demands creating the lock file when I need just to get a list of submodules

[yarikoptic]

which obviously fails if I have no write access to that repository. Here it is invoked from within datalad (trimmed the backtrace to the point where we access GitPython's Repo in self.repo)

  File "/home/yoh/proj/datalad/datalad/datalad/support/gitrepo.py", line 1543, in get_submodules
    submodules = self.repo.submodules
  File "/home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/git/repo/base.py", line 276, in submodules
    return Submodule.list_items(self)
  File "/home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/git/util.py", line 932, in list_items
    out_list.extend(cls.iter_items(repo, *args, **kwargs))
  File "/home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/git/objects/submodule/base.py", line 1185, in iter_items
    entry = index.entries[index.entry_key(p, 0)]
  File "/home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/gitdb/util.py", line 237, in __getattr__
    self._set_cache_(attr)
  File "/home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/git/index/base.py", line 120, in _set_cache_
    fd = lfd.open(write=False, stream=False)
  File "/home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/gitdb/util.py", line 314, in open
    raise IOError("Lock at %r could not be obtained" % self._lockfilepath())
IOError: Lock at '/tmp/physionet/.git/index.lock' could not be obtained
()
> /home/yoh/proj/datalad/datalad/venv/dev-gitpython/local/lib/python2.7/site-packages/gitdb/util.py(314)open()
-> raise IOError("Lock at %r could not be obtained" % self._lockfilepath())
(Pdb) import git
(Pdb) git.__version__
'2.1.1'

FWIW regular 'git submodule' command works fine.

[yarikoptic]

any ideas on either this one could be reasonably resolved, or this locking is way too deep within GitPython?

[Byron]

@yarikoptic I believe this can be resolved by making the read-lock optional. Technically it's correct to obtain a lock for reading, just to hold off other writers, yet I believe in most cases, not having a read-lock would not be a problem. If one would only add a special case like: 'ignore the lock if it fails due to a permission error', this should already improve life for those who have no write permissions.
The latter is a valid case, which should certainly be supported.

[yarikoptic]

And thank you in advance for that! :-)

[Byron]

You are welcome :)!