Clarify copyrights

[satta]

Hi,

I am packaging icdiff for Debian and I am looking to sort out some minor licensing issues.
The Python license linked by you in README.md specifies that the Python copyright:

Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006 Python Software Foundation; All Rights Reserved

must be retained in derivative based on it (section 2). This is missing in the script header. Also, section 3 states that

In the event Licensee prepares a derivative work that is based on
or incorporates Python or any part thereof, and wants to make
the derivative work available to others as provided herein, then
Licensee hereby agrees to include in any such work a brief summary of
the changes made to Python.

It would be nice to have such a brief summary.

Also, the UTF-8 decoder capability and stress test file (tests/b/1 and tests/input-3.txt) seems like it was not prepared by you but by Markus Kuhn. The version of the file you are distributing with icdiff does not include any license information, but a newer version does (https://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt, CC-BY). It would be nice to have this file updated or the license specified by you.

[jeffkaufman]

Added, with cdede39, and released as 1.7.6.

[satta]

Excellent, thanks! I wonder if you could also update the test files so they also contain a copyright and license info. See my original comment above, as well as #70.

[jeffkaufman]

All test files except input-3 (which you fixed) are trivial; do you really think it's worth it?

[satta]

The problem is that in Debian we need to be able to prove that a given file is under a free license, and that is only possible in the newer version of the UTF8 file (input-3.txt and tests/b/1) I added in #70 as only this one has 'CC-BY' explicitly mentioned in the text.
Please note that with your commit 440de77 the file tests/b/1 is still the old one without the license notice. This will still impede redistribution.
The rest of the changes in #70 was necessary to keep the tests passing as the new UTF8 file is slightly different than the previous one, causing slightly different output. I did a --regold to update the reference output. I hope that was OK.
AFAICS to fully solve this, all changes would need to be merged, unless I missed something.

[jeffkaufman]

Sorry, I've fixed tests/b/1 and regolded. Version 1.8.1 should be what you want. I think we're all set now?

[satta]

That was quick, many thanks! Unfortunately, there's one thing left: I now run into the problem described by #72, #73 and #75. I suggest you separate the icdiff tool from the module to be loaded so you do not have to rename the tool to icdiff.py -- globally available commands with script extensions are discouraged in Debian and I would have to work around it. Also, there are good arguments for not doing so (https://wiki.debian.org/UpstreamGuide#Language_extensions_in_scripts).
That indeed seems to be the last remaining request :)

[xrat]

Great work, guys! Looking forward to seeing icdiff packaged for Debian. Thanks a lot.

[satta]

Regarding my last comment, please also see PR #79.

[satta]

Just for the record, I would be really happy to see #75 merged as it is a better solution than my proposal. That would get me going again for finishing up the packaging.

[jeffkaufman]

The problem with #75 is the same as #79: I want to still have a icdiff on your PATH as icdiff and not icdiff.py but I also want to provide a simple copy-paste install for people who want that.

[satta]

Ah, I see and I do understand that. I was just asking because in the current 1.8.1 the setup.py is broken due to #72, so I won't be able to package this 'the Python way' without either locally reverting the module functionality (commit 46d5a72) or applying #75 to the Debian version. Since I usually like to stick with the original upstream I was looking for a way to sort it out...
Does anyone possibly have other suggestions?