Porting to Solaris from SunOS - response summary

From: Prasanna Kondapaneni (prasanna@chaitanya.tcs.com)
Date: Tue May 18 1993 - 12:23:33 CDT


Hi friends,

First of all, thanks to all of those who responded with their invaluable
contributions.

These are responses that I got for my questions:

|> Return-Path: <mikep@cfsmo.honeywell.com>
|> Posted-Date: Tue, 11 May 93 19:51:41 CDT
|> Date: Tue, 11 May 93 19:51:41 CDT
|> From: mikep@cfsmo.honeywell.com (Mike Polo)
|> To: prasanna@chaitanya.tcs.com
|> Subject: Re: porting Sun OS 4.1 software to Solaris 2.1
|>
|> I think Prentice Hall has a book about porting to Solaris 2.x. I got a
|> copy for around $40 at the Sunsoft Developer's Conference. It is VERY
|> thorough and covers just about everything you'd want to know. It also
|> comes with a floppy disk FULL of code. I think you can order it from
|> Sun Express. (1-800-USE-SUNX) I use the book quite extensively!!
|>
|> Hope this helps,
|> Mike
|>
|>
|>
|>
|>
|>
|> Date: Tue, 11 May 93 20:53:46 PDT
|> From: John.Matteson@gain.com (John Matteson)
|> To: prasanna@chaitanya.tcs.com
|> Subject: Re: porting Sun OS 4.1 software to Solaris 2.1
|> X-Sun-Charset: US-ASCII
|>
|> Hi Prasanna -
|>
|> I've been porting SunOS4.x code to Solaris 2.x for ~7 mo. (since
|> solaris v 2.0, bad decision). There's a number of problems to
|> be aware of which I'll just list out, you probably have thought
|> about many of them already, but...
|>
|> 1. signals. You will need to rewrite all of the signal code to
|> be "posix"/svr4-like instead of bsd. There's libucb to get around
|> this at first, but I've found problems with using the 'compatibility'
|> stuff, better just to bite the bullet and not depend on libucb.
|>
|> 2. A real neat gotcha. A vendor supplied lib that we linked in that
|> whose code was called thru a signal handler had a sscanf in it. Not
|> a bad thing in bsd-land, but Sun has seen fit to make malloc calls
|> within their standard i/o routines - not fun with sigs.
|>
|> 3. Many, many compile-time errors that need to be cleaned up, unless
|> the original code is _very_ clean. This is just a pain, no real
|> problem.
|>
|> There are a few others I can't recall right now, feel free to e-mail
|> me if you come up with any bizare problems with sol2s (that's our
|> id for solaris 2.x - sparc, the s. o. l. part really seems to fit :=) ).
|>
|> ::John Matteson
|> jmat@gain.com
|>
|>
|>
|> Date: Thu, 13 May 93 19:18:32 PDT
|> From: paula@grace.rt.cs.boeing.com (Paul Allen)
|> To: prasanna@chaitanya.tcs.com
|> Subject: Re: porting Sun OS 4.1 software to Solaris 2.1
|>
|> In article 11055@tcsi.com, prasanna@chaitanya.tcs.com (you) wrote:
|> >
|> >I am going to port some of our C++ based software to Solaris 2.1 (from Sun
|> >OS 4.1). Some people have already managed to scare me saying that it's
|> >going to be terrible. Can someone give me some pitfalls to look for while
|> >I am at it?
|>
|> Depending on what your software does, the port could be either simple
|> or quite time-consuming. The Solaris 2.x AnswerBook has a large section
|> on porting from Solaris 1.x, complete with long tables of system calls
|> and function calls. Sun also provides a thing called PipelineTool, which
|> somewhat automates the process of locating the places in your code that
|> need work.
|>
|> >Also, I heard that Solaris causes disks (over net) to crash once you
|> >install it on a machine that's on a network. Is this info accurate?
|>
|> I don't think so. We've had four of our ~200 Suns running Solaris 2.0
|> and Solaris 2.1 since January. They are all on the net and use NFS and
|> NIS. We haven't had any disks crashing.
|>
|> Paul Allen
|>
|> Paul L. Allen -- Paul.Allen@atc.boeing.com
|> Boeing Computer Services, Research & Technology, Computing Environment (whew
!)
|> "It ain't no fun sleepin' in unless you're awake to enjoy it!" (Bill Whistlec
od)
|>

--
Thanx.

--prasanna -- Thanx.

--prasanna



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:51 CDT