freetype build won't become local

[prhbrt]

I have a git checkout of matplotlib on which I would like to run the tests. In order to do so I have set local_freetype=True in setup.cfg and ran the following in a virtual environment:

(matplotlib) matplotlib$ python3.5 setup.py build
============================================================================
Edit setup.cfg to change the build options

BUILDING MATPLOTLIB
        matplotlib: yes [2.0.0b3+1919.g2dfef6a]
            python: yes [3.5.1+ (default, Mar 30 2016, 22:46:26)  [GCC
                    5.3.1 20160330]]
          platform: yes [linux]

REQUIRED DEPENDENCIES AND EXTENSIONS
             numpy: yes [version 1.11.1]
               six: yes [using six version 1.10.0]
          dateutil: yes [using dateutil version 2.5.3]
       functools32: yes [Not required]
      subprocess32: yes [Not required]
              pytz: yes [using pytz version 2016.6.1]
            cycler: yes [using cycler version 0.10.0]
           tornado: yes [tornado was not found. It is required for the
                    WebAgg backend. pip/easy_install may attempt to
                    install it after matplotlib.]
         pyparsing: yes [using pyparsing version 2.1.5]
            libagg: yes [pkg-config information for 'libagg' could not
                    be found. Using local copy.]
          freetype: yes [Using local version for testing]
               png: yes [version 1.2.54]
             qhull: yes [pkg-config information for 'qhull' could not be
                    found. Using local copy.]

OPTIONAL SUBPACKAGES
       sample_data: yes [installing]
          toolkits: yes [installing]
             tests: yes [using nose version 1.3.7 / using unittest.mock]
    toolkits_tests: yes [using nose version 1.3.7 / using unittest.mock]

OPTIONAL BACKEND EXTENSIONS
            macosx: no  [Mac OS-X only]
            qt5agg: yes [installing, Qt: 5.7.0, PyQt: 5.7.0; PySide2 not
                    found]
            qt4agg: yes [installing, Qt: 5.5.1, PyQt: 5.5.1; PySide not
                    found]
           gtk3agg: no  [Requires pygobject to be installed.]
         gtk3cairo: no  [Requires cairocffi or pycairo to be installed.]
            gtkagg: no  [Requires pygtk]
             tkagg: yes [installing; run-time loading from Python Tcl /
                    Tk]
             wxagg: no  [requires wxPython]
               gtk: no  [Requires pygtk]
               agg: yes [installing]
             cairo: no  [cairocffi or pycairo not found]
         windowing: no  [Microsoft Windows only]

OPTIONAL LATEX DEPENDENCIES
            dvipng: yes [version 1.15]
       ghostscript: yes [version 9.18]
             latex: yes [version 3.14159265]
           pdftops: yes [version 0.41.0]

OPTIONAL PACKAGE DATA
              dlls: no  [skipping due to configuration]

running build
running build_py
copying lib/matplotlib/mpl-data/matplotlibrc -> build/lib.linux-x86_64-3.5/matplotlib/mpl-data
UPDATING build/lib.linux-x86_64-3.5/matplotlib/_version.py
set build/lib.linux-x86_64-3.5/matplotlib/_version.py to '2.0.0b3+1919.g2dfef6a'
running build_ext
(matplotlib) matplotlib$ PYTHONPATH=build/lib.linux-x86_64-3.5/ python3.5 tests.py --nocapture
Python byte-compilation optimization level: 0
/home/herbert/Repositories/matplotlib/build/lib.linux-x86_64-3.5/matplotlib/__init__.py:1580: UserWarning: matplotlib is not built with the correct FreeType version to run tests.  Set local_freetype=True in setup.cfg and rebuild. Expect many image comparison failures below. Expected freetype version 2.6.1. Found freetype version 2.6.1. Freetype build type is not local
  "" if ft2font.__freetype_build_type__ == 'local' else "not "

as you can see, the build command reports that the used freetype should be local, but when running the tests, they are not local as reported by the error. Why is this the case?

[prhbrt]

I ran python3.5 setup.py clean before, but this was not enough. rm -rf build/ however, and then repeating the above seems to work though. This is not what I would expect.

[jenshnielsen]

We dont implement setup.py clean it's inherited from distribute/setuptools and as far as I know generally known to be broken. I don't think there is much we can do about that apart from perhaps replacing it with a no op that prints a warning.

[jenshnielsen]

It seems like if you pass all to setup.py clean it does the right thing. I.e. python setup.py clean --all
See
http://stackoverflow.com/questions/17804185/what-does-setup-py-clean-exactly-do

[tacaswell]

git clean -xfd is your (some what dangerous) friend 👿

[prhbrt]

For me the problem is solved, and this issue may be closed. Thank you.