[Met_help] [rt.rap.ucar.edu #79613] History for Fwd: Software install error, perhaps you can help?
John Halley Gotway via RT
met_help at ucar.edu
Tue Feb 28 15:14:54 MST 2017
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
Hi,
I'm trying to reinstall the MET tools and, upon executing make install,
consistently encountering the error:
In file included from nc_var_info.cc(24):
nc_var_info.h(24): catastrophic error: cannot open source file "netcdf.hh"
#include <netcdf.hh>
^
(Note that the MET install guide suggests "downloading and building [my]
own instance of ... NETCDF". This is why, attm, I have netcdf-c-4.4.1 in
my working directory, in addition to the netcdf we have installed for
general use on my shared server, and have been intermittently overriding my
.bashrc to set the MET_NETCDF env var to this directory instead of the
latter. Regardless of which NETCDF I use, I encounter the same error.
A colleague suggested that this may be an error related to the version of
NETCDF: that MET is not compatible with NETCDF4? However, I see in the
installation guide that it IS compatible, as long as one were to configure
with the flag --disable-netcdf-4: http://www.dtcenter.org/met/users/
support/online_tutorial/METv5.2/tutorial.php?name=
compilation&category=req_libs
Thank you,
Ryan
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Wed Feb 22 13:35:58 2017
Hello Ryan,
I see you're having trouble compiling MET using NetCDF. The NetCDF
library
gets a bit confusing. The 3-series versions of the NetCDF library
typically had the Fortran, C, and C++ interfaces all bundled together.
At
some point, Unidata split out those interfaces into different
libraries for
Fortran, C, and C++.
MET utilizes the C++ interface. And the current release of MET
(version
5.2) uses the classic data model from NetCDF version 3. In the next
version of MET (version 6.0), we'll support reading/writing NetCDF4
files,
which utilize compression.
I'd suggest the following...
- First, check to see if you can find the "netcdf.hh" header file
anywhere
on your system. If so, you may already have the NetCDF C++ interface
available. Try pointing MET_NETCDF at the location you found. For
example, if you find /usr/local/netcdf/netcdf.hh and
/usr/local/netcdf/lib/libnetcdf* then set
MET_NETCDF=/usr/local/netcdf.
- If you don't find that, just compile NetCDF version 3 locally and
point
at that one:
ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-3.6.2.tar.gz
Hopefully after the switch to NetCDF4 in MET version 6.0, you'll just
be
able to use the system-level NetCDF build.
Please let me know if you continue to have issues.
Thanks,
John Halley Gotway
met_help at ucar.edu
On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> Transaction: Ticket created by rconne01 at gmail.com
> Queue: met_help
> Subject: Fwd: Software install error, perhaps you can help?
> Owner: Nobody
> Requestors: rconne01 at gmail.com
> Status: new
> Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
>
> Hi,
>
> I'm trying to reinstall the MET tools and, upon executing make
install,
> consistently encountering the error:
>
> In file included from nc_var_info.cc(24):
> nc_var_info.h(24): catastrophic error: cannot open source file
"netcdf.hh"
> #include <netcdf.hh>
> ^
>
> (Note that the MET install guide suggests "downloading and building
[my]
> own instance of ... NETCDF". This is why, attm, I have netcdf-c-
4.4.1 in
> my working directory, in addition to the netcdf we have installed
for
> general use on my shared server, and have been intermittently
overriding my
> .bashrc to set the MET_NETCDF env var to this directory instead of
the
> latter. Regardless of which NETCDF I use, I encounter the same
error.
>
> A colleague suggested that this may be an error related to the
version of
> NETCDF: that MET is not compatible with NETCDF4? However, I see in
the
> installation guide that it IS compatible, as long as one were to
configure
> with the flag --disable-netcdf-4: http://www.dtcenter.org/met/users/
> support/online_tutorial/METv5.2/tutorial.php?name=
> compilation&category=req_libs
>
> Thank you,
> Ryan
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Wed Feb 22 14:39:52 2017
Hi John,
I located netcdf.hh and set the MET_NETCDF env var as suggested, but
the
problem persists.
I have also tried building NetCDF-3.6.2 locally as suggested, but
encounter
a different error there:
ncvalues.cpp(238): error: identifier "strncpy" is undefined
strncpy(s, (const char*)the_values + n, (int)the_number);
Not sure what that means or what to do about it, or why the first
solution
didn't work.
Ryan
On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Hello Ryan,
>
> I see you're having trouble compiling MET using NetCDF. The NetCDF
library
> gets a bit confusing. The 3-series versions of the NetCDF library
> typically had the Fortran, C, and C++ interfaces all bundled
together. At
> some point, Unidata split out those interfaces into different
libraries for
> Fortran, C, and C++.
>
> MET utilizes the C++ interface. And the current release of MET
(version
> 5.2) uses the classic data model from NetCDF version 3. In the next
> version of MET (version 6.0), we'll support reading/writing NetCDF4
files,
> which utilize compression.
>
> I'd suggest the following...
>
> - First, check to see if you can find the "netcdf.hh" header file
anywhere
> on your system. If so, you may already have the NetCDF C++
interface
> available. Try pointing MET_NETCDF at the location you found. For
> example, if you find /usr/local/netcdf/netcdf.hh and
> /usr/local/netcdf/lib/libnetcdf* then set
MET_NETCDF=/usr/local/netcdf.
>
> - If you don't find that, just compile NetCDF version 3 locally and
point
> at that one:
> ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-3.6.2.tar.gz
>
> Hopefully after the switch to NetCDF4 in MET version 6.0, you'll
just be
> able to use the system-level NetCDF build.
>
> Please let me know if you continue to have issues.
>
> Thanks,
> John Halley Gotway
> met_help at ucar.edu
>
> On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> > Transaction: Ticket created by rconne01 at gmail.com
> > Queue: met_help
> > Subject: Fwd: Software install error, perhaps you can help?
> > Owner: Nobody
> > Requestors: rconne01 at gmail.com
> > Status: new
> > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> >
> > Hi,
> >
> > I'm trying to reinstall the MET tools and, upon executing make
install,
> > consistently encountering the error:
> >
> > In file included from nc_var_info.cc(24):
> > nc_var_info.h(24): catastrophic error: cannot open source file
> "netcdf.hh"
> > #include <netcdf.hh>
> > ^
> >
> > (Note that the MET install guide suggests "downloading and
building [my]
> > own instance of ... NETCDF". This is why, attm, I have netcdf-c-
4.4.1 in
> > my working directory, in addition to the netcdf we have installed
for
> > general use on my shared server, and have been intermittently
overriding
> my
> > .bashrc to set the MET_NETCDF env var to this directory instead of
the
> > latter. Regardless of which NETCDF I use, I encounter the same
error.
> >
> > A colleague suggested that this may be an error related to the
version of
> > NETCDF: that MET is not compatible with NETCDF4? However, I see
in the
> > installation guide that it IS compatible, as long as one were to
> configure
> > with the flag --disable-netcdf-4:
http://www.dtcenter.org/met/users/
> > support/online_tutorial/METv5.2/tutorial.php?name=
> > compilation&category=req_libs
> >
> > Thank you,
> > Ryan
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Wed Feb 22 16:26:57 2017
Ryan,
I just went through the same steps... and it turns out I'm getting the
same
error from 3.6.2:
ncvalues.cpp: In member function 'virtual char*
NcValues_ncbyte::as_string(long int) const':
ncvalues.cpp:238:60: error: 'strncpy' was not declared in this scope
strncpy(s, (const char*)the_values + n, (int)the_number);
I'm guessing this is a new issue that's popping up with newer versions
of
the GNU compilers. I'm using 4.9.2:
[johnhg at number5]% g++ --version
g++ (Debian 4.9.2-10) 4.9.2
Let's try version 4.1.1:
wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-4.1.1.tar.gz
tar -xvzf netcdf-4.1.1.tar.gz
cd netcdf-4.1.1
setenv CC `which gcc`
setenv CXX `which g++`
setenv FC ''
setenv F90 ''
./configure --prefix=`pwd` --disable-netcdf-4
make >& make.log
make install >& make_install.log
That worked on my system. Are you still getting compilation errors?
Thanks,
John
On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
> Hi John,
>
> I located netcdf.hh and set the MET_NETCDF env var as suggested, but
the
> problem persists.
>
> I have also tried building NetCDF-3.6.2 locally as suggested, but
encounter
> a different error there:
>
> ncvalues.cpp(238): error: identifier "strncpy" is undefined
> strncpy(s, (const char*)the_values + n, (int)the_number);
>
> Not sure what that means or what to do about it, or why the first
solution
> didn't work.
>
> Ryan
>
> On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Hello Ryan,
> >
> > I see you're having trouble compiling MET using NetCDF. The
NetCDF
> library
> > gets a bit confusing. The 3-series versions of the NetCDF library
> > typically had the Fortran, C, and C++ interfaces all bundled
together.
> At
> > some point, Unidata split out those interfaces into different
libraries
> for
> > Fortran, C, and C++.
> >
> > MET utilizes the C++ interface. And the current release of MET
(version
> > 5.2) uses the classic data model from NetCDF version 3. In the
next
> > version of MET (version 6.0), we'll support reading/writing
NetCDF4
> files,
> > which utilize compression.
> >
> > I'd suggest the following...
> >
> > - First, check to see if you can find the "netcdf.hh" header file
> anywhere
> > on your system. If so, you may already have the NetCDF C++
interface
> > available. Try pointing MET_NETCDF at the location you found.
For
> > example, if you find /usr/local/netcdf/netcdf.hh and
> > /usr/local/netcdf/lib/libnetcdf* then set
MET_NETCDF=/usr/local/netcdf.
> >
> > - If you don't find that, just compile NetCDF version 3 locally
and point
> > at that one:
> > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-3.6.2.tar.gz
> >
> > Hopefully after the switch to NetCDF4 in MET version 6.0, you'll
just be
> > able to use the system-level NetCDF build.
> >
> > Please let me know if you continue to have issues.
> >
> > Thanks,
> > John Halley Gotway
> > met_help at ucar.edu
> >
> > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT <
> met_help at ucar.edu>
> > wrote:
> >
> > >
> > > Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> > > Transaction: Ticket created by rconne01 at gmail.com
> > > Queue: met_help
> > > Subject: Fwd: Software install error, perhaps you can help?
> > > Owner: Nobody
> > > Requestors: rconne01 at gmail.com
> > > Status: new
> > > Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
> >
> > >
> > >
> > > Hi,
> > >
> > > I'm trying to reinstall the MET tools and, upon executing make
install,
> > > consistently encountering the error:
> > >
> > > In file included from nc_var_info.cc(24):
> > > nc_var_info.h(24): catastrophic error: cannot open source file
> > "netcdf.hh"
> > > #include <netcdf.hh>
> > > ^
> > >
> > > (Note that the MET install guide suggests "downloading and
building
> [my]
> > > own instance of ... NETCDF". This is why, attm, I have netcdf-
c-4.4.1
> in
> > > my working directory, in addition to the netcdf we have
installed for
> > > general use on my shared server, and have been intermittently
> overriding
> > my
> > > .bashrc to set the MET_NETCDF env var to this directory instead
of the
> > > latter. Regardless of which NETCDF I use, I encounter the same
error.
> > >
> > > A colleague suggested that this may be an error related to the
version
> of
> > > NETCDF: that MET is not compatible with NETCDF4? However, I see
in the
> > > installation guide that it IS compatible, as long as one were to
> > configure
> > > with the flag --disable-netcdf-4:
http://www.dtcenter.org/met/users/
> > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > compilation&category=req_libs
> > >
> > > Thank you,
> > > Ryan
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> >
> >
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Thu Feb 23 10:58:52 2017
Hi John,
This returns the error:
In file included from ncdump.c:16:0:
/cm/shared/apps/intel/composer_xe/2013_sp1.3.174/compiler/include/math.h:29:3:
error: #error "This Intel <math.h> is for use with only the Intel
compilers!"
# error "This Intel <math.h> is for use with only the Intel
compilers!"
^
I switched CC to point to our icc compiler, and CXX to our c++, and
after
running make clean first, encountered a new error:
make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
TEXINPUTS=".:$TEXINPUTS" \
MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run makeinfo -I .'
\
texi2dvi -s --pdf --batch netcdf.texi
/bin/sh: texi2dvi: command not found
make[2]: *** [netcdf.pdf] Error 127
Thanks,
Ryan
On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Ryan,
>
> I just went through the same steps... and it turns out I'm getting
the same
> error from 3.6.2:
>
> ncvalues.cpp: In member function 'virtual char*
> NcValues_ncbyte::as_string(long int) const':
> ncvalues.cpp:238:60: error: 'strncpy' was not declared in this scope
> strncpy(s, (const char*)the_values + n, (int)the_number);
>
> I'm guessing this is a new issue that's popping up with newer
versions of
> the GNU compilers. I'm using 4.9.2:
>
> [johnhg at number5]% g++ --version
> g++ (Debian 4.9.2-10) 4.9.2
>
> Let's try version 4.1.1:
> wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-4.1.1.tar.gz
> tar -xvzf netcdf-4.1.1.tar.gz
> cd netcdf-4.1.1
> setenv CC `which gcc`
> setenv CXX `which g++`
> setenv FC ''
> setenv F90 ''
> ./configure --prefix=`pwd` --disable-netcdf-4
> make >& make.log
> make install >& make_install.log
>
> That worked on my system. Are you still getting compilation errors?
>
> Thanks,
> John
>
>
>
> On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> > Hi John,
> >
> > I located netcdf.hh and set the MET_NETCDF env var as suggested,
but the
> > problem persists.
> >
> > I have also tried building NetCDF-3.6.2 locally as suggested, but
> encounter
> > a different error there:
> >
> > ncvalues.cpp(238): error: identifier "strncpy" is undefined
> > strncpy(s, (const char*)the_values + n, (int)the_number);
> >
> > Not sure what that means or what to do about it, or why the first
> solution
> > didn't work.
> >
> > Ryan
> >
> > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Hello Ryan,
> > >
> > > I see you're having trouble compiling MET using NetCDF. The
NetCDF
> > library
> > > gets a bit confusing. The 3-series versions of the NetCDF
library
> > > typically had the Fortran, C, and C++ interfaces all bundled
together.
> > At
> > > some point, Unidata split out those interfaces into different
libraries
> > for
> > > Fortran, C, and C++.
> > >
> > > MET utilizes the C++ interface. And the current release of MET
> (version
> > > 5.2) uses the classic data model from NetCDF version 3. In the
next
> > > version of MET (version 6.0), we'll support reading/writing
NetCDF4
> > files,
> > > which utilize compression.
> > >
> > > I'd suggest the following...
> > >
> > > - First, check to see if you can find the "netcdf.hh" header
file
> > anywhere
> > > on your system. If so, you may already have the NetCDF C++
interface
> > > available. Try pointing MET_NETCDF at the location you found.
For
> > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > /usr/local/netcdf/lib/libnetcdf* then set
> MET_NETCDF=/usr/local/netcdf.
> > >
> > > - If you don't find that, just compile NetCDF version 3 locally
and
> point
> > > at that one:
> > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-3.6.2.tar.gz
> > >
> > > Hopefully after the switch to NetCDF4 in MET version 6.0, you'll
just
> be
> > > able to use the system-level NetCDF build.
> > >
> > > Please let me know if you continue to have issues.
> > >
> > > Thanks,
> > > John Halley Gotway
> > > met_help at ucar.edu
> > >
> > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT <
> > met_help at ucar.edu>
> > > wrote:
> > >
> > > >
> > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > Queue: met_help
> > > > Subject: Fwd: Software install error, perhaps you can
help?
> > > > Owner: Nobody
> > > > Requestors: rconne01 at gmail.com
> > > > Status: new
> > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> Ticket/Display.html?id=79613
> > >
> > > >
> > > >
> > > > Hi,
> > > >
> > > > I'm trying to reinstall the MET tools and, upon executing make
> install,
> > > > consistently encountering the error:
> > > >
> > > > In file included from nc_var_info.cc(24):
> > > > nc_var_info.h(24): catastrophic error: cannot open source file
> > > "netcdf.hh"
> > > > #include <netcdf.hh>
> > > > ^
> > > >
> > > > (Note that the MET install guide suggests "downloading and
building
> > [my]
> > > > own instance of ... NETCDF". This is why, attm, I have
> netcdf-c-4.4.1
> > in
> > > > my working directory, in addition to the netcdf we have
installed for
> > > > general use on my shared server, and have been intermittently
> > overriding
> > > my
> > > > .bashrc to set the MET_NETCDF env var to this directory
instead of
> the
> > > > latter. Regardless of which NETCDF I use, I encounter the
same
> error.
> > > >
> > > > A colleague suggested that this may be an error related to the
> version
> > of
> > > > NETCDF: that MET is not compatible with NETCDF4? However, I
see in
> the
> > > > installation guide that it IS compatible, as long as one were
to
> > > configure
> > > > with the flag --disable-netcdf-4:
http://www.dtcenter.org/met/users/
> > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > compilation&category=req_libs
> > > >
> > > > Thank you,
> > > > Ryan
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Thu Feb 23 11:17:38 2017
Ryan,
Boy NetCDF is causing all sorts of problems! If you're switching to
the
Intel compilers, I believe you should use:
CC = icc
CXX = icc
F77 = ifort
I believe icc handles both C and C++ source code... but am not
positive on
that. I really don't know about "c++"... but generally you should use
1
family of compilers together and I doubt "c++" is an Intel compiler.
John
On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
> Hi John,
>
> This returns the error:
>
> In file included from ncdump.c:16:0:
> /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> compiler/include/math.h:29:3:
> error: #error "This Intel <math.h> is for use with only the Intel
> compilers!"
> # error "This Intel <math.h> is for use with only the Intel
compilers!"
