*BSD News Article 84209


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.cs.su.oz.au!metro!metro!munnari.OZ.AU!news.ecn.uoknor.edu!paladin.american.edu!02-newsfeed.univie.ac.at!01-newsfeed.univie.ac.at!voskovec.radio.cz!news.msfc.nasa.gov!newsfeed.internetmci.com!howland.erols.net!EU.net!sun4nl!fwi.uva.nl!not-for-mail
From: casper@fwi.uva.nl (Casper H.S. Dik)
Newsgroups: comp.unix.solaris,comp.unix.bsd.misc,comp.unix.internals
Subject: Re: Solaris 2.6
Supersedes: <cancel.casper.32a40b7b@mail.fwi.uva.nl>
Date: 3 Dec 1996 12:14:04 +0100
Organization: Sun Microsystems, Netherlands
Lines: 51
Distribution: inet
Message-ID: <casper.32a40b7b@mail.fwi.uva.nl>
References: <32986299.AC7@mail.esrin.esa.it> <casper.32a1813c@mail.fwi.uva.nl> <57v6em$p73@helena.MT.net> <casper.32a3e5ae@mail.fwi.uva.nl> <580sgh$kpi@panix2.panix.com>
NNTP-Posting-Host: mail.fwi.uva.nl
Xref: euryale.cc.adfa.oz.au comp.unix.solaris:91320 comp.unix.bsd.misc:1671 comp.unix.internals:11443

tls@panix.com (Thor Lancelot Simon) writes:

>And this after hand-waving away the...slight problem...with the previous
>*unqualified* claim that the Sun approace "prevented" file corruption, yet!

>I give up.  Like I said, you folks suffer from inpenetrable NIH.




You have *never* addressed teh binary compatibility issue I raised
other than some hadnwaving "oh, just recompile everything".

You have not argued against it so I assume you agree with me there.

That's why I call you a "BSD bigot", you don't argue the important  points
I put forward but argue the lesser important file corruption argument.

My argument is best summarized as:
	We can't do it differently because of binary compatibily,
	and, oh, there's this hypothetical file corruption
	problem too.

I agree with you that:
	just making off_t bigger is nicer and cleaner
you fail to see that:
	it really doesn't matter whether you do it in the syscall table
	or the shared library, you have two different sets of functions anyway

The file corruption scenario is real, though, but I would not
claim we solved all of it, just the vast majority of that specific
case.  It is by no means as important as binary compatibility.

And with binary compatibility we don't mean just running binries,
but using libraries, .o and .so files as well.

Hell, if we could get people to recompile their .o files, we'd
have changed FILE->_cnt to a different size a long time ago.
(Barring the fact that that would break SV ABI compliance, as would
changing off_t, but that's a side issue)

You've failed to convince me that such compatibility could be gotten in
a different, acceptable way.  Even stronger, you've failed to put
forward even one convincing argument.

Casper
-- 
Casper Dik - Sun Microsystems - via my guest account at the University
of Amsterdam.  My work e-mail address is: Casper.Dik@Holland.Sun.COM
Statements on Sun products included here are not gospel and may
be fiction rather than truth.