Re: [Rd] make fails with no rule to make target `VR.ts'

2008-11-27 Thread Hiroyuki Kawakatsu 
On Mon, Nov 24, 2008 at 3:01 PM, I wrote:
> When trying to build R-devel.tar.bz2 (SVN-REVISION: 47015) I am
> getting the error:
>
[...]
> gmake[2]: *** No rule to make target `VR.ts', needed by
> `stamp-recommended'.  Stop.
[...]
> This is on FreeBSD 7.0-RELEASE (amd64) with
>
> ./configure MAKE=gmake CC=gcc44 F77=gfortran44 FC=gfortran44 CXX=g++44
> OBJC=gcc44 R_BROWSER=/usr/local/bin/firefox3
>
> This used to build fine until a few weeks ago. Any hints for what may
> be going wrong?

I have finally tracked down the source of error. A diff between the
2008-11-18-2008 (builds ok) and 2008-11-19 (build fails as above) tar
balls shows that the *.tgz link files are missing in
R_HOME/src/library/Recommended/
in the latter tar ball. Once I manually created the links, gmake runs
fine. Is this an omission or something to do with the way I am
unpacking the tar balls? (I presume most people who build R-devel
simply check out from trunk?)

h.
-- 
+---
| Hiroyuki Kawakatsu
| Business School, Dublin City University
| Dublin 9, Ireland. Tel +353 (0)1 700 7496

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Rd] Problem with installed.packages (PR#13332)

2008-11-27 Thread maechler 
> "LH" == lhansen  <[EMAIL PROTECTED]>
> on Tue, 25 Nov 2008 23:30:23 +0100 (CET) writes:

LH> Full_Name: Lars Hansen
LH> Version: 2.8.0
LH> OS: Linux
LH> Submission from: (NULL) (70.90.201.133)


LH> Hi,

LH> Calling installed.packages() on a directory with only one package 
causes an
LH> error if the priority argument is set to "NA":

>> installed.packages(lib.loc = "/home/lhansen/R/libs", priority = "NA")
LH> Error in mat[, "Package"] : incorrect number of dimensions

LH> This seems like a classic dropping from matrix to vector bug. 

Yes, indeed, it is : forgetting to use  ', drop = FALSE'

LH> Maybe the code

LH> should be mat[, "Package", drop=FALSE].
no, it's here :

 mat <- mat[keep, , drop=FALSE]


This will be fixed in R-devel and R-patched within the next few
hours.

Thank you very much, Lars, for the bug report!
Martin Maechler, ETH Zurich

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Rd] missing parentheses in FAQ (PR#13343)

2008-11-27 Thread spluque 
Hi,

A couple of missing parenthesis in R-devel's FAQ:


---<---cut here---start-->---
Index: doc/manual/R-FAQ.texi
===
--- doc/manual/R-FAQ.texi   (revision 47025)
+++ doc/manual/R-FAQ.texi   (working copy)
@@ -7962,7 +7962,7 @@
 
 @example
 eval(substitute(lm(y ~ x + variable),
-list(variable = as.name(varname[1]))
+list(variable = as.name(varname[1]
 @end example
 
 @noindent
---<---cut here---end>---



Cheers,

-- 
Seb

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Rd] A small bug in R code of the legend function of the Graphics pacakge (PR#13340)

2008-11-27 Thread asherman 
Full_Name: Arkady  Sherman
Version: 2.8.0
OS: Windows XP
Submission from: (NULL) (158.195.16.114)


Please consider initialization of the  "seg.len" variable. It's done only
if (do.lines) {  seg.len <- 2 ..,

but after the seg.len is used:
if (has.pch) {
pch <- rep(pch, length.out = n.leg)
pt.bg <- rep(pt.bg, length.out = n.leg)
pt.cex <- rep(pt.cex, length.out = n.leg)
pt.lwd <- rep(pt.lwd, length.out = n.leg)
ok <- !is.na(pch) & (is.character(pch) | pch >= 0)
x1 <- (if (merge) 
xt - (seg.len/2) * xchar
so if do.lines isn't specified, but has.pch and merge are specified an error is
produced.
Guess that seg.len should be initialized before the condition "if (do.lines)"
Thanks in advance.

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel