cvs commit: src/usr.sbin/sysinstall main.c

[Andrey Chernov]

On Tue, May 01, 2007 at 04:31:55AM -0700, Alfred Perlstein wrote:
> If the fallout from your changes broke a bunch of things in
> -current, then we can expect the fallout in -stable to be
> even worse.

They are not planning to go into -stable any soon until the moment when 
7.x becomes -stable, when libc version number will be bumped.

Do you mean under "bunch of things" few programs either mainly unchanged 
from V7 hacks time or suffering just because we have unportable poorly 
implemented putenv() in our libc which breaks its original SysV POLA? If 
yes, such changes are surely for good: backward-compatible, eliminating 
hacks and making our soft more portable.

> Your query about bug reports is a straw man as anticipating
> a lot of fallout which has already occured does not require
> that I actually have a bug report.

What you calls "fallouts" I call "portablity". 

> The end result is more users being bitten because a discussion
> regarding this has obviously not taken place.

It happens now, so they don't miss anything. At any moment those changes 
can be made non-default, under some sort of env. variable or old 
bug-to-bug functionality will be made by env. variable or some other 
ways.  It depends on problem reports I hit, but I don't see any yet.

> And yes, you're being a jerk. :)

Thanx, I hope that streak will help me to improve our code.

-- 
http://ache.pp.ru/