[Haskell-cafe] Re: getting crazy with character encoding

On Wed, Sep 12, 2007 at 04:18:43PM +0200, Andrea Rossato wrote a message of 60 lines which said:

Now, I would expect that the output of a shell command such as "ls ab*" would be a string/list of 5 chars.

I do not think this expectation is reasonable. I do not think that ls is Unicode-aware. It probably has only bytes semantic, not characters semantic.

I would expect that a file name set in an utf-8 locale should be read by locale aware application

"locale aware application" is too vague. An application can use the locale and still being unable to separate bytes from characters. ls may be "locale aware" but it is probably not "Unicode aware".

l <- fmap lines $ runProcessWithInput "/bin/bash" [] "ls ab*"

This is not an Haskell issue but a ls issue. use System.Directory.getDirectoryContents and we'll see.