filter out carriage returns for readfile

[gbwey]

This is related to this pull request
yesodweb/shakespeare#113 referrring to https://gist.github.com/gbwey/6833805

This change to readfile will filter out carriage returns as does Data.Text.IO.readFile.
Avoids the situation in windows where you read a file/then write it back and it is different.

Is there a reason why we don't just use Data.Text.IO.readFile?

Thanks,
Grant

[gregwebs]

I agree that using readFile from Text is probably best.

One potential difference though is the current use of lenientDecode: we don't want programs to crash from reading a file with an invalid UTF character.