R 2.2.1 is released

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

R 2.2.1 is released

Peter Dalgaard
I've rolled up R-2.2.1.tar.gz a short while ago. This is a maintenance
release containing mainly bugfixes.

See the full list of changes below.

You can get it from

http://cran.r-project.org/src/base/R-2/R-2.2.1.tar.gz

(give it some time to arrive there) or wait for it to be mirrored at a
CRAN site nearer to you. If you're *really* impatient,
http://www.biostat.ku.dk/~pd/R-release should work too. Binaries for
various platforms will appear in due course.
 
There is also a version split for floppies.

        For the R Core Team

        Peter Dalgaard

These are the md5sums for the freshly created files, in case you wish
to check that they are uncorrupted:

94d55d512a9ba36caa9b7df079bae19f  COPYING
fad9b3332be894bab9bc501572864b29  COPYING.LIB
de541086db1146c1595d5be1d94a1b94  FAQ
70447ae7f2c35233d3065b004aa4f331  INSTALL
2c832b91154f663c0f07930d5e2a3dee  NEWS
88bbd6781faedc788a1cbd434194480c  ONEWS
4f004de59e24a52d0f500063b4603bcb  OONEWS
42542290c6d1585af7ded330f811385c  R-2.2.1.tar.gz
a9126622c51bef60e3febb41b2e737e5  R-2.2.1.tar.gz-split.aa
e89b51f6dbc1ddffd646a3f1f203f4f1  R-2.2.1.tar.gz-split.ab
15d93b1b5c38b6178ddcf9485dbd8cf1  R-2.2.1.tar.gz-split.ac
282bbbdd00c0066b6042ce16844036be  R-2.2.1.tar.gz-split.ad
6d00347225140283e6c93f253cb7c180  R-2.2.1.tar.gz-split.ae
2441c58fc24a69dd5cc65517a0b5e285  R-2.2.1.tar.gz-split.af
4ddcd13639debd984e3664a6b0681bd4  R-2.2.1.tar.gz-split.ag
4a042d73ec00d7d46be1cc704063dc39  R-2.2.1.tar.gz-split.ah
0d5d54a78ab994bacb12a98bece1306f  R-2.2.1.tar.gz-split.ai
69b38c655e4c1b7c36cb0abaa4890253  R-2.2.1.tar.gz-split.aj
42542290c6d1585af7ded330f811385c  R-latest.tar.gz
56a780cdec835c5c598f8dfc0738f7f3  README
020479f381d5f9038dcb18708997f5da  RESOURCES

Here is the relevant part of the NEWS file:


USER-VISIBLE CHANGES

    o options("expressions") has been reduced to 1000: the limit
        of 5000 introduced in 2.1.0 was liable to give crashes from C
        stack overflow.


NEW FEATURES

    o Use of 'pch' (e.g. in points) in the symbol font 5 is now
        interpreted in the single-byte encoding used by that font.
        Similarly, strwidth now recognizes that font 5 has a different
        encoding from that of the locale.  (These are likely to affect
        the answer only in MBCS locales such as UTF-8.)

    o The URW font metrics have been updated to versions from late
        2002 which cover more glyphs, including Cyrillic.

    o New postscript encodings for CP1250 (Windows East European),
        ISO Latin-7 (8859-13, Latvian, Lithuanian and Maori), Cyrillic
        (8859-5), KOI8-R, KOI8-U and CP1251.

    o configure has more support for the Intel and Portland Group
        compilers on ix86 and x86_64 Linux.

    o R CMD INSTALL will clean up if interrupted (e.g. by ctrl-C from
        the keyboard).

    o There is now a comprehensive French translation of the messages,
        thanks to Philippe Grosjean.


DEPRECATED & DEFUNCT

    o The undocumented use of atan() with two arguments is deprecated:
        instead use atan2() (as documented).

    o The 'vfont' argument of axis() and mtext() is deprecated
        (it currently warns and does nothing).

    o   The function mauchley.test() is deprecated (was a misspelling)
        and replaced by mauchly.test()


BUG FIXES

    o The malloc's of AIX and OSF/1 which return NULL for size 0
        are now catered for in src/main/regex.c.

    o Names of list elements which are missing are now printed as
        $<NA> and not $"NA" (which is how the non-missing name "NA" is
        printed).  (Brought up in discussion of PR#8161.)

    o help.start() was not linking R.css for use by its front page and
        immediate links (2.2.0 only).

    o Indexing by character NA matched the name "NA".

    o The arith-true test used random inputs and did not set the seed, so
        it could fail very occasionally.

    o arima() with 'fixed' supplied and p=0 for the non-seasonal
        part could give spurious warnings about 'some AR parameters
        were fixed'.

    o summary.matrix() could give an infinite recursion on some
        classed objects (e.g. those of class "Surv").

    o The 255th character in an 8-bit character set was not handled
        correctly as a letter on some platforms where C char is
        signed: for example it was printed as \377 and not allowed in
        variable names.  (Spotted by Alexey Shipunov in Russian
        encodings.)

    o Conversion from POSIXct to POSIXlt is done more accurately
        around the change of DST in years not supported by the OS
        (pre-1970 on Windows and some others, and in the far past or
        future).

    o chisq.test(cbind(1:0, c(7,16)), simulate.p = TRUE) gave wrong
        P-values on some platforms. (PR#8224)

    o pdf() was not writing details of the encoding to the file
        correctly.  (Spotted by Alexey Shipunov in Russian encodings.)

