[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: itp: static bins / resolving static debian issues




On Tue, Aug 24, 1999 at 07:12:41AM -0400, Michael Stone wrote:

> One argument for static bins that I keep seeing is that "other unixen do
> it." That's a misleading half-truth. Yes, most provide a static /bin/sh.
> I don't know offhand of one that provides a static su or a static
> ifconfig or any of the other things that have been floated as essential
> to recovery. (Counting only those systems that support dynamic libs.)

For the record: RedHat (even a static rpm), FreeBSD, NetBSD, OpenBSD,
BSDi, and Solaris (not as much as you would like, but restore is
available) all do for sure (Solaris has some bins that are dynamic,
but only versus some minimal libs in /etc, so that sort of half
counts). I'd have to checking around to find out what other Unixes
provide, since I hardly ever use the others, but it's a well
established practice in the Free unixes anyway (except Slackware
apparently).

It wouldn't surprise me if HPUX and IRIX were totally broken; and
if SCO, OSF/1 and AIX provided unimpeachable, complete, and solid
live recovery functionality--but only in some obscure and non-obvious
way, involving hexadecimal numbers and odd prompts probably. Oops, 
too much hyperbole :-)

> Should we stop shipping
> a compiler because "that's what other unixes do"? Of course not: we're
> trying to build a new, reliable, streamlined, free operating system--not
> a mindless clone of someone else's OS.

However, there are two important points:

  1) People familiar with Unix expect it, and as a result do not 
     imagine that they have to do anything special to get it (this
     burned me the first time, I never imagined Debian wouldn't have
     static libraries.)

  2) It's there in other Unixes for good reason, 30 years of hard 
     won Unix knowledge. Debian shouldn't automatically turn its back
     on things that so many other OS's do either; if Debian is going
     to be non-standard, I submit it is the people advocating the
     non-standard behavior that carry the burden of proof.

Given the long "opinions" thread, I feel it necessary to point out 
that yes, there is a certain amount of opinion in (2), and no, I 
don't think that by itself it is a sufficient reason. Just another one.

Justin


Reply to: