cclib / cclib-data Goto Github PK

See #71 for an example of complications that would be avoided. It feels like it will be more valuable to have fixes and test together than tests and data.

This is both a current issue and copied from Sourceforge (#21). Namely, it would be prudent to provide some kind of license information concerning the data files we have here. We have had some discussion on the mailing list, but there doesn't seem to be one correct way to proceed.

On the one hand, the raw data -- which is what we are interested in -- is not subject to copyright. On the other hand, there are clear copyright statements in logfiles (referring to the software, not data), and any license cannot cover those portions of the logfiles.

It's common to add a regression file to this repo, and then write a regression test in the main repo at the same time. I feel like this process could be better, since we often just say "this needs to go in first" on one PR. Usually that's the data file, since the file is needed in the regression test itself.

Perhaps the right thing to do is move regressionfiles.txt to the main repo and not error out on missing files (or if there are less than N), so that here we just add files. Or maybe we just get rid of this file altogether, and track things in a different, smoother process. There needs to be a way.

See cclib/cclib#993.

mg_ct_sapt_gas.out.txt