Re: [Rd] Problem with configure's detection of glob on 2.6.0 (PR#10468)

From: <ripley_at_stats.ox.ac.uk>
Date: Mon, 26 Nov 2007 16:55:14 +0100 (CET)


I've no idea about the glob detection, but clearly this is the first system that allegedly lacks glob that we have encountered: the file needs to be

#if defined(HAVE_GLOB) || defined(Win32)

     glob_t globbuf;
#endif

Have fixed that for 2.6.1 patched.

On Mon, 26 Nov 2007, m.pacey_at_lancaster.ac.uk wrote:

> Full_Name: Mike Pacey
> Version: 2.6.0
> OS: SuSe Linux 9.3 x86_64
> Submission from: (NULL) (194.80.32.10)
>
>
>
> A "vanilla" version of R configures and compiles without a problem on my system.
> Switching to using the PGI compiler configures correctly (see the values at the
> bottom fo thismessage), but compilation aborts with:
>
> pgcc -I../../src/extra/zlib -I../../src/extra/bzip2 -I../../src/extra/pcre -I.
> -I../../src/include -I../../src/include -I/usr/local/include -DHAVE_CONFIG_H
> -Mvect=sse -Mscalarsse -Mcache_align -O2 -Kieee -c platform.c -o platform.o
> PGC-S-0040-Illegal use of symbol, glob_t (platform.c: 1141)
> PGC-W-0156-Type not specified, 'int' assumed (platform.c: 1141)
> PGC-W-0095-Type cast required for this conversion (platform.c: 1168)
> PGC/x86-64 Linux 7.0-2: compilation completed with severe errors
> make[3]: *** [platform.o] Error 2
>
>
> Running a diff on ./src/include/config.h files from both installs highlights a
> difference in the #define settings. For the vanilla install I get:
>
> /* Define to 1 if you have the `glob' function. */
> #define HAVE_GLOB 1
>
> /* Define to 1 if you have the <glob.h> header file. */
> #define HAVE_GLOB_H 1
>
> For the PGI version of the install I get:
>
> /* Define to 1 if you have the `glob' function. */
> /* #undef HAVE_GLOB */
>
> /* Define to 1 if you have the <glob.h> header file. */
> #define HAVE_GLOB_H 1
>
> I was a little puzzled as to why the latter would have the glob header file but
> not glob itself, so I modified the config.h file so that HAVE_GLOB was #defined
> to 1. The result compiles and passes the "make check" tests.
>
> Have I missed something or is configure/autoconf doing something odd with glob
> detection?
>
> -----
>
> Flags/variables for the PGI install are:
> export CC=pgcc
> export CFLAGS="-Mvect=sse -Mscalarsse -Mcache_align -O2 -Kieee"
> export F77=pgf77
> export FFLAGS="-Mvect=sse -Mscalarsse -Mcache_align -O2 -Kieee"
> export CXX=pgCC
> export CXXFLAGS="-Mvect=sse -Mscalarsse -Mcache_align -O2 -Kieee"
> export FC=pgf95
> export FCFLAGS="-Mvect=sse -Mscalarsse -Mcache_align -O2 -Kieee"
> export SHLIB_CXXLDFLAGS=-shared
> export SHLIB_LDFLAGS=-shared
> ./configure --prefix=/usr/local/packages/R-2.6.0-acml \
> --with-blas="-lacml" --x-includes=/usr/X11R6/include
> --x-libraries=/usr/X11R6/lib64
>
> ______________________________________________
> R-devel_at_r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Brian D. Ripley,                  ripley_at_stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

______________________________________________
R-devel_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Received on Mon 26 Nov 2007 - 16:01:12 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Mon 26 Nov 2007 - 16:30:33 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-devel. Please read the posting guide before posting to the list.