a quote, followed by a code chunk:
  >
  > ```{r}
  > x = 1:10
  > rev(x^2)
  > ```
```
the function current_input() gained a new argument dir = FALSE; when dir = TRUE, the absolute path of the input document is returned (thanks, @klmr, #950)
if you set options(knitr.duplicate.label = 'allow') before calling knit(), duplicate labels will be allowed, and an incremental suffix will be appended to the duplicate label, e.g. when there are two non-empty chunks with the same label foo, the second label may be converted to foo-2 automatically; please use this feature only if you understand the consequences (e.g. the plot generated by a chunk foo might be foo-2-1.pdf, which can be surprising) (thanks, @khughitt, #957)
added a new language engine named lein to support Clojure (thanks, @kindlychung, #940)

MAJOR CHANGES

the minimal required version of R is 3.0.2 now, and you are recommended to use the latest version of R when possible

BUG FIXES

fixed #942: the package option opts_knit$get('base.dir') does not work when the chunk option opts_chunk$get('fig.path') contains a directory (thanks, @daattali and @cheebingloh)
fixed the bug reported at http://stackoverflow.com/q/28954647/559676 (dev = 'png' failed to work with rmarkdown when the output format is pdf_document)

CHANGES IN knitr VERSION 1.9

NEW FEATURES

added a new function load_cache() so that we can read an object from a cached code chunk, even if the chunk appears later in the document and the object needs to be used earlier; see the example #114 at https://github.com/yihui/knitr-examples (related to #868)
added a new function clean_cache() to clean up the cache files that are probably no longer needed, e.g. after you rename or delete cached code chunks (thanks, @sjmgarnier, #933)
knit2wp() can update an existing post or create a page now (thanks, @jaredlander, #916, #917)
added an engine stan to support Stan through the rstan package; see http://rpubs.com/jrnold/knitr-stan for an example (thanks, @jrnold, #903)
for the tikz engine, the path to the convert utility of ImageMagick can be specified via the chunk option engine.opts, e.g. engine.opts = list(convert = 'path/to/convert') (thanks, @mienkoja, #897)
similarly, more command line arguments to convert can be specified via, for example, engine.opts = list(convert.opts = '-density 300') (#896)

MAJOR CHANGES

when the chunk option dev = 'png', grDevices::png() is used to record plots, instead of the default PDF null device (thanks, @yixuan, #729)
currently, R (3.1.2) does not really pass the vignette encoding to vignette engines (which is probably a bug), and knitr vignette engines will assume UTF-8 is the file encoding
when the chunk option tidy=FALSE, and eval takes a numeric vector, it used to mean the line numbers of the code chunk; now it means the indices of the R expressions in the code chunk, regardless of tidy=FALSE or TRUE (yihui/knitr-examples#39, thanks, @isomorphisms)

MINOR CHANGES

you need to upgrade the rgl package to at least v0.95.1201 if you use the hook function hook_webgl() (thanks, @dmurdoch, #901)
for the automatically generated figure caption and label of a figure, the label is placed outside of the caption now, i.e. the format was changed from \caption{\label{}} to \caption{}\label{} (thanks, @dr-moebius, @krlmlr, @ltorgo, #746)

BUG FIXES

fixed #898: kable() did not work on matrices with duplicate row names (thanks, @tomaskrehlik)
fixed #927: kable() did not work for matrices of zero row and/or zero column (thanks, @hadley)
fixed #929: opts_knit$restore() does not restore the animation.fun option (thanks, @julian-gehring)

CHANGES IN knitr VERSION 1.8

NEW FEATURES

when using knitr with the rmarkdown package, the internal output hook for plots will be automatically switched to the LaTeX plot hook when necessary; for example, when the chunk options fig.align, out.width, out.height, and/or out.extra are specified, raw LaTeX code will be generated to align/set the size of plots, because there is no support for figure alignment or setting the size in the native Markdown syntax; for Word output, these options are simply ignored (related issues: #626, rstudio/rmarkdown#86, rstudio/rmarkdown#148, rstudio/rmarkdown#303)
added a new function fig_chunk() to provide a public API to get the figure filenames produced from code chunks; since knitr 1.7 changed the figure file numbering scheme, it broke documents with hard-coded figure filenames, e.g. for Rnw documents, \includegraphics{foo.pdf} should be \includegraphics{foo-1.pdf} after knitr 1.7, and such problems can be avoided by \includegraphics{\Sexpr{fig_chunk('foo', 'pdf')}} (thanks, @edwardabraham, #870)
added an argument escape = TRUE to kable() to escape special characters in LaTeX and HTML tables (thanks, @juba, #852)
added a new function knit_filter() to filter out code chunks and inline R expressions; this function can be used as the filter for the spell check function utils::aspell(); see ?knit_filter for examples (#581)
added a new function spin_child() to spin child R scripts when we spin() a main script (thanks, @krlmlr, #621)
added a new function inline_expr() to help authors write the "source code" of the inline expression, e.g. inline_expr('1+1') generates `r 1+1` in R Markdown documents (#890)
the cache will attempt to preserve the order in which packages are stored on the search path (thanks, @dgrtwo, #867)
added a new argument table.envir to kable() for LaTeX tables only; if the table caption is specified (not NULL), the LaTeX environment table will be used by default (i.e. the table is generated in \begin{table} \end{table}), and you can specify alternative environments via kable(..., table.envir = '???') (thanks, @dalupus, #872)
chunk options are supported using the syntax # ---- label, options ---- in the R script passed to stitch() (thanks, @wibeasley, yihui/knitr-examples#35)
syntax highlighting for .Rnw and .Rhtml documents can be further customized by setting opts_knit$set(highr.opts = list(markup = cmd_mine)) where cmd_mine is a data frame for the markup argument of highr::hilight() (thanks, @lq, #869)
added a new language engine groovy (thanks, @vveitas, #876)

BUG FIXES

fixed #862: the YAML metadata in child R Markdown documents was not correctly removed (thanks, @krlmlr)
fixed #874: for the engines dot and tikz, the figure directory will be created recursively (thanks, @WilDoane)
fixed #654: sub figures were not aligned correctly in LaTeX when the chunk option fig.align was specified (thanks, @lionandoil)
the vignette engine knitr::rmarkdown_notangle did not really work (thanks, @bbolker, http://stackoverflow.com/q/26726388/559676)

CHANGES IN knitr VERSION 1.7

NEW FEATURES

added a simple Fortran engine after sitting with John Nash for a few minutes at UseR!2014; now we can use the chunk option engine = 'fortran' to include Fortran code in a source document, which will be compiled and loaded via R CMD SHILIB and dyn.load(), respectively
added vignette engines with the suffix _notangle, which have the same weave functions as those engines without this suffix but have disabled the tangle function, meaning there will not be R scripts generated from the vignettes during R CMD build or R CMD check (thanks, Carl Boettiger and Michael Koohafkan, #784)
added an argument col.names to kable(), so we can specify different column names (thanks, @jackflibb, #801)
added a new output hook called text, and its default value knit_hooks$get('text') is the identity function function(x) x; this hook is applied to the text chunks (recall knit_hooks$get('chunk') is applied to code chunks)
added a chunk option fig.showtext to support the showtext package; if fig.showtext = TRUE (which is what you should do if you use the showtext package), showtext::showtext.begin() is called before drawing plots (thanks, @yufree, #799)
added a new language engine node for Node.js (thanks, Jake Burkhead, #823)
added a package option global.par; if we set opts_knit$set(global.par = TRUE) (by default it is FALSE), the par() settings from the last code chunk will be preserved and applied to the next code chunk (thanks, Jim Winget)
language engines also write error messages (if there are any) in the output now (thanks, Fabian Hirschmann, #789)
added the envir argument to knit_child() so that users can specify a different environment to evaluate the child documents (thanks, Stéphane Laurent, http://stackoverflow.com/q/24009622/559676)
for set_parent(), the lines in the parent document that start with \bibliography are matched and inserted in the child document so LaTeX bibliography also works for the child document (thanks, Mark Heckmann, #819)
for the chunk option engine = 'cat', the code chunk can be displayed in the output if the chunk language is specified via engine.opts, e.g. engine.opts = list(lang = 'makefile')

BUG FIXES

fixed #779: when the chunk options tidy=FALSE and eval=FALSE, prompt=TRUE did not work for R expressions of multiple lines (thanks, Qijie Zhao)
fixed #788: there was no increment in the chunk counter when the code chunks were read through read_chunk(), which may lead to clashes of chunk labels (thanks, Jason Ackman)
fixed #790: when chunk A reuses code from chunk B via <<B>>, and only the first line of B is empty, chunk reuse can fail because A sees B as empty (thanks, @kingaa)
fixed #791: if one has specified the chunk option dev.args, only pointsize and bg in it can be passed to the default recording device (the pdf() device) (thanks, @M-Russell and @joelgombin)
fixed #822: cache.lazy = FALSE did not really work (thanks, Liz Ing-Simmons)
fixed rstudio/rmarkdown#205: when R marks the encoding of the input document as latin1, knit() can fail to convert its encoding (thanks, @ripkrizbi)
fixed #828: scientific notation for inline numbers did not work in R Markdown v2 when the output format is LaTeX (thanks, @nacnudus)
fixed #833: for the LaTeX output format, when fig.cap contains . in {}, the automatic short caption does not work (thanks, Roman Luštrik)
fixed #844: when the digits argument is a vector and x is a numeric matrix, kable(x) did not work (thanks, @dmenne, #844)

MAJOR CHANGES

the knit() function no longer modifies R's default options(digits) from 7 to 4, since it may lead to confusion especially when printing summary() output; for those who want the old behavior, you must set options(digits = 4) at the beginning of your document (thanks, John Honaker, #777)
the figure file numbering scheme has changed: for a chunk with a label foo, its figure files are named as foo-i where i ranges from 1 to n (the total number of plots in this chunk); previously, the figure file was named as foo instead of foo-1 when there was only one plot generated in this chunk, which has a potential bug: consider two chunks named foo and foo2, respectively; foo generates two figures foo1.png and foo2.png, and foo2 generates one figure foo2.png, which will overwrite the second figure generated from the chunk foo (thanks, @kevinushey, @kohske, @kforner, #704, #832)
for warnings and errors from code chunks, the call that produced them will be printed as part of the message, e.g. previously an error might just be Error: x must be positive, and now it may be Error in FUN(x = -1): x must be positive (thanks, @jennybc, #846)
for the engine coffee (CoffeeScript), the flag -p has been removed from the command line arguments, which means the default behaviour of this engine is to evaluate the code, instead of printing JavaScript; if you want the old behaviour, you need the chunk option engine.opts = '-p' (thanks, Jake Burkhead, #821)
when the chunk option results = 'hold', the text output blocks will be collapsed into a single block (thanks, Gavin Simpson, #798)
the video format for animations (when the chunk option fig.show='animate') was changed from OGG to WebM (http://www.webmproject.org), which has many benefits over other formats, especially for the web (thanks, @gaorongchao, #641)
the YAML metadata in Markdown child documents will be ignored (only the metadata in the top parent document is preserved)

MINOR CHANGES

