-----Original Message----- From: Boost On Behalf Of Paul A Bristow via Boost Sent: 27 July 2019 10:14 To: boost@lists.boost.org Cc: pbristow@hetp.u-net.com Subject: Re: [boost] Making copyright holders easier to parse

...

-----Original Message----- From: Boost On Behalf Of Antony Polukhin via Boost Sent: 27 July 2019 05:46 To: boost@lists.boost.org List Cc: Antony Polukhin Subject: Re: [boost] Making copyright holders easier to parse

On Sat, Jul 27, 2019, 04:44 Giovanni Mascellani via Boost < boost@lists.boost.org> wrote: <...>

...

My question is: are you interested in this kind of patches? Of course I would still go through the ordinary patch submission procedure. I am just asking if this kind of patches would be well received or not.

I have another suggestion. We already have the inspect program written in C++ which 'parses' and emits an html report on missing copyright (and many other transgressions of Boot guidelines). See https://www.boost.org/doc/libs/release/tools/inspect/inspect.cpp and copyright_check.cpp and .hpp I suspect that this could easily be altered to add an output to a file of copyright authors(s) and date(s) in whatever format and file type is easiest for Debian to deal with. For example a test file containing Library_name Author(s)_name Copyright_Date(s) ... The build tools are in I:\boost\tools\inspect If this works, I feel we could use this updated version in Boost itself. Other packagers and those needing to jump through copyright and GDPR hoops might find helpful. This would avoid a paroxysm in our extensive CI system 😊 Paul Paul A. Bristow Prizet Farmhouse Kendal, Cumbria LA8 8AB UK

On 27. Jul 2019, at 06:51, Bo Persson via Boost wrote:

Sounds like a good idea.

Old US copyright laws specifically mention "Copyright" and "Copr." as proper forms, but says nothing about "Copyright:".

Wouldn't want to stuble on such a technicality, would we? :-)

...

Although I'm all for making the attributions and licensing consistent :-)

Right.

The best proposal I have seen in this thread is to consistently apply the template from https://www.boost.org/users/license.html everywhere. If it is consistently applied, it is easy to parse automatically, even if the format is not particularly parser-friendly. As was said before, we need a new check in the Boost test matrix https://www.boost.org/development/tests/develop/developer/summary.html to enforce consistency, since the inspection reports http://boost.cowic.de/rc/docs-inspect-develop.html are currently not enforced. License and copyright errors are by far the most common problems found by the inspection tool, so by adding a check to the test matrix we could clean that up a lot. Best regards, Hans

On Mon, Aug 5, 2019 at 2:57 AM Hans Dembinski via Boost < boost@lists.boost.org> wrote:

...

On 27. Jul 2019, at 06:51, Bo Persson via Boost wrote:

Sounds like a good idea.

Old US copyright laws specifically mention "Copyright" and "Copr." as proper forms, but says nothing about "Copyright:".

Wouldn't want to stuble on such a technicality, would we? :-)

...

Although I'm all for making the attributions and licensing consistent :-)

Right.

The best proposal I have seen in this thread is to consistently apply the template from https://www.boost.org/users/license.html everywhere. If it is consistently applied, it is easy to parse automatically, even if the format is not particularly parser-friendly.

As was said before, we need a new check in the Boost test matrix https://www.boost.org/development/tests/develop/developer/summary.html to enforce consistency, since the inspection reports http://boost.cowic.de/rc/docs-inspect-develop.html are currently not enforced.

License and copyright errors are by far the most common problems found by the inspection tool, so by adding a check to the test matrix we could clean that up a lot.

Maybe.. It should be easy though. It just takes someone to add such a check to < https://github.com/boostorg/boost/blob/develop/status/boost_check_library.py

.

-- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net

On Mon, Aug 5, 2019 at 12:15 PM Vinnie Falco via Boost < boost@lists.boost.org> wrote:

Q: "How many Boost engineers does it take to change a copyright notice?"

A: "All of them."

Ah, but it's peer reviewed to be the best copyright notice available. - Trey

On 7. Aug 2019, at 05:50, jrmarsha via Boost wrote:

Can someone point me to where the license checking is used in CI? boost-ci doesn't have an example.

