[Scons-users] py3: string_escape?

[Mats Wichmann]

I finally tried our project, which now works with scons 3.0, using
Python 3 as well. Shock (sarcasm): it failed.

scons: Reading SConscript files ...
LookupError: unknown encoding: string_escape:
  File "/home/mats/iotivity.work/SConstruct", line 28:
    SConscript('build_common/SConscript')
  File
"/home/mats/virtualenv/scons3-py3/lib/python3.6/site-packages/scons-3.0.0/SCons/Script/SConscript.py",
line 607:
    return method(*args, **kw)
  File
"/home/mats/virtualenv/scons3-py3/lib/python3.6/site-packages/scons-3.0.0/SCons/Script/SConscript.py",
line 544:
    return _SConscript(self.fs, *files, **subst_kw)
  File
"/home/mats/virtualenv/scons3-py3/lib/python3.6/site-packages/scons-3.0.0/SCons/Script/SConscript.py",
line 253:
    call_stack[-1].globals)
  File "/home/mats/iotivity.work/build_common/SConscript", line 556:
    user_prefix = env.get('BUILD_DIR').encode('string_escape')


This was not something I expected, the experiment with conversion tools
hadn't suggested a problem here (unless I missed it). Of course, I also
didn't expect that some dev wasn't happy with fetching a string from the
environment without having to apply extra fiddling to it, either.  If I
change that to 'unicode_escape', have I caused problems for using the
project with Py2?  Is it better to just chop off the encode method call?