Hi All-
I want to thank Paul Sand, Andrew Cosgriff, Steve VanDevender and John
Speno for their responses. It looks like the long and the short of it
is that DEC managed to break ndbm under DU 4.0x, so if you are going
to use sendmail-8.8.x, it looks like the thing to do is compile it
with the -DNEWDB flag and link against -ldb. It look like -DNDBM
option _still_ works under DU3.2x. My solution seems to be to build
sendmail with the newdb for all of my 4.0x boxes and with the ndbm
for my 3.2x boxes and retire out the others as I "ugrade".
Also, John Speno correctly points out that m4 under DU4.0x will not
work on the sendmail mc files---either keep an old version from 3.2x
or build the gnu version.
Andrew points out that you can download the latest variant of the
new Berkeley database routines; however, here is an entry from
the sendmail READ_ME file (in the src subdirectory):
NEWDB The new Berkeley DB package. Some systems (e.g., BSD/OS and
Digital UNIX 4.0) have this package pre-installed. If your
system does not have NEWDB installed, get version 1.85
from
http://www.sleepycat.com/packages/db.1.85.tar.gz.
DO NOT use Berkeley DB version 2.X with sendmail. DO NOT
use the version from the Net2 distribution. If you are
still running BSD/386 1.x, you will also need to define
OLD_NEWDB.
Thanks
Sean
Yesterday after much frustration, I wrote:
>Hello-
>
>I have built sendmail-8.8.8 on a Digital Unix 4.0B machine with the
>re-release jumbo patch installed (not the latest one). It seems to
>work OK except that it cannot rebuild the aliases table with greater
>that 36 aliases when built with:
>
> DBMDEF= -DNDBM
>
>I get a segmentation fault and it dies. However, if I build
>it with the
>
> DBMDEF= -DNEWDBM
>
>option, newaliases seems to work just fine. I am curious if this is
>only relevant under 4.0x or if this will also happen under 3.2x (as my
>mailserver is actually 3.2c) which do not have the NEWDB installed on
>them. Is NDBM incompatible with the new cc (I haven't yet tried cc
>-oldc)?
>
>Just curious if anyone else has had a similar experience. Am I
>missing something obvious?
--
-------------------------------------------------------------------------
Sean O'Connell Email: sean_at_stat.Duke.EDU
Institute of Statistics and Decision Sciences Phone: (919) 684-5419
Duke University Fax: (919) 684-8594
Received on Wed Nov 26 1997 - 16:19:24 NZDT