R 3.1.0 (codename “Spring Dance“) was released today!

Photo credit: The Batsheva Dance Company in Ohad Naharin’s Hora. Photo by Gadi Dagon.

You can get the source code from
http://cran.r-project.org/src/base/R-3/R-3.1.0.tar.gz

or wait for it to be mirrored at a CRAN site nearer to you. Binaries for various platforms will appear in due course.

The full list of new features and bug fixes is provided below.

Upgrading to R 3.1.0

You can download the latest version from here.

If you are using Windows, it might take another 24 hours until you could update R. For convenience, you can upgrade to the latest version of R using the installr package. Simply run the following code:

# installing/loading the latest installr package:
install.packages("installr"); require(installr) #load / install+load installr
 
updateR()

After running “updateR()”, the function will detect that R is available for you, and will download+install it (etc.).

Note that the latest installr version (0.14.0) was released a week ago to CRAN, and it is recommended to upgrade to it, since it is now more robust for various extreme cases of upgrading R.
I try to keep the installr package updated and useful, so if you have any suggestions or remarks on the package – you are invited to leave a comment below.

If you use the global library system (as I do), you can run the following in the new version of R:

source("http://www.r-statistics.com/wp-content/uploads/2010/04/upgrading-R-on-windows.r.txt")
New.R.RunMe()

CHANGES IN R 3.1.0:

NEW FEATURES

• type.convert() (and hence by default read.table()) returns a character vector or factor when representing a numeric input as a double would lose accuracy. Similarly for complex inputs.

  If a file contains numeric data with unrepresentable numbers of decimal places that are intended to be read as numeric, specify colClasses in read.table() to be"numeric".

• tools::Rdiff(useDiff = FALSE) is closer to the POSIX definition of diff -b (as distinct from the description in the man pages of most systems).
• New function anyNA(), a version of any(is.na(.)) which is fast for atomic vectors, based on a proposal by Tim Hesterberg. (Wish of PR#15239.)
• arrayInd(*, useNames = TRUE) and, analogously, which(*, arr.ind = TRUE) now make use of names(.dimnames) when available.
• is.unsorted() now also works for raw vectors.
• The "table" method for as.data.frame() (also useful as as.data.frame.table()) now passes sep and base arguments to provideDimnames().
• uniroot() gets new optional arguments, notably extendInt, allowing to auto-extend the search interval when needed. The return value has an extra component,init.it.
• switch(f, ...) now warns when f is a factor, as this typically happens accidentally where the useR meant to pass a character string, but f is treated as integer (as always documented).
• The parser has been modified to use less memory.
• The way the unary operators (+ - !) handle attributes is now more consistent. If there is no coercion, all attributes (including class) are copied from the input to the result: otherwise only names, dims and dimnames are.
• colorRamp() and colorRampPalette() now allow non-opaque colours and a ramp in opacity via the new argument alpha = TRUE. (Suggested by Alberto Krone-Martins, but optionally as there are existing uses which expect only RGB values.)
• grid.show.layout() and grid.show.viewport() get an optional vp.ex argument.
• There is a new function find_gs_cmd() in the tools package to locate a GhostScript executable. (This is an enhanced version of a previously internal function there.)
• object.size() gains a format() method.
• There is a new family, "ArialMT", for the pdf() and postscript() devices. This will only be rendered correctly on viewers which have access to Monotype TrueType fonts (which are sometimes requested by journals).
• The text and PDF news files, including ‘NEWS’ and ‘NEWS.2’, have been moved to the ‘doc’ directory.
• combn(x, simplify = TRUE) now gives a factor result for factor input x (previously user error). (Related to PR#15442.)
• Added utils::fileSnapshot() and utils::changedFiles() functions to allow snapshots and comparison of directories of files.
• make.names(names, unique=TRUE) now tries to preserve existing names. (Suggestion of PR#15452.)
• New functions cospi(x), sinpi(x), and tanpi(x), for more accurate computation of cos(pi*x), etc, both in R and the C API. Using these gains accuracy in some cases, e.g., inside lgamma() or besselI(). (Suggested by Morten Welinder in PR#15529.)
• print.table(x, zero.print = ".") now also has an effect when x is not integer-valued.
• There is more support to explore the system’s idea of time-zone names. Sys.timezone() tries to give the current system setting by name (and succeeds at least on Linux, OS X, Solaris and Windows), and OlsonNames() lists the names in the system’s Olson database. Sys.timezone(location = FALSE) gives the previous behaviour.
• Platforms with a 64-bit time_t type are allowed to handle conversions between the "POSIXct" and "POSIXlt" classes for date-times outside the 32-bit range (before 1902 or after 2037): the existing workarounds are used on other platforms. (Note that time-zone information for post-2037 is speculative at best, and the OS services are tested for known errors and so not used on OS X.)

