Prettier error messages?

PrettyErrors is a wiki page summarising the state of play.

clang has very nice-looking error messages.

<pre><span style="font-weight: bold">pretty.c:6:7: <span style="color: pink">warning:</span> incompatible pointer to integer conversion passing 'char [14]' to parameter of type 'int' [-Wint-conversion]</span>
  foo("Hello, world!");
      <span style="color: green">^~~~~~~~~~~~~~~</span>
<span style="font-weight: bold">pretty.c:1:14: note: passing argument to parameter 'i' here</span>
void foo(int i) {
             <span style="color: green">^</span>
1 warning generated.</pre>

ghc's error messages are not so good.

<pre>ugly.hs:7:18:
    Couldn't match expected type &lsquo;()&rsquo; with actual type &lsquo;[Char]&rsquo;
    In the first argument of &lsquo;f&rsquo;, namely &lsquo;"Hello, world!"&rsquo;
    In the second argument of &lsquo;($)&rsquo;, namely &lsquo;f "Hello, world!"&rsquo;
    In the expression: print $ f "Hello, world!"</pre>

In my opinion, there are three independent improvements that could be made to GHC error messages and warnings: color, context and whitespace. Currently they're blobs of text.

Consider all three applied to error messages:

<pre><strong>ugly.hs:7:18: <span style="color: red">error:</span> Argument to 'f' is type '[Char]' but expected 'Int'</strong>
main = print $ f "Hello, world!"
                 <span style="color: green">^~~~~~~~~~~~~~~</span>

<strong>ugly.hs:3:1: note: type of 'f' is given here:</strong>
f :: () -> IO ()
     <span style="color: green">^~</span></pre>

or

ugly.hs: note: type of 'f' is inferred: f :: forall m. Monad m => () -> m () ^~

In my opinion, context and whitespace are more important that color. Even without color, compare this error message to the one shown above:

ugly.hs:7:18: error: Argument to 'f' is type '[Char]' but expected 'Int'
main = print $ f "Hello, world!"
                 ^~~~~~~~~~~~~~~

ugly.hs:3:1: note: type of 'f' is given here:
f :: () -> IO ()
     ^~

In my opinion this is much easier to visually process than GHC's current messages.