>
> ^
>
> I switched CC to point to our icc compiler, and CXX to our c++, and
after
> running make clean first, encountered a new error:
>
> make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> TEXINPUTS=".:$TEXINPUTS" \
> MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run makeinfo -I
.' \
> texi2dvi -s --pdf --batch netcdf.texi
> /bin/sh: texi2dvi: command not found
> make[2]: *** [netcdf.pdf] Error 127
>
>
> Thanks,
> Ryan
>
>
>
> On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Ryan,
> >
> > I just went through the same steps... and it turns out I'm getting
the
> same
> > error from 3.6.2:
> >
> > ncvalues.cpp: In member function 'virtual char*
> > NcValues_ncbyte::as_string(long int) const':
> > ncvalues.cpp:238:60: error: 'strncpy' was not declared in this
scope
> > strncpy(s, (const char*)the_values + n, (int)the_number);
> >
> > I'm guessing this is a new issue that's popping up with newer
versions of
> > the GNU compilers. I'm using 4.9.2:
> >
> > [johnhg at number5]% g++ --version
> > g++ (Debian 4.9.2-10) 4.9.2
> >
> > Let's try version 4.1.1:
> > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-4.1.1.tar.gz
> > tar -xvzf netcdf-4.1.1.tar.gz
> > cd netcdf-4.1.1
> > setenv CC `which gcc`
> > setenv CXX `which g++`
> > setenv FC ''
> > setenv F90 ''
> > ./configure --prefix=`pwd` --disable-netcdf-4
> > make >& make.log
> > make install >& make_install.log
> >
> > That worked on my system. Are you still getting compilation
errors?
> >
> > Thanks,
> > John
> >
> >
> >
> > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT
<met_help at ucar.edu
> >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > >
> > > Hi John,
> > >
> > > I located netcdf.hh and set the MET_NETCDF env var as suggested,
but
> the
> > > problem persists.
> > >
> > > I have also tried building NetCDF-3.6.2 locally as suggested,
but
> > encounter
> > > a different error there:
> > >
> > > ncvalues.cpp(238): error: identifier "strncpy" is undefined
> > > strncpy(s, (const char*)the_values + n, (int)the_number);
> > >
> > > Not sure what that means or what to do about it, or why the
first
> > solution
> > > didn't work.
> > >
> > > Ryan
> > >
> > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Hello Ryan,
> > > >
> > > > I see you're having trouble compiling MET using NetCDF. The
NetCDF
> > > library
> > > > gets a bit confusing. The 3-series versions of the NetCDF
library
> > > > typically had the Fortran, C, and C++ interfaces all bundled
> together.
> > > At
> > > > some point, Unidata split out those interfaces into different
> libraries
> > > for
> > > > Fortran, C, and C++.
> > > >
> > > > MET utilizes the C++ interface. And the current release of
MET
> > (version
> > > > 5.2) uses the classic data model from NetCDF version 3. In
the next
> > > > version of MET (version 6.0), we'll support reading/writing
NetCDF4
> > > files,
> > > > which utilize compression.
> > > >
> > > > I'd suggest the following...
> > > >
> > > > - First, check to see if you can find the "netcdf.hh" header
file
> > > anywhere
> > > > on your system. If so, you may already have the NetCDF C++
interface
> > > > available. Try pointing MET_NETCDF at the location you found.
For
> > > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > > /usr/local/netcdf/lib/libnetcdf* then set
> > MET_NETCDF=/usr/local/netcdf.
> > > >
> > > > - If you don't find that, just compile NetCDF version 3
locally and
> > point
> > > > at that one:
> > > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
3.6.2.tar.gz
> > > >
> > > > Hopefully after the switch to NetCDF4 in MET version 6.0,
you'll just
> > be
> > > > able to use the system-level NetCDF build.
> > > >
> > > > Please let me know if you continue to have issues.
> > > >
> > > > Thanks,
> > > > John Halley Gotway
> > > > met_help at ucar.edu
> > > >
> > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT <
> > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > >
> > > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> > > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > > Queue: met_help
> > > > > Subject: Fwd: Software install error, perhaps you can
help?
> > > > > Owner: Nobody
> > > > > Requestors: rconne01 at gmail.com
> > > > > Status: new
> > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > Ticket/Display.html?id=79613
> > > >
> > > > >
> > > > >
> > > > > Hi,
> > > > >
> > > > > I'm trying to reinstall the MET tools and, upon executing
make
> > install,
> > > > > consistently encountering the error:
> > > > >
> > > > > In file included from nc_var_info.cc(24):
> > > > > nc_var_info.h(24): catastrophic error: cannot open source
file
> > > > "netcdf.hh"
> > > > > #include <netcdf.hh>
> > > > > ^
> > > > >
> > > > > (Note that the MET install guide suggests "downloading and
building
> > > [my]
> > > > > own instance of ... NETCDF". This is why, attm, I have
> > netcdf-c-4.4.1
> > > in
> > > > > my working directory, in addition to the netcdf we have
installed
> for
> > > > > general use on my shared server, and have been
intermittently
> > > overriding
> > > > my
> > > > > .bashrc to set the MET_NETCDF env var to this directory
instead of
> > the
> > > > > latter. Regardless of which NETCDF I use, I encounter the
same
> > error.
> > > > >
> > > > > A colleague suggested that this may be an error related to
the
> > version
> > > of
> > > > > NETCDF: that MET is not compatible with NETCDF4? However, I
see in
> > the
> > > > > installation guide that it IS compatible, as long as one
were to
> > > > configure
> > > > > with the flag --disable-netcdf-4:
http://www.dtcenter.org/met/
> users/
> > > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > compilation&category=req_libs
> > > > >
> > > > > Thank you,
> > > > > Ryan
> > > > >
> > > > > --
> > > > > Ryan Connelly
> > > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > > University
> > > > > rconne01 at gmail.com
> > > > > ryan.connelly at stonybrook.edu
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Ryan Connelly
> > > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > > University
> > > > > rconne01 at gmail.com
> > > > > ryan.connelly at stonybrook.edu
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> >
> >
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Thu Feb 23 14:06:49 2017
Ok, that makes sense about the compilers. You can see I'm still new
at
some of this stuff.
With setting the compilers as such, I was able to get a bit farther
than
before, but I still encounter the error about the inability to find
netcdf.hh when using v4.1.1. At the moment, I can't even compile
v3.6.x,
even though I did earlier this morning (then removed it for space
purposes
as I'm nearly full, not thinking I'd still need to reference it). One
of
those days!
On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Ryan,
>
> Boy NetCDF is causing all sorts of problems! If you're switching to
the
> Intel compilers, I believe you should use:
>
> CC = icc
> CXX = icc
> F77 = ifort
>
> I believe icc handles both C and C++ source code... but am not
positive on
> that. I really don't know about "c++"... but generally you should
use 1
> family of compilers together and I doubt "c++" is an Intel compiler.
>
> John
>
> On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> > Hi John,
> >
> > This returns the error:
> >
> > In file included from ncdump.c:16:0:
> > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > compiler/include/math.h:29:3:
> > error: #error "This Intel <math.h> is for use with only the Intel
> > compilers!"
> > # error "This Intel <math.h> is for use with only the Intel
compilers!"
> >
> > ^
> >
> > I switched CC to point to our icc compiler, and CXX to our c++,
and after
> > running make clean first, encountered a new error:
> >
> > make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> > TEXINPUTS=".:$TEXINPUTS" \
> > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run makeinfo
-I .' \
> > texi2dvi -s --pdf --batch netcdf.texi
> > /bin/sh: texi2dvi: command not found
> > make[2]: *** [netcdf.pdf] Error 127
> >
> >
> > Thanks,
> > Ryan
> >
> >
> >
> > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Ryan,
> > >
> > > I just went through the same steps... and it turns out I'm
getting the
> > same
> > > error from 3.6.2:
> > >
> > > ncvalues.cpp: In member function 'virtual char*
> > > NcValues_ncbyte::as_string(long int) const':
> > > ncvalues.cpp:238:60: error: 'strncpy' was not declared in this
scope
> > > strncpy(s, (const char*)the_values + n, (int)the_number);
> > >
> > > I'm guessing this is a new issue that's popping up with newer
versions
> of
> > > the GNU compilers. I'm using 4.9.2:
> > >
> > > [johnhg at number5]% g++ --version
> > > g++ (Debian 4.9.2-10) 4.9.2
> > >
> > > Let's try version 4.1.1:
> > > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
4.1.1.tar.gz
> > > tar -xvzf netcdf-4.1.1.tar.gz
> > > cd netcdf-4.1.1
> > > setenv CC `which gcc`
> > > setenv CXX `which g++`
> > > setenv FC ''
> > > setenv F90 ''
> > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > make >& make.log
> > > make install >& make_install.log
> > >
> > > That worked on my system. Are you still getting compilation
errors?
> > >
> > > Thanks,
> > > John
> > >
> > >
> > >
> > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT <
> met_help at ucar.edu
> > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
>
> > > >
> > > > Hi John,
> > > >
> > > > I located netcdf.hh and set the MET_NETCDF env var as
suggested, but
> > the
> > > > problem persists.
> > > >
> > > > I have also tried building NetCDF-3.6.2 locally as suggested,
but
> > > encounter
> > > > a different error there:
> > > >
> > > > ncvalues.cpp(238): error: identifier "strncpy" is undefined
> > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > >
> > > > Not sure what that means or what to do about it, or why the
first
> > > solution
> > > > didn't work.
> > > >
> > > > Ryan
> > > >
> > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Hello Ryan,
> > > > >
> > > > > I see you're having trouble compiling MET using NetCDF. The
NetCDF
> > > > library
> > > > > gets a bit confusing. The 3-series versions of the NetCDF
library
> > > > > typically had the Fortran, C, and C++ interfaces all bundled
> > together.
> > > > At
> > > > > some point, Unidata split out those interfaces into
different
> > libraries
> > > > for
> > > > > Fortran, C, and C++.
> > > > >
> > > > > MET utilizes the C++ interface. And the current release of
MET
> > > (version
> > > > > 5.2) uses the classic data model from NetCDF version 3. In
the
> next
> > > > > version of MET (version 6.0), we'll support reading/writing
NetCDF4
> > > > files,
> > > > > which utilize compression.
> > > > >
> > > > > I'd suggest the following...
> > > > >
> > > > > - First, check to see if you can find the "netcdf.hh" header
file
> > > > anywhere
> > > > > on your system. If so, you may already have the NetCDF C++
> interface
> > > > > available. Try pointing MET_NETCDF at the location you
found. For
> > > > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > MET_NETCDF=/usr/local/netcdf.
> > > > >
> > > > > - If you don't find that, just compile NetCDF version 3
locally and
> > > point
> > > > > at that one:
> > > > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
3.6.2.tar.gz
> > > > >
> > > > > Hopefully after the switch to NetCDF4 in MET version 6.0,
you'll
> just
> > > be
> > > > > able to use the system-level NetCDF build.
> > > > >
> > > > > Please let me know if you continue to have issues.
> > > > >
> > > > > Thanks,
> > > > > John Halley Gotway
> > > > > met_help at ucar.edu
> > > > >
> > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT <
> > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> > > > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > > > Queue: met_help
> > > > > > Subject: Fwd: Software install error, perhaps you can
help?
> > > > > > Owner: Nobody
> > > > > > Requestors: rconne01 at gmail.com
> > > > > > Status: new
> > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > Ticket/Display.html?id=79613
> > > > >
> > > > > >
> > > > > >
> > > > > > Hi,
> > > > > >
> > > > > > I'm trying to reinstall the MET tools and, upon executing
make
> > > install,
> > > > > > consistently encountering the error:
> > > > > >
> > > > > > In file included from nc_var_info.cc(24):
> > > > > > nc_var_info.h(24): catastrophic error: cannot open source
file
> > > > > "netcdf.hh"
> > > > > > #include <netcdf.hh>
> > > > > > ^
> > > > > >
> > > > > > (Note that the MET install guide suggests "downloading and
> building
> > > > [my]
> > > > > > own instance of ... NETCDF". This is why, attm, I have
> > > netcdf-c-4.4.1
> > > > in
> > > > > > my working directory, in addition to the netcdf we have
installed
> > for
> > > > > > general use on my shared server, and have been
intermittently
> > > > overriding
> > > > > my
> > > > > > .bashrc to set the MET_NETCDF env var to this directory
instead
> of
> > > the
> > > > > > latter. Regardless of which NETCDF I use, I encounter the
same
> > > error.
> > > > > >
> > > > > > A colleague suggested that this may be an error related to
the
> > > version
> > > > of
> > > > > > NETCDF: that MET is not compatible with NETCDF4? However,
I see
> in
> > > the
> > > > > > installation guide that it IS compatible, as long as one
were to
> > > > > configure
> > > > > > with the flag --disable-netcdf-4:
http://www.dtcenter.org/met/
> > users/
> > > > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > compilation&category=req_libs
> > > > > >
> > > > > > Thank you,
> > > > > > Ryan
> > > > > >
> > > > > > --
> > > > > > Ryan Connelly
> > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> Valparaiso
> > > > > > University
> > > > > > rconne01 at gmail.com
> > > > > > ryan.connelly at stonybrook.edu
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Ryan Connelly
> > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> Valparaiso
> > > > > > University
> > > > > > rconne01 at gmail.com
> > > > > > ryan.connelly at stonybrook.edu
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Thu Feb 23 15:16:38 2017
Ryan,
OK, perhaps you haven't updated the MET_NETCDF environment variable to
tell
MET where to find that library.
Try running:
cd met-5.2
./configure --help
That'll dump out a bunch of info, including the setting of environment
variables. Once you get NetCDF compiled, look for the NetCDF header
files,
netcdf.h and netcdf.hh. Also, look for the library files named
libnetcdf.a
and libnetcdf_c++.a. Those are the header files that MET includes and
the
library files to which it links.
If those files are in /path/to/netcdf/include and /path/to/netcdf/lib
(and
they probably are), then just define:
MET_NETCDF /path/to/netcdf
Thanks,
John
On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
> Ok, that makes sense about the compilers. You can see I'm still new
at
> some of this stuff.
>
> With setting the compilers as such, I was able to get a bit farther
than
> before, but I still encounter the error about the inability to find
> netcdf.hh when using v4.1.1. At the moment, I can't even compile
v3.6.x,
> even though I did earlier this morning (then removed it for space
purposes
> as I'm nearly full, not thinking I'd still need to reference it).
One of
> those days!
>
> On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Ryan,
> >
> > Boy NetCDF is causing all sorts of problems! If you're switching
to the
> > Intel compilers, I believe you should use:
> >
> > CC = icc
> > CXX = icc
> > F77 = ifort
> >
> > I believe icc handles both C and C++ source code... but am not
positive
> on
> > that. I really don't know about "c++"... but generally you should
use 1
> > family of compilers together and I doubt "c++" is an Intel
compiler.
> >
> > John
> >
> > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT <
> met_help at ucar.edu>
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > >
> > > Hi John,
> > >
> > > This returns the error:
> > >
> > > In file included from ncdump.c:16:0:
> > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > compiler/include/math.h:29:3:
> > > error: #error "This Intel <math.h> is for use with only the
Intel
> > > compilers!"
> > > # error "This Intel <math.h> is for use with only the Intel
> compilers!"
> > >
> > > ^
> > >
> > > I switched CC to point to our icc compiler, and CXX to our c++,
and
> after
> > > running make clean first, encountered a new error:
> > >
> > > make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> > > TEXINPUTS=".:$TEXINPUTS" \
> > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run makeinfo
-I .'
> \
> > > texi2dvi -s --pdf --batch netcdf.texi
> > > /bin/sh: texi2dvi: command not found
> > > make[2]: *** [netcdf.pdf] Error 127
> > >
> > >
> > > Thanks,
> > > Ryan
> > >
> > >
> > >
> > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Ryan,
> > > >
> > > > I just went through the same steps... and it turns out I'm
getting
> the
> > > same
> > > > error from 3.6.2:
> > > >
> > > > ncvalues.cpp: In member function 'virtual char*
> > > > NcValues_ncbyte::as_string(long int) const':
> > > > ncvalues.cpp:238:60: error: 'strncpy' was not declared in this
scope
> > > > strncpy(s, (const char*)the_values + n, (int)the_number);
> > > >
> > > > I'm guessing this is a new issue that's popping up with newer
> versions
> > of
> > > > the GNU compilers. I'm using 4.9.2:
> > > >
> > > > [johnhg at number5]% g++ --version
> > > > g++ (Debian 4.9.2-10) 4.9.2
> > > >
> > > > Let's try version 4.1.1:
> > > > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
4.1.1.tar.gz
> > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > cd netcdf-4.1.1
> > > > setenv CC `which gcc`
> > > > setenv CXX `which g++`
> > > > setenv FC ''
> > > > setenv F90 ''
> > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > make >& make.log
> > > > make install >& make_install.log
> > > >
> > > > That worked on my system. Are you still getting compilation
errors?
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > >
> > > >
> > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT <
> > met_help at ucar.edu
> > > >
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > >
> > > > > Hi John,
> > > > >
> > > > > I located netcdf.hh and set the MET_NETCDF env var as
suggested,
> but
> > > the
> > > > > problem persists.
> > > > >
> > > > > I have also tried building NetCDF-3.6.2 locally as
suggested, but
> > > > encounter
> > > > > a different error there:
> > > > >
> > > > > ncvalues.cpp(238): error: identifier "strncpy" is undefined
> > > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > > >
> > > > > Not sure what that means or what to do about it, or why the
first
> > > > solution
> > > > > didn't work.
> > > > >
> > > > > Ryan
> > > > >
> > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > > Hello Ryan,
> > > > > >
> > > > > > I see you're having trouble compiling MET using NetCDF.
The
> NetCDF
> > > > > library
> > > > > > gets a bit confusing. The 3-series versions of the NetCDF
> library
> > > > > > typically had the Fortran, C, and C++ interfaces all
bundled
> > > together.
