Ticket History #489: 1.0.2 compile problem

From brian@karoshi.ucsd.edu Tue Nov 4 18:32:34 1997
Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU [18.72.1.2]) by rt-11.MIT.EDU (8.7.5/8.7.3) with SMTP id SAA01118 for <bugs@RT-11.MIT.EDU>; Tue, 4 Nov 1997 18:32:33 -0500
Received: from karoshi.ucsd.edu by MIT.EDU with SMTP
id AA05602; Tue, 4 Nov 97 18:32:30 EST
Received: (from brian@localhost)
by karoshi.ucsd.edu (8.8.5/8.8.5) id PAA12044
for krb5-bugs@mit.edu; Tue, 4 Nov 1997 15:32:29 -0800 (PST)
Message-Id: <199711042332.PAA12044@karoshi.ucsd.edu>
Date: Tue, 4 Nov 1997 15:32:29 -0800 (PST)
From: Brian Kantor <brian@karoshi.ucsd.edu>
To: krb5-bugs@MIT.EDU
Subject: too modern


State-Changed-From-To: open-closed
State-Changed-By: tlyu
State-Changed-When: Tue Nov 4 21:27:15 1997
State-Changed-Why:

user error; need ANSI C compiler

The 'signed' type does not exist in SunOS 4.1.3_U1 'cc'; the 'db' utilities
won't compile until it is defined to a null.

The 'const' type also does not exist. This causes a problem
in clib/memmove.c
- Brian

From brian@karoshi.ucsd.edu Wed Nov 5 00:28:47 1997
Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU [18.72.1.2]) by rt-11.MIT.EDU (8.7.5/8.7.3) with SMTP id AAA02705 for <bugs@RT-11.MIT.EDU>; Wed, 5 Nov 1997 00:28:47 -0500
Received: from karoshi.ucsd.edu by MIT.EDU with SMTP
id AA12379; Wed, 5 Nov 97 00:28:44 EST
Received: (from brian@localhost)
by karoshi.ucsd.edu (8.8.5/8.8.5) id VAA12560;
Tue, 4 Nov 1997 21:28:43 -0800 (PST)
Message-Id: <199711050528.VAA12560@karoshi.ucsd.edu>
Date: Tue, 4 Nov 1997 21:28:43 -0800 (PST)
From: Brian Kantor <brian@karoshi.ucsd.edu>
To: krb5-bugs@MIT.EDU, tlyu@MIT.EDU
Subject: Re: K&R vs Ansi C


State-Changed-From-To: open-closed
State-Changed-By: tlyu
State-Changed-When: Fri Nov 7 18:47:17 1997
State-Changed-Why:

Should have been appeneded to krb5-build/489

You're quite correct, in 1.0.2 the gssapi and other parts do require
an ANSI 'C' compiler. On our old suns, I had simply bypassed those
parts as they were not needed. My apologies. - Brian

From brian@karoshi.ucsd.edu Tue Nov 4 18:30:21 1997
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 SAA01098 for <bugs@RT-11.MIT.EDU>; Tue, 4 Nov 1997 18:30:20 -0500
Received: from karoshi.ucsd.edu by MIT.EDU with SMTP
id AA15905; Tue, 4 Nov 97 18:30:24 EST
Received: (from brian@localhost)
by karoshi.ucsd.edu (8.8.5/8.8.5) id PAA12041
for krb5-bugs@mit.edu; Tue, 4 Nov 1997 15:30:17 -0800 (PST)
Message-Id: <199711042330.PAA12041@karoshi.ucsd.edu>
Date: Tue, 4 Nov 1997 15:30:17 -0800 (PST)
From: Brian Kantor <brian@karoshi.ucsd.edu>
To: krb5-bugs@MIT.EDU
Subject: 1.0.2 compile problem


Responsible-Changed-From-To: gnats-admin->tlyu
Responsible-Changed-By: tlyu
Responsible-Changed-When: Tue Nov 4 21:25:05 1997
Responsible-Changed-Why:

refiled

State-Changed-From-To: open-closed
State-Changed-By: tlyu
State-Changed-When: Tue Nov 4 21:25:22 1997
State-Changed-Why:

user error


From: Tom Yu <tlyu@MIT.EDU>
To: brian@karoshi.ucsd.edu
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-build/489: 1.0.2 compile problem
Date: Tue, 4 Nov 1997 21:26:39 -0500

We're sorry, but you must use an ANSI-compliant C compiler in order to
build krb5-1.0.x. It should still be possible to compile applications
against the headers and libraries with non-ANSI compilers, though.

---Tom

From: Brian Kantor <brian@karoshi.ucsd.edu>
To: tlyu@MIT.EDU
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-build/489: 1.0.2 compile problem
Date: Tue, 4 Nov 1997 20:15:42 -0800 (PST)



Certainly the 'configure' code should check for this ANSI requirement
if indeed it is one. It is inexcusable that 'configure' should
not demand something that is supposed to be required - you can't
have it both ways.

Actually, with just the two changes I sent in previously, a non-ANSI
compiler is quite happy with the whole package.

Every single bit of the krb5 code (except the two files I mentioned)
*IS* K&R compatable. It would be really nice if you kept it that way.
Would you like the diffs? They're quite small.
- Brian


From: Tom Yu <tlyu@MIT.EDU>
To: Brian Kantor <brian@karoshi.ucsd.edu>
Cc: tlyu@MIT.EDU, krb5-bugs@MIT.EDU
Subject: Re: krb5-build/489: 1.0.2 compile problem
Date: Tue, 4 Nov 1997 23:23:29 -0500



We should probably fix this, yes... but it will take some work.


I disbelieve. We use ANSI C features such as string concatentation in
many places, and I believe some of the kadmin code uses
unconditionalized ANSI prototypes in places.

---Tom

From: Brian Kantor <brian@karoshi.ucsd.edu>
To: krb5-bugs@MIT.EDU, tlyu@MIT.EDU
Cc: Subject: Re: krb5-build/489: K&R vs Ansi C
Date: Tue, 4 Nov 1997 21:28:43 -0800 (PST)

You're quite correct, in 1.0.2 the gssapi and other parts do require
an ANSI 'C' compiler. On our old suns, I had simply bypassed those
parts as they were not needed. My apologies. - Brian
shs.c line 216 is an ANSI prototype and won't compile under the
SunOS 4.1.3_U1 'cc' compiler. Converting it to K&R prototype fixes it.
- Brian