[R] Shared Library Error

[Martin Maechler]

>>>>> "witn" == what's in the name? <born.to.b.wyld at gmail.com>
>>>>>     on Sat, 24 May 2008 02:47:02 -0500 writes:

    witn> R.version gives the following output
    >> R.version
    witn>                _ platform x86_64-unknown-linux-gnu
    witn> arch x86_64 os linux-gnu system x86_64, linux-gnu
    witn> status major 2 minor 6.0 year 2007 month 10 day 03 svn
    witn> rev 43063 language R version.string R version 2.6.0
    witn> (2007-10-03)


    witn> If you see my previous email, I did try .Fortran with
    witn> removing the last underscore, and I still got the same
    witn> error. I made the .so file using the f95 compiler. In
    witn> fact, 'nm *.so | grep xtabs', did show up the the name
    witn> "merge_xtabs_patterns_file_ ".  I am using f95 because
    witn> my source code is written in f90.

But in that case, I'd almost bet that

    .Fortran("merge_xtabs_patterns_file", ....)  

will work.  The idea being that you use the same name as in the
fortran code, namely with*OUT* trailing "_".

Regards,
Martin Maechler, ETH Zurich


    witn> Interestingly, doing the same works on another
    witn> machine, even though the f95 compiler on the two
    witn> machines is the same. R.version on that machine gives:
    >> R.version
    witn>                _ platform i686-pc-linux-gnu arch i686
    witn> os linux-gnu system i686, linux-gnu status major 2
    witn> minor 6.2 year 2008 month 02 day 08 svn rev 44383
    witn> language R version.string R version 2.6.2 (2008-02-08)



    witn> So, I am now wondering if this is a problem with the
    witn> older version of R which has now being fixed.

    witn> Please comment.

    witn> Thanks, A




    witn> On Sat, May 24, 2008 at 2:30 AM, Prof Brian Ripley
    witn> <ripley at stats.ox.ac.uk> wrote:

    >> The problem is that underscore is an illegal symbol in a
    >> Fortran 77 name (and R only requires a Fortran 77
    >> compiler).
    >> 
    >> You have not told us your OS nor R version nor
    >> compiler+version (because what a compiler does with this
    >> is a compiler-specific extension).
    >> 
    >> You should not be using the trailing underscore in
    >> .Fortran, so the failure was correct.
    >> 
    >> Some changes have been made fairly recently, and my
    >> belief is that this does work correctly in R-patched,
    >> even for old compilers such as gcc3's g77.  But it is not
    >> portable and not really supported.
    >> 
    >> So either try a current version of R-patched or write
    >> legal Fortran 77.
    >> 
    >> 
    >> 
    >> On Fri, 23 May 2008, what's in the name? wrote:
    >> 
    >> Can anyone help me resolve this? A part of the R function
    >> looks like this:
    >>> 
    >>> print(is.loaded('merge_xtabs_patterns_file'))
    >>> print(is.loaded('merge_xtabs_patterns_file_'))
    >>> 
    >>> 
    >>> .Fortran('merge_xtabs_patterns_file_',ydim[1],ydim[2],x=as.integer(as.matrix(y)),na=as.integer(c),
    >>> 
    >>> maxD=as.integer(maxD),lrowmem=length(rowmem),rowmem=as.integer(rowmem),
    >>> 
    >>> sequential=as.integer(Sequential),nclust=as.integer(nclust))
    >>> 
    >>> 
    >>> 
    >>> 
    >>> The corresponding output:
    >>> 
    >>> [1] FALSE [1] TRUE Error in
    >>> .Fortran("merge_xtabs_patterns_file_", ydim[1], ydim[2],
    >>> x = as.integer(as.matrix(y)), : Fortran symbol name
    >>> "merge_xtabs_patterns_file_" not in load table
    >>> 
    >>> 
    >>> 
    >>> 
    >>> 
    >>> Why is it that I get a TRUE for is.loaded with the same
    >>> symbol, but when using that under .Fortran I get the
    >>> above error?
    >>> 
    >>> 
    >>> I also tried replacing the third line with the following
    >>> (removing the last underscore from the symbol name):
    >>> 
    >>> 
    >>> .Fortran('merge_xtabs_patterns_file',ydim[1],ydim[2],x=as.integer(as.matrix(y)),na=as.integer(c),
    >>> 
    >>> maxD=as.integer(maxD),lrowmem=length(rowmem),rowmem=as.integer(rowmem),
    >>> 
    >>> sequential=as.integer(Sequential),nclust=as.integer(nclust))
    >>> 
    >>> 
    >>> The corresponding error:
    >>> 
    >>> Error in .Fortran("merge_xtabs_patterns_file", ydim[1],
    >>> ydim[2], x = as.integer(as.matrix(y)), : Fortran symbol
    >>> name "merge_xtabs_patterns_file" not in load table
    >>> 
    >>> 
    >>> 
    >>> Can anyone help?
    >>> 
    >>> Thanks, A
    >>> 
    >>> [[alternative HTML version deleted]]
    >>> 
    >>> ______________________________________________
    >>> R-help at r-project.org mailing list
    >>> https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do
    >>> read the posting guide
    >>> http://www.R-project.org/posting-guide.html and provide
    >>> commented, minimal, self-contained, reproducible code.
    >>> 
    >>> 
    >> --
> Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied
    >> Statistics,
    >> http://www.stats.ox.ac.uk/~ripley/<http://www.stats.ox.ac.uk/%7Eripley/>
    >> University of Oxford, Tel: +44 1865 272861 (self) 1 South
    >> Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax:
    >> +44 1865 272595
    >> 

	[[alternative HTML version deleted]]

    witn> ______________________________________________
    witn> R-help at r-project.org mailing list
    witn> https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do
    witn> read the posting guide
    witn> http://www.R-project.org/posting-guide.html and
    witn> provide commented, minimal, self-contained,
    witn> reproducible code.