scientific formatting for inline R output is only applied to objects of which the first class is numeric, e.g. chron::chron() objects will no longer be formatted using scientific notations (thanks, @sanfordweisberg, #806)
for R Markdown v2 documents, if the inline R output is formatted using the scientific notation, the output must be put in a math environment, e.g. $`r 2e10`$

CHANGES IN knitr VERSION 1.6

NEW FEATURES

added Textile (http://txstyle.org) support thanks to Richard Cotton; see the example 105 at https://github.com/yihui/knitr-examples (#623)
added an argument inline to spin() so that some R code (by default, code of the form {{code}}) can be treated as inline expressions (thanks, Kirill Müller, #620)
added an argument prefix to write_bib() so that we can customize the prefix for bib entries; we can also set options(knitr.bib.prefix = 'a_string') so that write_bib() uses this global option as the default value for prefix (thanks, Michael Friendly)
the chunk option dev.args will be passed to the pdf recording device if pdf is also in the chunk option dev (note the recording device can be different with the actual drawing device); for example, you can pass a different pointsize to the PDF recording device (thanks, Felix Klein)
intermediate files generated during spin() are deleted by default, this can be overridden by setting the new parameter precious to TRUE (thanks, Kirill Müller, #628)
exposed the previously internal chunk option code, which is used to store the source code of a chunk; now users also have control over this option, which means we can programmatically assign source code for a code chunk, e.g. code = capture.output(dump('fivenum', ''))
added a new chunk option collapse for Markdown output; if collapse = TRUE, knitr will try to merge the source and output blocks into one block; see example 039 at https://github.com/yihui/knitr-examples (thanks, Hadley Wickham)
added a new chunk option fig.retina for better display quality of images in HTML output; for example, the physical size of an image is doubled and its display size is halved when fig.retina = 2
added a new chunk option strip.white to trim the white lines in the beginning or end of a source chunk in the output (strip.white = TRUE by default)
added a new chunk option render for custom rendering of R objects in the output; by default, the render function is print() (or show() for S4 objects) (thanks, Milan Bouchet-Valat, #484)
added a new S3 generic function knit_print(), which is used as the default printing function for R objects in the code output now; users can define custom S3 methods to change the printing behaviour of R objects in knitr
added a new engine scala for Scala thanks to Simeon Fitch (#640)
added a new engine named asis to write the chunk content without processing it; it also respects the chunk options echo and eval -- when either one is FALSE, the chunk will be hidden; this makes it possible to write text conditionally (thanks, Simon, #622)
the Haskell engine supports multiline clauses using :set +m now, thanks to Adam Vogt (#633)
added a new hook function hook_purl() so that code can be more reliably extracted from input documents; see ?hook_purl for details (#607)
exported a function plot_crop() which was used in hook_pdfcrop() in previous versions to crop the white margin of plots
added a new chunk option cache.lazy (TRUE/FALSE) to decide whether to save and lazy load objects when cache is enabled; for really large objects, you may need cache.lazy=FALSE (thanks, Dario Strbenac, Scott Simpkins, and Mattrition, #572)
added a chunk option cache.comments (TRUE/FALSE) to decide whether updating R comments in a code chunk should invalidate the cache database (thanks, @knokknok, #718)
kable() supports alignment for HTML tables now (thanks, Joseph Larmarange, #656 and #683)
kable() supports table captions in LaTeX, HTML and Pandoc's Markdown now (thanks, Joseph Larmarange, #687)
the digits argument for kable(x) can also be a vector of length ncol(x), in which case the number of digits is set for each column separately, e.g. kable(data.frame(x=rnorm(10), y=rnorm(10)), digits = c(1, 4)) (thanks, @nacnudus, #771)
for Markdown tables, kable() gained a new argument padding to specify the inner padding of table cells using spaces (thanks, @gavril0, #699)
added a new vignette engine called rmarkdown, which uses rmarkdown::render() to create a package vignette from an R Markdown document; see https://rmarkdown.rstudio.com for more information about the rmarkdown package, and the vignette knit_print.Rmd in knitr for an example (basically you specify \VignetteEngine{knitr::rmarkdown} in your vignette)
indentation is preserved when using chunk references <<>>, i.e., if <<>> is indented, the spaces before it will be applied to the code that it refers to (thanks, Terry Therneau)
added a chunk option fig.process, which can be set as a function to take the path of a plot and process it, e.g. fig.process = function(x) knitr::plot_crop(x); note that the fig.process function must return a character string, such as the path of the figure, which might have been changed inside the function
added a chunk option R.options so that we can temporarily set local options() for the current code chunk, and the options will be restored after the chunk (thanks, @r2evans, #764)
the argument shortcode in knit2wp() can take a logical vector of length 2 now; the first element determines whether to highlight source code, and the second decides syntax highlighting for text output (thanks, Derek Ogle, #663)
Sweave2knitr() will report the line numbers indicating where the Sweave-specific syntax was used (thanks, Kirill Muller, #676)
added a function current_input() to return the filename of the input document (thanks, Kate Davis, #701)
for Markdown, if n (n >= 3) backticks are detected in the code output, the output will be wrapped in n+1 backticks to make sure the original backticks in the output are not interpreted as the token of fenced code blocks
for a chunk hook function, we can use a fourth optional argument name, which takes the value of the hook name, e.g. for knit_hooks$set(foo = hook_foo), hook_foo can be of the form function(before, options, envir, name), where name == 'foo' (thanks, Thell Fowler, #733)
all the four arguments before, options, envir, and name for a chunk hook are optional now, e.g. you can define a hook function of the form function(before, options), function(before, name, envir), and so on
added two read-only chunk options out.width.px and out.height.px, which are the numbers of pixels calculated from fig.width and fig.height (which have units in inches), respectively, and the chunk options out.width and out.height can override the calculated natural size
added a new argument encoding to the pandoc() function to specify the character encoding of the input file(s) (fixed the problem http://stackoverflow.com/q/22198832/559676)

BUG FIXES

due to the change in evaluate v0.5, evaluate() may return the raw values of expressions, but the S3 method wrap() does not know how to handle them; now these values are just ignored (thanks, Dan Tenenbaum)
fixed a bug for dep_auto() that may occur if old cache files generated from previous versions of knitr are used (thanks, Jeffrey Racine)
fixed the bug reported at http://stackoverflow.com/q/19166724/559676: the inline hook did not work well with non-numeric values, e.g. Date (thanks, Waldir Leoncio)
infinity is formatted as a symbol in inline LaTeX and HTML modes (thanks, Kirill Müller, #629)
kable() did not keep the row name when the data only has one row (thanks, @eev2, #636)
kable() did not recycle the align argument correctly (thanks, Adam Cooper, #638)
kable() chokes on tables with NA's (thanks, Harlan Harris and Benjamin Schiller, #720)
kable() did not work with non-numeric data frames (thanks, @talexand, #702)
for Markdown/reST tables, kable() should make sure there are at least one empty line before the table output (thanks, @talexand, #705)
fixed a bug related to child documents -- chunk options passed from a parent document may not be restored after the child document exits (thanks, Frank Harrell, http://bit.ly/17yitsD)
fixed a bug for hook_rgl() when the chunk option dev is a vector of length greater than one (thanks, Ashley Manton, #703)
hook_rgl() did not work with the chunk option fig.cap (thanks, Kohske Takahashi, #700)
%\documentclass{} was mistakenly treated as the specification of the document class instead of a comment in a LaTeX document (thanks, Jarad Niemi, #713)

MAJOR CHANGES

the default value for the chunk option tidy is FALSE now, which means the R source code in chunks will no longer be reformatted by formatR::tidy.source() by default; this feature must be explicitly turned on by tidy=TRUE, and it has brought a lot of confusion in the past, so it is perhaps a good idea not to reformat the source code by default
now we treat the closing mark (e.g. @ in Sweave or the three backticks in R Markdown) as part of the code chunk, instead of the beginning of a text chunk, and a consequence for this change is that knitr no longer adds blank lines to the beginning of the text chunks (thanks, Thell 'Bo' Fowler, #752)
inline R expressions will no longer be evaluated in try(), which means errors in inline R code will be emitted immediately
the first argument of the plot hook is the filename of the plot now; in previous versions, it was a vector of length 2 (basename and file extension); see ?hook_plot
the default value for the format argument in pandoc() has changed: it defaults to the t field in the configuration if found, otherwise, it defaults to 'html' (thanks, Kohske Takahashi, #697)
in the previous version, we can set options(knitr.foo = value) so that knitr can adjust the package options opts_knit$set(foo = value) before knitting a document; now the prefix for package options has been changed to knitr.package., i.e. we should set options(knitr.package.foo) to achieve opts_knit$set(foo); besides, it is also possible to change the default chunk options using options(knitr.chunk.foo) now, but you are warned that this may bring reproducibility issues, so please use with care

MINOR CHANGES

for R Markdown/AsciiDoc, line breaks are allowed in the inline R expressions now (thanks, Andrew MacDonald, #655)
for AsciiDoc input, the R inline expressions can be written using the same syntax as R Markdown, i.e. r expression
numbers containing characters other than digits 0-9, periods or commas are formatted as math in inline LaTeX, for negative numbers, infinity symbol, corner cases such as 10^{n}, ... (thanks, Jeffrey Racine and Kirill Müller, #635)
the column name id for row names in the kable() output is removed when the output format is Markdown (thanks, Artem Klevtsov, http://stackoverflow.com/q/19875529/559676)
for R Markdown, special characters in figure filenames are no longer replaced by _, since RStudio v0.98.490 has fixed the base64 encoding bug; please upgrade RStudio to the latest version: http://www.rstudio.com/ide/download/
the tikzMetricsDictionary option (for the tikzDevice package) is no longer overridden if it has already been set in options() (thanks, @rmatev, #708)

CHANGES IN knitr VERSION 1.5

NEW FEATURES

a new option value results='hold' to flush all text output to the end of a chunk like fig.show='hold' (thanks, Harlan Harris, #593)
when cache is enabled, automatic chunk dependencies can be truly automatic now; there is no need to call dep_auto() explicitly, and all we need to do is the chunk option autodep=TRUE; the chunk dependencies will be rebuilt after each chunk, so when new chunks are inserted into the document, knitr can also figure out the new dependencies automatically (thanks, @knokknok, #592)
for Sublime Text users, there is a SublimeKnitr package to support LaTeX and Markdown with knitr; thanks, Andrew Heiss (#449) (this is not really a new feature of knitr itself, though)
now the chunk options warning and message can also take numeric values as indices to select which warnings/messages to include in the output (thanks, Simon Urbanek, #590)

BUG FIXES

code changes in chunks should invalidate the cache when the chunk option cache < 3; fixed by @knokknok in #587
