Re: [netcdfgroup] why is ncdump a shell script in 4.6.2?

Ward, et al.,

A curious result: I saw Charlie Zender's thread about multiple definitions of nc_open_men
in 4.6.2:

https://sourceforge.net/p/nco/discussion/9831/thread/e6faa0b8c1/?limit=25#7606/4cf2

I decided, for kicks, to try the one-line patch he suggested (including ncexternal.h in
netcdf_mem.h).  I entered the commands "make clean", "make check" (no problems),
and "make install". Lo and behold, I no longer had the script version of ncdump. I don't really suspect the one-line modification had anything to do with it (although libnetcdf.so now only has one definition for nc_open_mem). I think the "make clean" fixed whatever
issue I was experiencing.

FYI,
-Hoop

On 12/10/18 4:37 PM, Mark Rivers wrote:
What does "make -j 12" mean, anyway?  I've never seen such in the instructions 
for building NetCDF.

It means run parallel make with up to 12 threads.  It does not affect the 
result, it just speeds it up.

Mark

________________________________
From: netcdfgroup-bounces@xxxxxxxxxxxxxxxx <netcdfgroup-bounces@xxxxxxxxxxxxxxxx> on 
behalf of Don Hooper (NOAA Affiliate) <don.k.hooper@xxxxxxxx>
Sent: Monday, December 10, 2018 4:09 PM
To: Ward Fisher; netcdfgroup@xxxxxxxxxxxxxxxx
Subject: Re: [netcdfgroup] why is ncdump a shell script in 4.6.2?

Ward,

I did configure with just a --prefix argument.  I did "make check" followed by
"make install".  So, very vanilla.  The bin directory under the directory
specified
in the --prefix argument is where I found the shell script for ncdump.  My
particular test
was a validated OPeNDAP access to the RDA TDS.  I had already found the
binary executable version of ncdump in the .libs directory.  It worked for my
test.  The script did not; it appeared to use the wrong version of the curl
library.
What does "make -j 12" mean, anyway?  I've never seen such in the instructions
for building NetCDF.

Thanks,
-Hoop

On 12/10/18 12:23 PM, Ward Fisher wrote:
Depending how netCDF has been configured using the `configure` script,
   'binaries' are often represented as scripts.  I believe it is only when
building as a shared library.  For example , looking at the
`netcdf-c/ncdump/ncdump` file after configuring/building with

      $ ./configure --prefix=$HOME/ttmp && make -j 12

I see the following header info:

#! /bin/sh

# ncdump - temporary wrapper script for .libs/ncdump
# Generated by libtool (GNU libtool) 2.4.6
#
# The ncdump program cannot be directly executed until all the libtool
# libraries that it depends on are installed.
#
# This wrapper script should never be moved out of the build directory.
# If it is, it will not operate correctly.

However, once I run 'make install' and look at $HOME/ttmp/bin/ncdump, I see
it's a binary file (as expected).

So I guess my question is, does this describe what you're seeing? Are you
running `make install` and seeing a shell script where you would expect to see
the binary?  If you are seeing a shell script in the build directory, it is
for the reasons outlined above; there is a binary ncdump file, but it is in
netcdf-c/ncdump/.libs, and cannot be executed directly until 'make install'
has been run.

-Ward
On Dec 7, 2018, 6:09 PM -0500, Don Hooper (NOAA Affiliate)
<don.k.hooper@xxxxxxxx>, wrote:
Subject says it all: why is ncdump a shell script in 4.6.2? -Hoop

_______________________________________________
NOTE: All exchanges posted to Unidata maintained email lists are
recorded in the Unidata inquiry tracking system and made publicly
available through the web. Users who post to any of the lists we
maintain are reminded to remove any personal information that they
do not want to be made public.


netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:
http://www.unidata.ucar.edu/mailing_lists/

_______________________________________________
NOTE: All exchanges posted to Unidata maintained email lists are
recorded in the Unidata inquiry tracking system and made publicly
available through the web.  Users who post to any of the lists we
maintain are reminded to remove any personal information that they
do not want to be made public.


netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe,  visit: 
http://www.unidata.ucar.edu/mailing_lists/




  • 2018 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdfgroup archives: