Compile errors

[dbkaplun]

 02:10:17  dan@elise:~/FHEW   master ✔ 
$ make
g++ -Wall -O3 -c distrib.cpp
In file included from distrib.cpp:2:
In file included from ./distrib.h:4:
./params.h:5:10: fatal error: 'fftw3.h' file not found
#include <fftw3.h>
         ^
1 error generated.
make: *** [distrib.o] Error 1

 02:10:44  ✘ 130  dan@elise:~/FHEW   master ✔ 
$ brew install fftw
==> Downloading https://homebrew.bintray.com/bottles/fftw-3.3.4_1.yosemite.bottle.tar.gz
######################################################################## 100.0%
==> Pouring fftw-3.3.4_1.yosemite.bottle.tar.gz
🍺  /usr/local/Cellar/fftw/3.3.4_1: 39 files, 11M

 02:10:53  dan@elise:~/FHEW   master ✔ 
$ make
g++ -Wall -O3 -c distrib.cpp
In file included from distrib.cpp:2:
In file included from ./distrib.h:4:
./params.h:26:31: warning: signed shift result (0x10000000800) requires 42 bits to represent, but 'ZmodQ' (aka 'int') only has 32 bits [-Wshift-overflow]
const ZmodQ vgprime[3] = {v, v<<11, v<<22};
                             ~^ ~~
./params.h:26:38: warning: signed shift result (0x8000000400000) requires 53 bits to represent, but 'ZmodQ' (aka 'int') only has 32 bits [-Wshift-overflow]
const ZmodQ vgprime[3] = {v, v<<11, v<<22};
                                    ~^ ~~
2 warnings generated.
g++ -Wall -O3 -c FFT.cpp
In file included from FFT.cpp:3:
In file included from ./FFT.h:4:
./params.h:26:31: warning: signed shift result (0x10000000800) requires 42 bits to represent, but 'ZmodQ' (aka 'int') only has 32 bits [-Wshift-overflow]
const ZmodQ vgprime[3] = {v, v<<11, v<<22};
                             ~^ ~~
./params.h:26:38: warning: signed shift result (0x8000000400000) requires 53 bits to represent, but 'ZmodQ' (aka 'int') only has 32 bits [-Wshift-overflow]
const ZmodQ vgprime[3] = {v, v<<11, v<<22};
                                    ~^ ~~
FFT.cpp:24:22: error: expected ')'
    res[k] = (double complex) out[2*k+1];
                     ^
FFT.cpp:24:14: note: to match this '('
    res[k] = (double complex) out[2*k+1];
             ^
FFT.cpp:24:14: error: C-style cast from 'double *' to 'double' is not allowed
    res[k] = (double complex) out[2*k+1];
             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
FFT.cpp:29:26: error: expected ')'
    out[2*k+1] = (double complex) val[k]/N;
                         ^
FFT.cpp:29:18: note: to match this '('
    out[2*k+1] = (double complex) val[k]/N;
                 ^
FFT.cpp:29:18: error: C-style cast from 'const double *' to 'double' is not allowed
    out[2*k+1] = (double complex) val[k]/N;
                 ^~~~~~~~~~~~~~~~~~~~~~~
FFT.cpp:30:26: error: expected ')'
    out[2*k]   = (double complex) 0;
                         ^
FFT.cpp:30:18: note: to match this '('
    out[2*k]   = (double complex) 0;
                 ^
FFT.cpp:30:16: error: array type 'fftw_complex' (aka 'double [2]') is not assignable
    out[2*k]   = (double complex) 0;
    ~~~~~~~~   ^
2 warnings and 6 errors generated.
make: *** [FFT.o] Error 1

[lducas]

This issue seems to be related to your OS (I had several report of the same issue on MacOS). This issue typically don't arise with linux systems.

[thomasdziedzic]

just encountered this same issue:

g++ is actually linked to clang on mac:

$ g++ -v
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 7.3.0 (clang-703.0.29)
Target: x86_64-apple-darwin15.4.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin

[lducas]

I am sorry I can mot reproduce the error to debug it, as I don't own this OS. (also, FHEW is a project I'm happy to share for reproducibility of scientific claims, but I have no time nor intention to maintain it). I'd be happy to accept a pull request if you have a fix though.

Best
-- Leo