* Python 3 uses unicode strings everywhere, but Python 2 has two different string types - str and unicode
* If you call str(s) on a unicode string with extended characters, it tries to convert to ascii (the default python 2 encoding) and it fails with the exception above
* If we use unicode types everywhere, we'd be OK.
* _ returns unicode strings (because we called install with unicode=1). This is good.
* I think this bug report relates to this call:
(nova.compute.manager): TRACE: raise Error(str(e))
* That call was removed in this commit, so this particular bug has been fixed:
git show 19b3b7a0c50dd6a68238daf8
* However, this does mean that everywhere we call str(x) is potentially problematic (?). I'm going to open another bug on this issue.
Here what I've diagnosed:
* Python 3 uses unicode strings everywhere, but Python 2 has two different string types - str and unicode
* If you call str(s) on a unicode string with extended characters, it tries to convert to ascii (the default python 2 encoding) and it fails with the exception above
* If we use unicode types everywhere, we'd be OK.
* _ returns unicode strings (because we called install with unicode=1). This is good.
* I think this bug report relates to this call: manager) : TRACE: raise Error(str(e))
(nova.compute.
* That call was removed in this commit, so this particular bug has been fixed: 68238daf8
git show 19b3b7a0c50dd6a
* However, this does mean that everywhere we call str(x) is potentially problematic (?). I'm going to open another bug on this issue.