You hit on a good way to do it in your original post. Pre-populating a hash is effective and fairly simple, even if it is a bit messy. By putting it in a module and putting that on CPAN, the whole Perl community can help with submissions on what their strange systems use for line endings. It might be a little bit of overkill to use a module computationally, but this situation is more data intensive than computationally intensive. Data gathering is always done best by the experts in the particular areas (in this case, differing platforms), with the results being reported to a central repository or person. CPAN or some other joint project system is ideal for this kind of task.
in reply to Quick and portable way to determine line-ending string?
Update: Changed some awkward wording and fixed a tpyo.