Steven and Robert,
Thanks for contributing this Steven. It is a really good idea. A map file seems much more flexible than the hard-coded case statements.
Robert, in your comment above you mention Solr includes a mapping based on the deprecated ISOLatin1AccentFilter.
Could you please point me to where I can find this mapping file for the deprecated ISOLatin1AccentFilter in SVN?
Otherwise, I'll just run or adapt the perl code here and run it against the ISOLatin1AccentFilter code.
We haven't switched to the newer ASCIIFoldingFilter and need to emulate the ISOLatin1AccentFilter in some custom non-java code until we make the switch and re-index all 6 million volumes.