    o image() was failing with an error when plotting a matrix
        of all NA values.  (PR#8228)

        image() could fail if called with add=FALSE (the default) and
        length(x)=1 for either x or y, as it uses the plot coordinates of
        the previous plot (if any).

    o tools::checkMD5sums was not accepting file names with spaces in.

    o The plot() method for TukeyHSD() needed updating after adding
        adjusted p-values.  (PR#8229)

    o read.fwf() did not work for header = TRUE.  (PR#8226)

    o diag() failed when its argument had NA values in its dimnames.

    o [g]sub(pcre=TRUE) did not work correctly with \U and \L in a
        UTF-8 locale, even on the example on the help page.

    o promptMethods() was failing if the "methods" argument was supplied.

    o is.loaded() now finds Fortran symbols whether or not the
        registration mechanism has been used.

    o ISODateTime() mistakenly corrected non-existent times (when
        DST was being started) in the current time zone.

    o Some replacement operations on data frames gave incorrect
        answers, e.g. DF[3:4, "y"] if column "y" did not exist or was
        a matrix.

    o getGraphicsEvent() would cause memory corruption if passed an
    empty prompt.

    o qr() and chol() now pivot the colnames of the result when
        pivoting is used.  (PR#8258)

    o example(points) omitted pch=0, although it was valid and
        said in the text to be illustrated.

    o plot.default() had an unused 'lab' argument, thereby preventing
        the 'lab' graphics parameter being passed through '...' .

    o Although polygon(col = NA) was the stated default, specifying
        NA was not equivalent to omitting the argument (but col=NULL
        was equivalent).

    o Im(-1) was pi. (PR#8272, a side effect from all previous
        versions of R returning the same value for Im and Arg of
        non-complex numbers.)

    o symbols(fg) defaulted to colour 1, not par("col") as documented.
        It does now defaults to par("col").

    o par("family") did not check the length of the value (up to 49
        bytes) and so could segfault.

    o aggregate.ts() did not allow for rounding in frequencies such
        as 1/5.

    o prcomp(tol=) was not dropping the sdev's corresponding to
        dropped columns.

    o Subassignment of a vector which increased the length of the
        vector _and_ had the wrong length of replacement could
        occasionally segfault.  (This has been there since at least
        mid 1997.)

    o The registration of .Fortran symbols was broken: these could
        only be looked up if there were also .Call symbols registered!

    o R CMD build was incorrectly rejecting the recommended form of
        name for a translation package, 'Translation-ll'.  (PR#8314)

    o numericDeriv() gave nonsense results unless the variables were
        real, which was not checked.

    o predict.prcomp() would sometimes give an error when predicting
        a single observation.  (PR#8324)

    o mapply() could segfault if MoreArgs was not a list.  (PR#8332)

    o The arith-true test used identical() on floating-point results,
        and this allowed a failure when the relative difference was
        less than .Machine$double.eps but non-zero.

    o qbinom() was not accepting p = -Inf when log.p = TRUE, although
        it is a legitimate value.

    o write.csv[2] only accepted logical constants for 'row.names', and
        now accepts variables.

    o Conversion of .Rd files did not correctly match braces
        enclosing a whole argument, e.g. \eqn{{\bf a}}{a}.

    o The C function pythag (used if hypot was not available) would
        infinite-loop on systems with effective optimizing compilers.

    o Writing long formats (more than 1000 bytes) with connections
        that use dummy_vfprintf could fail on some systems.  The limit
        has been changed to 100000 bytes pending a more complete fix in R
        2.3.0.

    o Making in src/nmath/standalone without making R was not
        making Rmath.h.

    o Both the R front-end and INSTALL could find the attempted
        temporary directory name already in use on platforms without
        mktemp (and a genuine Bourne shell /bin/sh, not bash).  Now
        both the process ID and a timestamp are used to create the
        directory name.

    o [dpqr]gamma now return NaN for an invalid 'shape' parameter
        (rather than throw an error), for consistency with other
        distribution functions.

    o t() now longer drops dimnames 'list(NULL,NULL)' or 'list(NULL)'.

    o Influence measures such as rstandard() and cooks.distance()
        could return infinite values rather than NaN for a case which
        was fitted exactly.  Similarly, plot.lm() could fail on such
        examples.  plot.lm(which = 5)  had to be modified to only plot
        cases with hat < 1.  (PR#8367)

        lm.influence() was incorrectly reporting 'coefficients' and
        'sigma' as NaN for cases with hat = 1, and on some platforms
        not detecting hat = 1 correctly.

    o Rmath.h for standalone Rmath was not recording HAVE_WORKING_LOG,
        so R_log was not available on platforms defining it.

    o HoltWinters() was using a slightly incorrect formula in the C code.

    o dir.create() could be confused by a trailing slash on the
        path, and by paths containing drives on Windows.

    o The search for tcl/tkConfig.sh looked in 'lib' before 'lib64'
        directories (and not at all in /usr/local/lib64) and so might
  prefer 32- to 64-bit versions if both are available.

    o nlminb() used an uninitialized variable unless bounds were supplied,
        and so failed on 64-bit Solaris.


--
   O__  ---- Peter Dalgaard             Ă˜ster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - ([hidden email])                  FAX: (+45) 35327907

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-announce

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html