fixed the bug reported at http://stackoverflow.com/q/18302179/559676; before evaluating inline R code, the chunk options eval must be evaluated to a logical value
fixed the bug reported at http://stackoverflow.com/q/18544045/559676; read_chunk() no longer excludes code without chunk headers, so stitch() will include all the code as expected
fixed the bug reported under http://stackoverflow.com/a/18541083/559676; now \ensuremath{} can be correctly applied to numbers of the form 10^{-n}
fixes a regression bug reported by Graham Williams: https://groups.google.com/forum/#!topic/knitr/_I5rlo9tOeA the global chunk options set in child documents are no longer restored
fixed #604: kable() did not work on data frames/matrices of one row (thanks, Kevin Ushey)
fixes the bug reported at http://stackoverflow.com/q/18992260/559676; render_jekyll('prettify') should have pasted the source code lines into one character string

MAJOR CHANGES

when the chunk option cache=2, the recorded plots (i.e. display lists) will no longer be cached, and the figure files will be assumed to exist the next time the document is compiled, otherwise the cache will be purged and the chunk needs to be recomputed; this will save some disk space and avoid bugs like #588 (thanks, @knokknok)

MINOR CHANGES

the Rnw chunk syntax is more tolerant on chunk headers in the sense that any characters after <<>>= are discarded, e.g. <<echo=TRUE>>===foo will be treated as <<echo=TRUE>>= (thanks, Michael Friendly)
knitr:::.onLoad() no longer modifies the PATH variable when /usr/texbin is not in PATH under Mac OS
when a message/warning/error contains line breaks, they are preserved and the message will not be re-wrapped (#602, thanks, Tyler Rinker)
read_chunk() tolerates white spaces at the end of the chunk headers now (suggested by John Maindonald, #606)
for R HTML documents, only the highlight component in opts_knit$get('header') was used in previous versions; now all components except framed and tikz are used; this makes it possible to further customize the HTML header (thanks, Wahlen Neuwirth and Erich Neuwirth)
in previous versions, the global option KNITR_WIDTH in R was used to set options(width); now this option is set from opts_knit$get('width'), which has the same default value 75 (#597)

CHANGES IN knitr VERSION 1.4

NOTE

if you are using Windows or Linux, you need to read the news about both versions 1.3 and 1.4, because the version 1.3 did not manage to survive on CRAN

NEW FEATURES

the cache system is more granular now: instead of the binary option cache = TRUE / FALSE, we can specify cache = 0, 1, 2, 3 (TRUE indicates 3, and FALSE indicates 0; for 1 and 2, see the documentation for cache: https://yihui.org/knitr/options); this means we may change the chunk options involved only with output rendering (e.g. from echo = TRUE to FALSE, or set fig.cap = 'a new caption') without breaking the cache (thanks, Jeroen Ooms, Clark Kogan, and Roman Lustrik, #396, #536)
added two new vignette engines called docco_linear and docco_classic using the Docco styles (http://jashkenas.github.io/docco/); see browseVignettes(package = 'knitr') for examples
added a function rocco() to compile R Markdown documents to HTML using the classic Docco style, i.e. a two-column layout, with text on the left and code on the right (thanks, Weicheng Zhu, #577)
added an argument comment in spin() to specify comment lines that will be ignored by spin(); by default, the block comment /* comment */ is recognized; thanks, Bryan Hanson http://stackoverflow.com/q/17664401/559676
it is possible to set package options prior to loading the knitr package now: for a package option foo, we can set options(knitr.foo = value) so that knitr will opts_knit$set(foo = value) when calling knit(); see ?opts_knit for details (thanks, Zhiguang Zhao)
added a new argument ext to the pandoc() function so that users can manually specify the output filename extensions (thanks, Baptiste, http://stackoverflow.com/q/17710249/559676)
for LaTeX and HTML output, syntax highlighting can be done for languages besides R (e.g. Python, Perl, ...); this is achieved by highr::hi_andre(), so Andre Simon's Highlight must be installed, otherwise knitr will fall back to verbatim output for source code; see https://github.com/yihui/knitr-examples/blob/master/098-highlight-python.Rnw for an example (#495)

MAJOR CHANGES

knitr formally depends on the highr package now (for syntax highlighting of LaTeX and HTML)
the package option stop_on_error has been deprecated; now it is much easier to specify whether you want to stop on errors or not by using the existing chunk option error; if you want to stop, use error=FALSE; see the documentation for details: https://yihui.org/knitr/options
the meanings of the chunk options warning and message when they take the value FALSE have also changed: FALSE means the warnings/messages will be printed in the R console and not recorded; this makes it easier to know when/where the warnings/messages were produced during knit()
syntax highlighting and adding prompts are done in the source hook now; in previous versions, they were done in knitr:::wrap.source; now the source hook receives the pure source code instead of syntax highlighted code

MINOR CHANGES

for the chunk options set in package option opts_knit$get('eval.after'), they will not be evaluated after a chunk if eval=FALSE for that chunk (#570, thanks, @knokknok)
for document formats that produce HTML output, the default graphical device is changed to png only if it is pdf; if the device has been changed to values other than pdf, knitr will no longer modify it internally; when the png device is not available, the svg device will be used instead
removed the global option KNITR_PROGRESS, which was introduced to suppress the progress bar, but now we have got knit(..., quiet = TRUE), so this option is redundant

BUG FIXES

in LaTeX output, the double quotes " in the messages, errors, and warnings are replaced by "{} because they might cause trouble to babel: http://stackoverflow.com/q/18125539/559676 (thanks, Thierry)

CHANGES IN knitr VERSION 1.3

NEW FEATURES

added support for AsciiDoc; see example 089 at https://github.com/yihui/knitr-examples/blob/master/089-minimal.Rasciidoc (thanks, Richard Cotton)
added support for reStructuredText vignettes in packages; now *.Rrst documents are recognized as vignettes as well, and they will be compiled to PDF via rst2pdf (#533) (thanks, Trevor L Davis)
a new function kable() to produce simple tables; see the vignette datatables in browseVignettes(package = 'knitr') for an application; also see a LaTeX example at https://github.com/yihui/knitr-examples/blob/master/091-knitr-table.Rnw
the chunk options fig.width and fig.height are vectorized according to the dev option, e.g. fig.wdith = c(7, 10) and dev = c('pdf', 'png') (#538) (thanks, @baptiste)
for purl(), code chunks with the option purl=FALSE will not be included in the R script (#519, thanks, Sebastian)
new 'hide' value to the fig.show option; the figures are created but not included in the output document (#532) (thanks, Simon)
the sas engine uses the listings output now (#541) and was tweaked for better LaTeX display (#562) (thanks, Nick Salkowski)
added a quick and dirty c engine (via R CMD SHLIB); see https://github.com/yihui/knitr-examples/blob/master/090-engine-c.Rmd for an example
added a new engine asy for Asymptote, a vector graphics language (http://asymptote.sourceforge.net); see examples 093 at https://github.com/yihui/knitr-examples (thanks, Thibaut Lamadon, #559)
added a new engine cat to write the content of a code chunk to a file; see the example 095 at https://github.com/yihui/knitr-examples (thanks, Stephen Eglen)
added a new function knit_exit() to allow knit() to exit early as if it had reached the end of the document (#524, thanks, Renaud)
the chunk option fig.align also works for Markdown output now, e.g., fig.align = 'center' will center images in HTML via the style attribute of the <img> tag (#387)
the argument format in the pandoc() function was vectorized, e.g. we can call pandoc(input, format = c('html', 'latex')) and the input file will be converted to HTML and LaTeX, respectively (#547, thanks, Jeroen Ooms)
added an argument options to knit_child() to set global chunk options for child documents; if a parent chunk calls a child document (via the child option), the chunk options of the parent chunk will be used as global options for the child document, e.g. for <<foo, child='bar.Rnw', fig.path='figure/foo-'>>=, the figure path prefix will be figure/foo- in bar.Rnw; see http://stackoverflow.com/q/17514055/559676 for an application
eclipse_theme() works with font weight (bold) and font style (italic) now when parsing themes from http://eclipsecolorthemes.org
added two package options latex.options.graphicx and latex.options.color to allow customization of LaTeX package options, e.g. opts_knit$set(latex.options.color = 'monochrome') generates \usepackage[monochrome]{color} in the LaTeX output (#546)
added a new package option unnamed.chunk.label to set the chunk labels for unnamed chunks, e.g. opts_knit$set(unnamed.chunk.label = 'fig') will generate chunk labels fig-1, fig-2, ... (#555) (thanks, Noam Ross)
when knit() is running, a global option knitr.in.progress will be set to TRUE; this allows other package authors (e.g. rCharts and googleVis) to adjust certain package options according to getOption('knitr.in.progress') (thanks, Ramnath V, #564)

BUG FIXES

fixed #502: using layout() and par() at the same time under R 3.0.0 may lead to a corrupt plot (thanks, Hong Xu http://tex.stackexchange.com/q/108335/9128)
fixed a bug in pandoc(): for single-lettered Pandoc arguments, the values are passed to them after spaces instead of equal signs (reported at http://stackoverflow.com/q/16569010/559676)
fixed #542: when a child document has a sub-child document and also uses set_parent(), the LaTeX header will be added to the wrong file (thanks, Johan Toloe)
stitch_rmd() was using a wrong R Markdown template
fixed #537: misleading error message when the graphical device does not exist (thanks, Scott Kostyshak)
fixed a bug in hook_rgl() reported at http://cos.name/cn/topic/110742 (incorrect LaTeX code when fig.align='center' and fig.show='hold')
fixed #565: added the envir argument to spin() so that the code is evaluated in the correct environment by default (thanks, @GillesSanMartin)

MAJOR CHANGES

purl() will write the chunk headers in # ---- label, opts ---- instead of # @knitr label, opts, i.e. @knitr was changed to four dashes ----; similarly, read_chunk() will use the pattern # ---- label ---- to read external code (the old syntax # @knitr is still preserved for compatibility); the main reason for this change is that RStudio uses four dashes in R comments as section headings, so that it is possible to fold sections of R
syntax highlighting is done by the highr package if it has been installed, otherwise, the old regular-expression-based syntax highlighting will still be used; the highr package does much better syntax highlighting than the regexp-based approach (#327)
the commands for syntax highlighting were changed for compatibility with Andre Simon's Highlight package; this will affect LaTeX and HTML users, e.g. \hlnumber was renamed to \hlnum; cached LaTeX and HTML will have to be rebuilt for the new syntax highlighting commands to work (#470)
the argument eval was removed in knit_child(); if we do not want to evaluate a child document, we can set eval=FALSE in its parent chunk
the script inst/bin/knit gains an option -o to specify the output filenames for knit() (#525, thanks, Aaron Wolen)
the default video format for animations is OGG (it is open and free) instead of MP4 (non-free) now; this means Internet Explorer under Windows may not work with the animations (consider Firefox, Chrome and other modern web browsers)
warnings and messages in adjacent output chunks are merged, respectively (#534)
when the package option verbose = TRUE, the time stamp will be printed after each chunk using timestamp(), but this will mess up with the R command history, so now knitr uses cat() to write the time stamp (#545) (thanks, @knokknok)