  Currently time_t is usually long and hence 64-bit on Unix-alike 64-bit platforms: however it several cases the time-zone database is 32-bit. On R for Windows it is 64-bit (for both architectures as from this version).

• The "save.defaults" option can include a value for compression_level. (Wish of PR#15579.)
• colSums() and friends now have support for arrays and data-frame columns with 2^31 or more elements.
• as.factor() is faster when f is an unclassed integer vector (for example, when called from tapply()).
• fft() now works with longer inputs, from the 12 million previously supported up to 2 billion. (PR#15593)
• Complex svd() now uses LAPACK subroutine ZGESDD, the complex analogue of the routine used for the real case.
• Sweave now outputs ‘.tex’ files in UTF-8 if the input encoding is declared to be UTF-8, regardless of the local encoding. The UTF-8 encoding may now be declared using a LaTeX comment containing the string %\SweaveUTF8 on a line by itself.
• file.copy() gains a copy.date argument.
• Printing of date-times will make use of the time-zone abbreviation in use at the time, if known. For example, for Paris pre-1940 this could be LMT, PMT, WET orWEST. To enable this, the "POSIXlt" class has an optional component "zone" recording the abbreviation for each element.

  For platforms which support it, there is also a component "gmtoff" recording the offset from GMT where known.

• (On Windows, by default on OS X and optionally elsewhere.) The system C function strftime has been replaced by a more comprehensive version with closer conformance to the POSIX 2008 standard.
• dnorm(x, log = FALSE) is more accurate (but somewhat slower) for |x| > 5; as suggested in PR#15620.
• Some versions of the tiff() device have further compression options.
• read.table(), readLines() and scan() have a new argument to influence the treatment of embedded nuls.
• Avoid duplicating the right hand side values in complex assignments when possible. This reduces copying of replacement values in expressions such as Z$a <- a0and ans[[i]] <- tmp: some package code has relied on there being copies.

  Also, a number of other changes to reduce copying of objects; all contributed by or based on suggestions by Michael Lawrence.

• The fast argument of KalmanLike(), KalmanRun() and KalmanForecast() has been replaced by update, which instead of updating mod in place, optionally returns the updated model in an attribute "mod" of the return value.
• arima() and makeARIMA() get a new optional argument SSinit, allowing the choice of a different state space initialization which has been observed to be more reliable close to non-stationarity: see PR#14682.
• warning() has a new argument noBreaks., to simplify post-processing of output with options(warn = 1).
• pushBack() gains an argument encoding, to support reading of UTF-8 characters using scan(), read.table() and related functions in a non-UTF-8 locale.
• all.equal.list() gets a new argument use.names which by default labels differing components by names (if they match) rather than by integer index. Saved Routput in packages may need to be updated.
• The methods for all.equal() and attr.all.equal() now have argument check.attributes after ... so it cannot be partially nor positionally matched (as it has been, unintentionally).

  A side effect is that some previously undetected errors of passing empty arguments (no object between commas) to all.equal() are detected and reported.

  There are explicit checks that check.attributes is logical, tolerance is numeric and scale is NULL or numeric. This catches some unintended positional matching.

  The message for all.equal.numeric() reports a "scaled difference" only for scale != 1.

