[Met_help] [rt.rap.ucar.edu #50702] History for Compilation problems, Thank you !

[John Halley Gotway via RT]

----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

Hi MET helper,

I'm from the Atmospheric Physics Group, the Uni of Auckland, New Zealand and
I'm trying to install MET on Fedora to verify my WRF results. However,
during the "make" process, something wrong happened and I really hope to get
assistance from your team.

It seems like that it should be the "netcdf" problem, but after NETCDF
re-compiled with PGI (same as compiler used in MET), problems were still
existed.

Could you give me some ideas ? Thank you very much !

Make log file attached.


Regards,


Sijin


----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: Re: [rt.rap.ucar.edu #50702] Compilation problems, Thank you !
From: John Halley Gotway
Time: Thu Oct 13 10:04:07 2011

Sijin,

Thank you for sending us your make log.  At the end of it, I see the
following error:
   /usr/bin/ld: cannot find -lm
   /usr/bin/ld: cannot find -lc
   make[3]: *** [pcp_combine] Error 2

Here the linker is telling you that it can't find the standard "m" and
"c" libraries - those stand for the math and standard c-compiler
library.  I suspect this is more of a problem with your
environment than it is with the MET code itself.  You have a couple of
options:

(1) Talk to a sys admin, tell them that when compiling with the PGI
C++ compiler, it can't find the math and C libraries.  They may be
able to reconfigure things to get it working.

(2) If that's not an option, you could try locating those libraries
for PGI on your system.  The files you're looking for are named
"libm.a" and "libc.a".  If you find them, you could modify the MET
Makefile to tell the linker where to find them.  Set the line:
   CXX_LIBS = -L/directory/containing/those/files
I'm not really hopeful for this approach.  If the linker can't find
those standard libraries, other problems will probably show up down
the road.

(3) Lastly, you could try building MET (and it's required libraries)
using the GNU compilers.  Examples of doing that can be found here:
   http://www.dtcenter.org/met/users/support/online_tutorial/METv3.0/compilation/index.php

Hope that helps.

John Halley Gotway


On 10/12/2011 10:17 PM, Zhang Sijin via RT wrote:
>
> Wed Oct 12 22:17:22 2011: Request 50702 was acted upon.
> Transaction: Ticket created by zsjzyhzp at gmail.com
>        Queue: met_help
>      Subject: Compilation problems, Thank you !
>        Owner: Nobody
>   Requestors: zsjzyhzp at gmail.com
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=50702 >
>
>
> Hi MET helper,
>
> I'm from the Atmospheric Physics Group, the Uni of Auckland, New
Zealand and
> I'm trying to install MET on Fedora to verify my WRF results.
However,
> during the "make" process, something wrong happened and I really
hope to get
> assistance from your team.
>
> It seems like that it should be the "netcdf" problem, but after
NETCDF
> re-compiled with PGI (same as compiler used in MET), problems were
still
> existed.
>
> Could you give me some ideas ? Thank you very much !
>
> Make log file attached.
>
>
> Regards,
>
>
> Sijin
>

------------------------------------------------