conda-smithy init fails inside existing repo

I don’t really remember if conda smithy init has ever worked for me. I remember that I have been doing it manually. Today, finally, I have decided to post the traceback.

$ conda smithy init some_dir/recipe
Traceback (most recent call last):
  File "/Users/shadow_walker/anaconda/bin/conda-smithy", line 6, in <module>
    sys.exit(conda_smithy.cli.main())
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/conda_smithy/cli.py", line 289, in main
    args.subcommand_func(args)
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/conda_smithy/cli.py", line 116, in __call__
    generate_feedstock_content(feedstock_directory, args.recipe_directory, meta)
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/conda_smithy/cli.py", line 27, in generate_feedstock_content
    configure_feedstock.copytree(source_recipe_dir, target_recipe_dir)
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/conda_smithy/configure_feedstock.py", line 519, in copytree
    copy_file(s, d)
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/conda_smithy/feedstock_io.py", line 87, in copy_file
    repo.index.add([dst])
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/git/index/base.py", line 745, in add
    entries_added.extend(self._entries_for_paths(paths, path_rewriter, fprogress, entries))
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/git/util.py", line 55, in wrapper
    return func(self, *args, **kwargs)
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/git/index/util.py", line 84, in set_git_working_dir
    return func(self, *args, **kwargs)
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/git/index/base.py", line 631, in _entries_for_paths
    entries_added.append(self._store_path(filepath, fprogress))
  File "/Users/shadow_walker/anaconda/lib/python2.7/site-packages/git/index/base.py", line 590, in _store_path
    st = os.lstat(filepath)     # handles non-symlinks as well
OSError: [Errno 2] No such file or directory: './szip-feedstock/recipe/bld.bat'

[jakirkham]

Can you clarify “inside of an existing directory”?

Please see the edit above.

[shadowwalkersb]

However, creating a full blown repo inside of another repo can result in confusion and problems down the line. Not to mention this will result in commits with binary blobs of this repo inside of the other repo for any change that occurs (including git gc). This will also lead to performance issues down the line.

I may not be a git expert and I don’t claim that my git setup is ideal. But, AFAIK nested repos that are not submodules or subtrees are simply ignored by the parent repo. So, I don’t believe nested repos are that dangerous or dangerous at all. At least, I haven’t had any problems with my setup.

Discouraging such nested git repos, in a way, sounds like, you can’t have your bashrc types of files under version control, because your whole user account would be a git repo and everything else under it would be a nested repo. Another use case is keeping IDE settings under version control. For example, Eclipse has the concept of workspaces where project data and settings are saved. Well, all those projects may be git repos. Then, the suggestion is that I can’t keep my eclipse under version control unless my other projects are made submodules of my settings repo. Just doesn’t sound natural.

Although, my use cases are, pretty legit if you ask me, 😉 , I’m not asking for any unnecessary efforts put into conda-smithy development. Will take a look at PR #460 and thanks for putting it together!