Björn, I've always found (with Samba 3) that this works for me after a clean 'configure'...
make proto_exists
make -j4 all
Maybe I've been lucky and have to confess to being largely ignorant of the Makefile's innards!
Regards,
Will.
William Charles,
Platform Engineering, Asia/Pacific.
Deutsche Bank AG Sydney, +61 (0)2 8258 1916.
Björn JACKE <bjoern at j3e.de>
Sent by: samba-technical-bounces+william.charles=db.com at lists.samba.org
To
samba-technical at samba.org
11/10/2006 05:30 AM cc
Subject
make samba3 make -j proof
Hi,
sitting in a train today waiting for a samba compile I thought it
might be quite a bit faster if it would be compilable parallelized. I
saw that the changes neccessary in the Makefile to make this work are
not very big. See attached patch. I tested the Makefile with make -j20
without a problem. On a dual core machine the build is in fact twice
as fast now with make -j2.
I needed to replace the suffix rule in the makefile with a implicit
pattern rule to be able to add dependencies to each resulting target.
VERY ancient make versions might not cope with that but all machines
that can cope with samba4's makefiles can cope with this, too. So I
thing there's nothing to worry about.
Any volunteer to check this in?
Cheers
Bjoern
[attachment "samba3-make-j-proof.patch" deleted by William Charles/Sydney/DBAustralia/DeuBa] [attachment "attqza0n.dat" deleted by William Charles/Sydney/DBAustralia/DeuBa]
---
This e-mail may contain confidential and/or privileged information. If you
are not the intended recipient (or have received this e-mail in error)
please notify the sender immediately and destroy this e-mail. Any
unauthorized copying, disclosure or distribution of the material in this
e-mail is strictly forbidden.