Created and added stack.yaml and .gitignore files.
Relaxed the version dependency on 'binary' package in cabal file. Is that OK?
Also brought the minimum cabal version to >=1.8, so I could add a test target that pulls in the library.
Changed all tabs to spaces - I don't know when the Haskell compiler started giving warnings about that.
This envolves adding the requirement, that every Encoding must also be an instance of Eq and Typeable to go into DynEncoding.
darcs-hash:20090225035150-a4fee-c7d902e28313929ee9ffe0c6a6b60d8ff4704ae9
Now it's possible to change the character encoding while de-/encoding.
Also, it's possible to use any data structure as a source or target of the de-/encoding process.
darcs-hash:20090221203100-a4fee-6da31f2e37c30a3f5cd5f10af71984209488bb0b
HPC doesn't like TemplateHaskell yet, so we have to comment things that require TemplateHaskell out using a preprocessor.
darcs-hash:20080203132156-a4fee-6587e14d7c57e0dfbc9f50746502e6b0a68b6632
This is a bit of a hack, because the static lookup data this encoding requires brings template haskell to it's knees. So I've got a program that generates a haskell module file from the XML mapping.
darcs-hash:20070823034759-a4fee-883359c8951d4376fc2d783cd7be352d6c5b2111
