Ticket History #260: master_princ multiply defined: kdb5_util.c, server_kdb.c

From bjaspan@MIT.EDU Wed Nov 27 17:39:56 1996
Received: from MIT.EDU (PACIFIC-CARRIER-ANNEX.MIT.EDU [18.69.0.28]) by rt-11.MIT.EDU (8.7.5/8.7.3) with SMTP id RAA28270 for <bugs@RT-11.MIT.EDU>; Wed, 27 Nov 1996 17:39:56 -0500
Received: from BEEBLEBROX.MIT.EDU by MIT.EDU with SMTP
id AA20093; Wed, 27 Nov 96 17:39:54 EST
Received: by beeblebrox.MIT.EDU (940816.SGI.8.6.9/4.7) id WAA03108; Wed, 27 Nov 1996 22:39:52 GMT
Message-Id: <199611272239.WAA03108@beeblebrox.MIT.EDU>
Date: Wed, 27 Nov 1996 22:39:52 GMT
From: bjaspan@MIT.EDU
Reply-To: bjaspan@MIT.EDU
To: krb5-bugs@MIT.EDU
Subject: master_princ multiply defined: kdb5_util.c, server_kdb.c
X-Send-Pr-Version: 3.99

mit

System: IRIX beeblebrox 5.3 11091812 IP22 mips



While compiling kdb5_util on Irix, I received the following error:

cc -L./../../lib -L./../../lib -o kdb5_util kdb5_util.o dump.o dumpv4.o loadv4.o kdb5_create.o kadm5_create.o string_table.o kdb5_stash.o kdb5_destroy.o ovload.o import_err.o strtok.o -lkadm5srv -lgssrpc -lgssapi_krb5 -lkdb5 -lkrb4 -ldes425 -lkrb5 -lcrypto -ldyn ./../../lib/libdb.a -lcom_err -lnsl -lgen
ld:
Warning: ./../../lib/libkadm5srv.a(server_kdb.o): master_princ: multiply defined
previous (used) definition from 'kdb5_util.o';
new (ignored) definition from './../../lib/libkadm5srv.a'
Warning: ./../../lib/libkadm5srv.a(server_kdb.o): master_encblock: multiply defined
previous (used) definition from 'kdb5_util.o';
new (ignored) definition from './../../lib/libkadm5srv.a'
Warning: ./../../lib/libkadm5srv.a(server_kdb.o): master_keyblock: multiply defined
previous (used) definition from 'kdb5_util.o';
new (ignored) definition from './../../lib/libkadm5srv.a'

I have confirmed that the symbol is, in fact, multiply defined in
those two files and, thinking back, I suspect it has been that way for
a long time. I have no idea why the linker is succeeding on all
platforms; I thought this was a fatal error on at least some systems.
Anyway, we should fix the problem.

I believe the right solution is to make the declaration in kdb5_util.c
extern, but I could be wrong.




From: "Barry Jaspan" <bjaspan@MIT.EDU>
To: bjaspan@MIT.EDU
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-admin/260: master_princ multiply defined: kdb5_util.c, server_kdb.c
Date: Wed, 27 Nov 1996 17:44:47 -0500

Actually, I the currently global master_princ and related variables
(master_encblock, etc) in libkadm5srv could probably be moved into the
API handle instead of being global. server_kdb.c could keep the info
in a static variable so that, if kadm5_init() is called multiple times
on a server without a stash file, the password is only needed once.

State-Changed-From-To: open-closed
State-Changed-By: bjaspan
State-Changed-When: Tue Dec 3 16:07:45 1996
State-Changed-Why:

Ooops, the previous message was supposed to be a state changed to
closed.


From: "Barry Jaspan" <bjaspan@MIT.EDU>
To: bjaspan@MIT.EDU
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-admin/260: master_princ multiply defined: kdb5_util.c, server_kdb.c
Date: Tue, 3 Dec 1996 21:06:43 GMT

I fixed the multiply-defined errors, and for now worrying about the
globals is more effort than is necessary.

Files:

kadmin/dbutil/ChangeLog
kadmin/dbutil/kdb5_util.c