Building GNU make 3.59 in ESIX. Help needed.

Chin Fang fangchin at elaine39.stanford.edu
Tue Dec 18 17:09:29 AEST 1990


Hi everyone,

I just moved GNU make 3.59 to my machine.  At first sight it didn't
look to involved to build.  I modified the Makefile and followed the
porting hints, putting in -DUSGr3 for defines, -DNO_LDAV for
LOAD_AVG, and -lPW for LOADLIBES. Typed make. After much disk churning
and plenty complains (warning: illegal pointer/integer combination, op=)
from cc (YES, I was using cc first to see if it would work), make quit
due to some unresolved symbols typically one would find in Berkeley
UNIX environment, eg. vfork, bcopy, NGROUPS, O_RDWR, O_RDONLY etc. 

I was puzzled so I went back, logged into a SUN sparc, pirated some
definitions and substituted these Berkeley calls using their System V
(ESIX to be precise) counterparts in these trouble-causing modules.
Still, I got nowhere.

I gave up -DUSGr3, changed it to -DUSG (for generic System V), then this
beast compiled.  Time to try it.  I renamed it to gmake first, then
rm *.o; gmake -f Makefile to see if it would work.

It core dumped!

To make this posting short, I won't put in some other (confessedly
silly) commbinations that I tried.  

I wonder could someone who has ever built GNU make (not necessarily
3.59, 3.58 is fine too) successfully give me some hints as to 

(1) what I did wrong?
(2) which compilation switches I should use? and
(3) Do I need any special patch to make GNU make work in ESIX
    environment?

Any help will be greatly appreciated.  I truely want to see how good
gmake is as comparing to AT&T make.  Without a working copy obviously
I can't do it.  Could someone help me out?

E-mails are most welcome.  However I read this newsgroup very often,
so postings are fine too.

Thanks in advance to anyone who helps

Chin Fang
Mechanical Engineering Dept.
Stanford University
fangchin at portia.stanford.edu



More information about the Comp.unix.sysv386 mailing list