> > > > > At
> > > > > > some point, Unidata split out those interfaces into
different
> > > libraries
> > > > > for
> > > > > > Fortran, C, and C++.
> > > > > >
> > > > > > MET utilizes the C++ interface. And the current release
of MET
> > > > (version
> > > > > > 5.2) uses the classic data model from NetCDF version 3.
In the
> > next
> > > > > > version of MET (version 6.0), we'll support
reading/writing
> NetCDF4
> > > > > files,
> > > > > > which utilize compression.
> > > > > >
> > > > > > I'd suggest the following...
> > > > > >
> > > > > > - First, check to see if you can find the "netcdf.hh"
header file
> > > > > anywhere
> > > > > > on your system. If so, you may already have the NetCDF
C++
> > interface
> > > > > > available. Try pointing MET_NETCDF at the location you
found.
> For
> > > > > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > MET_NETCDF=/usr/local/netcdf.
> > > > > >
> > > > > > - If you don't find that, just compile NetCDF version 3
locally
> and
> > > > point
> > > > > > at that one:
> > > > > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
3.6.2.tar.gz
> > > > > >
> > > > > > Hopefully after the switch to NetCDF4 in MET version 6.0,
you'll
> > just
> > > > be
> > > > > > able to use the system-level NetCDF build.
> > > > > >
> > > > > > Please let me know if you continue to have issues.
> > > > > >
> > > > > > Thanks,
> > > > > > John Halley Gotway
> > > > > > met_help at ucar.edu
> > > > > >
> > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT <
> > > > > met_help at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > >
> > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted upon.
> > > > > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > > > > Queue: met_help
> > > > > > > Subject: Fwd: Software install error, perhaps you
can
> help?
> > > > > > > Owner: Nobody
> > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > Status: new
> > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > Ticket/Display.html?id=79613
> > > > > >
> > > > > > >
> > > > > > >
> > > > > > > Hi,
> > > > > > >
> > > > > > > I'm trying to reinstall the MET tools and, upon
executing make
> > > > install,
> > > > > > > consistently encountering the error:
> > > > > > >
> > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > nc_var_info.h(24): catastrophic error: cannot open
source file
> > > > > > "netcdf.hh"
> > > > > > > #include <netcdf.hh>
> > > > > > > ^
> > > > > > >
> > > > > > > (Note that the MET install guide suggests "downloading
and
> > building
> > > > > [my]
> > > > > > > own instance of ... NETCDF". This is why, attm, I have
> > > > netcdf-c-4.4.1
> > > > > in
> > > > > > > my working directory, in addition to the netcdf we have
> installed
> > > for
> > > > > > > general use on my shared server, and have been
intermittently
> > > > > overriding
> > > > > > my
> > > > > > > .bashrc to set the MET_NETCDF env var to this directory
instead
> > of
> > > > the
> > > > > > > latter. Regardless of which NETCDF I use, I encounter
the same
> > > > error.
> > > > > > >
> > > > > > > A colleague suggested that this may be an error related
to the
> > > > version
> > > > > of
> > > > > > > NETCDF: that MET is not compatible with NETCDF4?
However, I
> see
> > in
> > > > the
> > > > > > > installation guide that it IS compatible, as long as one
were
> to
> > > > > > configure
> > > > > > > with the flag --disable-netcdf-4:
http://www.dtcenter.org/met/
> > > users/
> > > > > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > compilation&category=req_libs
> > > > > > >
> > > > > > > Thank you,
> > > > > > > Ryan
> > > > > > >
> > > > > > > --
> > > > > > > Ryan Connelly
> > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> > Valparaiso
> > > > > > > University
> > > > > > > rconne01 at gmail.com
> > > > > > > ryan.connelly at stonybrook.edu
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Ryan Connelly
> > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> > Valparaiso
> > > > > > > University
> > > > > > > rconne01 at gmail.com
> > > > > > > ryan.connelly at stonybrook.edu
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Ryan Connelly
> > > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > > University
> > > > > rconne01 at gmail.com
> > > > > ryan.connelly at stonybrook.edu
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> >
> >
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Fri Feb 24 14:48:53 2017
Hi John,
Even after verifying the existence of netcdf.hh in
$MET_NETCDF/include, and
libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and with the
$MET_NETCDF env var set properly, I still encounter the same error -
that
it can't open netcdf.hh. I'm going to talk to my system admin next;
at
this point, this can't be a MET problem.
Thank you for your help!
Ryan
On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Ryan,
>
> OK, perhaps you haven't updated the MET_NETCDF environment variable
to tell
> MET where to find that library.
>
> Try running:
> cd met-5.2
> ./configure --help
>
> That'll dump out a bunch of info, including the setting of
environment
> variables. Once you get NetCDF compiled, look for the NetCDF header
files,
> netcdf.h and netcdf.hh. Also, look for the library files named
libnetcdf.a
> and libnetcdf_c++.a. Those are the header files that MET includes
and the
> library files to which it links.
>
> If those files are in /path/to/netcdf/include and
/path/to/netcdf/lib (and
> they probably are), then just define:
> MET_NETCDF /path/to/netcdf
>
> Thanks,
> John
>
> On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> > Ok, that makes sense about the compilers. You can see I'm still
new at
> > some of this stuff.
> >
> > With setting the compilers as such, I was able to get a bit
farther than
> > before, but I still encounter the error about the inability to
find
> > netcdf.hh when using v4.1.1. At the moment, I can't even compile
v3.6.x,
> > even though I did earlier this morning (then removed it for space
> purposes
> > as I'm nearly full, not thinking I'd still need to reference it).
One of
> > those days!
> >
> > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Ryan,
> > >
> > > Boy NetCDF is causing all sorts of problems! If you're
switching to
> the
> > > Intel compilers, I believe you should use:
> > >
> > > CC = icc
> > > CXX = icc
> > > F77 = ifort
> > >
> > > I believe icc handles both C and C++ source code... but am not
positive
> > on
> > > that. I really don't know about "c++"... but generally you
should use
> 1
> > > family of compilers together and I doubt "c++" is an Intel
compiler.
> > >
> > > John
> > >
> > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT <
> > met_help at ucar.edu>
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
>
> > > >
> > > > Hi John,
> > > >
> > > > This returns the error:
> > > >
> > > > In file included from ncdump.c:16:0:
> > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > compiler/include/math.h:29:3:
> > > > error: #error "This Intel <math.h> is for use with only the
Intel
> > > > compilers!"
> > > > # error "This Intel <math.h> is for use with only the Intel
> > compilers!"
> > > >
> > > > ^
> > > >
> > > > I switched CC to point to our icc compiler, and CXX to our
c++, and
> > after
> > > > running make clean first, encountered a new error:
> > > >
> > > > make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run makeinfo
-I
> .'
> > \
> > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > /bin/sh: texi2dvi: command not found
> > > > make[2]: *** [netcdf.pdf] Error 127
> > > >
> > > >
> > > > Thanks,
> > > > Ryan
> > > >
> > > >
> > > >
> > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Ryan,
> > > > >
> > > > > I just went through the same steps... and it turns out I'm
getting
> > the
> > > > same
> > > > > error from 3.6.2:
> > > > >
> > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > ncvalues.cpp:238:60: error: 'strncpy' was not declared in
this
> scope
> > > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > > >
> > > > > I'm guessing this is a new issue that's popping up with
newer
> > versions
> > > of
> > > > > the GNU compilers. I'm using 4.9.2:
> > > > >
> > > > > [johnhg at number5]% g++ --version
> > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > >
> > > > > Let's try version 4.1.1:
> > > > > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
4.1.1.tar.gz
> > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > cd netcdf-4.1.1
> > > > > setenv CC `which gcc`
> > > > > setenv CXX `which g++`
> > > > > setenv FC ''
> > > > > setenv F90 ''
> > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > make >& make.log
> > > > > make install >& make_install.log
> > > > >
> > > > > That worked on my system. Are you still getting compilation
> errors?
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > >
> > > > >
> > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT <
> > > met_help at ucar.edu
> > > > >
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > > >
> > > > > > Hi John,
> > > > > >
> > > > > > I located netcdf.hh and set the MET_NETCDF env var as
suggested,
> > but
> > > > the
> > > > > > problem persists.
> > > > > >
> > > > > > I have also tried building NetCDF-3.6.2 locally as
suggested, but
> > > > > encounter
> > > > > > a different error there:
> > > > > >
> > > > > > ncvalues.cpp(238): error: identifier "strncpy" is
undefined
> > > > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > > > >
> > > > > > Not sure what that means or what to do about it, or why
the first
> > > > > solution
> > > > > > didn't work.
> > > > > >
> > > > > > Ryan
> > > > > >
> > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > > > Hello Ryan,
> > > > > > >
> > > > > > > I see you're having trouble compiling MET using NetCDF.
The
> > NetCDF
> > > > > > library
> > > > > > > gets a bit confusing. The 3-series versions of the
NetCDF
> > library
> > > > > > > typically had the Fortran, C, and C++ interfaces all
bundled
> > > > together.
> > > > > > At
> > > > > > > some point, Unidata split out those interfaces into
different
> > > > libraries
> > > > > > for
> > > > > > > Fortran, C, and C++.
> > > > > > >
> > > > > > > MET utilizes the C++ interface. And the current release
of MET
> > > > > (version
> > > > > > > 5.2) uses the classic data model from NetCDF version 3.
In the
> > > next
> > > > > > > version of MET (version 6.0), we'll support
reading/writing
> > NetCDF4
> > > > > > files,
> > > > > > > which utilize compression.
> > > > > > >
> > > > > > > I'd suggest the following...
> > > > > > >
> > > > > > > - First, check to see if you can find the "netcdf.hh"
header
> file
> > > > > > anywhere
> > > > > > > on your system. If so, you may already have the NetCDF
C++
> > > interface
> > > > > > > available. Try pointing MET_NETCDF at the location you
found.
> > For
> > > > > > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > >
> > > > > > > - If you don't find that, just compile NetCDF version 3
locally
> > and
> > > > > point
> > > > > > > at that one:
> > > > > > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
3.6.2.
> tar.gz
> > > > > > >
> > > > > > > Hopefully after the switch to NetCDF4 in MET version
6.0,
> you'll
> > > just
> > > > > be
> > > > > > > able to use the system-level NetCDF build.
> > > > > > >
> > > > > > > Please let me know if you continue to have issues.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > John Halley Gotway
> > > > > > > met_help at ucar.edu
> > > > > > >
> > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT <
> > > > > > met_help at ucar.edu>
> > > > > > > wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted
upon.
> > > > > > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > > > > > Queue: met_help
> > > > > > > > Subject: Fwd: Software install error, perhaps you
can
> > help?
> > > > > > > > Owner: Nobody
> > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > Status: new
> > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > Ticket/Display.html?id=79613
> > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > Hi,
> > > > > > > >
> > > > > > > > I'm trying to reinstall the MET tools and, upon
executing
> make
> > > > > install,
> > > > > > > > consistently encountering the error:
> > > > > > > >
> > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > nc_var_info.h(24): catastrophic error: cannot open
source
> file
> > > > > > > "netcdf.hh"
> > > > > > > > #include <netcdf.hh>
> > > > > > > > ^
> > > > > > > >
> > > > > > > > (Note that the MET install guide suggests "downloading
and
> > > building
> > > > > > [my]
> > > > > > > > own instance of ... NETCDF". This is why, attm, I
have
> > > > > netcdf-c-4.4.1
> > > > > > in
> > > > > > > > my working directory, in addition to the netcdf we
have
> > installed
> > > > for
> > > > > > > > general use on my shared server, and have been
intermittently
> > > > > > overriding
> > > > > > > my
> > > > > > > > .bashrc to set the MET_NETCDF env var to this
directory
> instead
> > > of
> > > > > the
> > > > > > > > latter. Regardless of which NETCDF I use, I encounter
the
> same
> > > > > error.
> > > > > > > >
> > > > > > > > A colleague suggested that this may be an error
related to
> the
> > > > > version
> > > > > > of
> > > > > > > > NETCDF: that MET is not compatible with NETCDF4?
However, I
> > see
> > > in
> > > > > the
> > > > > > > > installation guide that it IS compatible, as long as
one were
> > to
> > > > > > > configure
> > > > > > > > with the flag --disable-netcdf-4:
> http://www.dtcenter.org/met/
> > > > users/
> > > > > > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > > compilation&category=req_libs
> > > > > > > >
> > > > > > > > Thank you,
> > > > > > > > Ryan
> > > > > > > >
> > > > > > > > --
> > > > > > > > Ryan Connelly
> > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > Valparaiso
> > > > > > > > University
> > > > > > > > rconne01 at gmail.com
> > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > Ryan Connelly
> > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > Valparaiso
> > > > > > > > University
> > > > > > > > rconne01 at gmail.com
> > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Ryan Connelly
> > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> Valparaiso
> > > > > > University
> > > > > > rconne01 at gmail.com
> > > > > > ryan.connelly at stonybrook.edu
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Fri Feb 24 15:25:12 2017
Ryan,
OK, sorry for all this hassle. It sounds pretty frustrating. So
you've
found the NetCDF header/library files you need and you've set the
MET_NETCDF environment variable to tell MET where to find them.
But you get:
nc_var_info.h(24): catastrophic error: cannot open source file
"netcdf.hh"
I suppose it's possible that it's a permissions issue. Perhaps it's
trying
to read that file but you don't have permission to do so? So you
could
check the file permissions... but honestly that would be pretty odd.
I googled that error message and found the following met-help ticket
in
2011:
http://mailman.ucar.edu/pipermail/met_help/2011-
February/001522.html
That user was ultimately had success compiling NetCDF 3.6.3 using the
GNU
compilers.
John
On Fri, Feb 24, 2017 at 2:48 PM, Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
> Hi John,
>
> Even after verifying the existence of netcdf.hh in
$MET_NETCDF/include, and
> libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and with the
> $MET_NETCDF env var set properly, I still encounter the same error -
that
> it can't open netcdf.hh. I'm going to talk to my system admin next;
at
> this point, this can't be a MET problem.
>
> Thank you for your help!
> Ryan
>
> On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Ryan,
> >
> > OK, perhaps you haven't updated the MET_NETCDF environment
variable to
> tell
> > MET where to find that library.
> >
> > Try running:
> > cd met-5.2
> > ./configure --help
> >
> > That'll dump out a bunch of info, including the setting of
environment
> > variables. Once you get NetCDF compiled, look for the NetCDF
header
> files,
> > netcdf.h and netcdf.hh. Also, look for the library files named
> libnetcdf.a
> > and libnetcdf_c++.a. Those are the header files that MET includes
and
> the
> > library files to which it links.
> >
> > If those files are in /path/to/netcdf/include and
/path/to/netcdf/lib
> (and
> > they probably are), then just define:
> > MET_NETCDF /path/to/netcdf
> >
> > Thanks,
> > John
> >
> > On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT
<met_help at ucar.edu
> >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > >
> > > Ok, that makes sense about the compilers. You can see I'm still
new at
> > > some of this stuff.
> > >
> > > With setting the compilers as such, I was able to get a bit
farther
> than
> > > before, but I still encounter the error about the inability to
find
> > > netcdf.hh when using v4.1.1. At the moment, I can't even
compile
> v3.6.x,
> > > even though I did earlier this morning (then removed it for
space
> > purposes
> > > as I'm nearly full, not thinking I'd still need to reference
it). One
> of
> > > those days!
> > >
> > > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Ryan,
> > > >
> > > > Boy NetCDF is causing all sorts of problems! If you're
switching to
> > the
> > > > Intel compilers, I believe you should use:
> > > >
> > > > CC = icc
> > > > CXX = icc
> > > > F77 = ifort
> > > >
> > > > I believe icc handles both C and C++ source code... but am not
> positive
> > > on
> > > > that. I really don't know about "c++"... but generally you
should
> use
> > 1
> > > > family of compilers together and I doubt "c++" is an Intel
compiler.
> > > >
> > > > John
> > > >
> > > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT <
> > > met_help at ucar.edu>
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > >
> > > > > Hi John,
> > > > >
> > > > > This returns the error:
> > > > >
> > > > > In file included from ncdump.c:16:0:
> > > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > > compiler/include/math.h:29:3:
> > > > > error: #error "This Intel <math.h> is for use with only the
Intel
> > > > > compilers!"
> > > > > # error "This Intel <math.h> is for use with only the Intel
> > > compilers!"
> > > > >
> > > > > ^
> > > > >
> > > > > I switched CC to point to our icc compiler, and CXX to our
c++, and
> > > after
> > > > > running make clean first, encountered a new error:
> > > > >
> > > > > make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> > > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run
makeinfo -I
> > .'
> > > \
> > > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > > /bin/sh: texi2dvi: command not found
> > > > > make[2]: *** [netcdf.pdf] Error 127
> > > > >
> > > > >
> > > > > Thanks,
> > > > > Ryan
> > > > >
> > > > >
> > > > >
> > > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > > Ryan,
> > > > > >
> > > > > > I just went through the same steps... and it turns out I'm
> getting
> > > the
> > > > > same
> > > > > > error from 3.6.2:
> > > > > >
> > > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > > ncvalues.cpp:238:60: error: 'strncpy' was not declared in
this
> > scope
> > > > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > > > >
> > > > > > I'm guessing this is a new issue that's popping up with
newer
> > > versions
> > > > of
> > > > > > the GNU compilers. I'm using 4.9.2:
> > > > > >
> > > > > > [johnhg at number5]% g++ --version
> > > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > > >
> > > > > > Let's try version 4.1.1:
> > > > > > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
4.1.1.
