gcc 4.3 any known issues?

Hi,

Fedora is for Fedora 9 switching to gcc 4.3. Before I test it (rawhide) I want 
to be sure that R is running. So my question is whether there have been 
issues compiling R + packages using 4.3? 
  

I suspect that not many have tried.

There is an R-2.6.2 RPM in Fedora 9 alpha, so _something_ seems to work.

O__  ---- Peter Dalgaard             ?ster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 35327907

Hi,

Fedora is for Fedora 9 switching to gcc 4.3. Before I test it (rawhide) I want 
to be sure that R is running. So my question is whether there have been 
issues compiling R + packages using 4.3? 

Debian's "unstable" flavour already switched to 4.3 a while back. R,
and several CRAN packages, built just fine.  Mind you, those builds
used Debian's gcc build and not FC/RH but I'd say go ahead. 

Dirk

Three out of two people have difficulties with fractions.

Stefan Grosse wrote:

Hi,

Fedora is for Fedora 9 switching to gcc 4.3. Before I test it (rawhide) I want
to be sure that R is running. So my question is whether there have been
issues compiling R + packages using 4.3?

I suspect that not many have tried.

There is an R-2.6.2 RPM in Fedora 9 alpha, so _something_ seems to work.

But enough have to have sorted out many of the issues.  For example, the 
gcc 4.3 series uses C99-style inlining and that has been implemented 
(conditionally on GCC capabilities), and various bits of R have been 
rewritten to work around mis-compiles by the gcc 4.3 branch.

But note that gcc 4.3 is not released and not even branched with rather a 
lot of remaining regressions.  I'd be wary of using high levels of 
optimization: last time I looked R failed to build correctly at -O3 on 
x86_64, and there were more problems with packages.

Surely this is a topic that the posting guide directs to R-devel -- all 
the issues mentioned are programming ones, not R ones.

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