Date: Sat, 6 Feb 1999 22:29:05 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: wes@softweyr.com (Wes Peters) Cc: tlambert@primenet.com, witr@rwwa.com, dcs@newsguy.com, freebsd-hackers@FreeBSD.ORG Subject: Re: more modular rc/init/uninit system... Message-ID: <199902062229.PAA21963@usr02.primenet.com> In-Reply-To: <19990205212913.C6050@softweyr.com> from "Wes Peters" at Feb 5, 99 09:29:13 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > > :- Yep. I'm not against run states, just against run levels. > > > > > > I'm against both! See my earlier remarks about configuration mamagement. > > > > How do you propose to solve the Solaris binary compatability > > problem for commercial Solaris applications that install > > components into the rc.d directories in order to get them run > > at the correct time and in the correct order for dependent > > services requirements? > > By writing a port that will install the startup script in the > right place and modify it as necessary. We really don't have > to implement the entire brain-dead mess of the SysV init system > just to start a simple (or even not so simple) application. You're in Utah. I bet Bob would let you borrow his IBCS2 Sybase for the AT&T StartServer so you could install it on FreeBSD and make a "port". Of course, when I did it, it was absolute hell to get this working on that FreeBSD box for Dawn for the copy of the Hunam Genome database, but maybe youwill have better luck. While you are at it, you might consider the IBCS2 copies of Lotus 1-2-3 that I got working, with SEF's help on the math coprocessor probe using the sysi86()(sp?) call to call the "BIOS". The problem is that having an execution class (ABI type) loader is ony a tiny part of the battle in actually emulating foriegn environments for binaries. For the two examples above, you would effectively need to implement the entirety of IBCS2, including the packaging system (well, except SCO is now giving that away). The Lotus 1-2-3 especially will be a pain, in that it uses the uname return as part of the copy protection, after assembling a pare of simple cyphered halves of a binary file around its idea of the uname return during the install to make the install specific to the machine. For Solaris, the number of issues are doubled or tripled. This is not to discourage ABI compatability with Solaris; I think that, going forward, it's not optional. But at the same time, you can't solve a vast problem in a half-vast way. You either do it right, or you don't bother doing it at all. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199902062229.PAA21963>