similar to: sys.on.exit appears non-functional (PR#269)

I noticed this oddity about R profiling and setMethod. First, I "test out" Rprof. > require(methods) Loading required package: methods [1] TRUE > > Rprof("test.out") > data.frame("a") X.a. 1 a > Rprof(NULL) So far, so good. Next, I define myClass. > setClass("myClass", representation(mySlot = "numeric")) [1]

(1) ?cbind claims The generic functions `cbind' and `rbind' take a sequence of vector and/or matrix arguments and combine them as the columns or rows, respectively, of a matrix. Note: The method dispatching is not done via `UseMethod(..)', but by C-internal dispatching. Therefore, there's no need for, e.g., `rbind.default'. but my cbind.ts

The TITLE for the nnet package is garbled: it comes out as nnet Feed-forward neural networks and multinomial log-linear nnet Feed-forward neural networks and multinomial log-linear models The problem is in maketitle: auk% cat DESCRIPTION Bundle: VR Version: 6.1-9 Date: 2000/07/11 Depends: R (>= 1.1) Author: S original by Venables & Ripley. R port by Brian Ripley

B <- A <- factor(c("a", "b")) outer(A, B, "!=") Warning: "FUN" not meaningful for factors [,1] [,2] [1,] NA NA [2,] NA NA Now, this used to work in 0.63.2, but someone `improved' outer. There it did an implicit as.numeric. The problem is that get in match.fun does not understand group generics, and gets Browse[1]> FUN

Using zero-weight values in glm returns incorrect fitted values and linear predictors, the ninth value in the following. > example(glm) > fit <- glm(counts ~ outcome + treatment, family = poisson(), data=d.AD, weights=c(rep(1,8), 0)) > fit$linear.predictor 1 2 3 4 5 6 7 8 2.989646 2.535391 2.862201 2.989646

anova() on three or more objects behaves inconsistently in R. In R anovalist.lm does a sequential ANOVA using pairwise F tests, ignoring all the other objects, so the larger of the two models provides the denominator. In S anova.lmlist uses the denominator from the largest model (smallest residual df) in the set, as does anova.glmlist in both. I suggest that R's anovalist.lm is wrong (that

predict.lm has been broken by recent changes to the patched branch. It fails for all singular fits. An example: library(MASS) data(quine) quine.hi <- aov(log(Days + 2.5) ~ .^4, quine) quine.nxt <- update(quine.hi, . ~ . - Eth:Sex:Age:Lrn) predict(quine.nxt) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 NA NA NA NA NA NA NA NA NA NA NA NA NA

The inconsistency is that you use $<- to set the column, then [[<- to change it. Had you tried to set the column by x[[1]] <- as.character(x[[1]]) you would have seen the problem immediately (it does not work as you would have intended). If you want to be sure to turn off conversion to factor, you need to set the column to class "AsIs". My belief is that will behave

Version: platform = sparc-sun-solaris2.6 arch = sparc os = solaris2.6 system = sparc, solaris2.6 status = status.rev = 0 major = 0 minor = 64.2 year = 1999 month = July day = 3 language = R -- It is unclear to me that the handling of NAs is desirable, and it has problems with names: > z <- c(T,T,NA,F,T) > names(z) <- letters[1:5] > which(z) Error: names attribute

> From: rossini@blindglobe.net (A.J. Rossini) > Date: 01 Nov 2000 07:47:21 -0800 [...] > Thanks for the pointer to stack/unstack -- now, having been reminded, > I think I'd seen these float through on the list (still doesn't solve > the missing modeling routines (parametric GLMMs, some of the > econometrics stuff -- does R _easily_ do 3SLS?), but they'll appear >

On Mon, 28 Aug 2000, Rashid Nassar wrote: > Dear Professor Ripley, > > Thank you very much for your kind explanation. If I may lamely say > something in my defence, even as I apologize for my error: I mistook the > sentence "the (quoted) name of a function" to mean "optionally quoted" > because of the parentheses surrounding "quoted", and was

Environment variables set with Sys.putenv() "disappear" spontaneously (and somewhat randomly) when I load large datasets. This was reported in R-help by Dave Kane on June 4, 2001; I have not seen any further discussion. I'm not sure if I can report it as a "bug", since it is not entirely reproducible. Could someone try this and see if you have trouble too? The following

Environment variables set with Sys.putenv() disappear (i.e. become "") after a while, especially after heavy-duty I/O. Example: R> x <- matrix(1., 3000, 3000) R> save(x, file="myx.RData") R> Sys.putenv(HOME="/tmp") R> while (Sys.getenv("HOME") != "") {cat("ok\n"); load("myx.RData")} The loop prints

NAME Douglas Bates EMAIL bates@stat.wisc.edu VERSION 0.63.3 PLATFORM i386-unknown-linux SYSTEM Debian 2.1 CC/FC/MAKE egcs/g77/make NAME Martyn Plummer EMAIL plummer@iarc.fr VERSION 0.63.3 PLATFORM i386-unknown-linux SYSTEM Redhat 5.1 CC/FC/MAKE gcc/egcs-g77/make NAME Göran Broström EMAIL gb@stat.umu.se VERSION 0.63.3 PLATFORM

On Sun, 27 Aug 2000 rnassar@duke.edu wrote: > Dear R Team, > > # Summary of the problem: setting contrasts as > > contrasts(g) <- contr.treatment > or > contrasts(g) <- matrix(c(1,-1,0),ncol=1) > (i.e. without quotes around `contr.treatment' or `contr.sum', etc.) > and fitting an lm model without an intercept results in a model matrix > that lacks

This is not a bug in R: the documentation does say the result is OS-specific. `GMT' is a not a proper timezone on Windows, so NA is a valid answer. (Windows seems to use GMT to refer to the timezone of the UK, e.g. > Sys.time() [1] "2005-07-11 07:49:56 GMT Daylight Time" > Sys.timezone() [1] "GMT Daylight Time" although I am in British Summer Time not GMT.)

R 0.65.1 (and R-devel 09/11/99) > z <- data.frame(a=1:3) > b <- rep(NA, 3) > mode(b) [1] "logical" # how many of you expected that? I had forgotten! > zz <- cbind(z, b) > zz a b 1 1 NA 2 2 NA 3 3 NA > class(zz$b) [1] "factor" whereas in S it is NULL and zz$b is of mode "numeric". The same thing happens with data.frame zz <-

A change to diag() between 0.65.0 and 0.65.1 which seems undocumented in the logs (apparently revision 1.3.2.1, `diag names') introduced if (is.array(x)) stop("first argument is array, but not matrix.") which has broken the code in rpart. This looks wrong to me, as diag used to cope happily with a single-dimensional array. Such things occur quite often as the output

I'm building on cygwin. But the procedure for building the front-end on Cygwin is a bit daunting, to say the least,. Does anyone know a source for a pre-built version of the same?

Hello R developers, Platform: Windows 2000. Compiler: GNU GCC - 2.95.2 (CRTDLL) I have a C function for evaluating commands using eval(SEXP,SEXP) which is included in a personal make of R.dll. (Similiar to R_Proxy_evaluate() in rproxy_impl.c) Using SETJMP to catch errors works but there is a problem. For example, the following code: if (SETJMP(R_ToplevelContext->cjmpbuf)) //