From owner-freebsd-ports Fri Dec 1 12:05:13 1995 Return-Path: owner-ports Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id MAA21296 for ports-outgoing; Fri, 1 Dec 1995 12:05:13 -0800 Received: from time.cdrom.com (time.cdrom.com [192.216.222.226]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id MAA21290 for ; Fri, 1 Dec 1995 12:05:07 -0800 Received: from localhost (localhost [127.0.0.1]) by time.cdrom.com (8.6.12/8.6.9) with SMTP id MAA01773; Fri, 1 Dec 1995 12:04:30 -0800 To: Sergio Lenzi cc: ports@freebsd.org Subject: Re: INGRES In-reply-to: Your message of "Fri, 01 Dec 1995 09:12:21 GMT." Date: Fri, 01 Dec 1995 12:04:30 -0800 Message-ID: <1771.817848270@time.cdrom.com> From: "Jordan K. Hubbard" Sender: owner-ports@freebsd.org Precedence: bulk > > Ref documentation. > > The process I use to "trace bugs" in the source code > leads to "rewrite" all the code so an "patch" from the original, > will result in a file greater than the sources I offer. It sounds then like this is a case of an "orphaned port", or one that has changed significantly enough that it certainly can no longer really be called the same Ingres port we've seen from Berkeley. I still think it should be a "port", however, it just means that the original tarball will have to live in our LOCAL_PORTS. To get a working RDB back in ports/databases, I think this would be worth the price. Jordan > > It was done as: > 1) using an "indent" from unixware it rewrote all the source > code in a way it becomes K&R c. > 2) use of debug that can trace tasks (again in unixware). > 3) use of lint in the all the source code. > 4) Changed the structure that define tasks from structure to > pointer to typedefs. Of course after that nothing > worked at all. (That is the idea). > 5) With the help from the debugger and "cscope" (program that > makes references within c procedures and calls), I > started to make it running. After 2 weeks the > references to all procedures wrere fixed. > 6) Traced and removed references to a 16bit machine and it's > 16 bit integers. > 7) Implement of a program that calculates the offset in > the pages based in constans from the specific > machine "sparc,hpux,intel..." > 8) Almost rewritten the equel processor. It only let you > define a name say "STUDENT". When later in the program > you need to define "A.STUDENT" the scanner reports > dupplicate name. So I changed the scanner from tree > to list logic, working in the yacc also. > After that I wrote a program (system) to control video-shops > and with the same code compiled in Unixware, linux or BSD, > I installed in 3 video shops from a friend of mine. > The system is running by 6 months or so, without ANY problems. > In the shops it uses 386DX with 8Mb of memory and 500MB of disk, > with a multi serial of 4 ports running a console and 2 dumb > terminals. Now there is only Unixware and BSD (linux file systems was week, > as sometimes the power goes down and it losts some of the open files). > > If you're intereste, > I can send you a program as an example on how I've done a real > application (C written). > > Very Thanks, > > Sergio de Almeida Lenzi. >