I would like to see those packages that currently depend
on g95 dumped out into stable, if possible, so we can concentrate
on using gcc42 as the stock dependency in unstable. Gfortran
in gcc42 shouldn't be missing any major features compared to
g95. Also if there are any particular bugs in gcc 4.2 that
prevent it replacing g95, these should be identified and
that information passed back upstream to the gcc developers.
Currently, I see the following packages need to be ported
to use gcc42 instead of g95...
freefem++
hdf5
melina
modulef
narg
netcdf
cfitsio
pdftk
raster3d
atlas
fftw
fftw3
wip
biconv
r-base
scipy-py
octave
octave-forge
pgplot
scilib-atlas
I would like to point out that a number of the
middle-layer and optimizer gcc compiler developers
are now working on tweaking the gfortran front
end to better use GIMPLE, etc. Currently gfortran
in gcc42 already generates significantly faster
code than g95 (which doesn't, and is unlikely to
ever, directly use the middle-layer optimizers).
This speed advantage will only increase with
time so it is wise to transition over to gcc
for our fortran needs. Using gfortran also has
the advantage of having a coherent mechanism for
reporting and checking on bugs compared to g95
as well as proper testsuite to prevent regressions.
Jack
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier.
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Fink-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/fink-devel