> tar.gz
> > > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > > cd netcdf-4.1.1
> > > > > > setenv CC `which gcc`
> > > > > > setenv CXX `which g++`
> > > > > > setenv FC ''
> > > > > > setenv F90 ''
> > > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > > make >& make.log
> > > > > > make install >& make_install.log
> > > > > >
> > > > > > That worked on my system. Are you still getting
compilation
> > errors?
> > > > > >
> > > > > > Thanks,
> > > > > > John
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT <
> > > > met_help at ucar.edu
> > > > > >
> > > > > > wrote:
> > > > > >
> > > > > > >
> > > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
> >
> > > > > > >
> > > > > > > Hi John,
> > > > > > >
> > > > > > > I located netcdf.hh and set the MET_NETCDF env var as
> suggested,
> > > but
> > > > > the
> > > > > > > problem persists.
> > > > > > >
> > > > > > > I have also tried building NetCDF-3.6.2 locally as
suggested,
> but
> > > > > > encounter
> > > > > > > a different error there:
> > > > > > >
> > > > > > > ncvalues.cpp(238): error: identifier "strncpy" is
undefined
> > > > > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > > > > >
> > > > > > > Not sure what that means or what to do about it, or why
the
> first
> > > > > > solution
> > > > > > > didn't work.
> > > > > > >
> > > > > > > Ryan
> > > > > > >
> > > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway via
RT <
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > > > Hello Ryan,
> > > > > > > >
> > > > > > > > I see you're having trouble compiling MET using
NetCDF. The
> > > NetCDF
> > > > > > > library
> > > > > > > > gets a bit confusing. The 3-series versions of the
NetCDF
> > > library
> > > > > > > > typically had the Fortran, C, and C++ interfaces all
bundled
> > > > > together.
> > > > > > > At
> > > > > > > > some point, Unidata split out those interfaces into
different
> > > > > libraries
> > > > > > > for
> > > > > > > > Fortran, C, and C++.
> > > > > > > >
> > > > > > > > MET utilizes the C++ interface. And the current
release of
> MET
> > > > > > (version
> > > > > > > > 5.2) uses the classic data model from NetCDF version
3. In
> the
> > > > next
> > > > > > > > version of MET (version 6.0), we'll support
reading/writing
> > > NetCDF4
> > > > > > > files,
> > > > > > > > which utilize compression.
> > > > > > > >
> > > > > > > > I'd suggest the following...
> > > > > > > >
> > > > > > > > - First, check to see if you can find the "netcdf.hh"
header
> > file
> > > > > > > anywhere
> > > > > > > > on your system. If so, you may already have the
NetCDF C++
> > > > interface
> > > > > > > > available. Try pointing MET_NETCDF at the location
you
> found.
> > > For
> > > > > > > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > > >
> > > > > > > > - If you don't find that, just compile NetCDF version
3
> locally
> > > and
> > > > > > point
> > > > > > > > at that one:
> > > > > > > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
3.6.2.
> > tar.gz
> > > > > > > >
> > > > > > > > Hopefully after the switch to NetCDF4 in MET version
6.0,
> > you'll
> > > > just
> > > > > > be
> > > > > > > > able to use the system-level NetCDF build.
> > > > > > > >
> > > > > > > > Please let me know if you continue to have issues.
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > John Halley Gotway
> > > > > > > > met_help at ucar.edu
> > > > > > > >
> > > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via RT
<
> > > > > > > met_help at ucar.edu>
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > > >
> > > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted
upon.
> > > > > > > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > > > > > > Queue: met_help
> > > > > > > > > Subject: Fwd: Software install error, perhaps
you can
> > > help?
> > > > > > > > > Owner: Nobody
> > > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > > Status: new
> > > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > > Ticket/Display.html?id=79613
> > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Hi,
> > > > > > > > >
> > > > > > > > > I'm trying to reinstall the MET tools and, upon
executing
> > make
> > > > > > install,
> > > > > > > > > consistently encountering the error:
> > > > > > > > >
> > > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > > nc_var_info.h(24): catastrophic error: cannot open
source
> > file
> > > > > > > > "netcdf.hh"
> > > > > > > > > #include <netcdf.hh>
> > > > > > > > > ^
> > > > > > > > >
> > > > > > > > > (Note that the MET install guide suggests
"downloading and
> > > > building
> > > > > > > [my]
> > > > > > > > > own instance of ... NETCDF". This is why, attm, I
have
> > > > > > netcdf-c-4.4.1
> > > > > > > in
> > > > > > > > > my working directory, in addition to the netcdf we
have
> > > installed
> > > > > for
> > > > > > > > > general use on my shared server, and have been
> intermittently
> > > > > > > overriding
> > > > > > > > my
> > > > > > > > > .bashrc to set the MET_NETCDF env var to this
directory
> > instead
> > > > of
> > > > > > the
> > > > > > > > > latter. Regardless of which NETCDF I use, I
encounter the
> > same
> > > > > > error.
> > > > > > > > >
> > > > > > > > > A colleague suggested that this may be an error
related to
> > the
> > > > > > version
> > > > > > > of
> > > > > > > > > NETCDF: that MET is not compatible with NETCDF4?
However,
> I
> > > see
> > > > in
> > > > > > the
> > > > > > > > > installation guide that it IS compatible, as long as
one
> were
> > > to
> > > > > > > > configure
> > > > > > > > > with the flag --disable-netcdf-4:
> > http://www.dtcenter.org/met/
> > > > > users/
> > > > > > > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > > > compilation&category=req_libs
> > > > > > > > >
> > > > > > > > > Thank you,
> > > > > > > > > Ryan
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Ryan Connelly
> > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> University
> > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > Valparaiso
> > > > > > > > > University
> > > > > > > > > rconne01 at gmail.com
> > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Ryan Connelly
> > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> University
> > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > Valparaiso
> > > > > > > > > University
> > > > > > > > > rconne01 at gmail.com
> > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Ryan Connelly
> > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> > Valparaiso
> > > > > > > University
> > > > > > > rconne01 at gmail.com
> > > > > > > ryan.connelly at stonybrook.edu
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Ryan Connelly
> > > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > > University
> > > > > rconne01 at gmail.com
> > > > > ryan.connelly at stonybrook.edu
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> >
> >
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Mon Feb 27 15:45:15 2017
Hi John,
Do these netCDF files require execute permissions? They have r & w
already.
Thanks,
Ryan
On Fri, Feb 24, 2017 at 5:25 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Ryan,
>
> OK, sorry for all this hassle. It sounds pretty frustrating. So
you've
> found the NetCDF header/library files you need and you've set the
> MET_NETCDF environment variable to tell MET where to find them.
>
> But you get:
> nc_var_info.h(24): catastrophic error: cannot open source file
"netcdf.hh"
>
> I suppose it's possible that it's a permissions issue. Perhaps it's
trying
> to read that file but you don't have permission to do so? So you
could
> check the file permissions... but honestly that would be pretty odd.
>
> I googled that error message and found the following met-help ticket
in
> 2011:
> http://mailman.ucar.edu/pipermail/met_help/2011-
February/001522.html
>
> That user was ultimately had success compiling NetCDF 3.6.3 using
the GNU
> compilers.
>
> John
>
> On Fri, Feb 24, 2017 at 2:48 PM, Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> > Hi John,
> >
> > Even after verifying the existence of netcdf.hh in
$MET_NETCDF/include,
> and
> > libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and with the
> > $MET_NETCDF env var set properly, I still encounter the same error
- that
> > it can't open netcdf.hh. I'm going to talk to my system admin
next; at
> > this point, this can't be a MET problem.
> >
> > Thank you for your help!
> > Ryan
> >
> > On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Ryan,
> > >
> > > OK, perhaps you haven't updated the MET_NETCDF environment
variable to
> > tell
> > > MET where to find that library.
> > >
> > > Try running:
> > > cd met-5.2
> > > ./configure --help
> > >
> > > That'll dump out a bunch of info, including the setting of
environment
> > > variables. Once you get NetCDF compiled, look for the NetCDF
header
> > files,
> > > netcdf.h and netcdf.hh. Also, look for the library files named
> > libnetcdf.a
> > > and libnetcdf_c++.a. Those are the header files that MET
includes and
> > the
> > > library files to which it links.
> > >
> > > If those files are in /path/to/netcdf/include and
/path/to/netcdf/lib
> > (and
> > > they probably are), then just define:
> > > MET_NETCDF /path/to/netcdf
> > >
> > > Thanks,
> > > John
> > >
> > > On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT <
> met_help at ucar.edu
> > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
>
> > > >
> > > > Ok, that makes sense about the compilers. You can see I'm
still new
> at
> > > > some of this stuff.
> > > >
> > > > With setting the compilers as such, I was able to get a bit
farther
> > than
> > > > before, but I still encounter the error about the inability to
find
> > > > netcdf.hh when using v4.1.1. At the moment, I can't even
compile
> > v3.6.x,
> > > > even though I did earlier this morning (then removed it for
space
> > > purposes
> > > > as I'm nearly full, not thinking I'd still need to reference
it).
> One
> > of
> > > > those days!
> > > >
> > > > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Ryan,
> > > > >
> > > > > Boy NetCDF is causing all sorts of problems! If you're
switching
> to
> > > the
> > > > > Intel compilers, I believe you should use:
> > > > >
> > > > > CC = icc
> > > > > CXX = icc
> > > > > F77 = ifort
> > > > >
> > > > > I believe icc handles both C and C++ source code... but am
not
> > positive
> > > > on
> > > > > that. I really don't know about "c++"... but generally you
should
> > use
> > > 1
> > > > > family of compilers together and I doubt "c++" is an Intel
> compiler.
> > > > >
> > > > > John
> > > > >
> > > > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT <
> > > > met_help at ucar.edu>
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > > >
> > > > > > Hi John,
> > > > > >
> > > > > > This returns the error:
> > > > > >
> > > > > > In file included from ncdump.c:16:0:
> > > > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > > > compiler/include/math.h:29:3:
> > > > > > error: #error "This Intel <math.h> is for use with only
the Intel
> > > > > > compilers!"
> > > > > > # error "This Intel <math.h> is for use with only the
Intel
> > > > compilers!"
> > > > > >
> > > > > > ^
> > > > > >
> > > > > > I switched CC to point to our icc compiler, and CXX to our
c++,
> and
> > > > after
> > > > > > running make clean first, encountered a new error:
> > > > > >
> > > > > > make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> > > > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run
makeinfo
> -I
> > > .'
> > > > \
> > > > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > > > /bin/sh: texi2dvi: command not found
> > > > > > make[2]: *** [netcdf.pdf] Error 127
> > > > > >
> > > > > >
> > > > > > Thanks,
> > > > > > Ryan
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > > > Ryan,
> > > > > > >
> > > > > > > I just went through the same steps... and it turns out
I'm
> > getting
> > > > the
> > > > > > same
> > > > > > > error from 3.6.2:
> > > > > > >
> > > > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > > > ncvalues.cpp:238:60: error: 'strncpy' was not declared
in this
> > > scope
> > > > > > > strncpy(s, (const char*)the_values + n,
(int)the_number);
> > > > > > >
> > > > > > > I'm guessing this is a new issue that's popping up with
newer
> > > > versions
> > > > > of
> > > > > > > the GNU compilers. I'm using 4.9.2:
> > > > > > >
> > > > > > > [johnhg at number5]% g++ --version
> > > > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > > > >
> > > > > > > Let's try version 4.1.1:
> > > > > > > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
4.1.1.
> > tar.gz
> > > > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > > > cd netcdf-4.1.1
> > > > > > > setenv CC `which gcc`
> > > > > > > setenv CXX `which g++`
> > > > > > > setenv FC ''
> > > > > > > setenv F90 ''
> > > > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > > > make >& make.log
> > > > > > > make install >& make_install.log
> > > > > > >
> > > > > > > That worked on my system. Are you still getting
compilation
> > > errors?
> > > > > > >
> > > > > > > Thanks,
> > > > > > > John
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT <
> > > > > met_help at ucar.edu
> > > > > > >
> > > > > > > wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> Ticket/Display.html?id=79613
> > >
> > > > > > > >
> > > > > > > > Hi John,
> > > > > > > >
> > > > > > > > I located netcdf.hh and set the MET_NETCDF env var as
> > suggested,
> > > > but
> > > > > > the
> > > > > > > > problem persists.
> > > > > > > >
> > > > > > > > I have also tried building NetCDF-3.6.2 locally as
suggested,
> > but
> > > > > > > encounter
> > > > > > > > a different error there:
> > > > > > > >
> > > > > > > > ncvalues.cpp(238): error: identifier "strncpy" is
undefined
> > > > > > > > strncpy(s, (const char*)the_values + n,
> (int)the_number);
> > > > > > > >
> > > > > > > > Not sure what that means or what to do about it, or
why the
> > first
> > > > > > > solution
> > > > > > > > didn't work.
> > > > > > > >
> > > > > > > > Ryan
> > > > > > > >
> > > > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway
via RT <
> > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > >
> > > > > > > > > Hello Ryan,
> > > > > > > > >
> > > > > > > > > I see you're having trouble compiling MET using
NetCDF.
> The
> > > > NetCDF
> > > > > > > > library
> > > > > > > > > gets a bit confusing. The 3-series versions of the
NetCDF
> > > > library
> > > > > > > > > typically had the Fortran, C, and C++ interfaces all
> bundled
> > > > > > together.
> > > > > > > > At
> > > > > > > > > some point, Unidata split out those interfaces into
> different
> > > > > > libraries
> > > > > > > > for
> > > > > > > > > Fortran, C, and C++.
> > > > > > > > >
> > > > > > > > > MET utilizes the C++ interface. And the current
release of
> > MET
> > > > > > > (version
> > > > > > > > > 5.2) uses the classic data model from NetCDF version
3. In
> > the
> > > > > next
> > > > > > > > > version of MET (version 6.0), we'll support
reading/writing
> > > > NetCDF4
> > > > > > > > files,
> > > > > > > > > which utilize compression.
> > > > > > > > >
> > > > > > > > > I'd suggest the following...
> > > > > > > > >
> > > > > > > > > - First, check to see if you can find the
"netcdf.hh"
> header
> > > file
> > > > > > > > anywhere
> > > > > > > > > on your system. If so, you may already have the
NetCDF C++
> > > > > interface
> > > > > > > > > available. Try pointing MET_NETCDF at the location
you
> > found.
> > > > For
> > > > > > > > > example, if you find /usr/local/netcdf/netcdf.hh and
> > > > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > > > >
> > > > > > > > > - If you don't find that, just compile NetCDF
version 3
> > locally
> > > > and
> > > > > > > point
> > > > > > > > > at that one:
> > > > > > > > > ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
3.6.2.
> > > tar.gz
> > > > > > > > >
> > > > > > > > > Hopefully after the switch to NetCDF4 in MET version
6.0,
> > > you'll
> > > > > just
> > > > > > > be
> > > > > > > > > able to use the system-level NetCDF build.
> > > > > > > > >
> > > > > > > > > Please let me know if you continue to have issues.
> > > > > > > > >
> > > > > > > > > Thanks,
> > > > > > > > > John Halley Gotway
> > > > > > > > > met_help at ucar.edu
> > > > > > > > >
> > > > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly via
RT <
> > > > > > > > met_help at ucar.edu>
> > > > > > > > > wrote:
> > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was acted
upon.
> > > > > > > > > > Transaction: Ticket created by rconne01 at gmail.com
> > > > > > > > > > Queue: met_help
> > > > > > > > > > Subject: Fwd: Software install error, perhaps
you
> can
> > > > help?
> > > > > > > > > > Owner: Nobody
> > > > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > > > Status: new
> > > > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > > > Ticket/Display.html?id=79613
> > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > Hi,
> > > > > > > > > >
> > > > > > > > > > I'm trying to reinstall the MET tools and, upon
executing
> > > make
> > > > > > > install,
> > > > > > > > > > consistently encountering the error:
> > > > > > > > > >
> > > > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > > > nc_var_info.h(24): catastrophic error: cannot open
source
> > > file
> > > > > > > > > "netcdf.hh"
> > > > > > > > > > #include <netcdf.hh>
> > > > > > > > > > ^
> > > > > > > > > >
> > > > > > > > > > (Note that the MET install guide suggests
"downloading
> and
> > > > > building
> > > > > > > > [my]
> > > > > > > > > > own instance of ... NETCDF". This is why, attm, I
have
> > > > > > > netcdf-c-4.4.1
> > > > > > > > in
> > > > > > > > > > my working directory, in addition to the netcdf we
have
> > > > installed
> > > > > > for
> > > > > > > > > > general use on my shared server, and have been
> > intermittently
> > > > > > > > overriding
> > > > > > > > > my
> > > > > > > > > > .bashrc to set the MET_NETCDF env var to this
directory
> > > instead
> > > > > of
> > > > > > > the
> > > > > > > > > > latter. Regardless of which NETCDF I use, I
encounter
> the
> > > same
> > > > > > > error.
> > > > > > > > > >
> > > > > > > > > > A colleague suggested that this may be an error
related
> to
> > > the
> > > > > > > version
> > > > > > > > of
> > > > > > > > > > NETCDF: that MET is not compatible with NETCDF4?
> However,
> > I
> > > > see
> > > > > in
> > > > > > > the
> > > > > > > > > > installation guide that it IS compatible, as long
as one
> > were
> > > > to
> > > > > > > > > configure
> > > > > > > > > > with the flag --disable-netcdf-4:
> > > http://www.dtcenter.org/met/
> > > > > > users/
> > > > > > > > > > support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > > > > compilation&category=req_libs
> > > > > > > > > >
> > > > > > > > > > Thank you,
> > > > > > > > > > Ryan
> > > > > > > > > >
> > > > > > > > > > --
> > > > > > > > > > Ryan Connelly
> > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> > University
> > > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > > Valparaiso
> > > > > > > > > > University
> > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > --
> > > > > > > > > > Ryan Connelly
> > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> > University
> > > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > > Valparaiso
> > > > > > > > > > University
> > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > Ryan Connelly
> > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > Valparaiso
> > > > > > > > University
> > > > > > > > rconne01 at gmail.com
> > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Ryan Connelly
> > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> Valparaiso
> > > > > > University
> > > > > > rconne01 at gmail.com
> > > > > > ryan.connelly at stonybrook.edu
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Mon Feb 27 16:02:06 2017
Nope, just read permission on my machine:
-rw-r--r-- 1 root root 23 Feb 27 2016
/usr/local/netcdf/include/netcdf.hh
So permissions probably are not the problem.