the argument base in read_rforge() was removed and hard-coded inside the function instead
for Markdown output, the figure filenames no longer allow special characters like spaces, and special characters will be automatically replaced by _; this change will avoid problems when publishing to RPubs from RStudio: if figure paths contain special characters, the figures will not be uploaded (thanks, Sangsoon Woo)

MINOR CHANGES

the package vignettes uses \VignetteEngine{knitr::knitr} instead of \VignetteEngine{knitr} so that the next version of R can compile the vignettes out of the box (via R CMD Sweave) and no longer need to build the whole package in order to build the vignettes

MAINTENANCE

the package vignettes were moved to the vignettes directory from inst/doc since the former will be preferred by the future versions of R
the testing is done via the testit package now (https://cran.r-project.org/package=testit)

MISC

knitr becomes an affiliated project of FOAS (Foundation for Open Access Statistics): http://www.foastat.org/projects.html

CHANGES IN knitr VERSION 1.2

NEW FEATURES

added a Pandoc wrapper function pandoc() to convert Markdown documents to other formats that Pandoc support such as LaTeX/PDF, OpenDocument, HTML5 slides and unfortunately also Word (#206)
in the past, the chunk hooks were ignored when the chunk option engine != 'R' (i.e. code in the chunk is not R); now they are executed regardless of the language of the code chunk (#483) (thanks, @cdrv)
multiple WebGL objects from the rgl package can be generated in the same web page now; each object is uniquely identified by the corresponding chunk label; see https://dl.dropbox.com/u/15335397/misc/webgl-rmd.html for an example (#494) (thanks, Carson Sievert)
if multiple graphical devices are used (the chunk option dev is a vector of length greater than 1), the chunk option dev.args can be a named list of lists of device arguments, with each element a list of arguments to be passed to the single device; see https://yihui.org/knitr/options
as announced in the last version, R 3.0.0 will support non-Sweave vignettes; now it is also possible to compile R HTML vignettes via knitr since *.Rhtml files are also registered by knitr as vignette files
a new chunk option cache.vars to manually specify which variables to save in the cache database; by default all newly created and modified variables are identified and saved, but in some cases, knitr may not be able to identify the modified variables, e.g. DT[, foo:=value] in data.table (we can set cache.vars='DT' to force knitr to save a copy of DT)
added a new engine Rscript to run the R code in a new R session; see http://stackoverflow.com/q/15271406/559676 for an example
the executable script inst/bin/knit can accept multiple input files now (e.g. knit foo.Rnw bar.Rmd zzz.Rhtml)
knit() and knit2html() gained a quiet argument to suppress messages and the progress bar (thanks, Vince Buffalo)
added the text argument to spin() and stitch() respectively as an alternative way to provide the input like knit() (#509) (thanks, Craig Watson)
a new function wrap_rmd() to wrap long lines in Rmd files without affecting the code blocks (if there are any); this makes it easier for version control purposes
rst2pdf() will pass a default output filename to rst2pdf (if the input is foo.rst, the output will be foo.pdf)
knit2wp() gained a new argument publish = TRUE (thanks, Eric Nantz) (#512)

BUG FIXES

fixed the problem reported in http://stackoverflow.com/q/12448507/559676 now \usepackage{upquote} should appear after \usepackage{fontenc}, so single quotes in verbatim environments will no longer cause problems
fixed #487: stitch_rhtml() and stitch_rmd() should not use the chunk option out.width = '.6\\linewidth' (thanks, Tal Galili)
when the chunk option engine is not R, the code is also executed under the directory opts_knit$get('root.dir') (if specified); thanks, Winawer http://stackoverflow.com/q/15512545/559676
: is permitted in fig.path now (#513) (thanks, Sebastian)
fixed an encoding problem (CP950) for Hong Kong Windows users reported at http://bit.ly/16RQL5E

MAJOR CHANGES

all child documents are inserted into the parent document as character strings of the (compiled) content, instead of being saved into files (e.g. \input{foo-child.tex}); no matter how many child documents there are, only one main output file will be generated; the package option child.command was removed accordingly since it is no longer used
no longer generates concordance data for child documents; the past attempt did not really work well and the implementation was complicated, so now we only support concordance for the main document; the consequence of this change is the synchronization between PDF and Rnw for child documents no longer works at the line level (clicking in PDF will still bring the focus back to the child chunk)
in previous versions, cached chunks were evaluated in separate (empty) environments in order to capture the newly created variables, but this brings confusion when we use functions depending on the current environment such as ls() (which will return character(0)); now all chunks, cached or not, are evaluated in the same environment knit_global() (finally fixed #456)
knit2pdf() and knit2html() return the output filename when the input is a file (in previous versions, NULL was returned in this case)
the package option stop_on_error is set to 2 now when building package vignettes, which means R will stop on errors in vignettes; this makes it easier to find out possible problems in vignettes during R CMD build
the document hook hook_rjournal() was removed; it was too hackish (see https://yihui.org/en/2013/02/contribute-to-the-r-journal-with-lyx-knitr/ for how to write an article for The R Journal in a less hackish way)

MINOR CHANGES

the progress bar symbol was changed from > to . so it looks less intrusive (#395) (thanks, Michael Friendly)

DOCUMENTATION

the knitr book is forthcoming: http://www.crcpress.com/product/isbn/9781482203530 run citation('knitr') or toBibtex(citation('knitr')) in R to obtain the citation info
open help(package = 'knitr', help_type = 'html') to see the vignette examples (Rnw, R Markdown and R HTML)

CHANGES IN knitr VERSION 1.1

NEW FEATURES

(experimental) R 3.0.0 will support non-Sweave vignettes, e.g. Rnw documents can be compiled by knitr instead of Sweave; in addition, R Markdown vignettes have also become first-class citizens as R package vignettes; see https://yihui.org/knitr/demo/vignette/ for details
a new engine for CoffeeScript (i.e. the chunk option engine='coffee'); see https://github.com/yihui/knitr-examples/blob/master/080-engine-coffeescript.Rmd for an example (thanks, Nacho Caballero)
when the chunk option eval=FALSE, purl() will comment out the code when extracting code chunks (thanks, Randall Pruim)
the global option KNITR_PROGRESS can be used to set the package option progress in opts_knit, e.g. after options(KNITR_PROGRESS = FALSE), this option will be set to FALSE when the package is loaded (#395)
the global option KNITR_WIDTH can be use to set the R option width before knitr is called; in the past this option was hard-coded to 75, and now it is set as options(width = getOption('KNITR_WIDTH', 75L))
a new function knit2wp() which compiles R Markdown documents and publishes the results to WordPress; see https://yihui.org/knitr/demo/wordpress/ for details
a new hook hook_webgl() which writes the WebGL code of an rgl scene into the output using rgl::writeWebGL() so we can reproduce a 3D plot in the browser (thanks, Stephane Laurent http://stackoverflow.com/q/14879210/559676)

BUG FIXES

fixed #465: when eval=FALSE and echo is numeric, the code was incorrectly filtered by the indices in echo (thanks, @ateucher)
> was not allowed in HTML inline code expressions (http://stackoverflow.com/q/14360296/559676); now the regular expression works correctly to look for --> instead of >
set_parent() should not print NULL in the child document: http://stackoverflow.com/q/14487718/559676 (thanks, Thomas Holz)
child documents now inherit the encoding parameter from their parent document, i.e. the encoding value in knit(..., encoding = ?) is applied to knit_child(); note this assumes the parent document and the child documents use the same encoding (thanks, Henrik Nyhus)

MAJOR CHANGES

empty inline expressions are no longer recognized, e.g. \Sexpr{} will not be parsed; this allows one to write such markers in tutorials (e.g.  and `r `); internally all_patterns$foo$inline.code was changed
the function build_dep() was removed (the warning has been there for a long time); please use dep_auto() instead
the package option filter.chunk.end was removed; this means in Rnw documents, a single line @ has the meaning of terminating a chunk only if there is a chunk header <<>>= before it; otherwise it does not have any special meanings
the function run_chunk() was removed; it is redundant because we already have the chunk option ref.label as well as in-chunk reference <<label>>

MINOR CHANGES

the function imgur_upload() uses Imgur API version 3 now; if you are using the key obtained from version 2, you need to register for your own client id: http://api.imgur.com (#439)
allow users to pass a custom environment to Rcpp::sourceCpp() in the Rcpp engine; fixes http://stackoverflow.com/q/14882486/559676
slight improvement of encoding support in knit()

CHANGES IN knitr VERSION 1.0

NEW FEATURES

a new function knit_expand() which is similar to brew and mustache, e.g. it expands pi is {{pi}} to pi is 3.14; it can also be used for building child documents (see https://github.com/yihui/knitr-examples/blob/master/075-knit-expand.Rnw for example) (#397) (thanks, Frank Harrell)
knit() gained a new argument encoding to specify the encoding of the input file (multilingual support is complete now), e.g. knit(..., encoding = 'GBK') for Simplified Chinese
a new function Sweave2knitr() to convert Sweave documents to knitr; several automatic translations can be done, e.g. results=tex to results='asis', width=5 to fig.width=5, echo=true to echo=TRUE, keep.source=TRUE to tidy=FALSE, eps=TRUE to dev='postscript', \SweaveOpts{...} to opts_chunk$set(...) and so on; see the documentation in the package for details (#451)
if the Sweave syntax is detected in an Rnw document, a message box (tcltk) will pop up as a reminder on how to fix it
inline R code also respects the option opts_knti$get('stop_on_error') now, e.g. if we set this option to 2L, R will completely stop when error occurs in inline R code
a new function all_labels() to get all chunk labels in a document; see one application at https://github.com/yihui/knitr-examples/blob/master/073-code-appendix.Rnw
chunk hooks will be run (if exist) even if the code chunk is empty
added two wrapper functions stitch_rhtml() and stitch_rmd() which use the R HTML and R Markdown templates respectively when calling stitch()
the chunk label is used as the id of the div element in R HTML output, e.g. <div id='chunk-label'>...</div>

MAJOR CHANGES

(IMPORTANT) the internal compatibility with Sweave has been dropped as scheduled in the last version, and the knitr parser was greatly simplified accordingly; Sweave users can call the function Sweave2knitr() to convert old Rnw files to knitr files before running knit()
accordingly, the pattern elements global.options and inline.doc were removed from knit_patterns (\SweaveOpts{} and \SweaveInput{} will no longer be supported; please call Sweave2knitr() to convert incompatible Sweave documents)
chunk labels can be arbitrary characters now; in the past they had to be valid R symbols, e.g. 2a was an invalid label; this restriction has been removed, because chunk labels will be automatically quoted when necessary (<<2a>>= will become <<'2a'>>=, but <<'3-function'>>= will remain untouched)
if the chunk option include=FALSE, the evaluation will stop if errors occur (i.e. stop_on_error=2L for evaluate) because otherwise, it will be very difficult for authors to notice errors in chunks which have include=FALSE (#453) (thanks, Joshua Pritikin)
the function knit_env() is no longer available (it is not exported any more), and knit_global() has been exported now

