OpenVMS port status

[John Malmberg]

I am posting this here because it seems that there is at least 1 person 
monitoring this list that is interested in rsync on OpenVMS.

Build environment:

OpenVMS Alpha 7.3
Compaq C T6.5
Frontport 1.0
Compaq MMS.  Mad Goat Make (MMK) should also work.

RSYNC 2.5.2 sources.

The build is now compiling with no warnings and a minimal number of LINT 
type diagnostics.  It takes 4 minutes 52 seconds to do a clean build.

29 LINT type informational messages are being generated.

33 Modules are being patched by the build procedure to fix either 
compiler diagnostics or OpenVMS incompatabilites.

OpenVMS specific patches are:

FLIST.C - ino_t handling.
HLINK.C - ino_t handling.

FNMATCH.C - "../" in #include statement can not be used on OpenVMS.
           - Overriding prototype of getenv() in stdlib.h with
             incorrect version.  (Not really an OpenVMS specific patch)

MDFOUR.C - "../" in #include statement can not be used on OpenVMS.
ZUTIL.H - "../" in #include statement can not be used on OpenVMS.
         - Overriding prototype of strerror() in string.h with
           incorrect version.  (Not really an OpenVMS specific patch)


Build phases that apparently are working:

Building config.h from config.h.in

Converting rsync.1 and rsync.conf.5 into OpenVMS Digital Standard Runoff 
format, and help libraries are being built.

Building RSYNC.EXE and TLS.EXE.


I still need to deal with the fork() calls in SOCKET.C and UTIL.C in 
order to get a working program.

I am looking at converting at least one of the two processes into AST 
threads instead of a child and parent process.  To do this, I have to 
chart out their flow.

I also have not looked at doing anything with the test scripts, as I 
have not implemented any of the Unix compatable shells on my OpenVMS system.

-John
wb8tyw at qsl.network
Personal Opinion Only