John
On Mon, Feb 27, 2017 at 3:45 PM, Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
> Hi John,
>
> Do these netCDF files require execute permissions? They have r & w
> already.
>
> Thanks,
> Ryan
>
> On Fri, Feb 24, 2017 at 5:25 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Ryan,
> >
> > OK, sorry for all this hassle. It sounds pretty frustrating. So
you've
> > found the NetCDF header/library files you need and you've set the
> > MET_NETCDF environment variable to tell MET where to find them.
> >
> > But you get:
> > nc_var_info.h(24): catastrophic error: cannot open source file
> "netcdf.hh"
> >
> > I suppose it's possible that it's a permissions issue. Perhaps
it's
> trying
> > to read that file but you don't have permission to do so? So you
could
> > check the file permissions... but honestly that would be pretty
odd.
> >
> > I googled that error message and found the following met-help
ticket in
> > 2011:
> > http://mailman.ucar.edu/pipermail/met_help/2011-
February/001522.html
> >
> > That user was ultimately had success compiling NetCDF 3.6.3 using
the GNU
> > compilers.
> >
> > John
> >
> > On Fri, Feb 24, 2017 at 2:48 PM, Ryan Connelly via RT
<met_help at ucar.edu
> >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > >
> > > Hi John,
> > >
> > > Even after verifying the existence of netcdf.hh in
$MET_NETCDF/include,
> > and
> > > libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and with the
> > > $MET_NETCDF env var set properly, I still encounter the same
error -
> that
> > > it can't open netcdf.hh. I'm going to talk to my system admin
next; at
> > > this point, this can't be a MET problem.
> > >
> > > Thank you for your help!
> > > Ryan
> > >
> > > On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Ryan,
> > > >
> > > > OK, perhaps you haven't updated the MET_NETCDF environment
variable
> to
> > > tell
> > > > MET where to find that library.
> > > >
> > > > Try running:
> > > > cd met-5.2
> > > > ./configure --help
> > > >
> > > > That'll dump out a bunch of info, including the setting of
> environment
> > > > variables. Once you get NetCDF compiled, look for the NetCDF
header
> > > files,
> > > > netcdf.h and netcdf.hh. Also, look for the library files
named
> > > libnetcdf.a
> > > > and libnetcdf_c++.a. Those are the header files that MET
includes
> and
> > > the
> > > > library files to which it links.
> > > >
> > > > If those files are in /path/to/netcdf/include and
/path/to/netcdf/lib
> > > (and
> > > > they probably are), then just define:
> > > > MET_NETCDF /path/to/netcdf
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT <
> > met_help at ucar.edu
> > > >
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > >
> > > > > Ok, that makes sense about the compilers. You can see I'm
still
> new
> > at
> > > > > some of this stuff.
> > > > >
> > > > > With setting the compilers as such, I was able to get a bit
farther
> > > than
> > > > > before, but I still encounter the error about the inability
to find
> > > > > netcdf.hh when using v4.1.1. At the moment, I can't even
compile
> > > v3.6.x,
> > > > > even though I did earlier this morning (then removed it for
space
> > > > purposes
> > > > > as I'm nearly full, not thinking I'd still need to reference
it).
> > One
> > > of
> > > > > those days!
> > > > >
> > > > > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > > Ryan,
> > > > > >
> > > > > > Boy NetCDF is causing all sorts of problems! If you're
switching
> > to
> > > > the
> > > > > > Intel compilers, I believe you should use:
> > > > > >
> > > > > > CC = icc
> > > > > > CXX = icc
> > > > > > F77 = ifort
> > > > > >
> > > > > > I believe icc handles both C and C++ source code... but am
not
> > > positive
> > > > > on
> > > > > > that. I really don't know about "c++"... but generally
you
> should
> > > use
> > > > 1
> > > > > > family of compilers together and I doubt "c++" is an Intel
> > compiler.
> > > > > >
> > > > > > John
> > > > > >
> > > > > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT <
> > > > > met_help at ucar.edu>
> > > > > > wrote:
> > > > > >
> > > > > > >
> > > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
> >
> > > > > > >
> > > > > > > Hi John,
> > > > > > >
> > > > > > > This returns the error:
> > > > > > >
> > > > > > > In file included from ncdump.c:16:0:
> > > > > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > > > > compiler/include/math.h:29:3:
> > > > > > > error: #error "This Intel <math.h> is for use with only
the
> Intel
> > > > > > > compilers!"
> > > > > > > # error "This Intel <math.h> is for use with only the
Intel
> > > > > compilers!"
> > > > > > >
> > > > > > > ^
> > > > > > >
> > > > > > > I switched CC to point to our icc compiler, and CXX to
our c++,
> > and
> > > > > after
> > > > > > > running make clean first, encountered a new error:
> > > > > > >
> > > > > > > make[2]: Entering directory `/D2/ryan/netcdf-4.1.1/man4'
> > > > > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run
makeinfo
> > -I
> > > > .'
> > > > > \
> > > > > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > > > > /bin/sh: texi2dvi: command not found
> > > > > > > make[2]: *** [netcdf.pdf] Error 127
> > > > > > >
> > > > > > >
> > > > > > > Thanks,
> > > > > > > Ryan
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway via
RT <
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > > > Ryan,
> > > > > > > >
> > > > > > > > I just went through the same steps... and it turns out
I'm
> > > getting
> > > > > the
> > > > > > > same
> > > > > > > > error from 3.6.2:
> > > > > > > >
> > > > > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > > > > ncvalues.cpp:238:60: error: 'strncpy' was not declared
in
> this
> > > > scope
> > > > > > > > strncpy(s, (const char*)the_values + n,
> (int)the_number);
> > > > > > > >
> > > > > > > > I'm guessing this is a new issue that's popping up
with newer
> > > > > versions
> > > > > > of
> > > > > > > > the GNU compilers. I'm using 4.9.2:
> > > > > > > >
> > > > > > > > [johnhg at number5]% g++ --version
> > > > > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > > > > >
> > > > > > > > Let's try version 4.1.1:
> > > > > > > > wget ftp://ftp.unidata.ucar.edu/pub/netcdf/old/netcdf-
4.1.1.
> > > tar.gz
> > > > > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > > > > cd netcdf-4.1.1
> > > > > > > > setenv CC `which gcc`
> > > > > > > > setenv CXX `which g++`
> > > > > > > > setenv FC ''
> > > > > > > > setenv F90 ''
> > > > > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > > > > make >& make.log
> > > > > > > > make install >& make_install.log
> > > > > > > >
> > > > > > > > That worked on my system. Are you still getting
compilation
> > > > errors?
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > John
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via RT
<
> > > > > > met_help at ucar.edu
> > > > > > > >
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > > >
> > > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> > Ticket/Display.html?id=79613
> > > >
> > > > > > > > >
> > > > > > > > > Hi John,
> > > > > > > > >
> > > > > > > > > I located netcdf.hh and set the MET_NETCDF env var
as
> > > suggested,
> > > > > but
> > > > > > > the
> > > > > > > > > problem persists.
> > > > > > > > >
> > > > > > > > > I have also tried building NetCDF-3.6.2 locally as
> suggested,
> > > but
> > > > > > > > encounter
> > > > > > > > > a different error there:
> > > > > > > > >
> > > > > > > > > ncvalues.cpp(238): error: identifier "strncpy" is
undefined
> > > > > > > > > strncpy(s, (const char*)the_values + n,
> > (int)the_number);
> > > > > > > > >
> > > > > > > > > Not sure what that means or what to do about it, or
why the
> > > first
> > > > > > > > solution
> > > > > > > > > didn't work.
> > > > > > > > >
> > > > > > > > > Ryan
> > > > > > > > >
> > > > > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley Gotway
via RT
> <
> > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > >
> > > > > > > > > > Hello Ryan,
> > > > > > > > > >
> > > > > > > > > > I see you're having trouble compiling MET using
NetCDF.
> > The
> > > > > NetCDF
> > > > > > > > > library
> > > > > > > > > > gets a bit confusing. The 3-series versions of
the
> NetCDF
> > > > > library
> > > > > > > > > > typically had the Fortran, C, and C++ interfaces
all
> > bundled
> > > > > > > together.
> > > > > > > > > At
> > > > > > > > > > some point, Unidata split out those interfaces
into
> > different
> > > > > > > libraries
> > > > > > > > > for
> > > > > > > > > > Fortran, C, and C++.
> > > > > > > > > >
> > > > > > > > > > MET utilizes the C++ interface. And the current
release
> of
> > > MET
> > > > > > > > (version
> > > > > > > > > > 5.2) uses the classic data model from NetCDF
version 3.
> In
> > > the
> > > > > > next
> > > > > > > > > > version of MET (version 6.0), we'll support
> reading/writing
> > > > > NetCDF4
> > > > > > > > > files,
> > > > > > > > > > which utilize compression.
> > > > > > > > > >
> > > > > > > > > > I'd suggest the following...
> > > > > > > > > >
> > > > > > > > > > - First, check to see if you can find the
"netcdf.hh"
> > header
> > > > file
> > > > > > > > > anywhere
> > > > > > > > > > on your system. If so, you may already have the
NetCDF
> C++
> > > > > > interface
> > > > > > > > > > available. Try pointing MET_NETCDF at the
location you
> > > found.
> > > > > For
> > > > > > > > > > example, if you find /usr/local/netcdf/netcdf.hh
and
> > > > > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > > > > >
> > > > > > > > > > - If you don't find that, just compile NetCDF
version 3
> > > locally
> > > > > and
> > > > > > > > point
> > > > > > > > > > at that one:
> > > > > > > > > > ftp://ftp.unidata.ucar.edu/
> pub/netcdf/old/netcdf-3.6.2.
> > > > tar.gz
> > > > > > > > > >
> > > > > > > > > > Hopefully after the switch to NetCDF4 in MET
version 6.0,
> > > > you'll
> > > > > > just
> > > > > > > > be
> > > > > > > > > > able to use the system-level NetCDF build.
> > > > > > > > > >
> > > > > > > > > > Please let me know if you continue to have issues.
> > > > > > > > > >
> > > > > > > > > > Thanks,
> > > > > > > > > > John Halley Gotway
> > > > > > > > > > met_help at ucar.edu
> > > > > > > > > >
> > > > > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly
via RT <
> > > > > > > > > met_help at ucar.edu>
> > > > > > > > > > wrote:
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was
acted upon.
> > > > > > > > > > > Transaction: Ticket created by
rconne01 at gmail.com
> > > > > > > > > > > Queue: met_help
> > > > > > > > > > > Subject: Fwd: Software install error,
perhaps you
> > can
> > > > > help?
> > > > > > > > > > > Owner: Nobody
> > > > > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > > > > Status: new
> > > > > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > > > > Ticket/Display.html?id=79613
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > Hi,
> > > > > > > > > > >
> > > > > > > > > > > I'm trying to reinstall the MET tools and, upon
> executing
> > > > make
> > > > > > > > install,
> > > > > > > > > > > consistently encountering the error:
> > > > > > > > > > >
> > > > > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > > > > nc_var_info.h(24): catastrophic error: cannot
open
> source
> > > > file
> > > > > > > > > > "netcdf.hh"
> > > > > > > > > > > #include <netcdf.hh>
> > > > > > > > > > > ^
> > > > > > > > > > >
> > > > > > > > > > > (Note that the MET install guide suggests
"downloading
> > and
> > > > > > building
> > > > > > > > > [my]
> > > > > > > > > > > own instance of ... NETCDF". This is why, attm,
I have
> > > > > > > > netcdf-c-4.4.1
> > > > > > > > > in
> > > > > > > > > > > my working directory, in addition to the netcdf
we have
> > > > > installed
> > > > > > > for
> > > > > > > > > > > general use on my shared server, and have been
> > > intermittently
> > > > > > > > > overriding
> > > > > > > > > > my
> > > > > > > > > > > .bashrc to set the MET_NETCDF env var to this
directory
> > > > instead
> > > > > > of
> > > > > > > > the
> > > > > > > > > > > latter. Regardless of which NETCDF I use, I
encounter
> > the
> > > > same
> > > > > > > > error.
> > > > > > > > > > >
> > > > > > > > > > > A colleague suggested that this may be an error
related
> > to
> > > > the
> > > > > > > > version
> > > > > > > > > of
> > > > > > > > > > > NETCDF: that MET is not compatible with NETCDF4?
> > However,
> > > I
> > > > > see
> > > > > > in
> > > > > > > > the
> > > > > > > > > > > installation guide that it IS compatible, as
long as
> one
> > > were
> > > > > to
> > > > > > > > > > configure
> > > > > > > > > > > with the flag --disable-netcdf-4:
> > > > http://www.dtcenter.org/met/
> > > > > > > users/
> > > > > > > > > > >
support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > > > > > compilation&category=req_libs
> > > > > > > > > > >
> > > > > > > > > > > Thank you,
> > > > > > > > > > > Ryan
> > > > > > > > > > >
> > > > > > > > > > > --
> > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > University
> > > > > > > > > > > B.S. in Meteorology with Minors in Mathematics
and GIS,
> > > > > > Valparaiso
> > > > > > > > > > > University
> > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > --
> > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > University
> > > > > > > > > > > B.S. in Meteorology with Minors in Mathematics
and GIS,
> > > > > > Valparaiso
> > > > > > > > > > > University
> > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Ryan Connelly
> > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> University
> > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > Valparaiso
> > > > > > > > > University
> > > > > > > > > rconne01 at gmail.com
> > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Ryan Connelly
> > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> > Valparaiso
> > > > > > > University
> > > > > > > rconne01 at gmail.com
> > > > > > > ryan.connelly at stonybrook.edu
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Ryan Connelly
> > > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > > University
> > > > > rconne01 at gmail.com
> > > > > ryan.connelly at stonybrook.edu
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> >
> >
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Tue Feb 28 14:11:10 2017
Hi John,
Part of the problem was that I was assuming running make clean was
allowing
me to re-configure, but it appears re-configuring after that just
found the
config.h file and didn't do anything. So I'm a good bit further now.
Any
idea where I should be looking, or telling my computer to look, for
-lbufr?
ld: cannot find -lbufr
Thanks,
Ryan
On Mon, Feb 27, 2017 at 6:02 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Nope, just read permission on my machine:
>
> -rw-r--r-- 1 root root 23 Feb 27 2016
/usr/local/netcdf/include/netcdf.hh
>
> So permissions probably are not the problem.
>
> John
>
> On Mon, Feb 27, 2017 at 3:45 PM, Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> > Hi John,
> >
> > Do these netCDF files require execute permissions? They have r &
w
> > already.
> >
> > Thanks,
> > Ryan
> >
> > On Fri, Feb 24, 2017 at 5:25 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Ryan,
> > >
> > > OK, sorry for all this hassle. It sounds pretty frustrating.
So
> you've
> > > found the NetCDF header/library files you need and you've set
the
> > > MET_NETCDF environment variable to tell MET where to find them.
> > >
> > > But you get:
> > > nc_var_info.h(24): catastrophic error: cannot open source file
> > "netcdf.hh"
> > >
> > > I suppose it's possible that it's a permissions issue. Perhaps
it's
> > trying
> > > to read that file but you don't have permission to do so? So
you could
> > > check the file permissions... but honestly that would be pretty
odd.
> > >
> > > I googled that error message and found the following met-help
ticket in
> > > 2011:
> > > http://mailman.ucar.edu/pipermail/met_help/2011-
> February/001522.html
> > >
> > > That user was ultimately had success compiling NetCDF 3.6.3
using the
> GNU
> > > compilers.
> > >
> > > John
> > >
> > > On Fri, Feb 24, 2017 at 2:48 PM, Ryan Connelly via RT <
> met_help at ucar.edu
> > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
>
> > > >
> > > > Hi John,
> > > >
> > > > Even after verifying the existence of netcdf.hh in
> $MET_NETCDF/include,
> > > and
> > > > libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and with
the
> > > > $MET_NETCDF env var set properly, I still encounter the same
error -
> > that
> > > > it can't open netcdf.hh. I'm going to talk to my system admin
next;
> at
> > > > this point, this can't be a MET problem.
> > > >
> > > > Thank you for your help!
> > > > Ryan
> > > >
> > > > On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Ryan,
> > > > >
> > > > > OK, perhaps you haven't updated the MET_NETCDF environment
variable
> > to
> > > > tell
> > > > > MET where to find that library.
> > > > >
> > > > > Try running:
> > > > > cd met-5.2
> > > > > ./configure --help
> > > > >
> > > > > That'll dump out a bunch of info, including the setting of
> > environment
> > > > > variables. Once you get NetCDF compiled, look for the
NetCDF
> header
> > > > files,
> > > > > netcdf.h and netcdf.hh. Also, look for the library files
named
> > > > libnetcdf.a
> > > > > and libnetcdf_c++.a. Those are the header files that MET
includes
> > and
> > > > the
> > > > > library files to which it links.
> > > > >
> > > > > If those files are in /path/to/netcdf/include and
> /path/to/netcdf/lib
> > > > (and
> > > > > they probably are), then just define:
> > > > > MET_NETCDF /path/to/netcdf
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > > On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT <
> > > met_help at ucar.edu
> > > > >
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > > >
> > > > > > Ok, that makes sense about the compilers. You can see I'm
still
> > new
> > > at
> > > > > > some of this stuff.
