Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!sun-barr!cs.utexas.edu!uunet!mcsun!hp4nl!star.cs.vu.nl!wundt!valke
From: valke@wundt.psy.vu.nl (Peter Valkenburg)
Newsgroups: comp.os.minix
Subject: Re: Revision of PC 1.5.10 ps.c
Summary: It's more efficient to coordinate big changes in programs
Message-ID: <1454@wundt.psy.vu.nl>
Date: 3 Sep 90 18:11:43 GMT
References: <4217@crash.cts.com>
Organization: VU Psychologie, Amsterdam, the Netherlands
Lines: 49

In article <4217@crash.cts.com> cwr@pnet01.cts.com (Will Rose) writes:
>This is a revision to ps.c for the IBM PC 1.5.10.  The present version
>works, but is a bear to change when recompiling the system; partly because
>it uses a non-ASCII database, and partly because it either expects permanent
>symbol tables in kernel, mm and fs, or the images on a specific path.
> 
>I changed it to use an ASCII database, produced by the Makefile in tools;
>I also added an option to use an alternate database, handy if you are just
>switching kernels briefly.  There was a logic bug in the selection of
>options which has gone, and processes which ps can't get to grips with now
>have [ ] around them.
> 
>More doubtfully, I renamed the options to V7 standards - I think.
> 

Yes, well of course it's okay for anyone to patch up programs that come with
Minix, indeed it is one of the reasons Minix is improving in quality rather
fast.

What bothers me (being the author of ps ;-) in this case, however, is that a
petite letter to either me or perhaps some of the testers would coordinate
efforts to improve programs like this one in a far more efficient way than is
possible when everybody just hacks away at the sources and posts the results.
I still think that such tactics can work when concerned with fixes for
evident bugs, but in this case there are a few problems:
	1) Will Rose has provided a fix for a problem which isn't there -
	   Ps neither requires kernel/mm/fs to be in standard places, nor
	   that they always contain symbol tables.  You can pass the paths
	   for kernel/mm/fs along with a ps -U, and they don't need to
	   contain symbol tables afterwards.
	2) I have included in ps a mechanism that uses /dev to get terminal
	   names from, and fixed some other minor problems as well.  This was
	   posted a while ago, but the changes are not incorporated in Will's
	   version.
	3) I have received a change to make ps run on 386-Minix, and planned
	   to include that one, with an improved database mechanism, in a
	   new posting.

Again, I'm not saying that nobody should lay his/her hands on "my baby" or
anything silly like that, but I just don't see the sense in having multiple
incompatible versions and out-of-phase updates around.  It's just wasted
effort if nobody knows what the other person is doing (and this is of course
a problem the Minix community is somewhat susceptible to).

Bye,
	Peter Valkenburg (valke@psy.vu.nl).

[P.S.: of course I will go through your source and check out the fixes, Will.
       This article was *not* meant to critisize your work in itself].