[Rd] Error report upon using the help.search( ) command (PR#8144)

[maechler@stat.math.ethz.ch]

>>>>> "paltr" == paltr  <paltr at bio.uio.no>
>>>>>     on Tue, 20 Sep 2005 11:24:18 +0200 (CEST) writes:

    paltr> Full_Name: Pål Trosvik Version: 2.0.1 
    paltr> OS: Windows XP Submission from: (NULL) (129.240.91.220)


    paltr> Upon entering the command: help.search("linear") 
    paltr> I get the following error report: Error in rbind(...) :
    paltr> number of columns of matrices must match (see arg 8)

    paltr> I get the same error message whichever search term I
    paltr> enter between the quotation marks. This function was
    paltr> working fine just a few weeks ago, and in the mean
    paltr> time I haven't done anything unusual. All I have done
    paltr> is download a few packages from CRAN, namely 'pls'
    paltr> and 'splines', as well as 'gplots','gtools' and
    paltr> 'gdata'. I have no idea what is causing this problem,
    paltr> but I guess it might be a bug.

it's a broken package in your library; i.e., one of the packages
you have added to your library is not properly installed.
Something which should not have happened if you installed
packages via the GUI or  install.packages() -- unless you have
interrupted the installation.
It also can happen very well if you use manual unzipping and
other non-supported ways of "putting packages in your library".

If you upgrade your version of R, the problem also goes away:
help.search() will produce warnings and continue searching
instead of producing an error when encountering incorrectly
installed packages.  So, please do upgrade.


    paltr> I shold perhaps mention that I saw an almost
    paltr> identical bug report entered in the 'not-repoducible'
    paltr> directory of previous bug reports by one Sabyasaschi
    paltr> Ray on 020805.

    paltr> His problem remains unsolved.

Well, I think this is about 3rd or 4th time the topic has come up.

Note that you are *NOT* supposed to send bug reports to R-bugs
when you use outdated versions of R.
It's fine to ask about it on R-help or R-devel if appropriate,
but sending a bug report for an outdated version is "no no!"

Regards,
Martin Maechler, ETH Zurich

    paltr> Best regards

    paltr> Pål Trosvik Ph.D. student University of Oslo -
    paltr> Department of Biology/Norwegian Food Research
    paltr> Institute