Maybe you should read this whole thread again carefully, the answers are here. There is currently no license checking done in CI. There is the inspect tool https://github.com/boostorg/inspect which has a license check, but it is not run in CI and it was not written to be run as a unit test. boost-ci is a new project, it is not used by all Boost projects. You probably want to add the check to __boost_check_library__, a special check that is run together with the library-local tests in the boost test matrix https://www.boost.org/development/tests/develop/developer/summary.html You can find the code for boost_check_library here: https://github.com/boostorg/boost/blob/master/status/boost_check_library.py Best regards, Hans

Hi, Il 27/07/19 10:22, Glen Fernandes ha scritto:

+1.

Thank you for everybody's feedback, which seems to be mostly positive!

Instead of inventing a new format now, if anything, why not make them consistent with https://www.boost.org/users/license.html prescribes? After all, many of our libraries are already consistent with it, and as Rene (and the page) conveys, some effort when into deciding things like that.

The page has the format "Copyright Joe Coder 2004 - 2006" and most libraries have that, or "Copyright (C) Joe Coder 2004 - 2006". Both are easy to parse without needing to introduce a colon after "Copyright:".

I wasn't actually pushing for any specific header format, sorry for not making this clear. To me, anything that can be easily automatically parsed is fine, including of course the one mentioned in the FAQs (which I had not previously noticed). However, that template does not cover the case of more than one copyright holder. Would something like this be acceptable? // Copyright Joe Coder 2004 - 2006. // Copyright Bob Hacker 2010 - 2015. // Copyright Department of Writing Very Long Names, Newline // Company Inc. 2017 - 2019. // Distributed under the Boost Software License, Version 1.0. // (See accompanying file LICENSE_1_0.txt or copy at // https://www.boost.org/LICENSE_1_0.txt) If not, what other? (the thing with a very long name is not pretentious; there is already a "Institute of Transport, Railway Construction and Operation, University of Hanover" in Boost). If this proposal seems appropriate for you, I can start to patch a few files and submit them here, so that they can be evaluated more carefully (this might not immediate, as I still have to write the code to do so). Thanks again, Giovanni. -- Giovanni Mascellani Postdoc researcher - Université Libre de Bruxelles

-----Original Message----- From: Boost On Behalf Of Vinnie Falco via Boost Sent: 28 July 2019 00:43 To: boost@lists.boost.org List Cc: Vinnie Falco Subject: Re: [boost] Making copyright holders easier to parse

On Sat, Jul 27, 2019 at 6:23 AM Glen Fernandes via Boost wrote:

...

...

Although I'm all for making the attributions and licensing consistent :-)

I am very much in favor of this as long as it does not require any of my source files to change.

+1 Because we have a massive Continuous Integration system with many compilers and platform that picks up changes to source files, *any* change to source, test or documentation files is like to trigger a massive recompilation and rebuild and retest and redocumentation. That will cause a load of machine time (already insufficient).. I think that we are pretty much following the guidelines https://www.boost.org/users/license.html (though they do not mention the common case of multiple authors). We (and anyone) can check the results of the inspect program to confirm that all files have a Boost copyright claim. All copyright lines have either a name or a date after the word copyright. Nobody has a digit in their name? Surely this isn't too difficult to parse? Any that cause trouble can be fixed individually if you tell us? Paul Paul A. Bristow Prizet Farmhouse Kendal, Cumbria LA8 8AB UK

On 7/28/19 11:27 AM, Paul A Bristow via Boost wrote:

Because we have a massive Continuous Integration system with many compilers and platform that picks up changes to source files, *any* change to source, test or documentation files is like to trigger a massive recompilation and rebuild and retest and redocumentation.

I don't think we build docs during our CI jobs, do we?

That will cause a load of machine time (already insufficient)..

If you want a commit to not trigger a CI job, you can add "[ci skip]" to the commit title line. Unfortunately, in case of PRs, that is on the submitter's concience.

Both of those changes, to the attribution and licensing, would almost certainly require Boost to get legal consult.

I don't think that is the case.  All the same information is conveyed with very similar context, manner, and detail. What I'd be interested in is adding this automated check to the general CI infrastructure.