• all.equal() now has a "POSIXt" method replacing the "POSIXct" method.
• The "Date" and "POSIXt" methods of seq() allows by = "quarter" for completeness (by = "3 months" always worked).
• file.path() removes any trailing separator on Windows, where they are invalid (although sometimes accepted). This is intended to enhance the portability of code written by those using POSIX file systems (where a trailing / can be used to confine path matching to directories).
• New function agrepl() which like grepl() returns a logical vector.
• fifo() is now supported on Windows. (PR#15600)
• sort.list(method = "radix") now allows negative integers (wish of PR#15644).
• Some functionality of print.ts() is now available in .preformat.ts() for more modularity.
• mcparallel() gains an option detach = TRUE which allows execution of code independently of the current session. It is based on a new estranged = TRUEargument to mcfork() which forks child processes such that they become independent of the parent process.
• The pdf() device omits circles and text at extremely small sizes, since some viewers were failing on such files.
• The rightmost break for the "months", "quarters" and "years" cases of hist.POSIXlt() has been increased by a day. (Inter alia, fixes PR#15717.)
• The handling of DF[i,] <- a where i is of length 0 is improved. (Inter alia, fixes PR#15718.)
• hclust() gains a new method "ward.D2" which implements Ward’s method correctly. The previous "ward" method is "ward.D" now, with the old name still working. Thanks to research and proposals by Pierre Legendre.
• The sunspot.month dataset has been amended and updated from the official source, whereas the sunspots and sunspot.year datasets will remain immutable. The documentation and source links have been updated correspondingly.
• The summary() method for "lm" fits warns if the fit is essentially perfect, as most of the summary may be computed inaccurately (and with platform-dependent values).

  Programmers who use summary() in order to extract just a component which will be reliable (e.g. $cov.unscaled) should wrap their calls insuppressWarnings().

INSTALLATION and INCLUDED SOFTWARE

• The included version of LAPACK has been updated to 3.5.0.
• There is some support for parallel testing of an installation, by setting TEST_MC_CORES to an integer greater than one to indicate the maximum number of cores to be used in parallel. (It is worth specifying at least 8 cores if available.) Most of these require a make program (such as GNU make and dmake) which supports the$MAKE -j nproc syntax.

  Except on Windows: the tests of standard package examples in make check are done in parallel. This also applies to runningtools::testInstalledPackages().

  The more time-consuming regression tests are done in parallel.

  The package checks in make check-devel and make check-recommended are done in parallel.

• More of make check will work if recommended packages are not installed: but recommended packages remain needed for thorough checking of an R build.
• The version of tzcode included in ‘src/extra/tzone’ has been updated. (Formerly used only on Windows.)
• The included (64-bit) time-zone conversion code and Olson time-zone database can be used instead of the system version: use configure option –with-internal-tzcode. This is the default on Windows and OS X. (Note that this does not currently work if a non-default rsharedir configure variable is used.)

  (It might be necessary to set environment variable TZ on OSes where this is not already set, although the system timezone is deduced correctly on at least Linux, OS X and Windows.)

  This option also switches to the version of strftime included in directory ‘src/extra/tzone’.

• configure now tests for a C++11-compliant compiler by testing some basic features. This by default tries flags for the compiler specified by CXX, but an alternative compiler, options and standard can be specified by variables CXX1X, CXX1XFLAGS and CXX1XSTD (e.g. -std=gnu++11).
• R can now optionally be compiled to use reference counting instead of the NAMED mechanism by defining SWITCH_TO_REFCNT in ‘Rinternals.h’. This may become the default in the future.
• There is a new option –use-system-tre to use a suitable system tre library: at present this means a version from their git repository, after corrections. (Wish ofPR#15660.)

PACKAGE INSTALLATION

• The CRANextra repository is no longer a default repository on Windows: all the binary versions of packages from CRAN are now on CRAN, although CRANextracontains packages from Omegahat and elsewhere used by CRAN packages.
• Only vignettes sources in directory ‘vignettes’ are considered to be vignettes and hence indexed as such.
• In the ‘DESCRIPTION’ file,

      License: X11

  is no longer recognized as valid. Use MIT or BSD_2_clause instead, both of which need + file LICENSE.

