EMBOSS, the European Molecular Biology Open Software Suite, contains a
forked (and thus, bundled) copy of plplot, and has, for nearly 10 years.
Initially, they made attempts to merge their changes with upstream, but
were unsuccessful, so they forked.

= Overview =

Upstream noticed that we were disabling their fork of plplot (renamed by
upstream EMBOSS to eplplot), and asked that we stop, because their code is
functionally dependent on the forked copy. They've pointed out specific
bugs caused by this, so it is not theoretical.

I asked them if they would be willing to revisit the issue and try to work
with plplot to merge their changes, and they considered it, and replied
accordingly:

"We, the core EMBOSS developers, have recently discussed the PLPLOT
library position.
Early attempts (approximately 10 years ago) at coordinating with
the PLPLOT developers, regarding integrating changes for the correct
functioning of EMBOSS into their codebase, proved unsuccessful.
It is our firm intention to investigate the current plplot code, as
supplied by the PLPLOT developers, with a view to EMBOSS utilising it in
an unmodified form. As EMBOSS is grant-funded we have existing
development
commitments under the current funding round. If such investigations
suggest that extra funding is required then we will apply for it in the
next round. In the interim we request that the libeplplot library
be used in your build of the EMBOSS package.

Dr A.J. Bleasby
EMBOSS project
European Bioinformatics Institute
Hinxton
Cambridge
CB10 1SD
UK

On behalf of the EMBOSS core developers"

This satisfies me, in that there is a legitimate need for the forked copy
of plplot, and upstream is willing to make every attempt to remove the
necessity for it. I'm asking FESCo to permit this bundling as an
exception.