> > > > > >
> > > > > > With setting the compilers as such, I was able to get a
bit
> farther
> > > > than
> > > > > > before, but I still encounter the error about the
inability to
> find
> > > > > > netcdf.hh when using v4.1.1. At the moment, I can't even
compile
> > > > v3.6.x,
> > > > > > even though I did earlier this morning (then removed it
for space
> > > > > purposes
> > > > > > as I'm nearly full, not thinking I'd still need to
reference it).
> > > One
> > > > of
> > > > > > those days!
> > > > > >
> > > > > > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > > > Ryan,
> > > > > > >
> > > > > > > Boy NetCDF is causing all sorts of problems! If you're
> switching
> > > to
> > > > > the
> > > > > > > Intel compilers, I believe you should use:
> > > > > > >
> > > > > > > CC = icc
> > > > > > > CXX = icc
> > > > > > > F77 = ifort
> > > > > > >
> > > > > > > I believe icc handles both C and C++ source code... but
am not
> > > > positive
> > > > > > on
> > > > > > > that. I really don't know about "c++"... but generally
you
> > should
> > > > use
> > > > > 1
> > > > > > > family of compilers together and I doubt "c++" is an
Intel
> > > compiler.
> > > > > > >
> > > > > > > John
> > > > > > >
> > > > > > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT <
> > > > > > met_help at ucar.edu>
> > > > > > > wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> Ticket/Display.html?id=79613
> > >
> > > > > > > >
> > > > > > > > Hi John,
> > > > > > > >
> > > > > > > > This returns the error:
> > > > > > > >
> > > > > > > > In file included from ncdump.c:16:0:
> > > > > > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > > > > > compiler/include/math.h:29:3:
> > > > > > > > error: #error "This Intel <math.h> is for use with
only the
> > Intel
> > > > > > > > compilers!"
> > > > > > > > # error "This Intel <math.h> is for use with only the
Intel
> > > > > > compilers!"
> > > > > > > >
> > > > > > > > ^
> > > > > > > >
> > > > > > > > I switched CC to point to our icc compiler, and CXX to
our
> c++,
> > > and
> > > > > > after
> > > > > > > > running make clean first, encountered a new error:
> > > > > > > >
> > > > > > > > make[2]: Entering directory `/D2/ryan/netcdf-
4.1.1/man4'
> > > > > > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > > > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing --run
> makeinfo
> > > -I
> > > > > .'
> > > > > > \
> > > > > > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > > > > > /bin/sh: texi2dvi: command not found
> > > > > > > > make[2]: *** [netcdf.pdf] Error 127
> > > > > > > >
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > Ryan
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway
via RT <
> > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > >
> > > > > > > > > Ryan,
> > > > > > > > >
> > > > > > > > > I just went through the same steps... and it turns
out I'm
> > > > getting
> > > > > > the
> > > > > > > > same
> > > > > > > > > error from 3.6.2:
> > > > > > > > >
> > > > > > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > > > > > ncvalues.cpp:238:60: error: 'strncpy' was not
declared in
> > this
> > > > > scope
> > > > > > > > > strncpy(s, (const char*)the_values + n,
> > (int)the_number);
> > > > > > > > >
> > > > > > > > > I'm guessing this is a new issue that's popping up
with
> newer
> > > > > > versions
> > > > > > > of
> > > > > > > > > the GNU compilers. I'm using 4.9.2:
> > > > > > > > >
> > > > > > > > > [johnhg at number5]% g++ --version
> > > > > > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > > > > > >
> > > > > > > > > Let's try version 4.1.1:
> > > > > > > > > wget ftp://ftp.unidata.ucar.edu/
> pub/netcdf/old/netcdf-4.1.1.
> > > > tar.gz
> > > > > > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > > > > > cd netcdf-4.1.1
> > > > > > > > > setenv CC `which gcc`
> > > > > > > > > setenv CXX `which g++`
> > > > > > > > > setenv FC ''
> > > > > > > > > setenv F90 ''
> > > > > > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > > > > > make >& make.log
> > > > > > > > > make install >& make_install.log
> > > > > > > > >
> > > > > > > > > That worked on my system. Are you still getting
> compilation
> > > > > errors?
> > > > > > > > >
> > > > > > > > > Thanks,
> > > > > > > > > John
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via
RT <
> > > > > > > met_help at ucar.edu
> > > > > > > > >
> > > > > > > > > wrote:
> > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> > > Ticket/Display.html?id=79613
> > > > >
> > > > > > > > > >
> > > > > > > > > > Hi John,
> > > > > > > > > >
> > > > > > > > > > I located netcdf.hh and set the MET_NETCDF env var
as
> > > > suggested,
> > > > > > but
> > > > > > > > the
> > > > > > > > > > problem persists.
> > > > > > > > > >
> > > > > > > > > > I have also tried building NetCDF-3.6.2 locally as
> > suggested,
> > > > but
> > > > > > > > > encounter
> > > > > > > > > > a different error there:
> > > > > > > > > >
> > > > > > > > > > ncvalues.cpp(238): error: identifier "strncpy" is
> undefined
> > > > > > > > > > strncpy(s, (const char*)the_values + n,
> > > (int)the_number);
> > > > > > > > > >
> > > > > > > > > > Not sure what that means or what to do about it,
or why
> the
> > > > first
> > > > > > > > > solution
> > > > > > > > > > didn't work.
> > > > > > > > > >
> > > > > > > > > > Ryan
> > > > > > > > > >
> > > > > > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley
Gotway via
> RT
> > <
> > > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > > >
> > > > > > > > > > > Hello Ryan,
> > > > > > > > > > >
> > > > > > > > > > > I see you're having trouble compiling MET using
NetCDF.
> > > The
> > > > > > NetCDF
> > > > > > > > > > library
> > > > > > > > > > > gets a bit confusing. The 3-series versions of
the
> > NetCDF
> > > > > > library
> > > > > > > > > > > typically had the Fortran, C, and C++ interfaces
all
> > > bundled
> > > > > > > > together.
> > > > > > > > > > At
> > > > > > > > > > > some point, Unidata split out those interfaces
into
> > > different
> > > > > > > > libraries
> > > > > > > > > > for
> > > > > > > > > > > Fortran, C, and C++.
> > > > > > > > > > >
> > > > > > > > > > > MET utilizes the C++ interface. And the current
> release
> > of
> > > > MET
> > > > > > > > > (version
> > > > > > > > > > > 5.2) uses the classic data model from NetCDF
version 3.
> > In
> > > > the
> > > > > > > next
> > > > > > > > > > > version of MET (version 6.0), we'll support
> > reading/writing
> > > > > > NetCDF4
> > > > > > > > > > files,
> > > > > > > > > > > which utilize compression.
> > > > > > > > > > >
> > > > > > > > > > > I'd suggest the following...
> > > > > > > > > > >
> > > > > > > > > > > - First, check to see if you can find the
"netcdf.hh"
> > > header
> > > > > file
> > > > > > > > > > anywhere
> > > > > > > > > > > on your system. If so, you may already have the
NetCDF
> > C++
> > > > > > > interface
> > > > > > > > > > > available. Try pointing MET_NETCDF at the
location you
> > > > found.
> > > > > > For
> > > > > > > > > > > example, if you find /usr/local/netcdf/netcdf.hh
and
> > > > > > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > > > > > >
> > > > > > > > > > > - If you don't find that, just compile NetCDF
version 3
> > > > locally
> > > > > > and
> > > > > > > > > point
> > > > > > > > > > > at that one:
> > > > > > > > > > > ftp://ftp.unidata.ucar.edu/
> > pub/netcdf/old/netcdf-3.6.2.
> > > > > tar.gz
> > > > > > > > > > >
> > > > > > > > > > > Hopefully after the switch to NetCDF4 in MET
version
> 6.0,
> > > > > you'll
> > > > > > > just
> > > > > > > > > be
> > > > > > > > > > > able to use the system-level NetCDF build.
> > > > > > > > > > >
> > > > > > > > > > > Please let me know if you continue to have
issues.
> > > > > > > > > > >
> > > > > > > > > > > Thanks,
> > > > > > > > > > > John Halley Gotway
> > > > > > > > > > > met_help at ucar.edu
> > > > > > > > > > >
> > > > > > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan Connelly
via RT
> <
> > > > > > > > > > met_help at ucar.edu>
> > > > > > > > > > > wrote:
> > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was
acted
> upon.
> > > > > > > > > > > > Transaction: Ticket created by
rconne01 at gmail.com
> > > > > > > > > > > > Queue: met_help
> > > > > > > > > > > > Subject: Fwd: Software install error,
perhaps
> you
> > > can
> > > > > > help?
> > > > > > > > > > > > Owner: Nobody
> > > > > > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > > > > > Status: new
> > > > > > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > > > > > Ticket/Display.html?id=79613
> > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > Hi,
> > > > > > > > > > > >
> > > > > > > > > > > > I'm trying to reinstall the MET tools and,
upon
> > executing
> > > > > make
> > > > > > > > > install,
> > > > > > > > > > > > consistently encountering the error:
> > > > > > > > > > > >
> > > > > > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > > > > > nc_var_info.h(24): catastrophic error: cannot
open
> > source
> > > > > file
> > > > > > > > > > > "netcdf.hh"
> > > > > > > > > > > > #include <netcdf.hh>
> > > > > > > > > > > > ^
> > > > > > > > > > > >
> > > > > > > > > > > > (Note that the MET install guide suggests
> "downloading
> > > and
> > > > > > > building
> > > > > > > > > > [my]
> > > > > > > > > > > > own instance of ... NETCDF". This is why,
attm, I
> have
> > > > > > > > > netcdf-c-4.4.1
> > > > > > > > > > in
> > > > > > > > > > > > my working directory, in addition to the
netcdf we
> have
> > > > > > installed
> > > > > > > > for
> > > > > > > > > > > > general use on my shared server, and have been
> > > > intermittently
> > > > > > > > > > overriding
> > > > > > > > > > > my
> > > > > > > > > > > > .bashrc to set the MET_NETCDF env var to this
> directory
> > > > > instead
> > > > > > > of
> > > > > > > > > the
> > > > > > > > > > > > latter. Regardless of which NETCDF I use, I
> encounter
> > > the
> > > > > same
> > > > > > > > > error.
> > > > > > > > > > > >
> > > > > > > > > > > > A colleague suggested that this may be an
error
> related
> > > to
> > > > > the
> > > > > > > > > version
> > > > > > > > > > of
> > > > > > > > > > > > NETCDF: that MET is not compatible with
NETCDF4?
> > > However,
> > > > I
> > > > > > see
> > > > > > > in
> > > > > > > > > the
> > > > > > > > > > > > installation guide that it IS compatible, as
long as
> > one
> > > > were
> > > > > > to
> > > > > > > > > > > configure
> > > > > > > > > > > > with the flag --disable-netcdf-4:
> > > > > http://www.dtcenter.org/met/
> > > > > > > > users/
> > > > > > > > > > > >
support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > > > > > > compilation&category=req_libs
> > > > > > > > > > > >
> > > > > > > > > > > > Thank you,
> > > > > > > > > > > > Ryan
> > > > > > > > > > > >
> > > > > > > > > > > > --
> > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > > University
> > > > > > > > > > > > B.S. in Meteorology with Minors in Mathematics
and
> GIS,
> > > > > > > Valparaiso
> > > > > > > > > > > > University
> > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > --
> > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > > University
> > > > > > > > > > > > B.S. in Meteorology with Minors in Mathematics
and
> GIS,
> > > > > > > Valparaiso
> > > > > > > > > > > > University
> > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > --
> > > > > > > > > > Ryan Connelly
> > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> > University
> > > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > > Valparaiso
> > > > > > > > > > University
> > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > Ryan Connelly
> > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > Valparaiso
> > > > > > > > University
> > > > > > > > rconne01 at gmail.com
> > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Ryan Connelly
> > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> Valparaiso
> > > > > > University
> > > > > > rconne01 at gmail.com
> > > > > > ryan.connelly at stonybrook.edu
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: John Halley Gotway
Time: Tue Feb 28 14:56:53 2017
Ryan,
Great, I'm glad you were able to make progress. I should have
emphasized
that each time you change your environment (i.e. MET_NETCDF) you
should
rerun the configure script. My apologies.
As for compiling BUFRLIB, please follow the instructions posted on
this
page:
http://www.dtcenter.org/met/users/support/online_tutorial/METv5.2/tutorial.php?name=compilation&category=req_libs
Thanks
John
On Tue, Feb 28, 2017 at 2:23 PM Ryan Connelly via RT
<met_help at ucar.edu>
wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
>
> Hi John,
>
> Part of the problem was that I was assuming running make clean was
allowing
> me to re-configure, but it appears re-configuring after that just
found the
> config.h file and didn't do anything. So I'm a good bit further
now. Any
> idea where I should be looking, or telling my computer to look, for
-lbufr?
>
> ld: cannot find -lbufr
>
>
> Thanks,
> Ryan
>
> On Mon, Feb 27, 2017 at 6:02 PM, John Halley Gotway via RT <
> met_help at ucar.edu> wrote:
>
> > Nope, just read permission on my machine:
> >
> > -rw-r--r-- 1 root root 23 Feb 27 2016
> /usr/local/netcdf/include/netcdf.hh
> >
> > So permissions probably are not the problem.
> >
> > John
> >
> > On Mon, Feb 27, 2017 at 3:45 PM, Ryan Connelly via RT
<met_help at ucar.edu
> >
> > wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > >
> > > Hi John,
> > >
> > > Do these netCDF files require execute permissions? They have r
& w
> > > already.
> > >
> > > Thanks,
> > > Ryan
> > >
> > > On Fri, Feb 24, 2017 at 5:25 PM, John Halley Gotway via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > > Ryan,
> > > >
> > > > OK, sorry for all this hassle. It sounds pretty frustrating.
So
> > you've
> > > > found the NetCDF header/library files you need and you've set
the
> > > > MET_NETCDF environment variable to tell MET where to find
them.
> > > >
> > > > But you get:
> > > > nc_var_info.h(24): catastrophic error: cannot open source file
> > > "netcdf.hh"
> > > >
> > > > I suppose it's possible that it's a permissions issue.
Perhaps it's
> > > trying
> > > > to read that file but you don't have permission to do so? So
you
> could
> > > > check the file permissions... but honestly that would be
pretty odd.
> > > >
> > > > I googled that error message and found the following met-help
ticket
> in
> > > > 2011:
> > > > http://mailman.ucar.edu/pipermail/met_help/2011-
> > February/001522.html
> > > >
> > > > That user was ultimately had success compiling NetCDF 3.6.3
using the
> > GNU
> > > > compilers.
> > > >
> > > > John
> > > >
> > > > On Fri, Feb 24, 2017 at 2:48 PM, Ryan Connelly via RT <
> > met_help at ucar.edu
> > > >
> > > > wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > >
> > > > > Hi John,
> > > > >
> > > > > Even after verifying the existence of netcdf.hh in
> > $MET_NETCDF/include,
> > > > and
> > > > > libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and with
the
> > > > > $MET_NETCDF env var set properly, I still encounter the same
error
> -
> > > that
> > > > > it can't open netcdf.hh. I'm going to talk to my system
admin
> next;
> > at
> > > > > this point, this can't be a MET problem.
> > > > >
> > > > > Thank you for your help!
> > > > > Ryan
> > > > >
> > > > > On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT <
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > > Ryan,
> > > > > >
> > > > > > OK, perhaps you haven't updated the MET_NETCDF environment
> variable
> > > to
> > > > > tell
> > > > > > MET where to find that library.
> > > > > >
> > > > > > Try running:
> > > > > > cd met-5.2
> > > > > > ./configure --help
> > > > > >
> > > > > > That'll dump out a bunch of info, including the setting of
> > > environment
> > > > > > variables. Once you get NetCDF compiled, look for the
NetCDF
> > header
> > > > > files,
> > > > > > netcdf.h and netcdf.hh. Also, look for the library files
named
> > > > > libnetcdf.a
> > > > > > and libnetcdf_c++.a. Those are the header files that MET
> includes
> > > and
> > > > > the
> > > > > > library files to which it links.
> > > > > >
> > > > > > If those files are in /path/to/netcdf/include and
> > /path/to/netcdf/lib
> > > > > (and
> > > > > > they probably are), then just define:
> > > > > > MET_NETCDF /path/to/netcdf
> > > > > >
> > > > > > Thanks,
> > > > > > John
> > > > > >
> > > > > > On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT <
> > > > met_help at ucar.edu
> > > > > >
> > > > > > wrote:
> > > > > >
> > > > > > >
> > > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
> >
> > > > > > >
> > > > > > > Ok, that makes sense about the compilers. You can see
I'm
> still
> > > new
> > > > at
> > > > > > > some of this stuff.
> > > > > > >
> > > > > > > With setting the compilers as such, I was able to get a
bit
> > farther
> > > > > than
> > > > > > > before, but I still encounter the error about the
inability to
> > find
> > > > > > > netcdf.hh when using v4.1.1. At the moment, I can't
even
> compile
> > > > > v3.6.x,
> > > > > > > even though I did earlier this morning (then removed it
for
> space
> > > > > > purposes
> > > > > > > as I'm nearly full, not thinking I'd still need to
reference
> it).
> > > > One
> > > > > of
> > > > > > > those days!
> > > > > > >
> > > > > > > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway via
RT <
> > > > > > > met_help at ucar.edu> wrote:
> > > > > > >
> > > > > > > > Ryan,
> > > > > > > >
> > > > > > > > Boy NetCDF is causing all sorts of problems! If
you're
> > switching
> > > > to
> > > > > > the
> > > > > > > > Intel compilers, I believe you should use:
> > > > > > > >
> > > > > > > > CC = icc
> > > > > > > > CXX = icc
> > > > > > > > F77 = ifort
> > > > > > > >
> > > > > > > > I believe icc handles both C and C++ source code...
but am
> not
> > > > > positive
> > > > > > > on
> > > > > > > > that. I really don't know about "c++"... but
generally you
> > > should
> > > > > use
> > > > > > 1
> > > > > > > > family of compilers together and I doubt "c++" is an
Intel
> > > > compiler.