MINOR CHANGES

for inline R code, the value is returned only if the R code prints a visible value, e.g. \Sexpr{x <- 1} will be empty, and \Sexpr{pi} will return the value of pi

BUG FIXES

fixed #432: no longer uses \\\\ in LaTeX output; only a single line break is converted to \\ (thanks, Kevin Wright)
render_html() guarantees that the R source code is highlighted when the chunk option highlight = TRUE (#447) (thanks, Ramnath Vaidyanathan)
dep_auto() was unable to find the cache files if the input document is not under the current working directory (thanks, Hui Yao)

Documentation

because Github has deprecated downloads, all downloads were moved to Bitbucket, and the links in the package website, as well as all examples, have been updated (#438)

CHANGES IN knitr VERSION 0.9

NEW FEATURES

added a demo named notebook which is an R notebook based on the shiny package (https://github.com/rstudio/shiny); use demo('notebook', package = 'knitr') to see it, or visit http://glimmer.rstudio.com/yihui/knitr
for numeric inline output in LaTeX, the I() trick is no longer needed, e.g. $x=\Sexpr{1.2e10}$ is safe for LaTeX now due to \ensuremath{} (#137) (thanks, Randall Pruim)
the chunk option eval can take numeric values now, specifying which expressions to evaluate in a chunk (#383) (thanks, Jared Lander)
a new package option stop_on_error which specifies the behavior of errors in code chunks; this option is passed to the evaluate package, e.g. opts_knit$set(stop_on_error = 2L) will make knitr completely stop on errors (the default value is 0L which means to move on even if errors occurred); this makes it possible to see the call stacks via traceback() in an interactive R session when an error occurs (#344) (thanks, Hadley Wickham and Dzidorius Martinaitis)
added support to the Rcpp package through the chunk option engine='Rcpp' so that we can write C++ source code in the document; see https://github.com/yihui/knitr-examples/blob/master/029-engine-Rcpp.Rmd for an example (#415) (thanks, JJ Allaire)
knitr throws a warning when a cached chunk depends on an uncached chunk because this kind of dependency will be ignored (#431) (thanks, @ghostwheel)
a list of arguments can be passed to formatR::tidy.source() as the chunk option tidy.opts now, e.g. tidy.opts=list(width.cutoff=60, keep.blank.line=FALSE) (#429)
some chunk options are recycled for plots such as fig.env, out.width and out.extra, etc; this means if there are multiple plots per chunk, we can specify different output options for them individually (e.g. out.width=c('2in', '.4\\linewidth') for two plots); see https://github.com/yihui/knitr-examples/blob/master/067-graphics-options.Rnw for an example (motivated by #430) (thanks, @autumnlin)
added a new chunk option fig.subcap for captions of subfigures in LaTeX; when there are multiple plots in a chunk, and neither fig.subcap nor fig.cap is NULL, \subfloat{} will be used for individual plots (you need to add \usepackage{subfig} in the preamble); also see https://github.com/yihui/knitr-examples/blob/master/067-graphics-options.Rnw for an example (#388) (thanks, @skipperhoyer)
stitch() accepts labelled R scripts now; if an R script contains chunk headers of the form ## @knitr label, options, they will be used in the template (#411) (thanks, @jamiefolson)
the function read_chunk() gained a few new arguments so that we can reference external code chunks in another way, which was an idea from the SweaveListingUtils package (thanks, Peter Ruckdeschel)
a new function read_demo() based on read_chunk() to read demo scripts in R packages
a new convenience function read_rforge() to read code from R-Forge repositories; combined with read_chunk(), it can insert R code from R-Forge into knitr dynamically; see https://github.com/yihui/knitr-examples/blob/master/046-read-rforge.Rmd for an example (thanks, Peter Ruckdeschel)
chunk options are also written after ## @knitr along with chunk labels when tangling R scripts via purl()
purl() gained a new argument documentation to also write documentation lines into the R script (#391 and #401) (thanks, Noam Ross and Fernando Mayer)
knit_rd() generates a navigation frame on the left and builds links now; this is like the CHM help in old days (thanks, Michael Friendly)
a new function knit_rd_all() to build static HTML help pages for all the packages installed
we can also use ## @knitr to write chunk options for spin() now (#+ and #- still work)
added new language support for Perl and Z Shell (zsh); see an example at https://github.com/yihui/knitr-examples/blob/master/028-engine-perl.Rmd (#406) (thanks, Jim Hester)
render_jekyll() gained an argument highlight to specify which highlighting engine to use (Pygments or Prettify.js) (thanks, Yanping Chen)
two new chunk options for language engines: engine.path and engine.opts; the former can be used to specify the path of the program (e.g. <<engine='ruby', engine.path='/usr/bin/ruby1.9.1'>>=); the latter can be used to pass additional arguments to the engine program
added new engines for GraphViz (engine='dot') and TikZ (engine='tikz'); see https://github.com/yihui/knitr-examples/blob/master/057-engine-dot.Rmd and https://github.com/yihui/knitr-examples/blob/master/058-engine-tikz.Rmd for examples (#419) (thanks, Michel Kuhlmann)
added a preliminary engine for SAS which is basically a call like system('sas chunk-code.sas') (#354)
a new document hook to post-process the LaTeX output document to move code chunks out of figure/table environments so that code chunks will not float with the environments; see ?hook_movecode for details
chunk hooks are called in the reverse order after a chunk (and natural order before a chunk); this allows one to, e.g. write an opening environment before a chunk and close it properly after a chunk
all language engines also respect the comment option when writing output just like R code chunks (by default the output is commented out by ##)
added a new function set_alias() as a wrapper to opts_knit$set(aliases = ...), e.g. set_alias(w = 'fig.width') sets w as an alias for the chunk option fig.width

MAJOR CHANGES

global options are strongly recommended to be set via real R code opts_chunk$set(opt = value) in a code chunk instead of the old syntax in text chunks like \SweaveOpts{opt = value}, or , etc, which will be deprecated in the next version; this will make it cleaner and safer to parse the source documents, e.g. we can write arbitrarily complicated expressions like opts_chunk$set(fig.width = if (foo == 'beamer') { 5 } else { 7 }) which is impossible in the old syntax; if you still use the old syntax like \SweaveOpts{}, you will see a warning with a pause of 10 seconds
based on the same reason, it is recommended to use the chunk option child to input child documents; old syntax like \SweaveInput{} will be deprecated
for markdown output, results from inline R code will no longer be put in a pair of backticks (#379)
the package option opts_knit$get('cache.extra') was removed because this option should really be a chunk option instead; see https://yihui.org/knitr/demo/cache/ for the updated documentation (#404 and #405) (thanks, Jim Hester)
the chunk option highlight.opts was deprecated and renamed to engine.opts; this affects users who use Andre Simon's highlight through the highlight engine in knitr
the chunk option file for Awk was deprecated; we can also use engine.opts to specify the file for Awk; see https://github.com/yihui/knitr-examples/blob/master/024-engine-awk.Rmd for example
the pattern knit_pattern$get('ref.label') was deprecated since it is unlikely to be customized; a fixed pattern '^#+\\s*@knitr(.*)$' will be used instead

MINOR CHANGES

when opts_knit$get('verbose') is TRUE, a timestamp() will be printed before each code chunk is evaluated (#377) (thanks, Carl Boettiger)
stitch() will no longer copy the template over to the current working directory (thanks, Michael Friendly)
stitch() will no longer open the PDF/HTML output automatically (#411) (thanks, Michel Kuhlmann)
the script inst/bin/knit can also convert R Markdown documents to HTML now; the argument --pdf was removed and a new argument --no-convert was added
dots in figure filenames will not be replaced with _ when the output is not LaTeX (thanks, Stewart Macarthur)

BUG FIXES

fixed #410: when the inline R code returns NA_real_, the scientific notation of numbers will run into errors (thanks, Rafik)
the syntax pattern for Rnw documents was not quite right: all_patterns$rnw$chunk.end = '^\\s*@\\s*%*' actually allows any characters after @, but only LaTeX comments and white spaces are allowed; it has been fixed to ^\\s*@\\s*(%+.*|)$ now

DOCUMENTATION

an example of combining R, knitr and D3 to draw a contour plot: https://yihui.org/knitr/demo/javascript/

CHANGES IN knitr VERSION 0.8

NEW FEATURES

output from other languages (e.g. python, awk, ...) can also be cached like R output when cache=TRUE; see 023-engine-python.Rmd for an example
added support for bash/shell scripts; use the chunk option engine='bash' or engine='sh' to write shell scripts in code chunks (#361)
a new function knit_rd() to knit package documentation (run examples code and insert output in the HTML documentation) (#227) (thanks, Taiyun Wei)
added LuaTeX support for tikz graphics with the tikzDevice package (set options(tikzDefaultEngine = 'luatex')); this feature requires tikzDevice version > 0.6.2 (#358) (thanks, Alastair Andrew)
a new chunk option fig.env to set which environment to use for figures in LaTeX, e.g. we can set fig.env='marginfigure' to use \begin{marginfigure} (#364) (thanks, Bryan Hanson)
added a new package option global.device (default FALSE) which specifies whether to use a global graphics device to capture plots; if TRUE, it is possible to write plot(1:10) in a previous chunk and abline(0, 1) in a latter chunk because all code chunks share the same device, however, this may also bring unexpected consequences (in particular, using par() can bring redundant plots)

BUG FIXES

dots in figure paths are more safely replaced with _ now, e.g. fig.path='../figure' will no longer be replaced by __/figure (#346) (thanks, @ralfer)
the c() syntax for the chunk option dependson did not actually work, e.g. dependson=c('foo', 'bar') (#350) (thanks, Cassio Pereira)
fixed a bug when eval=FALSE and prompt=TRUE (the continuation character was used in some places where there should be the prompt character) (thanks, Derek Ogle)
persp() plots were not recognized in the last version (thanks, Jeffrey Racine)

MAJOR CHANGES

leading spaces are allowed in chunk headers now, e.g. in the past <<>>= must appear in the beginning of a line, but now we can indent the chunk header by a number of white spaces; this amount of spaces will be stripped off the chunk if the whole chunk is indented (#236) (thanks, @jamiefolson and Vitalie Spinu)
markdown output will be indented if the original code chunk is indented; this allows chunk output to be nested within its parent environment, e.g. inside an ordered list (see 001-minimal.Rmd for example)
when the global chunk option eval=FALSE, inline R code will not be evaluated, and ?? is returned for inline R expressions (#367)

MINOR CHANGES

if getOption('OutDec') is not ., inline numeric output will be put inside \text{} in LaTeX to avoid situations like #348 (the normal math mode may add a space after the comma in 3,1415)
if the chunk option external==FALSE (default is TRUE), knitr will no longer automatically add \usepackage{tikz} to the LaTeX preamble; you need to add it manually (but it is recommended to use external=TRUE with cache=TRUE for the sake of speed, because compilation of tikz graphics may be slow)
*.brew generates *.txt by default (instead of *-out.brew)
knit(text = ...) will no longer write output in the console (the output is only returned as a character string)

DOCUMENTATION

added a simple reference card: http://cran.rstudio.com/web/packages/knitr/vignettes/knitr-refcard.pdf

CHANGES IN knitr VERSION 0.7

NEW FEATURES

added a new chunk option out.extra to write extra graphics output options, e.g. <<out.extra='angle=90'>>= to rotate the figure by 90 degrees; see https://yihui.org/knitr/options (#301) (thanks, @knokknok)
