Comment 4 for bug 6614

# I've done some investigating and found a bit of a rats-nest,
# unfortunately. (Defoma is rather strange inside and gs.defoma reaches
# right inside Defoma's private data structures, for example.)
#
# In any case, I think I've established that:
# * This warning is generated when trying to produce a `/CID'
# value which is a character set identifier, which ought to
# consist (in this case) of a registry and a `supplement'.
# * The registry value is invented by a fixed mapping from the
# `Location' field in the hints file supplied by the font package.
# This seems wrong - at best, a guess - but I can't find any
# relevant documentation for the model so I'm not sure.
# * There is no mechanism for discovering the `supplement' (which
# indicates which version of the registry is in use). The machinery
# in gs.defoma which is intended to calculate the supplement doesn't
# work at all in my tests.
# * I have no idea at the moment how /correct/ registry and supplement
# values might be obtained.
#
# So I have produced a version of gs-common which does not attempt to
# produce a /CID setting at all. This definitely solves the warning.
# The package can be found at
# http://www.chiark.greenend.org.uk/~ian/d/
# as gs-common_0.3.9ubuntu1.0iwjtest1_all.deb (source, etc.,
# in the same directory).
#
# But it may break all users of affected CJK fonts. For a list of
# affected fonts, run
# find /var/lib/defoma/gs.d ! -name '*.ttf' ! -name '*.pfb' | xargs grep /CSI
# before installing the test version of gs-common. (Afterwards there
# will be no output.)
#
# So I would like one or two CJK users to try out this package and let
# us know if it breaks the fonts listed by the command above. If it
# doesn't then the code that I've removed was probably ill-advised. If
# it does then we need to think again.
#
# Ian.
#
# (Lines prefixed with `#' to avoid a Malone command parsing bug.)