*BSD News Article 31938


Return to BSD News archive

Xref: sserve comp.os.386bsd.questions:11050 gnu.misc.discuss:15322
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!bunyip.cc.uq.oz.au!munnari.oz.au!spool.mu.edu!sdd.hp.com!elroy.jpl.nasa.gov!netline-fddi.jpl.nasa.gov!nntp-server.caltech.edu!news.claremont.edu!bridge2!usenet.coe.montana.edu!bsd.coe.montana.edu!nate
From: nate@bsd.coe.montana.edu (Nate Williams)
Newsgroups: comp.os.386bsd.questions,gnu.misc.discuss
Subject: Re: Taylor UUCP on FreeBSD???
Message-ID: <2ua21l$d4g@pdq.coe.montana.edu>
Date: 22 Jun 94 19:05:25 GMT
References: <1994Jun13.040754.17764@kosman.uucp> <RSANDERS.94Jun21155032@hrothgar.mindspring.com> <2u7tcb$6mr@pdq.coe.montana.edu> <RSANDERS.94Jun22110639@hrothgar.mindspring.com>
Organization: Montana State University, Bozeman  Montana
Lines: 59
NNTP-Posting-Host: 153.90.192.29

In article <RSANDERS.94Jun22110639@hrothgar.mindspring.com>,
Robert Sanders <rsanders@mindspring.com> wrote:
>In article <2u7tcb$6mr@pdq.coe.montana.edu> nate@bsd.coe.montana.edu (Nate Williams) writes:
>
>   >Well, that's a valid argument on Linux, but FreeBSD (and I suppose
>   >NetBSD) think that shared libraries belong in /usr/lib.  I disagree,
>
>   As a FreeBSD developer, I'll give a quick WHY we did it that way.
>   3) Breaking up a standalone library for the root partition is wasteful
>      since the code will be also needed for other binaries.
>
>Could you explain this a little more?  I'm not sure I understand.

If we were going for small space one solution would be to break up the
library into those parts needed for the root programs, and another for
the rest of the routines.

>   4) In the case of a bad crash using the Mach VM system in BSD systems you
>      *CAN* corrupt the running binary, which in the case of shared binaries
>      can be the shared library.  (The way paging is done)
>
>You mean a file mapped into memory in such a way that changes
>shouldn't be committed to disk can still get corrupted?  That's a bug.

Yes, it is. :-)  (It may have been fixed, but things like this existed
in 1.0+shlibs before all the VM work was done by David and John)

>Although I normally find the FreeBSD developers as reasonable as any
>people on Usenet, this is definitely from left field.  Let's compact
>all this discussion into a quick dialogue:
>
>user:	"Hey, ash doesn't run some standard 'sh' scripts, but bash
>         does."
>nate:   "ash is smaller.  You don't need the interactive bloat for
>         /bin/sh."
>user:   "Smaller, yes, but it doesn't work."
>nate:   "So fix ash."
>
>It seems to me that as one of the principal developers of a respected
>and important operating system, it should be more important to you
>that things work than is apparent here.

I get lots of 'it's broken', but no details.  I can't fix a bug that I
don't know it's symptoms.  'It works for me' is my excuse.

How about this.  If you have a reproducible bug with ash, send me th
bug-report with enough info. to track down the bug.  I was given an
individuals name in private email who was 'volunteered' to fix the bugs
in ash.

Good enough?  I can't fix something that's not broken. :-)


Nate
-- 
nate@bsd.coe.montana.edu     |  FreeBSD core member and all around tech.
nate@cs.montana.edu          |  weenie.
work #: (406) 994-4836       | 
home #: (406) 586-0579       |  Available for contract/otherwise work.