similar to: anomalies with call objects (PR#412)

I have committed a set of changes to the tcltk package for R-1.8.0. The most important changes are 1) Support for Tcl arrays, which should come in handy for people playing with things like the tkTable extension. Basically these work like tclVar objects but can be subscripted (notice that these beasts are associative - like a Perl hash - rather than objects of set dimensions). Basic

A couple of bugs turned up immediately after the release, one of them causing nasty memory corruption when strsplit() was used. Therefore, a set of patches has been supplied on ftp://cvs.r-project.org/pub/R/src/base/R-0.99.0-fixes.gz and the rest of CRAN should soon follow. The recipe for patching is (GNU patch assumed) gunzip -dc R-0.99.0-fixes.gz | patch -p 1 For those with broken patch

A couple of bugs turned up immediately after the release, one of them causing nasty memory corruption when strsplit() was used. Therefore, a set of patches has been supplied on ftp://cvs.r-project.org/pub/R/src/base/R-0.99.0-fixes.gz and the rest of CRAN should soon follow. The recipe for patching is (GNU patch assumed) gunzip -dc R-0.99.0-fixes.gz | patch -p 1 For those with broken patch

"Marco Taboga" <mtaboga@tiscalinet.it> writes: > OK, it works. xlab and ylab had to be set properly. > If you want to see what actually goes on, try: > a<-chron(1:2000) > b<-1:2000 > plot(a,b,type="l") > One of the standard blunders with R's lazy evaluation and substitute is to change the value of part of a substitute expression before it

OK folks, We're warming up to the release of R-2.0.0 on Oct. 4 (aka 2004-10-04). There will be two weeks of alpha releases followed by two weeks of beta releases, starting today. The releases are made available via CRAN in http://cran.r-project.org/src/base-prerelease and the first one is there already. In case of urgency or CRAN failure, the releases will also be found at

Marcus Eger <marcus.eger@physik.uni-marburg.de> writes: > > time(sqrt(arrts)) > Time Series: > Start = c(1, 1) > End = c(5, 1) > Frequency = 1 > [1] 1 2 3 4 5 Looks like a bug... > ---------------------------------------------------------------------------- > 2. (At least) boolean indexing with matrices does not seem to work > properly: >

Here is a copy of the last few lines in base-Ex.Rout.fail: > x <- seq(3,500);yl <- c(-.3, .2) > plot(x,x, ylim = yl, ylab="",type='n', main = "Bessel Functions Y_nu(x)") > for(nu in nus){xx <- x[x > .6*nu]; lines(xx,besselY(xx,nu=nu), col = nu+2)} > legend(300,-.08, leg=paste("nu=",nus), col = nus+2, lwd=1) > > x <-

I'm trying to make a habit of announcing upcoming releases on r-devel, to give package maintainers a chance of staying in sync, and to encourage everyone to report bugs in the prerelease versions. So: R-1.3.1 is scheduled for release on August 31st. In order to be able to get a solid round of testing on different platforms, maintainers of the group of recommended packages are requested to

If I'm not mistaken, "bundle" is really only useful as a concept for distribution and installation. You distribute and install a bundle, but load the individual packages when you want to use them. Once you install the bundle, you won't see the name of the bundle in the list of installed packages, but you see the constituent packages, and those are what you load when you want to

I agree with the the comments below on "Numerical recipes". We developed some commercial software using the code and tried to get a licence. They never responded to our corresponce :-) We also mailed them fixes to some bugs !!! but never heard anything. So we threw their code away and wrote our own... it was a waste of time using the book. Kim On Thursday, 25 November 1999 6:41,

For small datasets it would be useful to be able to copy a block of data from a spreadsheet then toggle over to R and just paste it in. If it's possible, having read.table access the clipboard (e.g. x <- read.table(file="clipboard", ...) would do the trick. Of course, exporting to a file and the reading into R is pretty easy but usring the clipboard would save a couple extra

bhoel@server.python.net (Berthold Höllmann) writes: > Peter Dalgaard BSA <p.dalgaard@biostat.ku.dk> writes: > > > I've rolled up R-1.0.0.tgz a short while ago. > > > I've build R-1.0.0 on my > > >uname -a > Linux pchoel 2.2.14 #3 Mit Jan 5 08:57:39 MET 2000 i686 unknown > > box. Calling "make check" fails with .... > >

bhoel@server.python.net (Berthold Höllmann) writes: > Peter Dalgaard BSA <p.dalgaard@biostat.ku.dk> writes: > > > I've rolled up R-1.0.0.tgz a short while ago. > > > I've build R-1.0.0 on my > > >uname -a > Linux pchoel 2.2.14 #3 Mit Jan 5 08:57:39 MET 2000 i686 unknown > > box. Calling "make check" fails with .... > >

TASK: problem with "glm" with binomial errors STATUS: Open FROM: p.dalgaard@kubism.ku.dk in glm(,binomial) it's possible that loss of significant digits make expected values 0 or 1 even though there's no divergence of the fit. (Happened to me with menarche data, infants and grown-ups included) [ Need the example data. Glm needs a

Faheem Mitha <faheem@email.unc.edu> writes: > Dear Dr. Dalgaard, > > I am sending this message only to you since it is probably not of general > interest. Forward it to anyone you like. > > Congratulations to the R development team on the release of version 1.0.0 > of R. If you will excuse me asking, I was wondering if there are any plans > to release a rpm of

Just to update: What I originally thought was a glm problem appears to be a memory problem that occurs only when R is compiled with the DEC C compiler. Some variables that are still in use get clobbered during garbage collection. No problems if I compile with gcc though. I've made some attempts to see if I can identify a specific compiler optimization that is responsible, but so far no luck.

Kjetil Kjernsmo <kjetil.kjernsmo@astro.uio.no> writes: > On 8 Nov 2000, Peter Dalgaard BSA wrote: > > >I'm not at all happy with this: > > > >Solaris : > >> range(sapply(1:2000, function(n) mean(rpois(10000, 15.0)))) > >[1] 15.0524 15.3403 > > Hm, OK, so it isn't just me.... I guess it is time to file a bug report, > should I do it,

I can confirm this bug has been solved, by the expedient of removing the incorrect functions! I think the simplest way to get a bug closed is to send the correction to the original reporter for confirmation, then as Peter suggests send a follow-up to R-bugs. On Fri, 6 Apr 2001 p.dalgaard@biostat.ku.dk wrote: > Paul Gilbert <pgilbert@bank-banque-canada.ca> writes: > > > Peter

The release schedule has now been set with a release date on April 18. The detailed procedure can be found at http://developer.r-project.org/ (it is not quite there yet, but will appear after a short propagation delay.) The main point for non-developers is that we start making alpha tarballs on March 21 and beta tarballs on April 4. Only very simple and/or critical bugs are fixed in the last

Thanks to Peter Dalgaard and Suchandra Thapa for answering my question! Before I settle on a particular option, I'd like to ask one more question if I may: Are there any practical advantages to compiling R to 64-bit vs 32-bit? Regards, Andy > -----Original Message----- > From: Peter Dalgaard BSA [mailto:p.dalgaard at biostat.ku.dk] > Sent: Thursday, April 18, 2002 4:19 PM > To: