henry@utzoo.uucp (Henry Spencer) (08/25/89)
There is (sigh) a bug in the "updatemin" program that came out with the C News patch two days ago. It fails to unlink its temporary from active, which is big trouble when the *next* invocation of it re-creates the temporary. I was only running it once in regression test, in a messy directory, so I missed this. My apologies. There will be a patch, probably today, to fix this. -- V7 /bin/mail source: 554 lines.| Henry Spencer at U of Toronto Zoology 1989 X.400 specs: 2200+ pages. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
seth@ctr.columbia.edu (Seth Robertson) (08/25/89)
(Those easily bored may skip the first paragraph)
Well after getting my active file trashed by updatemin (and thus
spending several hours watching the dust settle on my terminal while
my program was taking a half-assed stab at rebuilding it) I decided
this was as good a time as any to get dbz working. I compiled
everything and installed it. I then noticed that I hadn't gotten any
new news since I installed it. Nothing in the errlog, nothing, in
fact, in any log. I tried posting only to have relaynews hang during
startup. In desperation, I tried rebuilding the history file &&
suddenly news started to flow. Lets here it for documentation!!!
Summary: If you are upgrading to dbz from dbm YOU MUST REMAKE THE
HISTORY FILE!!!
Other hints include:
cc -O dbz.c -c
ar rcv libdbz.a dbz.o
ranlib libdbz
&& during the config process, tell the script that your dbm library is
called by '-ldbz' (instead of -ldbm) and tell it to include the
directory which libdbz.a resides in the search path (e.x.
'-L/this/way/to/dbz')
and, one more once, rebuild the history file!!!
--
-Seth Robertson
seth@ctr.columbia.edulmb@vicom.com (Larry Blair) (08/25/89)
In article <1989Aug24.191942.5491@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes:
=There is (sigh) a bug in the "updatemin" program that came out with the
=C News patch two days ago. It fails to unlink its temporary from active,
=which is big trouble when the *next* invocation of it re-creates the
=temporary. I was only running it once in regression test, in a messy
=directory, so I missed this. My apologies. There will be a patch,
=probably today, to fix this.
This makes patch # 5 (I think; hard to tell without numbering:-)). So much
for "not enough patches to worry about numbering."
Henry, why don't you consider adding patch numbering to your current scheme?
It would make the rest of us happy, show that you have the capacity for
change, and still allow you to experiment with date-based numbering.
--
Larry Blair ames!vsi1!lmb lmb@vicom.com