BF: there is no exc variable, raising NotASurrogateError if that is the right thing todo

[yarikoptic]

a blind fix for a blind (not tested) line of code which has some variable lost from old pre-refactoring code I guess ;)

PS FWIW -- we have managed to run into it in quite a strange place, so not really sure how yet, here is our backtrace with 2.1.0 GitPython on Debian stretch

======================================================================
ERROR: Failure: NameError (global name 'exc' is not defined)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/nose/loader.py", line 418, in loadTestsFromName
    addr.filename, addr.module)
  File "/usr/lib/python2.7/dist-packages/nose/importer.py", line 47, in importFromPath
    return self.importFromDir(dir_path, fqname)
  File "/usr/lib/python2.7/dist-packages/nose/importer.py", line 94, in importFromDir
    mod = load_module(part_fqname, fh, filename, desc)
  File "/home/buildbot/datalad-pr-docker-dl-nd90/build/datalad/tests/test__main__.py", line 17, in <module>
    from .. import __main__, __version__
  File "/home/buildbot/datalad-pr-docker-dl-nd90/build/datalad/__main__.py", line 13, in <module>
    from .auto import AutomagicIO
  File "/home/buildbot/datalad-pr-docker-dl-nd90/build/datalad/auto.py", line 20, in <module>
    import h5py
  File "/usr/lib/python2.7/dist-packages/h5py/__init__.py", line 60, in <module>
    from .tests import run_tests
  File "/usr/lib/python2.7/dist-packages/h5py/tests/__init__.py", line 15, in <module>
    from . import old, hl
  File "/usr/lib/python2.7/dist-packages/h5py/tests/old/__init__.py", line 4, in <module>
    from . import ( test_attrs,
  File "/usr/lib/python2.7/dist-packages/h5py/tests/old/test_group.py", line 41, in <module>
    fsencode(u"α")
  File "/usr/lib/python2.7/dist-packages/h5py/_hl/compat.py", line 68, in fsencode
    return filename.encode(encoding, errors)
  File "/usr/lib/python2.7/dist-packages/git/compat.py", line 180, in surrogateescape_handler
    decoded = replace_surrogate_encode(mystring)
  File "/usr/lib/python2.7/dist-packages/git/compat.py", line 207, in replace_surrogate_encode
    raise exc
NameError: global name 'exc' is not defined

[yarikoptic]

looking at that traceback -- are you overloading some str.encode builtin methods? then better make sure they do not blow, and if they blow -- just relay to original/non-overloaded ;)

[codecov-io]

Codecov Report

Merging #598 into master will decrease coverage by -7.48%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master     #598      +/-   ##
==========================================
- Coverage   94.48%   87.01%   -7.48%     
==========================================
  Files          63       63              
  Lines        9958     9956       -2     
==========================================
- Hits         9409     8663     -746     
- Misses        549     1293     +744

Impacted Files	Coverage Δ
git/compat.py	40.74% <0%> (-22.45%)	❌
git/test/test_tree.py	18.33% <0%> (-81.67%)	❌
git/test/test_remote.py	31.33% <0%> (-67.03%)	❌
git/test/test_submodule.py	59.92% <0%> (-39.3%)	❌
git/odict.py	20% <0%> (-30%)	❌
git/objects/submodule/root.py	64.39% <0%> (-28.04%)	❌
git/remote.py	78.08% <0%> (-14.89%)	❌
git/test/test_base.py	90.8% <0%> (-8.05%)	❌
git/test/lib/asserts.py	61.53% <0%> (-7.7%)	❌
... and 23 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 72dddc7...de4999b. Read the comment docs.

[Byron]

Thanks a lot!

[yarikoptic]

@Byron -- I am still curious about why this all happened since we are importing another module entirely, so smells like GitPython is messing some unrelated things up ;) for us it happens only on that debian buildbot

[Byron]

@yarikoptic To me all this belongs to the field of ... python is an inherently unsafe language which by now is plagued by plenty of systemic problems, many of which GitPython suffers from too. Understanding how it happened doesn't matter (to me) knowing that it will never be fixed or functioning properly in all cases. Maybe a bit dark, but the opinion I arrived at over the years.

[yarikoptic]

ok @Byron -- now had some conscious look at it, #603 should fix my fix and the original issue ;)