Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!ncar!gatech!mcnc!rti!tijc02!cgh018
From: cgh018@tijc02.UUCP (Calvin Hayden    )
Newsgroups: comp.unix.wizards
Subject: Re: Bugs in the AT&T Toolchest program 'nmake'
Message-ID: <464@tijc02.UUCP>
Date: 9 May 89 09:22:30 GMT
References: <1640@internal.Apple.COM>
Organization: Texas Instr., Johnson City TN
Lines: 37

> I have the AT&T Toolchest 'nmake' program, version 1.0 and was wondering if 
> there's a poor soul out there who's found some bugs in this product and
> would be willing to tell me about them so I don't go crazy trying to
> build these new makefiles...

Well, here goes.  First, we are using nmake now, and I have talked with 
Mark Forman (At&t toolchest admin) about the bugs (features :->) associated
with nmake.  I've been told that At&t has made major changes to nmake
internally, and that they are not going to release this to Mark, despite 
all of his efforts -- this conversation took place around Feburary, so dont
hold your breath for another update to come from the toolchest.

As far as bugs... I have a list of bug fixes, but unfortunately, these
were supplied by engineers using nmake, and having source to it -- I don't
qualify as a C guru yet.  Most of the changes to be made involved the
source to cpp.  One bug is that when compiling using the cpp supplied
with nmake, often there are warnings about extra characters being ignored.
Another I seem to remember is a problem with matching up #ifdef/#endif's -
initially didn't want to do this.  And another interesting one I remember
is that when making a change to cpp code, and compiling again, I had to 
use the Sys V cpp, the nmake cpp would not work on its own output.  I also
remember some problem with cpp being able to 'eat its own output' - a file
would be run through cpp, and the compiled, and would bomb when the compiler
called cpp itself.  

As I said earlier, I only have an inkling of the problems seen, but I under-
stand that it is being used here now.  There are still some outstanding
problems that have not been fixed -- makes you a bit uncomfortable when you
make a kernel and get 10-15 warnings about extra characters being ignored :->.

I know this isn't too specific, but hopefully, it'll let you know that
you are going to encounter several problems -- mostly with cpp.  I could 
better answer your question if I directly used nmake myself, I am but
a sys admin, not a programmer/engineer.  Hope it helps in any way.

Calvin
uucp:  ...mcnc!rti!tijc03!cgh018