when opts_knit$get('verbose') is TRUE, logs (messages, warnings and errors) along with the corresponding R code will be printed after knit() is done; this might help users figure out possible problems in R code quickly (#276)
.Random.seed is cached again for the sake of reproducibility; see https://yihui.org/knitr/demo/cache/ for how to maintain reproducibility when the computation involves with random number generation (#274) (thanks, Renaud)
the package option opts_knit$get('cache.extra') can be an unevaluated R expression now, e.g. opts_knit$set(cache.extra = quote(.Random.seed)); see the cache page above for a concrete example
added a new package option 'root.dir' (default NULL) which can be used to set the root directory to evaluate code chunks in a document; by default, the root directory is the directory of the input document, and this option enables users to set other directories as the working directory for code chunks (#277) (thanks, Ken Williams)
spin() will add \documentclass{article} for Rnw output if no document class is specified in the R script so that the LaTeX output will be a complete document (#295) (thanks, Christiaan Klijn)
added Ruby support in the engine option; see the example https://github.com/yihui/knitr/blob/master/inst/examples/knitr-lang.Rmd (#294) (thanks, Ramnath Vaidyanathan)
also added Haskell support in the option engine='haskell' through calling ghc (#336) (thanks, Michel Kuhlmann)
added support to Andre Simon's highlight through the option engine='highlight'; see https://gist.github.com/3114112 for an example of highlighting Matlab code in LaTeX (thanks, Dirk Eddelbuettel and Renaud Gaujoux)
the output hooks for LaTeX, HTML, Markdown and reST will respect the engine option now, so these hooks can be readily used for output when the language is not R, e.g. render_markdown(strict = TRUE) also works for Python output (#251) (thanks, Chris Fonnesbeck)
the chunk options eval, echo and results are also respected when the language is not R, e.g. for a Python code chunk with eval=FALSE, the code will not be evaluated, or for a Ruby chunk with results='hide', the output will be hidden (#293) (thanks, Ramnath Vaidyanathan)
chunk options out.width, out.height and out.extra also work for plots in HTML and Markdown output now, e.g. out.width='200px' or out.extra='style="display:block;"' (#297) (thanks, Taiyun Wei and Alan Severini)
the hook function to create animations in HTML output is exported now as hook_ffmpeg_html()
added a package option opts_knit$get('animation.fun') which defaults to hook_ffmpeg_html; this option is used to create animations in HTML output so that we do not have to use FFmpeg
added two functions hook_r2swf() and hook_scianimator() which can be set as the package option opts_knit$get('animation.fun') and create animations via the R2SWF package or the SciAnimator library (see animation::saveHTML) (thanks, Taiyun Wei)
a new function image_uri() to create data URIs for image files; we can set opts_knit$set(upload.fun = image_uri) so that images are embedded in the HTML output as data URIs (hence the HTML page does not depend on external images) (#298, #324) (thanks, Wush Wu)
added a new object opts_template which can be used to set a group of chunk options and they can be referenced later with the new chunk option opts.label; see ?opts_template for examples; this makes it easy to reuse groups of frequently used chunk options (#316, #320) (thanks, Cassio Pereira)
a new function dep_prev() to build chunk cache dependencies so that all later chunks will depend on previous chunks; if any of a previous chunk is updated, the cache of all chunks after it will be updated as well (#285) (thanks, @muelleki)
a new chunk hook function hook_optipng() to optimize PNG images using optipng (#272) (thanks, Winston Chang)
added a new output hook named document in knit_hooks (see knit_hooks$get('document')); this hook function is used to process the output of the whole document; it can be useful when we want to post-process the whole output before writing it to the output file
a new function rst2pdf() which uses the program rst2pdf to convert reST documents to PDF; it is also supported by knit2pdf() when compiler='rst2pdf' (#300) (thanks, Alex Zvoleff)

BUG FIXES

fixed #286: messages (including warnings and errors) are guaranteed to be ended by \n, so even when chunk option comment=NA, messages will also be rendered correctly (thanks, Carl Boettiger)
fixed #273: when knitting a file under another directory with cache turned on (e.g. knit('foo/bar.Rnw')), lazyLoad() is unable to load the cache files under a relative path because the working directory has been changed to the directory of the input file during the evaluation
fixed #292: layout() may cause the capture of unwanted plots (thanks, Austen Wallace Head)
fixed #302: when there are multiple instances of \documentclass in the source document, knitr will be unable to insert the LaTeX preamble
fixed #308: if options('OutDec') was set to a character other than ., the LaTeX code can be malformed (#308) (thanks, Cassio Pereira)
opts_chunk$set() in a child document was only working in that child document, but was expected to change chunk options globally; now it works everywhere, and will affect all chunks after this setting, no matter where opts_chunk$set() is (thanks, Guy Lebanon) (http://bit.ly/MexHXd)
fixed #332: calling purl() inside a source document when knit() the same document could cause clashes; now it is safe to put purl() inside a source document and knit() it
fixed #342: when eval=FALSE, line breaks in the source code were missing

MAJOR CHANGES

if the chunk label contains non-alphanumeric characters (except - and _), these characters will be replaced by _ in the figure filenames (if there are any) and a warning will be issued; this is to guarantee the figure filenames are valid to LaTeX (#321) (thanks, (Roman Lustrik)
the highlight package is not enabled by default; use opts_knit$set(use.highlight = TRUE) to enable it
the default LaTeX output will put messages, warnings and errors in special LaTeX environments: errors are red, warnings are magenta, and messages are italic; in previous versions, they were in the verbatim environment (#264) (thanks, @muelleki)
unnamed chunks are named sequentially in a single call of knit() according to the order of their appearance no matter where they are, e.g. if there are two unnamed chunks in two child documents respectively, they will be named unnamed-chunk-1 and unnamed-chunk-2; in previous versions, both will be named as unnamed-chunk-1 which can cause clashes of cache and figure files
the function build_dep() was renamed to dep_auto() which better reflects what this function really does; it is still available in this package but may be removed in future versions
the package tikzDevice was removed from the Suggests field, but this will not affect users who have already installed tikzDevice; for those who have not, this package has been archived on CRAN (hopefully only temporarily), so you have to install from the source

MINOR CHANGES

the LaTeX environment kframe was updated so that it can be used in other environments such as center or tabular (#283) (thanks, @muelleki)
the OpenCPU demo is more robust to double quotes in the text (#271); see http://public.opencpu.org/userapps/opencpu/knitr/
for Sweave output, the results will not be put inside the Schunk environment when results='asis'
stitch() templates use more sensible figure paths by default: the path for figures is 'figure/input-script-name' now, i.e. it will be different for different input R scripts to avoid possible clashes
stitch() templates no longer use default title and author names; if the user did not set them in the R script (as meta comments # title: and # author:), there will not be titles or authors in the output
knitr will no longer use scientific notations for integers in inline R code output; sci notation only applies to double-precision numbers (#296) (thanks, @knokknok)
options() set in the main document will apply to its child documents (e.g. options('digits')) (#306) (thanks, Cassio Pereira)
the ... argument in knit2html() is passed to markdown::markdownToHTML() instead of knit(); this allows us to pass more arguments to control the rendering of HTML output, e.g. knit2html(input, fragment.only = TRUE) (#333) (thanks, @Bart6114)

DOCUMENTATION

the example using other languages was updated to show how some chunk options (eval, echo and results) also work for foreign languages: https://github.com/yihui/knitr/blob/master/inst/examples/knitr-lang.Rmd

CHANGES IN knitr VERSION 0.6.3

MAJOR CHANGES

this is an urgent patch version for CRAN: the dependencies on highlight and parser were removed because these two packages were orphaned on CRAN; now knitr uses a naive syntax highlighter for LateX and HTML output if highlight is not available, which has a similar appearance with highlight; when the parser package is not available, users should be careful with the chunk option tidy=TRUE: replace.assign may not work as expected; see the NEWS of formatR for details: https://github.com/yihui/formatR/blob/master/NEWS; you are welcome to improve the naive highlighter: https://github.com/yihui/knitr/tree/master/R/highlight.R

NEW FEATURES

spin() gained a new argument report for Rnw/Rtex/Rmd documents to be compiled to PDF/HTML respectively (#287) (thanks, kohske takahashi)

BUG FIXES

stitch() and spin() should use parent.frame() to evaluate code by default; in 0.6, the default parent frame for the inner knit() was a wrong environment to use (thanks, Michael Nelson)
use unnamed-chunk-i as the label when chunk label == '' (#280) (thanks, Josh Paulson)
fixed #279 and #281, both of which are about concordance
'\\maxwidth' does not apply to LaTeX/PDF animations, so the default value of out.width for animations in LaTeX is still NULL; you will have to set out.width manually for a reasonable width (#282) (thanks, Ramnath Vaidyanathan)

MISC

knitr and RStudio had an invited talk at useR! 2012; slides at https://yihui.org/slides/2012-knitr-RStudio.html (source at https://github.com/yihui/knitr-talks)

CHANGES IN knitr VERSION 0.6

NEW FEATURES

for LaTeX output, the chunk option out.width defaults to '\\maxwidth' now; see https://yihui.org/knitr/demo/framed/ for its definition; this makes sure the figures do not overflow the page margins (#221)
the chunk option size now defines the font size of the whole chunk instead of only some special characters in it, so that the old trick of redefining the knitrout environment for different font sizes is no longer necessary; see updates in the beamer examples: https://yihui.org/knitr/demo/beamer/ (thanks, Baptiste Auguie)
added a new chunk option dev.args which can be used to pass more arguments to the graphical device (#254) (thanks, knokknok)
warnings, messages and errors will be wrapped according to options('width'); this will make long messages stay within the page margin when the width option is appropriately small (#259) (thanks, @muelleki)
added a new function spin() to turn a specially formatted R script to a report; see https://yihui.org/knitr/demo/stitch/ (#223) (thanks, Rich FitzJohn)
knit() gained a new argument envir to specify the environment in which to evaluate code chunks (by default in the parent frame); this will allow users to run code in a specified environment so that the global environment will not be cluttered by objects in chunks, which can hopefully make the documents more reproducible (#228)
added a new component inline.comment in knit_patterns to strip off tokens of inline R code from inline comments, e.g. a comment line like % x is \Sexpr{x} will become % x is x (#110); this will only happen to LaTeX documents because HTML does not have inline comments (it only has block comments )
concordance works for child documents as well now (try RStudio), although it is not very precise (#225); note when concordance is enabled, the results from child documents will be written into the parent output instead of individual tex files; also note you have to set opts_knit$set(self.contained = FALSE) for concordance to work better
