Ticket #3058: Add 'hex' to the pretty package (and other
thoughts)
Duncan Coutts
duncan.coutts at worc.ox.ac.uk
Mon Mar 2 07:55:29 EST 2009
On Mon, 2009-03-02 at 01:53 -0800, Thomas DuBuisson wrote:
> All,
>
> http://hackage.haskell.org/trac/ghc/ticket/3058
>
> Attached is a patch for Text.PrettyPrint.HughesPJ that adds a 'hex'
> function to print hexidecimal numbers.
> The only point that I exepect to be contended is it varies slightly from
> the surrounding functions in that it allows one to control the number of
> characters printed:
There are a number of other functions from the Numeric module we should
also consider. These are all the ones that return ShowS, the equivalent
of the Doc type.
showEFloat :: (RealFloat a) => Maybe Int -> a -> ShowS
showFFloat :: (RealFloat a) => Maybe Int -> a -> ShowS
showGFloat :: (RealFloat a) => Maybe Int -> a -> ShowS
showHex :: (Integral a) => a -> ShowS
showInt :: (Integral a) => a -> ShowS
showIntAtBase :: (Integral a) => a -> (Int -> Char) -> a -> ShowS
showOct :: (Integral a) => a -> ShowS
showFloat :: (RealFloat a) => a -> ShowS
showSigned :: (Real a) => (a -> ShowS) -> Int -> a -> ShowS
> While we can argue about the consistancy issues, I almost always want to
> control the number of digits when dealing with hex. Hence I feel this
> is a reasonable special case.
For some reason I thought I remembered that the Numeric show functions
also allowed fixed width 0-padded display but it appears they do not,
except by manually rendering and adding leading 0's.
Note the show float functions take an optional precision. Perhaps that's
a sensible approach for optional fixed width octal and hexadecimal
display.
> Secondary Issues:
> 1) Is there a repo for pretty? I didn't see one on code.haskell.org
It's on darcs.haskell.org along with ghc and the other core libs.
> 2) If I submit a ticket+patch to move Text.PrettyPrint.HughesPJClass
> (from prettyclass, Augustsson) into Text.PrettyPrint (of the 'pretty'
> package) would anyone object? It feels mis-placed / I feel things need
> consolidated.
That's a much bigger change. I suggest a separate ticket and discussion
for that. There's a rather large design space and range of use cases.
Duncan
More information about the Libraries
mailing list