[Spread-users] Spread release 3.16.0 -- New Open-Source license

[Jonathan Stanton]

On Thu, Jun 28, 2001 at 12:53:38PM +0200, Julien Dufour wrote:
> 
> On jeudi, juin 28, 2001, at 05:00 , George Schlossnagle wrote:
> 
> > MacOSX makefiles and arch.h seem to be broken.  There are extra 
> > linefeeds and improper spacing (instead of tabbing in the 
> > Makefile), and the arch.h file seems to require types not 
> > present on OSX (int32).
> >
> > On Wednesday, June 27, 2001, at 05:29 PM, Jonathan Stanton wrote:
> 
> You are right about the Makefile. There are some issues with 
> linefeeds and spacing. I fixed them and I also added the 
> "-traditional-cpp" option to the CFLAGS and TCFLAGS. This option 
> does not seem to change anything when compiling spread, but it 
> is often needed to avoid conflict with ObjectiveC. That is 
> preventive.

I think the makefile got corrupted by email, as I used one of the versions
of the MacOSX makefiles that were mailed to me by other people. We don't
have a MacOSX machine in the lab so I wasn't able to test the build.

> About compilation errors, "arch.h" is guilty. The Darwin 
> architecture check stands in the i386 section that prevent 
> ARCH_PPC_DARWIN being defined... Moving the check out of the 
> i386 section fix the bug. A ARCH_i386_DARWIN section should also 
> be made, if someone has installed Darwin for i386...

I think this must have been a cut and paste error on my part. I had
thought someone I talked with tested the MacOSX build before we released 
it but I guess I misunderstood.

> Moreover, I met some other issues during linking (lrand48 is not 
> defined). This can be easily fixed by removing the "#define 
> HAVE_LRAND48" from the ARCH_PPC_DARWIN section of "arch.h".
> 
> Spread does compile now. It still remains some warning in the 
> parser (functions defined but not used), in spread.c (implicit 
> declarations of "null_init()", "ip_init()" and "permit_init()") 
> and in "user.c" (funtion "ip_authenticate()" defined but not 
> used), but those warnings should occur on every plateform.

Yes those occuur on all platforms. I knew about the warnings but I didn't
want to just remove them since the right way to remove them takes some 
restructuring work and doing dynamic loading of modules.

> 
> I have also made a very few tests. "spread" and "spuser" 
> binaries seem to be working. I ran two servers on two computers, 
> and two clients. I made the clients connect on the same or on 
> different servers, join a group and send some message. 
> Everything was fine, but deeper test are needed.
> 
> Modified files are attached to this mail. Just replace the 
> original ones and it should work.

I will include them in the next release.

Jonathan
-- 
-------------------------------------------------------
Jonathan R. Stanton         jonathan at cs.jhu.edu
Dept. of Computer Science   
Johns Hopkins University    
-------------------------------------------------------