> > > > > > > >
> > > > > > > > John
> > > > > > > >
> > > > > > > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via RT
<
> > > > > > > met_help at ucar.edu>
> > > > > > > > wrote:
> > > > > > > >
> > > > > > > > >
> > > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> > Ticket/Display.html?id=79613
> > > >
> > > > > > > > >
> > > > > > > > > Hi John,
> > > > > > > > >
> > > > > > > > > This returns the error:
> > > > > > > > >
> > > > > > > > > In file included from ncdump.c:16:0:
> > > > > > > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > > > > > > compiler/include/math.h:29:3:
> > > > > > > > > error: #error "This Intel <math.h> is for use with
only the
> > > Intel
> > > > > > > > > compilers!"
> > > > > > > > > # error "This Intel <math.h> is for use with only
the
> Intel
> > > > > > > compilers!"
> > > > > > > > >
> > > > > > > > > ^
> > > > > > > > >
> > > > > > > > > I switched CC to point to our icc compiler, and CXX
to our
> > c++,
> > > > and
> > > > > > > after
> > > > > > > > > running make clean first, encountered a new error:
> > > > > > > > >
> > > > > > > > > make[2]: Entering directory `/D2/ryan/netcdf-
4.1.1/man4'
> > > > > > > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > > > > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing
--run
> > makeinfo
> > > > -I
> > > > > > .'
> > > > > > > \
> > > > > > > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > > > > > > /bin/sh: texi2dvi: command not found
> > > > > > > > > make[2]: *** [netcdf.pdf] Error 127
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Thanks,
> > > > > > > > > Ryan
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley Gotway
via RT
> <
> > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > >
> > > > > > > > > > Ryan,
> > > > > > > > > >
> > > > > > > > > > I just went through the same steps... and it turns
out
> I'm
> > > > > getting
> > > > > > > the
> > > > > > > > > same
> > > > > > > > > > error from 3.6.2:
> > > > > > > > > >
> > > > > > > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > > > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > > > > > > ncvalues.cpp:238:60: error: 'strncpy' was not
declared in
> > > this
> > > > > > scope
> > > > > > > > > > strncpy(s, (const char*)the_values + n,
> > > (int)the_number);
> > > > > > > > > >
> > > > > > > > > > I'm guessing this is a new issue that's popping up
with
> > newer
> > > > > > > versions
> > > > > > > > of
> > > > > > > > > > the GNU compilers. I'm using 4.9.2:
> > > > > > > > > >
> > > > > > > > > > [johnhg at number5]% g++ --version
> > > > > > > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > > > > > > >
> > > > > > > > > > Let's try version 4.1.1:
> > > > > > > > > > wget ftp://ftp.unidata.ucar.edu/
> > pub/netcdf/old/netcdf-4.1.1.
> > > > > tar.gz
> > > > > > > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > > > > > > cd netcdf-4.1.1
> > > > > > > > > > setenv CC `which gcc`
> > > > > > > > > > setenv CXX `which g++`
> > > > > > > > > > setenv FC ''
> > > > > > > > > > setenv F90 ''
> > > > > > > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > > > > > > make >& make.log
> > > > > > > > > > make install >& make_install.log
> > > > > > > > > >
> > > > > > > > > > That worked on my system. Are you still getting
> > compilation
> > > > > > errors?
> > > > > > > > > >
> > > > > > > > > > Thanks,
> > > > > > > > > > John
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly via
RT <
> > > > > > > > met_help at ucar.edu
> > > > > > > > > >
> > > > > > > > > > wrote:
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> > > > Ticket/Display.html?id=79613
> > > > > >
> > > > > > > > > > >
> > > > > > > > > > > Hi John,
> > > > > > > > > > >
> > > > > > > > > > > I located netcdf.hh and set the MET_NETCDF env
var as
> > > > > suggested,
> > > > > > > but
> > > > > > > > > the
> > > > > > > > > > > problem persists.
> > > > > > > > > > >
> > > > > > > > > > > I have also tried building NetCDF-3.6.2 locally
as
> > > suggested,
> > > > > but
> > > > > > > > > > encounter
> > > > > > > > > > > a different error there:
> > > > > > > > > > >
> > > > > > > > > > > ncvalues.cpp(238): error: identifier "strncpy"
is
> > undefined
> > > > > > > > > > > strncpy(s, (const char*)the_values + n,
> > > > (int)the_number);
> > > > > > > > > > >
> > > > > > > > > > > Not sure what that means or what to do about it,
or why
> > the
> > > > > first
> > > > > > > > > > solution
> > > > > > > > > > > didn't work.
> > > > > > > > > > >
> > > > > > > > > > > Ryan
> > > > > > > > > > >
> > > > > > > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley
Gotway via
> > RT
> > > <
> > > > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > > > >
> > > > > > > > > > > > Hello Ryan,
> > > > > > > > > > > >
> > > > > > > > > > > > I see you're having trouble compiling MET
using
> NetCDF.
> > > > The
> > > > > > > NetCDF
> > > > > > > > > > > library
> > > > > > > > > > > > gets a bit confusing. The 3-series versions
of the
> > > NetCDF
> > > > > > > library
> > > > > > > > > > > > typically had the Fortran, C, and C++
interfaces all
> > > > bundled
> > > > > > > > > together.
> > > > > > > > > > > At
> > > > > > > > > > > > some point, Unidata split out those interfaces
into
> > > > different
> > > > > > > > > libraries
> > > > > > > > > > > for
> > > > > > > > > > > > Fortran, C, and C++.
> > > > > > > > > > > >
> > > > > > > > > > > > MET utilizes the C++ interface. And the
current
> > release
> > > of
> > > > > MET
> > > > > > > > > > (version
> > > > > > > > > > > > 5.2) uses the classic data model from NetCDF
version
> 3.
> > > In
> > > > > the
> > > > > > > > next
> > > > > > > > > > > > version of MET (version 6.0), we'll support
> > > reading/writing
> > > > > > > NetCDF4
> > > > > > > > > > > files,
> > > > > > > > > > > > which utilize compression.
> > > > > > > > > > > >
> > > > > > > > > > > > I'd suggest the following...
> > > > > > > > > > > >
> > > > > > > > > > > > - First, check to see if you can find the
"netcdf.hh"
> > > > header
> > > > > > file
> > > > > > > > > > > anywhere
> > > > > > > > > > > > on your system. If so, you may already have
the
> NetCDF
> > > C++
> > > > > > > > interface
> > > > > > > > > > > > available. Try pointing MET_NETCDF at the
location
> you
> > > > > found.
> > > > > > > For
> > > > > > > > > > > > example, if you find
/usr/local/netcdf/netcdf.hh and
> > > > > > > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > > > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > > > > > > >
> > > > > > > > > > > > - If you don't find that, just compile NetCDF
> version 3
> > > > > locally
> > > > > > > and
> > > > > > > > > > point
> > > > > > > > > > > > at that one:
> > > > > > > > > > > > ftp://ftp.unidata.ucar.edu/
> > > pub/netcdf/old/netcdf-3.6.2.
> > > > > > tar.gz
> > > > > > > > > > > >
> > > > > > > > > > > > Hopefully after the switch to NetCDF4 in MET
version
> > 6.0,
> > > > > > you'll
> > > > > > > > just
> > > > > > > > > > be
> > > > > > > > > > > > able to use the system-level NetCDF build.
> > > > > > > > > > > >
> > > > > > > > > > > > Please let me know if you continue to have
issues.
> > > > > > > > > > > >
> > > > > > > > > > > > Thanks,
> > > > > > > > > > > > John Halley Gotway
> > > > > > > > > > > > met_help at ucar.edu
> > > > > > > > > > > >
> > > > > > > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan
Connelly via
> RT
> > <
> > > > > > > > > > > met_help at ucar.edu>
> > > > > > > > > > > > wrote:
> > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613 was
acted
> > upon.
> > > > > > > > > > > > > Transaction: Ticket created by
rconne01 at gmail.com
> > > > > > > > > > > > > Queue: met_help
> > > > > > > > > > > > > Subject: Fwd: Software install error,
perhaps
> > you
> > > > can
> > > > > > > help?
> > > > > > > > > > > > > Owner: Nobody
> > > > > > > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > > > > > > Status: new
> > > > > > > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > > > > > > Ticket/Display.html?id=79613
> > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > > Hi,
> > > > > > > > > > > > >
> > > > > > > > > > > > > I'm trying to reinstall the MET tools and,
upon
> > > executing
> > > > > > make
> > > > > > > > > > install,
> > > > > > > > > > > > > consistently encountering the error:
> > > > > > > > > > > > >
> > > > > > > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > > > > > > nc_var_info.h(24): catastrophic error:
cannot open
> > > source
> > > > > > file
> > > > > > > > > > > > "netcdf.hh"
> > > > > > > > > > > > > #include <netcdf.hh>
> > > > > > > > > > > > > ^
> > > > > > > > > > > > >
> > > > > > > > > > > > > (Note that the MET install guide suggests
> > "downloading
> > > > and
> > > > > > > > building
> > > > > > > > > > > [my]
> > > > > > > > > > > > > own instance of ... NETCDF". This is why,
attm, I
> > have
> > > > > > > > > > netcdf-c-4.4.1
> > > > > > > > > > > in
> > > > > > > > > > > > > my working directory, in addition to the
netcdf we
> > have
> > > > > > > installed
> > > > > > > > > for
> > > > > > > > > > > > > general use on my shared server, and have
been
> > > > > intermittently
> > > > > > > > > > > overriding
> > > > > > > > > > > > my
> > > > > > > > > > > > > .bashrc to set the MET_NETCDF env var to
this
> > directory
> > > > > > instead
> > > > > > > > of
> > > > > > > > > > the
> > > > > > > > > > > > > latter. Regardless of which NETCDF I use, I
> > encounter
> > > > the
> > > > > > same
> > > > > > > > > > error.
> > > > > > > > > > > > >
> > > > > > > > > > > > > A colleague suggested that this may be an
error
> > related
> > > > to
> > > > > > the
> > > > > > > > > > version
> > > > > > > > > > > of
> > > > > > > > > > > > > NETCDF: that MET is not compatible with
NETCDF4?
> > > > However,
> > > > > I
> > > > > > > see
> > > > > > > > in
> > > > > > > > > > the
> > > > > > > > > > > > > installation guide that it IS compatible, as
long
> as
> > > one
> > > > > were
> > > > > > > to
> > > > > > > > > > > > configure
> > > > > > > > > > > > > with the flag --disable-netcdf-4:
> > > > > > http://www.dtcenter.org/met/
> > > > > > > > > users/
> > > > > > > > > > > > >
support/online_tutorial/METv5.2/tutorial.php?name=
> > > > > > > > > > > > > compilation&category=req_libs
> > > > > > > > > > > > >
> > > > > > > > > > > > > Thank you,
> > > > > > > > > > > > > Ryan
> > > > > > > > > > > > >
> > > > > > > > > > > > > --
> > > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > > > University
> > > > > > > > > > > > > B.S. in Meteorology with Minors in
Mathematics and
> > GIS,
> > > > > > > > Valparaiso
> > > > > > > > > > > > > University
> > > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > > --
> > > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > > > University
> > > > > > > > > > > > > B.S. in Meteorology with Minors in
Mathematics and
> > GIS,
> > > > > > > > Valparaiso
> > > > > > > > > > > > > University
> > > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > --
> > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > University
> > > > > > > > > > > B.S. in Meteorology with Minors in Mathematics
and GIS,
> > > > > > Valparaiso
> > > > > > > > > > > University
> > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > Ryan Connelly
> > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> University
> > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > Valparaiso
> > > > > > > > > University
> > > > > > > > > rconne01 at gmail.com
> > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Ryan Connelly
> > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> > Valparaiso
> > > > > > > University
> > > > > > > rconne01 at gmail.com
> > > > > > > ryan.connelly at stonybrook.edu
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Ryan Connelly
> > > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > > University
> > > > > rconne01 at gmail.com
> > > > > ryan.connelly at stonybrook.edu
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > > --
> > > Ryan Connelly
> > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > University
> > > rconne01 at gmail.com
> > > ryan.connelly at stonybrook.edu
> > >
> > >
> >
> >
>
>
> --
> Ryan Connelly
> M.S. Student in Atmospheric Sciences, Stony Brook University
> B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> University
> rconne01 at gmail.com
> ryan.connelly at stonybrook.edu
>
>
------------------------------------------------
Subject: Fwd: Software install error, perhaps you can help?
From: Ryan Connelly
Time: Tue Feb 28 15:02:58 2017
Hi John,
It appears to have finally installed! I was re-configuring, but for
some
reason even that wasn't enough. It actually took rm-ing the directory
and
re-tar-ing the tar file. Nonetheless, that and recompiling BUFRLIB,
seems
to have made it happy, so thank you for all your help!
Ryan
On Tue, Feb 28, 2017 at 4:56 PM, John Halley Gotway via RT <
met_help at ucar.edu> wrote:
> Ryan,
>
> Great, I'm glad you were able to make progress. I should have
emphasized
> that each time you change your environment (i.e. MET_NETCDF) you
should
> rerun the configure script. My apologies.
>
> As for compiling BUFRLIB, please follow the instructions posted on
this
> page:
>
> http://www.dtcenter.org/met/users/support/online_tutorial/
> METv5.2/tutorial.php?name=compilation&category=req_libs
>
> Thanks
> John
>
> On Tue, Feb 28, 2017 at 2:23 PM Ryan Connelly via RT
<met_help at ucar.edu>
> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> >
> > Hi John,
> >
> > Part of the problem was that I was assuming running make clean was
> allowing
> > me to re-configure, but it appears re-configuring after that just
found
> the
> > config.h file and didn't do anything. So I'm a good bit further
now.
> Any
> > idea where I should be looking, or telling my computer to look,
for
> -lbufr?
> >
> > ld: cannot find -lbufr
> >
> >
> > Thanks,
> > Ryan
> >
> > On Mon, Feb 27, 2017 at 6:02 PM, John Halley Gotway via RT <
> > met_help at ucar.edu> wrote:
> >
> > > Nope, just read permission on my machine:
> > >
> > > -rw-r--r-- 1 root root 23 Feb 27 2016
> > /usr/local/netcdf/include/netcdf.hh
> > >
> > > So permissions probably are not the problem.
> > >
> > > John
> > >
> > > On Mon, Feb 27, 2017 at 3:45 PM, Ryan Connelly via RT <
> met_help at ucar.edu
> > >
> > > wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613
>
> > > >
> > > > Hi John,
> > > >
> > > > Do these netCDF files require execute permissions? They have
r & w
> > > > already.
> > > >
> > > > Thanks,
> > > > Ryan
> > > >
> > > > On Fri, Feb 24, 2017 at 5:25 PM, John Halley Gotway via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > > Ryan,
> > > > >
> > > > > OK, sorry for all this hassle. It sounds pretty
frustrating. So
> > > you've
> > > > > found the NetCDF header/library files you need and you've
set the
> > > > > MET_NETCDF environment variable to tell MET where to find
them.
> > > > >
> > > > > But you get:
> > > > > nc_var_info.h(24): catastrophic error: cannot open source
file
> > > > "netcdf.hh"
> > > > >
> > > > > I suppose it's possible that it's a permissions issue.
Perhaps
> it's
> > > > trying
> > > > > to read that file but you don't have permission to do so?
So you
> > could
> > > > > check the file permissions... but honestly that would be
pretty
> odd.
> > > > >
> > > > > I googled that error message and found the following met-
help
> ticket
> > in
> > > > > 2011:
> > > > > http://mailman.ucar.edu/pipermail/met_help/2011-
> > > February/001522.html
> > > > >
> > > > > That user was ultimately had success compiling NetCDF 3.6.3
using
> the
> > > GNU
> > > > > compilers.
> > > > >
> > > > > John
> > > > >
> > > > > On Fri, Feb 24, 2017 at 2:48 PM, Ryan Connelly via RT <
> > > met_help at ucar.edu
> > > > >
> > > > > wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=79613 >
> > > > > >
> > > > > > Hi John,
> > > > > >
> > > > > > Even after verifying the existence of netcdf.hh in
> > > $MET_NETCDF/include,
> > > > > and
> > > > > > libnetcdf.a and libnetcdf_c++.a in $MET_NETCDF/lib, and
with the
> > > > > > $MET_NETCDF env var set properly, I still encounter the
same
> error
> > -
> > > > that
> > > > > > it can't open netcdf.hh. I'm going to talk to my system
admin
> > next;
> > > at
> > > > > > this point, this can't be a MET problem.
> > > > > >
> > > > > > Thank you for your help!
> > > > > > Ryan
> > > > > >
> > > > > > On Thu, Feb 23, 2017 at 5:16 PM, John Halley Gotway via RT
<
> > > > > > met_help at ucar.edu> wrote:
> > > > > >
> > > > > > > Ryan,
> > > > > > >
> > > > > > > OK, perhaps you haven't updated the MET_NETCDF
environment
> > variable
> > > > to
> > > > > > tell
> > > > > > > MET where to find that library.
> > > > > > >
> > > > > > > Try running:
> > > > > > > cd met-5.2
> > > > > > > ./configure --help
> > > > > > >
> > > > > > > That'll dump out a bunch of info, including the setting
of
> > > > environment
> > > > > > > variables. Once you get NetCDF compiled, look for the
NetCDF
> > > header
> > > > > > files,
> > > > > > > netcdf.h and netcdf.hh. Also, look for the library
files named
> > > > > > libnetcdf.a
> > > > > > > and libnetcdf_c++.a. Those are the header files that
MET
> > includes
> > > > and
> > > > > > the
> > > > > > > library files to which it links.
