At 10:38 AM -0700 1/15/04, Michael Downey wrote:
>
>The real problem we're running into due to this bug is that we have an
>install script that sets up anonymous FTP via calling the
>SYS$MANAGER:TCPIP$CONFIG dcl script.  I don't believe we could change that
>script since it is a system provided script.
>
>A while back I posted some tests I ran on using the system() call.  I
>believe we can get this to work on 7.3+ systems using the system() and
>popen() calls.  I took a bit of a look at the code but wasn't too sure where
>to start as most of the PP_ code is fairly complicated and not much inline
>documentation is provided.  Also my experience with the perl code is
>practially none.  I could probably hack up a fix for this but it will likely
>break things for everyone else.  If someone would be willing to review the
>changes I make and maybe give me an idea of what things likely would need to
>be changed then I probably can handle this task.  Also I'm not too sure what
> defines we have to determine the version of VMS we are on.  If I could do
>something like:
>if (VMS_VERSION > 7.3)
>{
>    all the changes needed to do this
>}
>
>else
>{
>    normal stuff
>}
>
>Then I think I could put these changes in.

I would steer clear of the "PP_ code".  As you say, it's pretty
twisty stuff, and it's also the wrong place to address the fact that
we don't use the CRTL's popen and system functions.  Take a look at
Perl_do_spawn in [.vms]vms.c and try replacing the call to safe_popen
with a call to popen.  There's probably more to it than that, but
that would be a start.

You can also try reversing the following patch:

http://public.activestate.com/cgi-bin/perlbrowse?patch=16029

and see if that gets you what you want.  You won't be able to use
perldoc or anything else that depends on access to the PPFs of the
parent.

Just for the record, both of these suggestions will probably result
in all sorts of problems with the normal operation of Perl.  They
may, however, force your subprocess to choose a different default for
SYS$COMMAND.
-- 
________________________________________
Craig A. Berry
mailto:[EMAIL PROTECTED]

"... getting out of a sonnet is much more
 difficult than getting in."
                 Brad Leithauser

Reply via email to