Re: Purpose of rawsock module

[Sam S. <sd...@gn...>]

> * Daniel Jour <qna...@tz...> [2016-05-27 00:40:35 +0200]:
>
> I'm currently working on getting the rawsock module up-to-date and running
> on the major target platforms. I got a bit unsure about the real purpose,
> the "goal" of that module, though:
>
> The documentation suggests that it's providing access to what sys/socket.h
> provides, i.e. "raw" access to sockets.

Indeed, rawsock originated with Fred & Don Cohens, who wrote it for some
very low level access to networking.

> Windows doesn't have sys/socket.h, but rather WinSock2. While there are
> a lot of similarities between those (IIRC that was one of the design goals
> of WinSock) they're certainly different in some parts.

Sockets are the OS-level interface to TCP/IP.
Thus all the functions in rawsock can, should(must!) and did work on all
"modern" operating systems (i.e., linux, *bsd including mac os x,
windows xp and later, all other unixes released after 2000 - but not on
atari and amiga).

Yes, windows is an oddball system, this is why we have a problem now,
and this is precisely the problem this project is supposed to address.

> Unifying both sys/socket.h and WinSock2 behind the same interface
> seems to be a contradiction to the "raw" access to me, though. If
> that's the goal of the module, wouldn't it be much better to have a
> module providing access to some "socket library"?

This module _IS_ that "socket library".

> Or would it be useful to split the rawsock module in platform dependent
> modules, just like it's done with bindings/glibc and bindings/win32?

The idea of the module is to have it working on _all_ modern OS OOTB,
without "os-specific" interfaces all over the place.

-- 
Sam Steingold (http://sds.podval.org/) on Ubuntu 16.04 (xenial) X 11.0.11803000
http://www.childpsy.net/ http://honestreporting.com
http://memri.org http://dhimmi.org http://mideasttruth.com
If you know that you know nothing, you know too much.