Develop Reference

Trouble in using Matplotlib-cpp

I encounter a problem on using the matplotlib-cpp in MacOS. Recently I download the matplotlibb-cpp into my macbook and try to run the sample code to see whether it works properly. I compile the sample code using
g++ test.cpp -std=c++17 \
-I /Users/pangkachun/matplotlib-cpp \
-I /Users/pangkachun/miniforge3/include/python3.9 \
-I /Users/pangkachun/miniforge3/lib/python3.9/site-packages/numpy/core/include \
-L /Users/pangkachun/miniforge3/lib \
-lpython3.9 \
-o plot
However, when I tried to execute the output file ./plot, I got an error message from terminal
dyld[47656]: Library not loaded: #rpath/libpython3.9.dylib
Referenced from: /Users/pangkachun/plot
Reason: tried: '/usr/local/lib/libpython3.9.dylib' (no such file), '/usr/lib/libpython3.9.dylib' (no such file)
zsh: abort ./plot
Could someone give me some guidelines on how to solve this problem?
""" Update"""
I use an alternative method to compile ./plot which symlink the miniforge3/lib that containing libpython3.9.dylib to /usr/local/lib. This /usrlocal/lib is created by my own under symlink. Although this works, I hope there is a "normal" method to deal with this issue.

uwsgi can't build plugin.so for python3.6

I was trying to build uwsgi+django server. I need to build uwsgi python36_plugin.so for python3.6 using this command:
make PROFILE=nolang
PYTHON=python3 ./uwsgi --build-plugin "plugins/python python36"
It didn't work:
[root#izuf64a9gck81cz uwsgi-2.0.15]# PYTHON=python ./uwsgi --build-plugin "plugins/python python36"
*** uWSGI building and linking plugin from plugins/python ***
[gcc -pthread] python36_plugin.so
/usr/bin/ld: /usr/python3/lib/python3.6/config-3.6m-x86_64-linux-gnu/libpython3.6m.a(abstract.o): relocation R_X86_64_32S against `_Py_NotImplementedStruct' can not be used when making a shared object; recompile with -fPIC
/usr/python3/lib/python3.6/config-3.6m-x86_64-linux-gnu/libpython3.6m.a: error adding symbols: Bad value
collect2: error: ld returned 1 exit status
*** unable to build python36 plugin ***
When i use build python2.7 version,it turns out right
PYTHON=python2.7 ./uwsgi --build-plugin "plugins/python python27"

My problem is I have two version of python 3, the default python3 is python3.4, and the second one is python3.6, and I want to use python3.6 for some of my django project so I need to build the plugins and I got same problem you got.
It took me couple hours to solved this issue. Here is what I did from here:
Reconfigure my python3.6
cd Python-3.6.4/
./configure --enable-shared --enable-optimizations --prefix=/usr/local LDFLAGS="-Wl,--rpath=/usr/local/lib"
Get in to your uwsgi folder and try to build again
PYTHON=python3.6 ./uwsgi --build-plugin "plugins/python python36"
Goodluck! May it solve your prob too!

f2py: command not found

Mac OS Sierra.
In terminal, while trying to execute a .sh file that involves a Python script with Fortran portions, I got this message:
lenscov-master Roger$ ./run_covariances.sh recompile
rm *.o *.so
rm: *.so: No such file or directory
make: [clean] Error 1 (ignored)
gfortran -fPIC -c *.f -lgfortran -lifcore
f2py-2.7 -c loop_lensing.f90 *.o -m loop_lensing --opt=-ffixed-line-length-
none --opt=-O3
/bin/sh: f2py-2.7: command not found
make: *** [loop_lensing] Error 127
However, it is fine just test-running the .f90 file:
f2py -c loop_lensing.F90 -m test
returns a bunch of normal-looking information.
Does anyone maybe know what is going on?

it is fine just test-running the .f90 file
means you have f2py installed. But the script use f2py-2.7
You need to check the version of installed f2py, and make sure f2py-2.7 redirects to the corrent f2py.

The script (or a Makefile) tries to invoke f2py-2.7 instead of f2py. If it is your script or you can easily adjust it, rename all f2py-2.7 to f2py. Otherwise you have to tell us more about the script and show the code of the script.
As foodtooth says, make sure you have f2py 2.7 and not f2py-3, but I don't expect that to be a problem.
You could also make a symlink f2py -> f2py-2.7 for the script.

cannot compile cpp files under the python pyc script file

I am trying to compile & execute my hw cpp file under the python script file which we are given by lecturer. the how-to-manual.pdf he sent us it says use:
c:\>python ./submit.pyc problemID -u username -p password -b //submit.pyc is already given to us
and here is the manifest.txt we are given:
[main]
problem = gc
build =
g++ main.cpp -o solver
run =
./solver %f
my cpp file works normally like this:
./solver input_file
However, I am trying (I have to) to do this under the windows OS. I have Python 2.7.x installed and python.exe is in the Command PATH. I can't run it under the linux ssh sytem because there is 2.4.x python installed and I can't touch it (school's system).
Anyway, when I execute the line above, it returns me:
Command execution failed:
g++ solver.cpp -o solver
I think I told everything I can. So, any idea that what I have to do else? except asking to lecturer:)

For the above to work it needs to be able to find g++ so you need to add the directory that it resides in to the PATH environment variable. This can be done from within your python script or on the command line with:
path=Where\g++\lives;%path%
This will only apply within the current DOS session.
Or you can add it permanenty through system settings->advanced settings->environmental variables
You could also look at using a python virtual environments on the schools linux system.

Python missing or unusable while compiling gdb under Cent OS 5

I checked this question and it is not applicable.
Python missing or unusable error while cross compiling GDB
I'm trying to build gdb 7.5 with python support on Cent OS 5.
configure --with-expat --with-python
Does not log any errors and the Python devel files are installed.
I set LDFLAGS="L/usr/local/expat-2.1 -L/opt/python27/lib
/opt/python27/lib contains:
rwxrwxrwx 1 root root 19 Mar 14 15:35 libpython2.7.so -> libpython2.7.so.1.0
-r-xr-xr-x 1 root root 1893448 Aug 30 2012 libpython2.7.so.1.0
drwxr-xr-x 2 root root 4096 Mar 14 15:35 pkgconfig
drwxr-xr-x 27 root root 20480 Mar 14 21:49 python2.7
Running make produces this error:
checking whether to use python... yes
checking for python... /opt/python27/bin/python
checking for python2.7... no
configure: error: python is missing or unusable
/opt/python27/bin/python exists has the correct permissions and is executable.
Why can't the link process find python?

The next line AFTER the configure: error: line should be something like:
make[1]: [configure-gdb]: ....
telling you that this error comes from running configure in the gdb subdirectory. So you should look in the config.log file there. This should show you something that configure tried to do that failed -- probably a chunk of C code that failed to compile, with an error message from the compiler something like 'can't find include file "Python.h"'. This tells you what the problem is -- it couldn't find the python header file used to embed python in gdb.
Now your problem may be slightly different, but its probably related to not having the python embedding components properly installed.

Why can't the link process find python?
It's not the link process; it's the configure process in one of the subdirectories.
Find out which subdirectory it is (make should say Leaving directory XXXX), and look in config.log in that subdirectory for what went wrong.
The reason I posted is that config.log does not point to a problem, yet when I run make I receive the error from the original post
The make process runs additional configure processes. Your error is clearly from that additional process. You claim to have checked all config.logs in all subdirectories, but that appears unlikely given the symptoms you've described.

Sorry for making you wait 10 years for an answer.
In my case I looked under the gdb/config.log file and found this:
configure:10369: checking for python2.7
configure:10387: gcc -o conftest -g -O2 -I/home/user/stevie-bsp/toolchain-build/builds/destdir/aarch64-host-linux-gnu/include -I/usr/include/python2.7 -I/usr/include/python2.7 -static-libstdc++ -static-libgcc conftest.c -ldl -lncursesw -lm -ldl -lpthread -ldl -lutil -lm -lpython2.7 -Xlinker -export-dynamic -Wl,-O1 -Wl,-Bsymbolic-functions >&5
conftest.c:54:10: fatal error: Python.h: No such file or directory
54 | #include "Python.h"
| ^~~~~~~~~~
compilation terminated.
Looks like it is a misleading error message when the Python 2 headers aren't installed. I fixed that on Ubuntu with apt install python2-dev; on CentOS the package name would vary.
Secondly, the command python must start the Python 2 interpreter. The build process will not give a clear error about this and silently fails to query the include path for the Python headers. I resolved it by using the python-is-python2 package on Ubuntu. I'm not sure about CentOS but in the worst case you can create a symlink.