R devel - Aug 2004 - Re: [R] R on gentoo amd64 (gcc 3.3.3) is unstable --- no!

On Sat, 21 Aug 2004 ivo_welch-rstat8783@mailblocks.com wrote:
> 
> peter/brian:  thank you for the help.  i can now report that gentoo 
> amd64 can compile R just fine, too; it requires the f77 USE flag and a 
> gcc compiler rebuild first, though.  I also went to gcc 3.4.1.  my 
> segfault troubles earlier were caused by my use of f2c.
> 
> suggestion: would it be possible to add a code snippet  that tests 
> whether the user is trying to compile with f2c---and if so, bombs the 
> installation?  (if "f77 --version == f2c" then die)  I know the 
> problems are not R's fault, but it can percolate into R and it can 
> protect people like myself.
The thing is, it does work on some platforms -- I do test i686 RH8.0
(updated with gcc 3.4.1) from time to time and it looks like I tested
1.9.0 during beta test.  So `make check' is the only reliable test.
(I've just run the current R-devel to cross-check.  It runs into the 
infinite-loop in LAPACK problem with gcc >= 3.4.0 though, with the same 
fix.)

I suspect f2c on a 64-bit platform is a bad idea as int != long.
We could detect that.

-- 
Brian D. Ripley,                  ripley@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

Prof Brian Ripley wrote:
>On Sat, 21 Aug 2004 ivo_welch-rstat8783@mailblocks.com wrote:
>
>  
>
>>peter/brian:  thank you for the help.  i can now report that gentoo 
>>amd64 can compile R just fine, too; it requires the f77 USE flag and a 
>>gcc compiler rebuild first, though.  I also went to gcc 3.4.1.  my 
>>segfault troubles earlier were caused by my use of f2c.
>>
>>suggestion: would it be possible to add a code snippet  that tests 
>>whether the user is trying to compile with f2c---and if so, bombs the 
>>installation?  (if "f77 --version == f2c" then die)  I know
the
>>problems are not R's fault, but it can percolate into R and it can 
>>protect people like myself.
>>    
>>
>
>The thing is, it does work on some platforms -- I do test i686 RH8.0
>(updated with gcc 3.4.1) from time to time and it looks like I tested
>1.9.0 during beta test.  So `make check' is the only reliable test.
>(I've just run the current R-devel to cross-check.  It runs into the 
>infinite-loop in LAPACK problem with gcc >= 3.4.0 though, with the same 
>fix.)
>
>I suspect f2c on a 64-bit platform is a bad idea as int != long.
>We could detect that.
>
>  
>
gentlemen:  the folks on the gentoo mailing list did some further 
investigation into this f2c incompatibility with blas (boy, they are 
good!).  apparently, this is likely to be a gcc bug:
    http://bugs.gentoo.org/show_bug.cgi?id=61042
they also have a fix for the particular show-stopper, which occurs in 
function blas.c.  (a "const" addition to the parameters cures the
problem.)

just in case you are still interested...

regards,

/ivo