> > > > > > >
> > > > > > > If those files are in /path/to/netcdf/include and
> > > /path/to/netcdf/lib
> > > > > > (and
> > > > > > > they probably are), then just define:
> > > > > > > MET_NETCDF /path/to/netcdf
> > > > > > >
> > > > > > > Thanks,
> > > > > > > John
> > > > > > >
> > > > > > > On Thu, Feb 23, 2017 at 2:06 PM, Ryan Connelly via RT <
> > > > > met_help at ucar.edu
> > > > > > >
> > > > > > > wrote:
> > > > > > >
> > > > > > > >
> > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> Ticket/Display.html?id=79613
> > >
> > > > > > > >
> > > > > > > > Ok, that makes sense about the compilers. You can see
I'm
> > still
> > > > new
> > > > > at
> > > > > > > > some of this stuff.
> > > > > > > >
> > > > > > > > With setting the compilers as such, I was able to get
a bit
> > > farther
> > > > > > than
> > > > > > > > before, but I still encounter the error about the
inability
> to
> > > find
> > > > > > > > netcdf.hh when using v4.1.1. At the moment, I can't
even
> > compile
> > > > > > v3.6.x,
> > > > > > > > even though I did earlier this morning (then removed
it for
> > space
> > > > > > > purposes
> > > > > > > > as I'm nearly full, not thinking I'd still need to
reference
> > it).
> > > > > One
> > > > > > of
> > > > > > > > those days!
> > > > > > > >
> > > > > > > > On Thu, Feb 23, 2017 at 1:17 PM, John Halley Gotway
via RT <
> > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > >
> > > > > > > > > Ryan,
> > > > > > > > >
> > > > > > > > > Boy NetCDF is causing all sorts of problems! If
you're
> > > switching
> > > > > to
> > > > > > > the
> > > > > > > > > Intel compilers, I believe you should use:
> > > > > > > > >
> > > > > > > > > CC = icc
> > > > > > > > > CXX = icc
> > > > > > > > > F77 = ifort
> > > > > > > > >
> > > > > > > > > I believe icc handles both C and C++ source code...
but am
> > not
> > > > > > positive
> > > > > > > > on
> > > > > > > > > that. I really don't know about "c++"... but
generally you
> > > > should
> > > > > > use
> > > > > > > 1
> > > > > > > > > family of compilers together and I doubt "c++" is an
Intel
> > > > > compiler.
> > > > > > > > >
> > > > > > > > > John
> > > > > > > > >
> > > > > > > > > On Thu, Feb 23, 2017 at 10:58 AM, Ryan Connelly via
RT <
> > > > > > > > met_help at ucar.edu>
> > > > > > > > > wrote:
> > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> > > Ticket/Display.html?id=79613
> > > > >
> > > > > > > > > >
> > > > > > > > > > Hi John,
> > > > > > > > > >
> > > > > > > > > > This returns the error:
> > > > > > > > > >
> > > > > > > > > > In file included from ncdump.c:16:0:
> > > > > > > > > > /cm/shared/apps/intel/composer_xe/2013_sp1.3.174/
> > > > > > > > > > compiler/include/math.h:29:3:
> > > > > > > > > > error: #error "This Intel <math.h> is for use with
only
> the
> > > > Intel
> > > > > > > > > > compilers!"
> > > > > > > > > > # error "This Intel <math.h> is for use with only
the
> > Intel
> > > > > > > > compilers!"
> > > > > > > > > >
> > > > > > > > > > ^
> > > > > > > > > >
> > > > > > > > > > I switched CC to point to our icc compiler, and
CXX to
> our
> > > c++,
> > > > > and
> > > > > > > > after
> > > > > > > > > > running make clean first, encountered a new error:
> > > > > > > > > >
> > > > > > > > > > make[2]: Entering directory `/D2/ryan/netcdf-
4.1.1/man4'
> > > > > > > > > > TEXINPUTS=".:$TEXINPUTS" \
> > > > > > > > > > MAKEINFO='/bin/sh /D2/ryan/netcdf-4.1.1/missing
--run
> > > makeinfo
> > > > > -I
> > > > > > > .'
> > > > > > > > \
> > > > > > > > > > texi2dvi -s --pdf --batch netcdf.texi
> > > > > > > > > > /bin/sh: texi2dvi: command not found
> > > > > > > > > > make[2]: *** [netcdf.pdf] Error 127
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > Thanks,
> > > > > > > > > > Ryan
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > On Wed, Feb 22, 2017 at 6:26 PM, John Halley
Gotway via
> RT
> > <
> > > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > > >
> > > > > > > > > > > Ryan,
> > > > > > > > > > >
> > > > > > > > > > > I just went through the same steps... and it
turns out
> > I'm
> > > > > > getting
> > > > > > > > the
> > > > > > > > > > same
> > > > > > > > > > > error from 3.6.2:
> > > > > > > > > > >
> > > > > > > > > > > ncvalues.cpp: In member function 'virtual char*
> > > > > > > > > > > NcValues_ncbyte::as_string(long int) const':
> > > > > > > > > > > ncvalues.cpp:238:60: error: 'strncpy' was not
declared
> in
> > > > this
> > > > > > > scope
> > > > > > > > > > > strncpy(s, (const char*)the_values + n,
> > > > (int)the_number);
> > > > > > > > > > >
> > > > > > > > > > > I'm guessing this is a new issue that's popping
up with
> > > newer
> > > > > > > > versions
> > > > > > > > > of
> > > > > > > > > > > the GNU compilers. I'm using 4.9.2:
> > > > > > > > > > >
> > > > > > > > > > > [johnhg at number5]% g++ --version
> > > > > > > > > > > g++ (Debian 4.9.2-10) 4.9.2
> > > > > > > > > > >
> > > > > > > > > > > Let's try version 4.1.1:
> > > > > > > > > > > wget ftp://ftp.unidata.ucar.edu/
> > > pub/netcdf/old/netcdf-4.1.1.
> > > > > > tar.gz
> > > > > > > > > > > tar -xvzf netcdf-4.1.1.tar.gz
> > > > > > > > > > > cd netcdf-4.1.1
> > > > > > > > > > > setenv CC `which gcc`
> > > > > > > > > > > setenv CXX `which g++`
> > > > > > > > > > > setenv FC ''
> > > > > > > > > > > setenv F90 ''
> > > > > > > > > > > ./configure --prefix=`pwd` --disable-netcdf-4
> > > > > > > > > > > make >& make.log
> > > > > > > > > > > make install >& make_install.log
> > > > > > > > > > >
> > > > > > > > > > > That worked on my system. Are you still getting
> > > compilation
> > > > > > > errors?
> > > > > > > > > > >
> > > > > > > > > > > Thanks,
> > > > > > > > > > > John
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > On Wed, Feb 22, 2017 at 2:39 PM, Ryan Connelly
via RT <
> > > > > > > > > met_help at ucar.edu
> > > > > > > > > > >
> > > > > > > > > > > wrote:
> > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > <URL: https://rt.rap.ucar.edu/rt/
> > > > > Ticket/Display.html?id=79613
> > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > Hi John,
> > > > > > > > > > > >
> > > > > > > > > > > > I located netcdf.hh and set the MET_NETCDF env
var as
> > > > > > suggested,
> > > > > > > > but
> > > > > > > > > > the
> > > > > > > > > > > > problem persists.
> > > > > > > > > > > >
> > > > > > > > > > > > I have also tried building NetCDF-3.6.2
locally as
> > > > suggested,
> > > > > > but
> > > > > > > > > > > encounter
> > > > > > > > > > > > a different error there:
> > > > > > > > > > > >
> > > > > > > > > > > > ncvalues.cpp(238): error: identifier "strncpy"
is
> > > undefined
> > > > > > > > > > > > strncpy(s, (const char*)the_values + n,
> > > > > (int)the_number);
> > > > > > > > > > > >
> > > > > > > > > > > > Not sure what that means or what to do about
it, or
> why
> > > the
> > > > > > first
> > > > > > > > > > > solution
> > > > > > > > > > > > didn't work.
> > > > > > > > > > > >
> > > > > > > > > > > > Ryan
> > > > > > > > > > > >
> > > > > > > > > > > > On Wed, Feb 22, 2017 at 3:35 PM, John Halley
Gotway
> via
> > > RT
> > > > <
> > > > > > > > > > > > met_help at ucar.edu> wrote:
> > > > > > > > > > > >
> > > > > > > > > > > > > Hello Ryan,
> > > > > > > > > > > > >
> > > > > > > > > > > > > I see you're having trouble compiling MET
using
> > NetCDF.
> > > > > The
> > > > > > > > NetCDF
> > > > > > > > > > > > library
> > > > > > > > > > > > > gets a bit confusing. The 3-series versions
of the
> > > > NetCDF
> > > > > > > > library
> > > > > > > > > > > > > typically had the Fortran, C, and C++
interfaces
> all
> > > > > bundled
> > > > > > > > > > together.
> > > > > > > > > > > > At
> > > > > > > > > > > > > some point, Unidata split out those
interfaces into
> > > > > different
> > > > > > > > > > libraries
> > > > > > > > > > > > for
> > > > > > > > > > > > > Fortran, C, and C++.
> > > > > > > > > > > > >
> > > > > > > > > > > > > MET utilizes the C++ interface. And the
current
> > > release
> > > > of
> > > > > > MET
> > > > > > > > > > > (version
> > > > > > > > > > > > > 5.2) uses the classic data model from NetCDF
> version
> > 3.
> > > > In
> > > > > > the
> > > > > > > > > next
> > > > > > > > > > > > > version of MET (version 6.0), we'll support
> > > > reading/writing
> > > > > > > > NetCDF4
> > > > > > > > > > > > files,
> > > > > > > > > > > > > which utilize compression.
> > > > > > > > > > > > >
> > > > > > > > > > > > > I'd suggest the following...
> > > > > > > > > > > > >
> > > > > > > > > > > > > - First, check to see if you can find the
> "netcdf.hh"
> > > > > header
> > > > > > > file
> > > > > > > > > > > > anywhere
> > > > > > > > > > > > > on your system. If so, you may already have
the
> > NetCDF
> > > > C++
> > > > > > > > > interface
> > > > > > > > > > > > > available. Try pointing MET_NETCDF at the
location
> > you
> > > > > > found.
> > > > > > > > For
> > > > > > > > > > > > > example, if you find
/usr/local/netcdf/netcdf.hh
> and
> > > > > > > > > > > > > /usr/local/netcdf/lib/libnetcdf* then set
> > > > > > > > > > > MET_NETCDF=/usr/local/netcdf.
> > > > > > > > > > > > >
> > > > > > > > > > > > > - If you don't find that, just compile
NetCDF
> > version 3
> > > > > > locally
> > > > > > > > and
> > > > > > > > > > > point
> > > > > > > > > > > > > at that one:
> > > > > > > > > > > > > ftp://ftp.unidata.ucar.edu/
> > > > pub/netcdf/old/netcdf-3.6.2.
> > > > > > > tar.gz
> > > > > > > > > > > > >
> > > > > > > > > > > > > Hopefully after the switch to NetCDF4 in MET
> version
> > > 6.0,
> > > > > > > you'll
> > > > > > > > > just
> > > > > > > > > > > be
> > > > > > > > > > > > > able to use the system-level NetCDF build.
> > > > > > > > > > > > >
> > > > > > > > > > > > > Please let me know if you continue to have
issues.
> > > > > > > > > > > > >
> > > > > > > > > > > > > Thanks,
> > > > > > > > > > > > > John Halley Gotway
> > > > > > > > > > > > > met_help at ucar.edu
> > > > > > > > > > > > >
> > > > > > > > > > > > > On Wed, Feb 22, 2017 at 12:44 PM, Ryan
Connelly via
> > RT
> > > <
> > > > > > > > > > > > met_help at ucar.edu>
> > > > > > > > > > > > > wrote:
> > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Wed Feb 22 12:44:15 2017: Request 79613
was acted
> > > upon.
> > > > > > > > > > > > > > Transaction: Ticket created by
> rconne01 at gmail.com
> > > > > > > > > > > > > > Queue: met_help
> > > > > > > > > > > > > > Subject: Fwd: Software install error,
> perhaps
> > > you
> > > > > can
> > > > > > > > help?
> > > > > > > > > > > > > > Owner: Nobody
> > > > > > > > > > > > > > Requestors: rconne01 at gmail.com
> > > > > > > > > > > > > > Status: new
> > > > > > > > > > > > > > Ticket <URL: https://rt.rap.ucar.edu/rt/
> > > > > > > > > > > Ticket/Display.html?id=79613
> > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Hi,
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > I'm trying to reinstall the MET tools and,
upon
> > > > executing
> > > > > > > make
> > > > > > > > > > > install,
> > > > > > > > > > > > > > consistently encountering the error:
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > In file included from nc_var_info.cc(24):
> > > > > > > > > > > > > > nc_var_info.h(24): catastrophic error:
cannot
> open
> > > > source
> > > > > > > file
> > > > > > > > > > > > > "netcdf.hh"
> > > > > > > > > > > > > > #include <netcdf.hh>
> > > > > > > > > > > > > > ^
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > (Note that the MET install guide suggests
> > > "downloading
> > > > > and
> > > > > > > > > building
> > > > > > > > > > > > [my]
> > > > > > > > > > > > > > own instance of ... NETCDF". This is why,
attm,
> I
> > > have
> > > > > > > > > > > netcdf-c-4.4.1
> > > > > > > > > > > > in
> > > > > > > > > > > > > > my working directory, in addition to the
netcdf
> we
> > > have
> > > > > > > > installed
> > > > > > > > > > for
> > > > > > > > > > > > > > general use on my shared server, and have
been
> > > > > > intermittently
> > > > > > > > > > > > overriding
> > > > > > > > > > > > > my
> > > > > > > > > > > > > > .bashrc to set the MET_NETCDF env var to
this
> > > directory
> > > > > > > instead
> > > > > > > > > of
> > > > > > > > > > > the
> > > > > > > > > > > > > > latter. Regardless of which NETCDF I use,
I
> > > encounter
> > > > > the
> > > > > > > same
> > > > > > > > > > > error.
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > A colleague suggested that this may be an
error
> > > related
> > > > > to
> > > > > > > the
> > > > > > > > > > > version
> > > > > > > > > > > > of
> > > > > > > > > > > > > > NETCDF: that MET is not compatible with
NETCDF4?
> > > > > However,
> > > > > > I
> > > > > > > > see
> > > > > > > > > in
> > > > > > > > > > > the
> > > > > > > > > > > > > > installation guide that it IS compatible,
as long
> > as
> > > > one
> > > > > > were
> > > > > > > > to
> > > > > > > > > > > > > configure
> > > > > > > > > > > > > > with the flag --disable-netcdf-4:
> > > > > > > http://www.dtcenter.org/met/
> > > > > > > > > > users/
> > > > > > > > > > > > > > support/online_tutorial/METv5.
> 2/tutorial.php?name=
> > > > > > > > > > > > > > compilation&category=req_libs
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Thank you,
> > > > > > > > > > > > > > Ryan
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > --
> > > > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > > > M.S. Student in Atmospheric Sciences,
Stony Brook
> > > > > > University
> > > > > > > > > > > > > > B.S. in Meteorology with Minors in
Mathematics
> and
> > > GIS,
> > > > > > > > > Valparaiso
> > > > > > > > > > > > > > University
> > > > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > --
> > > > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > > > M.S. Student in Atmospheric Sciences,
Stony Brook
> > > > > > University
> > > > > > > > > > > > > > B.S. in Meteorology with Minors in
Mathematics
> and
> > > GIS,
> > > > > > > > > Valparaiso
> > > > > > > > > > > > > > University
> > > > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > > > >
> > > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > --
> > > > > > > > > > > > Ryan Connelly
> > > > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony
Brook
> > > > University
> > > > > > > > > > > > B.S. in Meteorology with Minors in Mathematics
and
> GIS,
> > > > > > > Valparaiso
> > > > > > > > > > > > University
> > > > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > --
> > > > > > > > > > Ryan Connelly
> > > > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
> > University
> > > > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > > > Valparaiso
> > > > > > > > > > University
> > > > > > > > > > rconne01 at gmail.com
> > > > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > Ryan Connelly
> > > > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > > > B.S. in Meteorology with Minors in Mathematics and
GIS,
> > > Valparaiso
> > > > > > > > University
> > > > > > > > rconne01 at gmail.com
> > > > > > > > ryan.connelly at stonybrook.edu
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Ryan Connelly
> > > > > > M.S. Student in Atmospheric Sciences, Stony Brook
University
> > > > > > B.S. in Meteorology with Minors in Mathematics and GIS,
> Valparaiso
> > > > > > University
> > > > > > rconne01 at gmail.com
> > > > > > ryan.connelly at stonybrook.edu
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Ryan Connelly
> > > > M.S. Student in Atmospheric Sciences, Stony Brook University
> > > > B.S. in Meteorology with Minors in Mathematics and GIS,
Valparaiso
> > > > University
> > > > rconne01 at gmail.com
> > > > ryan.connelly at stonybrook.edu
> > > >
> > > >
> > >
> > >
> >
> >
> > --
> > Ryan Connelly
> > M.S. Student in Atmospheric Sciences, Stony Brook University
> > B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
> > University
> > rconne01 at gmail.com
> > ryan.connelly at stonybrook.edu
> >
> >
>
>
--
Ryan Connelly
M.S. Student in Atmospheric Sciences, Stony Brook University
B.S. in Meteorology with Minors in Mathematics and GIS, Valparaiso
University
rconne01 at gmail.com
ryan.connelly at stonybrook.edu
------------------------------------------------
More information about the Met_help
mailing list