[Nagiosplug-devel] Roadmap to 1.3.0

[Karl DeBisschop]

On Thu, 2002-09-26 at 23:24, Subhendu Ghosh wrote:
> On 25 Sep 2002, Karl DeBisschop wrote:
> 
> > I'd like to consider creating a clearcut roadmap to a 1.3.0 release.
> > 
> > I agree that plugin stablility is key. There are also some vague
> > thoughts in the ROADMAP file.
> > 
> > But right now, AFAICT, there is no real threshold for saying when we
> > should call it a release.
> > 
> > Have people been thinking about this?
> > 
> > If so, care to share your thoughts?
> > 
> 
> I think the threshold for 1.3.0 should be fixes for the known 
> issues/patches that have been posted. (form mailing list and patch list on 
> sourceforge)
> 
> - check_snmp memory allocation
> - check_snmp - mibdir and snmpv3 support
> - check_smtp - remove ifdef and move functionality to an option.
> - check_hpjd - not NULL community
> - check_http - client certificate
> - check_http - Russel's patches
> - check_snmp_interface - merge check_ifoperstatus and check_xinterface
> - check_nntp - -g support
> 
> 
> 
> Optional - merge check_smtp and check_nntp changes back into check_tcp and
> create a symlink target.  This creates a question for check_smtp - can we 
> provide additional options for specific versions?

We can, but only to the extent that is help code management. As an
example, it's pretty clear that we will not merger check_http with
check_tcp. So it pretty much depends on how many options, and how
complex they are.

> Note: the ROADMAP says 1.3.x is a devel branch and fixes are going to be 
> added to the 1.2.9 branch.  Since we started nagiosplug as a new 
> CVS project, is this still true?

I really doesn't matter to me. I was only going that way because other
people on the list had suggested it, and a slim majority seemed to
prefer it.

> Other comments?

I'm looking into the possibility of removing all sprintf code, or at
least doing an audit to ensure there are no potential buffer overfolws
possibilities.

Also, I need to be sure that all of the reverse compatibility option
processing ind ripped out. THat caode was bug ridden. May already be
done, but I need to check.