HOW TO GENERATE THE COPYRIGHT FILE
----------------------------------

The debian/copyright file for Boost is very complicated, because the
library has accumulated a huge number of files and contributors over
the years. It would be intractable to fill it by hand, and FTP masters
do not like the previous choice to just generically indicate the Boost
Software License and see each individual file for the differences.

Fortunately Boost itself already contains a small program, called bcp,
that searches for common regular expressions in each file to find the
file's license and copyright holders. The program has some
deficiencies, but is overall a good foundation to automatically
generate a debian/copyright file.

A patched fork of the bcp program is maintained by Giovanni Mascellani
in the repository [1].

 [1] https://salsa.debian.org/gio/boost-copyright

It can be compiled with qmake and make. Then you change dir to where
you have unpacked the Boost tarball and run something like:

    find -type f | grep -v ^./debian/ | grep -v ^./.pc/ | bcp --report-debian --read-from-stdin debian/copyright

This will erase and regenerate the debian/copyright file. The
boost-copyright repository already contains a skeleton of the
copyright file (called manual_copyright), which is automatically
included at the beginning of the generated file.

Since bcp cannot identify the copyright status for each single file,
because some of them do not actually contain such information, the
manual_copyright file contains some blanket stanzas for the main Boost
directories, so that the copyright for each not better identified file
is (reasonably) assigned to the people responsible for that
subdirectory. The subdirectories loosely correspond to the individual
Boost libraries. Also, a public domain dedication is presumed for the
few and very simple files that are neither recognized by the program,
neither contained in one of the well-known subdirectories.

The script generator.py, in the boost-copyright repository, can be
used to manually regenerate most of manual_copyright based on the
meta/repositories.json files available in the Boost
distribution. However it is recommended to carefully check its output,
because there will be some inaccuracies.