Skip to contents

Changelog

Source: NEWS.md

Binary files

  • webr_repl_project() and webr_repl_directory() now carry a binary file in a link, in the data field webR’s own sharelink format defines. A small image or data file can travel beside the code that uses it. Bytes do not compress, so livelink warns once a file passes 32 KB.

  • decode_webr_link() and decode_shinylive_link() gain a binary argument, FALSE by default. A preview can show you the text a link holds but not its bytes, so a binary is skipped with a warning until you ask for it with binary = TRUE.

  • webr_repl_link() still takes a script only. A file holding a NUL points you to webr_repl_project(), and one that is not valid UTF-8 says so, rather than both failing the same way.

CRAN release: 2026-07-31

Security

Encoding

  • webR links now use msgpack rather than JSON, so a link ends in mz where it used to end in jz. This is the same form the webR share button writes. A one-line snippet comes out about 11% shorter, and links you have already shared keep working, since both forms still decode.

Bug fixes

  • webr_repl_link() now handles a braced expression whose code cannot be read back, which failed in an unsaved buffer in Positron, when the file was shorter than the expression, and when the block was empty. Where the code can be read its comments survive, and where it cannot the code is rebuilt without them.

  • webr_repl_link() and webr_repl_project() now refuse input that cannot become a script, rather than building a link that will not open. That covers several file paths given to a single script, a file that is not valid UTF-8 or holds an embedded NUL, NA, and empty input.

  • webr_repl_project() now warns when autorun_files names a file webR cannot run on opening, and print() marks the files that really do run.

  • decode_webr_link() writes a binary file back exactly as it was, rather than as text listing its byte values.

  • decode_webr_link() and preview_webr_link() now read links built against a webR site you host yourself, and report a malformed link when the ending of the link does not describe the code inside it.

  • shinylive_directory() now includes the .css and .js files an app needs, leaves out only files that cannot be read as text, and checks app_file.

  • link.only = TRUE no longer leaves an empty code block above the link.

Error messages

  • webr_repl_project() and shinylive_project() now explain a missing input instead of reporting the name of an internal variable.

  • Errors about an invalid argument now name the value at fault. NA given to version, panels, or mode failed with missing value where TRUE/FALSE needed, an invalid panels did not say which component was wrong, and NULL left a blank You provided: line.

  • preview_webr_link() and preview_shinylive_link() now keep the explanation from the underlying error rather than replacing it.

Documentation

  • Every help page now lists what a returned object holds, one entry at a time, rather than describing it only as an object with metadata.

CRAN release: 2026-07-24

Initial CRAN release. livelink packs R code into a shareable URL that runs in the browser, with no server and nothing to upload. Two destinations are supported: webR, a full R REPL, and Shinylive, which runs R and Python Shiny apps.

Features

Decoding and preview

  • decode_webr_link() and decode_shinylive_link(): extract the embedded files to a directory. They default to the session temporary directory, so a stray call cannot litter your project.
  • preview_webr_link() and preview_shinylive_link(): inspect a link in memory without writing anything, which is the safe way to open a link from someone else.
  • Decoded results expose the same fields whether a link came from webR or Shinylive, so one piece of code can handle both.

Input flexibility

  • Code can be an R expression (no quotes or escaping), a character string, a file path, or whatever is on the clipboard.
  • Multi-file input via named lists or file path vectors. The primary argument is input across every link-creating function, whatever form the code takes.

Configuration

  • Choose which webR panels appear (plot, files, terminal, editor) with panels, and Shinylive’s editor-or-app view with mode.
  • autorun runs the code the moment a link opens.
  • set_webr_base_url(): point webR links at a custom or self-hosted deployment.

Working with the results

  • Every returned object answers to repl_urls() and as.character() for the bare URL(s), and to format() for its summary as a character vector.
  • knit_print() renders a link object as a clickable link when it is the value of a ‘knitr’ or ‘Quarto’ chunk.
  • as.data.frame() turns a directory or a decoded batch into a tidy data frame.

Documents

  • use_livelink_engine(): registers a livelink chunk engine for ‘knitr’ and ‘Quarto’, so a code chunk in a document becomes a runnable link.

    ```{livelink}
#| autorun: true
# Load the data
data(mtcars)
plot(mtcars$mpg, mtcars$wt)
```

    Use this rather than expression input inside a knitted document. ‘knitr’ runs chunks through evaluate::evaluate(), which discards source references, so comments inside a { } expression are silently dropped when a document is rendered, and no keep.source setting recovers them. The engine receives the chunk’s verbatim source, so comments survive. Chunk options cover autorun, panels, mode, filename, link.text, link.only, and engine.target ("webr", "shinylive-r", or "shinylive-py").

  • use_livelink_hook(): a ‘knitr’ chunk hook, so an ordinary r chunk can carry its own link. The chunk still runs, its output and plots appear in the rendered page, and a link is added underneath:

    ```{r}
#| livelink: true
#| autorun: true
data(mtcars)
plot(mtcars$mpg, mtcars$wt)   # a comment that survives

    ```

    The engine (```{livelink}) replaces execution, so it produces a link but no R output. The hook is what you want when the reader should see the result and also be able to open it; the engine is for code the session cannot or should not run, such as a Shiny app or anything needing a package you have not installed.

    Both are registered on load, and both receive the chunk’s verbatim source, so comments reach the link either way.

Projects as code

  • webr_repl_project(), shinylive_project(), and shinylive_r_link() accept a file’s contents as a braced expression, so a multi-file project can be written as R rather than as strings full of escaped newlines:

    webr_repl_project(list(
  "main.R"    = { source("utils.R"); summarise(mtcars) },
  "utils.R"   = { summarise <- function(d) summary(d) },
  "README.md" = "# Analysis"
))

    The blocks are captured, never evaluated. They are source to ship, not code to run, so an assignment inside one leaves nothing behind in the session. Strings and expressions mix freely, which is what a project holding both code and a README.md needs.

    This only works when the list is written inside the call. list() is an ordinary function, so assigning it to a variable first evaluates its arguments and the block runs there and then. livelink detects that case and errors, rather than encoding whatever the block returned.