Inline calculations storing source 1026

[lekah]

This would resolve #1026 by removing the storage of the file source code.

[sphuber]

Looks fine to me, although I did notice that this was the most "hardcore" solution of all the potential solutions that were discussed in the issue. Would it not still be desirable to keep the source file in the repository. Personally I am not sure it would be useful given the problem of imports, but I just wanted to make sure that everyone agreed on this solution since I was not involved in the original discussion

[nmounet]

We really need the source code. We should put it in the repository. Otherwise, what the point of keeping provenance if one cannot find back what the inline calc/workfunction actually did?

[lekah]

@nmounet The source code of the function is still being stored in the database, and can be queried. I only removed the storage of the entire file. This was actually redundant in a way, because in this way the inline function code was stored twice. Was there a misunderstanding? If you need the entire source file, the repository is indeed the right place....

[giovannipizzi]

@nmounet Indeed saving the full file was just a 'backup solution', still not enough if the actual code is imported from somewhere else. The storage of the function source code is still there:
https://github.com/aiidateam/aiida_core/pull/1082/files#diff-fca7583529982daff057e6983b21e26dL31

For me it is ok to go like this (even if in the future we could move the source code into the repo, if we still have problems).

[nmounet]

Ok ! Apologies that I did not see the source code was indeed still stored as an Attribute. My misunderstanding.
Yet, I still thing we should put the full source code file in the repo, instead of putting the code of the inline in an Attribute. But that's for another issue.