added an OpenCPU app so that we can run knitr in the cloud now; see system.file('opencpu', 'apps', 'index.html', package = 'knitr') or http://public.opencpu.org/apps/knitr (thanks, Jeroen Ooms)
all messages, errors and warnings during evaluation are recorded in an internal object knit_log (use knitr:::knit_log$get() to get all messages) and they are printed if the package option verbose is TRUE (i.e. opts_knit$get('verbose')) (#224)
child documents are also supported in other document formats besides LaTeX, e.g. Markdown and HTML, etc; please use the chunk option child to specify the child documents; see system.file('examples', 'child', 'knitr-main.Rmd', package = 'knitr') for an example in markdown (#268) (thanks, @martinaryee)

BUG FIXES

the templates for stitch() used results=hide which should really be results='hide' (#219) (thanks, @r2d3)
format numbers with the reST syntax instead of HTML (#218) (thanks, Jeffrey Arnold)
hook_pdfcrop() should work better under Windows now (#209) (thanks @DCCKC and @r2d3)
tikz compilation fails on Windows network drives
FFmpeg does not really work for HTML/Markdown output because the dot in figure filenames was omitted (thanks, Ming Kuo)
child documents can fail when they are in different subdirectories (thanks, Christoph J)
set_parent() failed to work in the last version due to a bug when inserting the parent preamble into the child document (#240)
better preserve plot sizes in interactive sessions (#258)

MAJOR CHANGES

.Random.seed is not cached any more because of weird problems due to lazy loading (#248 and #253); users should use set.seed() to make sure reproducibility of random simulations; the chunk output is cached in a .RData database instead of a lazy-load database to avoid problems in #253
the default graphics device is set to the null PDF device before evaluating code chunks, i.e. pdf(file = NULL), so that neither Rplots.pdf nor plot windows will be opened during knitting

MINOR CHANGES

knitr will show a message when a chunk is empty; this helps users who do not actually want a chunk to be empty realize the problem like #229; in the past, knitr just silently returns an empty string from such chunks
knitr will show a message when the cache is loaded and the option opts_knit$get('verbose') is TRUE (#249) (thanks, Carl Boettiger)
the filename extensions Snw and Stex are also supported (foo.Snw/foo.Stex produces foo.tex)
the HTML output hooks are changed according to the suggestion of Barry Rowlingson (#250) and the default CSS is also slightly modified
knit() will no longer try to remove the file NA generated by pdf(file = NULL) before R 2.14.2 (which was a bug and fixed later); you should update R if you see this file

DOCUMENTATION

added a minimal brew example under system.file('examples', package = 'knitr')

CHANGES IN knitr VERSION 0.5

NEW FEATURES

white spaces are allowed before <<>> when using chunk references, and this approach of references is supported in tex and html documents as well
added a new pattern list named md so that R code blocks can be written more naturally in extended markdown (e.g. GFM and pandoc): use ```{r label, opts} to begin a chunk and ``` (three or more backticks) to begin normal text, and write inline R code in `r code.here`; it is the default syntax for markdown input, or you can call pat_md() before knit() so knitr can make use of this pattern list to process the input document
RStudio has added full support to knitr: we can knit HTML and Markdown documents easily now, and markdown output will be automatically converted to an HTML preview page just like TeX to PDF
if the pattern list is not set in advance, knitr will try to detect the syntax automatically by matching all built-in pattern lists against the input document, e.g. if the input contains <<>>=, the Rnw syntax will be used, and if ```{r} is detected, the markdown syntax will be used; this feature enables us to use different sets of syntax freely, e.g. we can use Sweave syntax in markdown files and knitr will be able to recognize it (#189)
new filename extensions with a prefix R or r are introduced: *.Rhtml, *.Rhtm, *.Rtex, *.Rmd and *.Rmarkdown will be recognized as source documents like *.Rnw, and the output filename will be automatically determined by removing the R prefix, e.g. foo.Rmd produces foo.md by default; the old clumsy naming convention foo_knit_.md is still preserved but not recommended any more (#187)
new function knit2html() to knit an Rmd file (R with markdown) and convert to HTML in one step using the markdown package
new functions pat_rst() and render_rst() to support reStructuredText; use .. {r label, options} and .. .. to write R code chunks; see https://github.com/yihui/knitr/tree/master/inst/examples/knitr-minimal.Rrst (thanks, Jeffrey Arnold and Ramnath Vaidyanathan)
new package option self.contained which decides whether to write style definitions (highlighting) in external files or put them in the output document; the highlighting definitions in LaTeX output is often too long, so opts_knit$set(self.contained = FALSE) can help in this case (#176) (thanks, Ramnath Vaidyanathan)
new package option filter.chunk.end which decides if the chunk.end pattern really means chunk.end; see https://yihui.org/knitr/options (thanks, Joe Cheng)
syntax highlighting themes are also available to HTML output now; the usage is the same as in LaTeX (#179) (thanks, Ramnath Vaidyanathan)
the chunk option fig.cap is also used in markdown output now
the random seed .Random.seed is also cached for the sake of reproducibility in random simulations
the function call read_chunk() will be evaluated when tangling R code via purl() (#175) (thanks, Carl Boettiger)
the default LaTeX output will use the upquote package if it exists so that single quotes are straight in the output (thanks, Mathematical Coffee http://bit.ly/IKjluw)
the chunk option engine is back but it is used to run code from other languages instead of just ignoring the code in chunks which have engine != 'R'; currently this option is still in rough edges and only supports python and awk; other languages may be added in future versions, but users can also do it by themselves by knit_engines$set(language = function(options) {...}); see an example at system.file('examples', 'knitr-lang.Rmd') (#201)
new function write_bib() to write Bibtex citations for R packages; see the main manual for a sample usage (#13)
hook_pdfcrop() also supports cropping other image formats like PNG or JPEG through ImageMagick (convert -trim) (#203) (thanks, @r2d3)

MAJOR CHANGES

knitr will completely stop when duplicated labels are found and the corresponding chunks are non-empty; in the previous version, only a warning is given and R code in later chunks will override previous chunks (#185) (thanks, Randall Pruim)
the default graphical device for HTML and markdown output is png() now to avoid the possible unexpected consequence that PDF plots are included in an HTML page which will not work normally in browsers
markdown output will use the extended markdown hooks by default now: render_markdown(strict = FALSE); in the past the default was render_jekyll(); the major difference is that the code blocks are put in ```r and ```; if you want the strict markdown syntax, you can all render_markdown(strict = TRUE) which will indent code blocks by 4 spaces
render_gfm() has been removed because the name can be misleading; the main purpose of this function was to put code blocks in ```, and now we can replace it by render_markdown(FALSE); other markdown flavors also support such fenced code blocks (e.g. pandoc) -- it is not limited to Github only
the default value for the fig.path option is 'figure/' now so that plots will be put under this directory by default; the default was './' in the past which makes the current directory messy when there are many plots
knitr will fully stop when an error is encountered in knit(); in the past, only a message was issued in this case in an interactive R session
the package option all.patterns has been dropped; please use the objects all_patterns or knit_patterns directly if you want to tweak the syntax

BUG FIXES

the compilation of tikz graphics can hang up when there are TeX errors in the tikz file; now we use \nonstopmode to avoid hiccup (#188)
multiple devices per chunk was buggy (#181)
S4 objects will be printed by show() instead of print(); this is a bug of the evaluate package, which has been fixed (please update it to be > 0.4.2)

MISC

it is recommended to use opts_chunk$set() to set global chunk options now instead of \SweaveOpts{}; all documentation has been updated (#216)
number of downloads (https://github.com/yihui/knitr/downloads) of knitr documentation before I removed and updated them on GitHub: c(main = ?, graphics = 275+)

CHANGES IN knitr VERSION 0.4

NEW FEATURES

Sweave concordance was finally implemented: when opts_knit$get('concordance') is TRUE, knitr will write a file named 'input-concordance.tex' which contains the mapping between input Rnw and output tex line numbers; this feature is mainly for (but not limited to) RStudio to provide better error navigations: you can jump from the TeX error message to the Rnw source directly to know where the error comes from (the line number of the source of the error may not be accurate but should be in the ballpark) (#133) (thanks, JJ Allaire and Joe Cheng)
if output hooks have been set before calling knit(), they will be respected, i.e. knitr will no longer override them by default hooks; you need to make sure all output hooks are set appropriately, e.g. you can start by render_latex() and change some individual hooks later (#165) (thanks, Andrew Redd)
newly created objects in the global environment will also be cached if cache is turned on (cache=TRUE); in previous versions knitr is unaware of objects created in globalenv(), e.g. setGeneric() creates S4 generic functions in globalenv() and knitr was unable to capture them (#138) (thanks, syoh)
chunk options dev, fig.ext and dpi can be vectors now; this allows one to save a plot to multiple formats, e.g. <<foo, dev=c('pdf', 'png')>>= creates two files for the same plot: foo.pdf and foo.png (#168) (thanks, MYaseen208)
an experimental feature for animations created by FFmpeg in HTML/markdown output when fig.show='animate' (#166) (thanks, gabysbrain)
the chunk option fig.cap supports multiple figure captions in LaTeX now, e.g. if a chunk produces two plots, we can use fig.cap = c('first caption', 'second caption') to assign two different captions to them respectively when fig.show = 'asis' (#155) (thanks, Jonathan Kennel)
new package option opts_knit$get('upload.fun') which is a function that takes a plot file to upload to a certain host and returns the link to the image; by default it is imgur_upload(), and you can use your own function to upload images to other hosts like Flickr (#159) (thanks, Carl Boettiger)
all packages loaded in the current session are also cached, so as long as a package has been loaded previously, it will be available to all following chunks (#160)
new chunk option autodep and function build_dep() to build cache dependencies among cached chunks automatically by analyzing object names in all cached chunks; this is a loose alternative to the dependson option (see main manual and ?build_dep for details) (#72) (thanks, Seth Falcon)
input and output in knit() are no longer restricted to files; they can be stdin()/stdout() or other types of connections (#162; see yihui#162) (thanks, gabysbrain)
as-is output (results='asis') and plots are no longer put in the framed environments because of incompatibilities (#163) (thanks, DCCKC, Murray Logan and Jennym Hutchison)

BUG FIXES

for plots in LaTeX output, centering should be done with {\centering } instead of \centering{} (#156) (thanks, Ramnath Vaidyanathan)
the recorded plot is a more precise representation of the expected plot now because the recording device also takes the plot size into consideration (#157) (thanks, Christiaan Klijn and Frank Harrell)
format_sci() now correctly formats 0; this function is used for inline R code to format numbers in scientific notation (#161) (thanks, Kihoro J. M.)
