Flag --edit locks nb systemwide until editor is closed about nb

[voxelv]

@eblume I noticed your use case is like what I requested under #300

[eblume]

Thanks @voxelv. It looks like my reading of #300's solution on lines 144-146, this plugin is not using --edit but rather nb edit. My issue is specifically related to using either nb add or nb edit while using the --edit flag to indicate that the editor should be invoked. In particular, my problem is that the resulting editor session seems to hold a "lock" of some sort on either nb itself (globally across all processes) or else on the underlying git repo. (Note the obscure corner case where nb edit is used with --edit which corresponds to the case that we want to append content to a document AND THEN, afterwards, open the resulting document in an editor.)

Can you tell me if the plugin exhibits this behavior as well? From my reading it doesn't seem to invoke --edit in any case so I don't know if a test can be made.

[eblume]

(An extra word aside here for my use case: I am trying to design an 'interactive log' of sorts where a long-lived editor session is set to autoload changes from scripted prompts while allowing the user to attach prose and links in context buffers. It's all very hacked and alpha right now, but the code is working and can be seen here.)