• For consistency, entries in ‘.Rinstignore’ are now matched case-insensitively on all platforms.
• Help for S4 methods with very long signatures now tries harder to split the description in the Usage field to no more than 80 characters per line (some packages had over 120 characters).
• R CMD INSTALL --build (not Windows) now defaults to the internal tar() unless R_INSTALL_TAR is set.
• There is support for compiling C++11 code in packages on suitable platforms: see ‘Writing R Extensions’.
• Fake installs now install the contents of directory ‘inst’: some packages use this to install e.g. C++ headers for use by other packages that are independent of the package itself. Option –no-inst can be used to get the previous behaviour.

DEBUGGING

• The behaviour of the code browser has been made more consistent, in part following the suggestions in PR#14985.
• Calls to browser() are now consistent with calls to the browser triggered by debug(), in that Enter will default to n rather than c.
• A new browser command s has been added, to “step into” function calls.
• A new browser command f has been added, to “finish” the current loop or function.
• Within the browser, the command help will display a short list of available commands.

UTILITIES

• Only vignettes sources in directory ‘vignettes’ are considered to be vignettes by R CMD check. That has been the preferred location since R 2.14.0 and is now obligatory.
• For consistency, R CMD build now matches entries in ‘.Rbuildignore’ and ‘vignettes/.install_extras’ case-insensitively on all platforms (not just on Windows).
• checkFF() (called by R CMD check by default) can optionally check foreign function calls for consistency with the registered type and argument count. This is the default for R CMD check --as-cran or can be enabled by setting environment variable _R_CHECK_FF_CALLS_ to registration (but is in any case suppressed by –install=no). Because this checks calls in which .NAME is an R object and not just a literal character string, some other problems are detected for such calls.

  Functions suppressForeignCheck() and dontCheck() have been added to allow package authors to suppress false positive reports.

• R CMD check --as-cran warns about a false value of the ‘DESCRIPTION’ field BuildVignettes for Open Source packages, and ignores it. (An Open Source package needs to have complete sources for its vignettes which should be usable on a suitably well-equipped system).
• R CMD check --no-rebuild-vignettes is defunct:
  R CMD check --no-build-vignettes has been preferred since R 3.0.0.
• R CMD build --no-vignettes is defunct:
  R CMD build --no-build-vignettes has been preferred since R 3.0.0.
• R CMD Sweave and R CMD Stangle now process both Sweave and non-Sweave vignettes. The tools::buildVignette() function has been added to do the same tasks from within R.
• The flags returned by R CMD config --ldflags and (where installed) pkg-config --libs libR are now those needed to link a front-end against the (shared or static) R library.
• ‘Sweave.sty’ has a new option [inconsolata].
• R CMD check customizations such as _R_CHECK_DEPENDS_ONLY_ make available packages only in LinkingTo only for installation, and not for loading/runtime tests.
• tools::checkFF() reports on .C and .Fortran calls with DUP = FALSE if argument check_DUP is true. This is selected by R CMD check by default.
• R CMD check --use-gct can be tuned to garbage-collect less frequently using gctorture2() via the setting of environment variable _R_CHECK_GCT_N_.
• Where supported, tools::texi2dvi() limits the number of passes tried to 20.

C-LEVEL FACILITIES

• (Windows only) A function R_WaitEvent() has been added (with declaration in header‘R.h’) to block execution until the next event is received by R.
• Remapping in the ‘Rmath.h’ header can be suppressed by defining R_NO_REMAP_RMATH.
• The remapping of rround() in header ‘Rmath.h’ has been removed: use fround() instead.
• ftrunc() in header ‘Rmath.h’ is now a wrapper for the C99 function trunc(), which might as well be used in C code: ftrunc() is still needed for portable C++ code.
• The never-documented remapping of prec() to fprec() in header ‘Rmath.h’ has been removed.
• The included LAPACK subset now contains ZGESDD and ZGELSD.
• The function LENGTH() now checks that it is only applied to vector arguments. However, in packages length() should be used. (In R itself LENGTH() is a macro without the function overhead of length().)
• Calls to SET_VECTOR_ELT() and SET_STRING_ELT() are now checked for indices which are in-range: several packages were writing one element beyond the allocated length.
• allocVector3 has been added which allows custom allocators to be used for individual vector allocations.

