Install the man pages

[lmoureaux]

Description

This PR modifies the Makefile and setup script to install the man pages. They are installed at $BASE/share/man, where $BASE is as follows:

• $VIRTUAL_ENV when installed within a virtualenv
• /usr when installed system-wide

Man pages in the virtualenv can be accessed by setting the correct $MANPATH.

[outdated] A side-effect of this change is that I figured out how to modify the default setuptools install location. All scripts and macros are now installed in $BASE/bin. This means that virtualenv can take care of updating the $PATH (system-wide installs still need a setup script). [/outdated]

Another side-effect is that I corrected the license information in the package metadata. It was saying "MIT" while the LICENSE file always contained the text of GPLv3 (there is no way to change back to MIT without contacting every single contributor).

The setup script and developer documentation will need to be updated to use the $MANPATH.

Downsides

• The man pages produced when the package is not installed may not be completely accurate. This is a problem for make rpm since it doesn't work inside a virtualenv.
• [outdated]Macros and scripts are present twice in the package: under bin/ and under gempython/.... I couldn't figure out why.[/outdated]
• CI is broken :cry:

As a result of these downsides and my inability to fix them shortly, this PR is provided more as a list of commits to be cherry-picked.

Types of changes

• [ ] Bug fix (non-breaking change which fixes an issue)
• [ ] New feature (non-breaking change which adds functionality)
• [x] Breaking change (fix or feature that would cause existing functionality to change)

Motivation and Context

• 138

• [outdated] Comment in setup.py:

  would prefer to use this for executables, if one can control the install location
  to be user defined rather than /usr/bin

  The preferred method (native support from setuptools) is now used. [/outdated]

How Has This Been Tested?

• The pip and rpm packages were built, the former from a virtualenv and the latter on gem904daq01.
• The pip package was installed; scripts and man pages are correctly installed and have the correct permissions.
• It was checked that the rpm package contains all files.
• The man pages built in each environment were (quickly) checked.

Since I don't have access to a rpm-based machine and I don't want to install one, I couldn't check that the rpm package installs cleanly.

Checklist:

• [x] My code follows the code style of this project.
• [x] My change requires a change to the documentation.
• [ ] I have updated the documentation accordingly.
• [ ] I have added tests to cover my changes.
• [ ] All new and existing tests passed.

[lmoureaux]

avoiding putting things in macros into /opt/cmsgemos/bin

May I ask what's the rationale for the difference between scripts and macros in this project?

[lmoureaux]

I would have to go back in time to figure out why a given license was (probably arbitrarily) chosen for this project

Commit 3d58f5d2 by... @jsturdy. The commit message ("Initial commit") isn't helpful in telling how the license was chosen.

FYI:

• CMSSW uses Apache 2.0.
• XDAQ is under a 3-clause BSD-style license

[jsturdy]

May I ask what's the rationale for the difference between scripts and macros in this project?

Mainly, it is to isolate the key executable parts (analyzing scans, which should go in /opt/cmsgemos/bin) from random plotting and style macros (can still be on the PATH for ease of use, provided PATH is modified to include the macros directory)

Further down the road, much of this will be replaced with console_scripts and entry_points, but that is a minor priority

[lmoureaux]

Removed the controversial move to a more standard behavior using standard tools; the rpm package will now install man pages in /usr/share. Updated the first post accordingly.

[lmoureaux]

Comments have been addressed and I'm waiting for another review.