| news {utils} | R Documentation |
Build and query the news data base for R or add-on packages.
news(query, package = "R", lib.loc = NULL, format = NULL,
reader = NULL, db = NULL)
## S3 method for class 'news_db'
print(x, doBrowse = interactive(),
browser = getOption("browser"), ...)
query |
an expression for selecting news entries |
package |
a character string giving the name of an installed
add-on package, or |
lib.loc |
a character vector of directory names of R libraries,
or |
format |
Not yet used. |
reader |
Not yet used. |
db, x |
a news db obtained from |
doBrowse |
logical specifying that the news should be opened in
the browser (by |
browser |
the browser to be used, see |
... |
potentially further arguments passed to |
If package is "R" (default), a news db is built with the
news since the 3.0.0 release of R (corresponding to R's top-level
‘NEWS’ file). Otherwise, if the given add-on package can be
found in the given libraries, it is attempted to read its news in
structured form from files ‘inst/NEWS.Rd’, ‘NEWS.md’
(since R version 3.6.0, needs packages commonmark and
xml2 to be available), ‘NEWS’ or ‘inst/NEWS’
(in that order).
File ‘inst/NEWS.Rd’ should be an Rd file given the entries as Rd
\itemize lists, grouped according to version using
section elements with names starting with a suitable prefix
(e.g. “Changes in version”) followed by a space and the version
number, and optionally followed by a space and a parenthesized ISO
8601 (%Y-%m-%d, see strptime) format date, and
possibly further grouped according to categories using
\subsection elements named as the categories.
At the very end of \section{..}, the date may also be specified
as (%Y-%m-%d, <note>), i.e., including parentheses.
File ‘NEWS.md’ should contain the news in Markdown (following the CommonMark (https://commonmark.org/) specification), with the primary heading level giving the version number after a prefix followed by a space, and optionally followed by a space and a parenthesized ISO 8601 format date. Where available, secondary heading are taken to indicate categories. To accommodate for common practice, news entries are only split down to the category level.
The plain text ‘NEWS’ files in add-on packages use a variety of different formats; the default news reader should be capable to extract individual news entries from a majority of packages from the standard repositories, which use (slight variations of) the following format:
Entries are grouped according to version, with version header “Changes in version” at the beginning of a line, followed by a version number, optionally followed by an ISO 8601 format date, possibly parenthesized.
Entries may be grouped according to category, with a category header (different from a version header) starting at the beginning of a line.
Entries are written as itemize-type lists, using one of o, *, - or + as item tag. Entries must be indented, and ideally use a common indentation for the item texts.
Additional formats and readers may be supported in the future.
Package tools provides an (internal) utility function
news2Rd to convert plain text ‘NEWS’ files to Rd. For
‘NEWS’ files in a format which can successfully be handled by the
default reader, package maintainers can use
tools:::news2Rd(dir, "NEWS.Rd"),
possibly with additional argument codify = TRUE,
with dir a character string specifying the path to a package's
root directory. Upon success, the ‘NEWS.Rd’ file can further be
improved and then be moved to the ‘inst’ subdirectory of the
package source directory.
The news db built is a character data frame inheriting from
"news_db" with variables Version, Category,
Date and Text, where the last contains the entry texts
read, and the other variables may be NA if they were missing or
could not be determined.
Using query, one can select news entries from the db. If
missing or NULL, the complete db is returned. Otherwise,
query should be an expression involving (a subset of) the
variables Version, Category, Date and
Text, and when evaluated within the db returning a logical
vector with length the number of entries in the db. The entries for
which evaluation gave TRUE are selected. When evaluating,
Version and Date are coerced to
numeric_version and Date objects,
respectively, so that the comparison operators for these classes can
be employed.
A data frame inheriting from class "news_db", with
attributes "package" (and "subset" if the
query lead to proper subsetting).
## Build a db of all R news entries.
db <- news()
## Bug fixes with PR number in 3.0.1.
db3 <- news(Version == "3.0.1" & grepl("^BUG", Category) & grepl("PR#", Text),
db = db)
## News from a date range ('Matrix' is there in a regular R installation):
if(length(iM <- find.package("Matrix", quiet=TRUE)) && nzchar(iM)) {
dM <- news(package="Matrix")
stopifnot(identical(dM, news(db=dM)))
dM2014 <- news("2014-01-01" <= Date & Date <= "2014-12-31", db = dM)
stopifnot(paste0("1.1-", 2:4) %in% dM2014[,"Version"])
}
## Which categories have been in use? % R-core maybe should standardize a bit more
sort(table(db[, "Category"]), decreasing = TRUE)
## Entries with version >= 3.0.0 (including "3.0.0 patched"):
table(news(Version >= "3.0.0", db = db)$Version)