Accept source files in encodings other than UTF-8

[simonmar]

We need to support at least Latin-1, for backwards compatibility.

One possibility is to run iconv as a preprocessor. Alternatively, call iconv directly from the compiler. What about Windows? Should there be a pragma specifying encoding, or take it from the locale, or a command-line flag, or all three?

[igloo]

Duncan Coutts points out that, as a workaround, you can pass

-F -pgmF convert.sh

to GHC, where convert.sh contains

#!/bin/sh
( echo "{-# LINE 1 \"$2\" #-}" ; iconv -f l1 -t utf-8 $2 ) > $3

[simonmar]

We documented the workaround using iconv for 6.6, this is unfortunately not a solution for Windows users. Punting a better solution to 6.6.1, or possibly 6.8.

[igloo]

See also 782.

[simonmar]

We don't plan to do anything here. Having source files in UTF-8 only means that source files are always shareable, which is a good thing.