fixed a bug for the case in which the chunk option only contains the label like <<label=foobar>>=; knitr 0.3 was unable to parse the label correctly (<<foobar>>= is OK) (thanks, Muhammad Yaseen)

MINOR CHANGES

imgur_upload() returns the link to the image directly, with the XML list as its attribute (in v0.3 the list was returned)
more verbose messages in case of chunk errors: both line numbers of the source and chunk info will be printed

DOCUMENTATION

website updated as usual: https://yihui.org/knitr
added an example for the subfloat environment: https://github.com/yihui/knitr/releases/download/doc/knitr-subfloats.pdf
most manuals (main or graphics manuals) have been updated

MISC

number of downloads (https://github.com/yihui/knitr/downloads) of knitr documentation before I removed and updated them on GitHub: c(main = 400, graphics = 177)

CHANGES IN knitr VERSION 0.3

NEW FEATURES

a fundamental and important new feature for writing chunk options: they can be written as valid R code now, just like we write function arguments (e.g. echo=c(1, 3, 5), fig.cap="my figure caption"); all options will be parsed and evaluated as R code by default; see https://yihui.org/knitr/options for details (#142) (thanks, Baptiste Auguie)
chunk references using <<label>> is supported now (#86); thanks to Kevin R. Coombe and Terry Therneau for the discussion
new function run_chunk() to run the code in a specified chunk, which is an alternative to the chunk reference in Sweave; see https://yihui.org/knitr/demo/reference/
a executable script knit under system.files('bin', package = 'knitr') which makes it easier to call knitr via command line under *nix (call knit input [output] [--pdf])
the inline hooks respect getOption('digits') and getOption('scipen') now (see ?options); numbers returned from inline R code will be formatted according to these two options (see a demo at https://yihui.org/knitr/demo/output/)
if you still use old Sweave syntax for chunk options, it is possible to write literal commas in chunk options now -- they have to be escaped by \, e.g. caption=hello\, world; this will be parsed to 'hello, world' as a character string; of course this looks ugly and has limited power, so please please consider the new syntax!
knit2pdf() gained another argument compiler which can be used to specify the program to compile the tex document to PDF, such as xelatex (#131) (thanks, Ramnath Vaidyanathan and Dennis Murphy)
a new function imgur_upload() to upload images to imgur.com; it can be used in HTML or Markdown hooks so the output is a self-contained document which does not need additional image files; opts_knit$get('upload.fun') can use this function (#66) (thanks, Ramnath Vaidyanathan)
a child document can be compiled individually with the LaTeX preamble borrowed automatically from a parent document using a new function set_parent(); see the help page for details (#136) (thanks, Helder Correia)
to avoid $$ around numbers in the inline output, we can use I() to protect the numeric inline output, e.g. $x = \Sexpr{I(10^7)}$ gives $x = 10^7$ whereas \Sexpr{10^7} gives $10^7$ (thanks, Kevin Middleton)
the listings package is formally supported now (see ?render_listings); the default style is borrowed from Sweavel.sty written by Frank Harrell (#101) (thanks, Frank)
new package option cache.extra which allows more objects to affect cache; see https://yihui.org/knitr/demo/cache/ (#134)
new package option child.path to specify the search path of child documents relative to the parent document (#141)
new package option aliases to set aliases for chunk options; see https://yihui.org/knitr/options (#144)
new chunk options fig.cap, fig.scap and fig.lp to write captions, short captions, label prefix for the figure environment in LaTeX (#145) (thanks, Frank Harrell)
new package option eval.after to set a character vector of chunk options which should be evaluated after a chunk is executed (thanks, Frank Harrell)
a series of convenience functions pat_rnw(), pat_tex(), pat_brew() and pat_html() to set built-in patterns (syntax) to read input

MINOR CHANGES

package option eval.opts has been dropped: all options of classes symbol or language will be evaluated, so there is no need to specify which options to evaluate manually; remember, the chunk options are similar to function arguments, so you can use any valid R code there
the default value for the output argument in knit() is NULL now, so we can also provide output filenames to stitch() and knit2pdf() (#119)
standard LaTeX messages are suppressed when a tikz plot is compiled to PDF so that we can see the knitr process more clearly
%\SweaveInput{} will be ignored now (#150)
results=asis will no longer affect the chunk hook (in the past, the chunk output was not wrapped in the kframe environment when results=asis); it only affects the output hook now
the package website allows comments now

MAJOR CHANGES

the starting pattern of normal texts in an Rnw document is ^@\\s*%* instead of ^@\\s*$ now, meaning you can write @ % a comment to end a code chunk (this is consistent with Sweave)
the default value of the argument output of knit() will be a filename under the current working directory; in previous versions, the output file will be under the same directory as the input file; this change makes it possible to completely separate the input files and output files into different places, and hopefully will give users better experience in managing a whole collection of files (including child documents): put all source files in one place and output files in another place
the package homepage is https://yihui.org/knitr now (the previous URL yihui.github.com/knitr will be automatically redirected to the new address)

BUG FIXES

the object opts_current does not give the evaluated version of the current chunk options because it was created before the options are evaluated; this has been fixed and opts_current$get() will give the expected values of options (thanks, Frank Harrell)

MISC

number of downloads (https://github.com/yihui/knitr/downloads) of knitr documentation before I removed and updated them on GitHub: c(main = 1300, graphics = 549, themes = 130, beamer = 565, listings = 240, minimal = 160)

CHANGES IN knitr VERSION 0.2

NEW FEATURES

added support for including child documents in the main document (like \SweaveInput{} but with different implementations); see https://yihui.org/knitr/demo/child/ (#92)
for inline R code, character results are returned as-is now (without \texttt{})
new function purl() as a wrapper to knit(..., tangle = TRUE) which extracts R code from the input document (thanks to Dieter Menne's wife who suggested the function name)
the error hook applies to inline R code when an error occurs in the inline R code, in which case knitr will not stop by default; instead, it writes the error message into the output (#85)
chunk option split also works for HTML output now using <iframe></iframe> (#82)
knit() gained an argument text as an alternative to input (#88)
new chunk option child to include child documents into the main document (#92)
chunk option external defaults to TRUE now (was FALSE by default in the last version)
added a new demo to show how to build package vignettes with knitr: https://yihui.org/knitr/demo/vignette/
added support to the quartz() device under Mac (#103); now the dev option has more choices (see https://yihui.org/knitr/options)
chunk option echo can take a numeric vector to select which R expressions to echo into the output now (#108); see https://yihui.org/knitr/options
a new function stitch() which is a convenience function to insert an R script into a template and compile (to quickly create a report based on an R script)
for a chunk hook to run, the corresponding chunk option no longer has to be TRUE; it can be any non-null values; this enables us to make use of the option value directly instead of only knowing it is TRUE (see https://yihui.org/knitr/demo/cache/ for an example)
knit() will no longer write figure or cache files in the same directory as the input document; instead, these files are written in the current working directory (see ?knit)
a new function knit_env() that makes the environment of the current chunk accessible to the user

BUG FIXES

the code used to merge global chunk options and local options was buggy for cache; it has been fixed now, so cache is more stable (#105), but users may see previously cached chunks being re-evaluated with this version, which should be regarded as a normal phenomenon, and on the second run, the cached chunks will not be evaluated again
fixed a buglet when using both options out.width and out.height in Rnw (#113)

CHANGES IN knitr VERSION 0.1

NEW FEATURES

first version of knitr: it covers most features in Sweave, cacheSweave and pgfSweave; see package homepage for documentation and examples: https://yihui.org/knitr/

MISC

knitr won an Honorable Mention prize (before it was formally released to CRAN) in the Applications of R in Business Contest hosted by Revolution Analytics: http://bit.ly/wP1Dii http://bit.ly/wDRCPV
in this NEWS file, #n means the issue number on GitHub, e.g. #142 is yihui#142

Files

NEWS.md

Latest commit

History

NEWS.md

File metadata and controls

CHANGES IN knitr VERSION 1.35

BUG FIXES

CHANGES IN knitr VERSION 1.34

NEW FEATURES

MAJOR CHANGES

MINOR CHANGES

BUG FIXES

CHANGES IN knitr VERSION 1.33

NEW FEATURES

BUG FIXES

MAJOR CHANGES

MINOR CHANGES

CHANGES IN knitr VERSION 1.32

NEW FEATURES

BUG FIXES

MINOR CHANGES

CHANGES IN knitr VERSION 1.31

NEW FEATURES

BUG FIXES

MAJOR CHANGES

MINOR CHANGES

CHANGES IN knitr VERSION 1.30

NEW FEATURES

BUG FIXES

MINOR CHANGES

CHANGES IN knitr VERSION 1.29

NEW FEATURES

BUG FIXES

MAJOR CHANGES

MINOR CHANGES

CHANGES IN knitr VERSION 1.28

BUG FIXES

MAJOR CHANGES

CHANGES IN knitr VERSION 1.27

NEW FEATURES

BUG FIXES

MAJOR CHANGES

MINOR CHANGES

CHANGES IN knitr VERSION 1.26

NEW FEATURES

MINOR CHANGES

CHANGES IN knitr VERSION 1.25

NEW FEATURES

MAJOR CHANGES

MINOR CHANGES

BUG FIXES

CHANGES IN knitr VERSION 1.24

MAJOR CHANGES

BUG FIXES

CHANGES IN knitr VERSION 1.23

NEW FEATURES

BUG FIXES

MAJOR CHANGES

CHANGES IN knitr VERSION 1.22

NEW FEATURES

BUG FIXES

MAJOR CHANGES

MINOR CHANGES

CHANGES IN knitr VERSION 1.21

NEW FEATURES

BUG FIXES

MAJOR CHANGES

MINOR CHANGES

CHANGES IN knitr VERSION 1.20

NEW FEATURES

BUG FIXES

CHANGES IN knitr VERSION 1.19

NEW FEATURES

MAJOR CHANGES

BUG FIXES

CHANGES IN knitr VERSION 1.18

NEW FEATURES

MAJOR CHANGES

BUG FIXES