swarm-support
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: RPMs for R; question about HDf5 compiling, gcc egcs etc


From: Steve Emsley
Subject: Re: RPMs for R; question about HDf5 compiling, gcc egcs etc
Date: Tue, 9 Feb 1999 00:25:07 GMT

>>>>> "Marcus" == Marcus G Daniels <address@hidden> writes:

    Marcus> I don't know how Redhat is set up, but there is nothing
    Marcus> about GCC that prevents you from having gcc 2.7.2.3, gcc
    Marcus> 2.8.1, egcs-1.0.3, and egcs-1.1 all on the same system.

Perhaps the RPM dependencies are incorrect or irrelevant but when I
upgraded to Swarm1.2 on my Redhat5.0 box I upgraded gcc2.7.2.3 to
gcc2.8.1 because the Swarm binary build used it. Unfortunately I hit
the pow() bug and have had to use -fno-inline since. I looked into
egcs but the dependencies change re:

failed dependencies:
        binutils >= 2.8.1.0.15 is needed by egcs-1.0.3a-1

Redhat5.0 has binutils-2.8.1.0.1-1

There lies a quagmire that you enter at your peril since it's easier
to buy Redhat5.2! Anyway, I'll wait till I get a new machine (and
finish my thesis)!

However, during that period I had 3 gcc compilers installed and a
script that would link each one in as required. However, they were never
coexistent. When you suggest that's possible are you suggesting
renaming gcc to gcc2.7.2 and gcc2.8.1 etc. O.K. the gcc-lib libs will
be in subdirectories relevant to the release number but other things
(like the g++ include files for instance) aren't.

If you could advise me how it's possible to have gcc2.8.1 and gcc2.7.2
coexist I'd be appreciative. I have to use f2c since the g77 RPM will
not work with gcc2.8.1 and so I can't install octave. Hence I have to
drop into Windows to use MATLAB. Not that I have ANYTHING against
MATLAB of course ... it's contour graphs are the best in the
business. It's just the OS that gives me a sour taste!

Anyway, if there are any Redhat5.0 gcc2.8.1 (and probably 2.8.0) users
out there I installed Swarm1.4 O.K. The SantaFe R and NCSA HDF5
installed BUT only by setting compiler flags -fno-inline. This was
definitely the pow() bug at work. By the way, it might help to make a
note that HDF5 should be installed prior to R. The config script goes
by pretty quick ... I missed it.

Cheers.

-- 

  Steve Emsley             Avoid the Gates of Hell.  Use Linux
-----------------------------------------------------------------
  address@hidden                    www.oikos.ac.uk/~sme

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]