John Van Enk wrote:
I want to group them with Network functions because traditionally these
specific functions *are* in networking packages.
I agree. It is true that these functions are in the <arpa/inet.h> header on my
machine. The ntoh* and hton* are just data manipulation, just as Network.URI is
all about constructing and manipulating URI data.
The question I think we should answer is: Do we wish to follow the precedent of
putting URI under Network and not Data?
Pragmatically, programmers will expect to find these functions near the 'net
ones.
A "network-byteorder" package which has the "Network.ByteOrder" module is the
most flexible thing to create. It can be updated and reinstalled independently
of the "network" package.
On *nix systems, the prototypes reside in <netinet/in.h>. On Windows
they reside in <winsock2.h> and one must link against libws2_32.
The package I have locally supports all platforms. It just does a
foreign import of the function (and uses a preprocessor definition to
pick the calling convention).
One could make an argument to add these sorts of functions to Word and
Int packages.
/jve
On Sat, Dec 27, 2008 at 7:14 PM, Chris Kuklewicz
<hask...@list.mightyreason.com <mailto:hask...@list.mightyreason.com>>
wrote:
Are these going to be available on Windows or just Posix systems?
And these are NOT performing network communication. They are bit
manipulations.
I could imagine putting them in/under the Data.Word or Data.Bits
more than Network.*
And they should also be made to work with the corresponding
Foreign.C.Types of CShort, etc.
John Van Enk wrote:
While working on a project, I discovered that hton[sl] and
ntoh[sl] are missing from the networking libraries. It appears
there was some justification for this, but I've fallen upon a
few instances where things were *not* working as expected.
(Specifically when newtype'ing around HostAddress.)
As such, I'm putting together a few extra functions that I think
should exist in Network somewhere. I'm calling the package
Network.Util (but am open to changes).
Currently this only has:
* htons
* htonl
* ntohs
* ntohl
Before I drop this on hackage:
1. Is there a reason I shouldn't do this?
2. Are there other suggestions for missing functions? (preferably
cross-platform suggestions)
Thanks all.
/jve
------------------------------------------------------------------------
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org <mailto:Haskell-Cafe@haskell.org>
http://www.haskell.org/mailman/listinfo/haskell-cafe
------------------------------------------------------------------------
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe
