Extremely slow installation

[homm]

$ time pip install pgmagick
Collecting pgmagick
  Downloading pgmagick-0.6.7.tar.gz (357kB)
    100% |████████████████████████████████| 358kB 1.2MB/s 
Building wheels for collected packages: pgmagick
  Running setup.py bdist_wheel for pgmagick ... done
  Stored in directory: /home/cuda/.cache/pip/wheels/38/9b/ea/8558bf96d0534e2b5242a2f78a8e0d72c391788b8b07cfeaab
Successfully built pgmagick
Installing collected packages: pgmagick
Successfully installed pgmagick-0.6.7

real	5m2.436s
user	4m37.144s
sys	0m17.372s

Comparing with competitors:

$ time pip install wand
Collecting wand
  Downloading Wand-0.4.4-py2-none-any.whl (71kB)
    100% |████████████████████████████████| 81kB 881kB/s 
Installing collected packages: wand
Successfully installed wand-0.4.4

real	0m0.855s
user	0m0.504s
sys	0m0.044s

$ time pip install opencv-python
Collecting opencv-python
  Downloading opencv_python-3.3.0.10-cp27-cp27mu-manylinux1_x86_64.whl (15.4MB)
    100% |████████████████████████████████| 15.5MB 107kB/s 
Collecting numpy>=1.11.1 (from opencv-python)
  Downloading numpy-1.13.3-cp27-cp27mu-manylinux1_x86_64.whl (16.6MB)
    100% |████████████████████████████████| 16.7MB 101kB/s 
Installing collected packages: numpy, opencv-python
Successfully installed numpy-1.13.3 opencv-python-3.3.0.10

real	0m7.949s
user	0m4.204s
sys	0m0.348s

$ time pip install Pillow-simd
Collecting Pillow-simd
  Downloading Pillow-SIMD-4.3.0.post0.tar.gz (13.8MB)
    100% |████████████████████████████████| 13.8MB 117kB/s 
Requirement already satisfied: olefile in /home/cuda/envs/pillow/lib/python2.7/site-packages (from Pillow-simd)
Building wheels for collected packages: Pillow-simd
  Running setup.py bdist_wheel for Pillow-simd ... done
  Stored in directory: /home/cuda/.cache/pip/wheels/88/8d/86/f591134ede4f540e900631c35d69f284f2928f4a60f88cacdf
Successfully built Pillow-simd
Installing collected packages: Pillow-simd
Successfully installed Pillow-simd-4.3.0.post0

real	0m9.814s
user	0m13.552s
sys	0m1.028s

[komackaj]

Related to #9 - installation speed will improve (and no compiler and development libraries are needed) after providing a wheel.