[OpenAFS-port-freebsd] Re: [OpenAFS-devel] coding conventions / standards?

Derrick J Brashear shadow@dementia.org
Wed, 8 Nov 2000 00:59:14 -0500 (EST)

On Tue, 7 Nov 2000, Bill Sommerfeld wrote:

> I'm currently working on porting openafs to NetBSD.  [I'm most of the
> way through getting the userland code to build at the moment..]
> Have the gatekeepers come up with any coding conventions, standards,
> baseline assumptions, etc?  

The only baseline assumption thus far is the codebase will be converted to
use autoconf; Portability assumptions based on an autoconf model
(HAVE_MUMBLE) are thus encouraged; Obviously at some point we'll need to
branch a "stable" (e.g. it compiles) and keep working on "devel" (e.g.
it's broken and we're still working on it)

> It appears that a bunch of the accreted layers of ifdefs would be
> significantly cleaned up if we could just assume a baseline of ANSI C
> and POSIX.1 .. i.e., use <termios.h> for terminal manipuation,
> strerror() instead of sys_errlist[], assume vfprintf exists rather
> than defining it in terms of _doprnt(), etc., etc., etc.

A portability library for systems which don't comply with such
requirements combined with these assumptions is the current target.

> Are the gatekeepers likely to accept changes which assume an ANSI
> compiler and a POSIX.1 environment in userland?


> Oh, and is anyone working on replacing the archaic Steve Miller DES
> implementation with something better/faster/smaller/easier to build?

As yet unknown to me, but if no one does, I will.