Binary IO

[Glynn Clements]

Simon Marlow wrote:

> Does anyone have any *objections* to introducing
> 
>   System.IO.hGetWord8 :: Handle -> IO Word8
>   System.IO.hPutWord8 :: Word8 -> Handle -> IO Word8
> 
> (I prefer these names, please speak up if you prefer hGetByte/hPutByte
> instead, or something else like hGetOctet/hPutOctet).

No objections (other than the argument order, as Henning pointed out),
but I would prefer byte/octet; I don't really care which.

> One constraint would be that the Handle has to be in Binary mode.
> Lifting this constraint in GHC is a bit tricky because we currenly use
> the OS's text<->binary translation to do I/O (doesn't matter on Unix,
> right now).

Byte I/O should work on both text and binary streams.

> I presume that hPutWord8 (fromIntegral (ord '\n')) should not flush a
> line-buffered Handle, which is another slight complication.

No, it should flush it.

> Perhaps line-buffering should be the same as block-buffering on
> binary Handles.

Buffering mode and EOL conversions should be orthogonal.

With regard to the last three points, byte I/O isn't just for binary
files. It's equally useful for "8-bit text" I/O, i.e. the historical
Unix (and ANSI C stdio) model, where and you just shuffle bytes around
without paying attention to the encoding.

-- 
Glynn Clements <glynn at gclements.plus.com>