This was giving many warnings like this one in the SciPy build:
```
scipy/special/_specfunmodule.c: In function 'complex_double_from_pyobj':
scipy/special/_specfunmodule.c:198:47: warning: passing argument 1 of 'PyArray_DATA' from incompatible pointer type [-Wincompatible-pointer-types]
  198 |         (*v).r = ((npy_cdouble *)PyArray_DATA(arr))->real;
      |                                               ^~~
      |                                               |
      |                                               PyObject * {aka struct _object *}
In file included from /home/rgommers/code/numpy/numpy/core/include/numpy/ndarrayobject.h:12,
                 from /home/rgommers/code/numpy/numpy/core/include/numpy/arrayobject.h:5,
                 from /home/rgommers/code/numpy/numpy/f2py/src/fortranobject.h:16,
                 from scipy/special/_specfunmodule.c:22:
/home/rgommers/code/numpy/numpy/core/include/numpy/ndarraytypes.h:1524:29: note: expected 'PyArrayObject *' {aka 'struct tagPyArrayObject *'} but argument is of type 'PyObject *' {aka 'struct _object *'}
 1524 | PyArray_DATA(PyArrayObject *arr)
      |              ~~~~~~~~~~~~~~~^~~
```

Fixing pointer mismatches is important for Pyodide/Emscripten.

TST: added test for issue #18684

ENH: f2py opens files with correct encoding, fixes #635

TST: added test for issue #6308

TST: added test for issue #4519

TST: added test for issue #3425

ENH: Implement user-defined hooks support for post-processing f2py data structure. Implement character BC hook.

ENH: Add support for detecting utf-16 and utf-32 encodings.

Closes https://github.com/rgommers/scipy/issues/125
This is a follow-up of gh-18910 (included in 1.20.3 and 1.21.0), which
looks incorrect.

This fixes warnings like these when building SciPy with Mingw-w64:
```
[94/1557] Compiling C object scipy/linalg/_flapack.cp39-win_amd64.pyd.p/meson-generated_..__flapackmodule.c.obj
In file included from C:\hostedtoolcache\windows\Python\3.9.9\x64\lib\site-packages\numpy\core\include/numpy/ndarraytypes.h:4,
                 from C:\hostedtoolcache\windows\Python\3.9.9\x64\lib\site-packages\numpy\core\include/numpy/ndarrayobject.h:12,
                 from C:\hostedtoolcache\windows\Python\3.9.9\x64\lib\site-packages\numpy\core\include/numpy/arrayobject.h:4,
                 from C:\hostedtoolcache\windows\Python\3.9.9\x64\lib\site-packages\numpy\f2py\src/fortranobject.h:13,
                 from scipy/linalg/_flapackmodule.c:16:
scipy/linalg/_flapackmodule.c:1248:1: warning: 'thread' attribute directive ignored [-Wattributes]
 1248 | static F2PY_THREAD_LOCAL_DECL cb_cselect_in_gees__user__routines_t *_active_cb_cselect_in_gees__user__routines = NULL;
      | ^~~~~~
scipy/linalg/_flapackmodule.c:1410:1: warning: 'thread' attribute directive ignored [-Wattributes]
 1410 | static F2PY_THREAD_LOCAL_DECL cb_zselect_in_gees__user__routines_t *_active_cb_zselect_in_gees__user__routines = NULL;
      | ^~~~~~
scipy/linalg/_flapackmodule.c:1572:1: warning: 'thread' attribute directive ignored [-Wattributes]
 1572 | static F2PY_THREAD_LOCAL_DECL cb_sselect_in_gees__user__routines_t *_active_cb_sselect_in_gees__user__routines = NULL;
      | ^~~~~~
scipy/linalg/_flapackmodule.c:1738:1: warning: 'thread' attribute directive ignored [-Wattributes]
 1738 | static F2PY_THREAD_LOCAL_DECL cb_dselect_in_gees__user__routines_t *_active_cb_dselect_in_gees__user__routines = NULL;
      | ^~~~~~

...
```

Also fixes gh-19437, where `__STDC_NO_THREADS__` should be defined but
isn't (this seems to be more common, e.g. one can reports for ICC about
this).

Co-authored-by: Sebastian Berg <sebastian@sipsolutions.net>