DEPRECATED AND DEFUNCT

• chol(pivot = TRUE, LINPACK = TRUE) is defunct.

  Arguments EISPACK for eigen() and LINPACK for chol(), chol2inv(), solve() and svd() are ignored: LAPACK is always used.

• .find.package() and .path.package() are defunct: only the versions without the initial dot introduced in R 2.13.0 have ever been in the API.
• Partial matching when using the $ operator on data frames now throws a warning and may become defunct in the future. If partial matching is intended, replacefoo$bar by foo[["bar", exact = FALSE]].
• The long-deprecated use of \synopsis in the Usage section of ‘.Rd’ files has been removed: such sections are now ignored (with a warning).
• package.skeleton()‘s deprecated argument namespace has been removed.
• Many methods are no longer exported by package stats. They are all registered on their generic, which should be called rather than calling a method directly.
• Functions readNEWS() and checkNEWS() in package tools are defunct.
• download.file(method = "lynx") is deprecated.
• .C(DUP = FALSE) and .Fortran(DUP = FALSE) are now deprecated, and may be disabled in future versions of R. As their help has long said, .Call() is much preferred.

  R CMD check notes such usages (by default).

• The workaround of setting R_OSX_VALGRIND has been removed: it is not needed in current valgrind.

BUG FIXES

• Calling lm.wfit() with no non-zero weights gave an array-overrun in the Fortran code and a not very sensible answer. It is now special-cased with a simpler answer (no qr component).
• Error messages involving non-syntactic names (e.g. as produced by `\r` when that object does not exist) now encode the control characters. (Reported by Hadley Wickham.)
• getGraphicsEvent() caused 100% usage of one CPU in Windows. (PR#15500)
• nls() with no start argument may now work inside another function (scoping issue).
• pbeta() and similar work better for very large (billions) ncp.
• Where time zones have changed abbreviations over the years, the software tries to more consistently use the abbreviation appropriate to the time or if that is unknown, the current abbreviation. On some platforms where the C function localtime changed the tzname variables the reported abbreviation could have been that of the last time converted.
• all.equal(list(1), identity) now works.
• Bug fix for pushing viewports in grid (reported by JJ Allaire and Kevin Ushey).

  NOTE for anyone poking around within the graphics engine display list (despite the warnings not to) that this changes what is recorded by grid on the graphics engine display list.

• Extra checks have been added for unit resolution and conversion in grid, to catch instances of division-by-zero. This may introduce error messages in existing code and/or produce a different result in existing code (but only where a non-finite location or dimension may now become zero).
• Some bugs in TRE have been corrected by updating from the git repository. This allows R to be installed on some platforms for which this was a blocker (PR#15087 suggests Linux on ARM and HP-UX).
• ? applied to a call to an S4 generic failed in several cases. (PR#15680)
• The implicit S4 generics for primitives with ... in their argument list were incorrect. (PR#15690)
• Bug fixes to methods::callGeneric(). (PR#15691)
• The bug fix to aggregrate() in PR#15004 introduced a new bug in the case of no grouping variables. (PR#15699)
• In rare cases printing deeply nested lists overran a buffer by one byte and on a few platforms segfaulted. (PR#15679)
• The dendrogram method of as.dendrogram() was hidden accidentally, (PR#15703), and order.dendrogram(d) gave too much for a leaf d. (PR#15702)
• R would try to kill processes on exit that have pids ever used by a child process spawned by mcparallel even though the current process with that pid was not actually its child.
• cophenetic() applied to a "dendrogram" object sometimes incorrectly returned a "Labels" attribute with dimensions. (PR#15706)
• printCoefmat() called from quite a few print() methods now obeys small getOption("width") settings, line wrapping the ”signif. codes” legend appropriately. (PR#15708)
• model.matrix() assumed that the stored dimnames for a matrix was NULL or length 2, but length 1 occurred.
• The clipping region for a device was sometimes used in base graphics before it was set.