Ok.. So which is the latest stable version .. OpenLDAP and BDB ?
Thanks, Sumith

On Thu, Apr 10, 2008 at 10:20 AM, Aaron Richton <[EMAIL PROTECTED]>
wrote:

> Is it just me or do people on this list never support problems related to
> > anything but the latest release? Not everyone can immediately upgrade to the
> > latest release to fix any issues they have, and they shouldn't have to
> > either. Previous versions (to an extent) should be supported just like any
> > other product.
> >
>
> Well, most "I didn't expect output x" questions end up with:
>
> (0) This behavior is correct; we just need to figure out why/how it came
> to be. This is upgrade-independent, and I think we're pretty good with
> prompting users to figure out if it's the case. As an example,
>
>  Do you have reason to believe the error is wrong? i.e., does the parent
> > > of the entry you are trying to create exist, and do you have appropriate
> > > access (even "disclose" comes to mind in this case, let alone write)?
> > >
> >
> along with the excellent JLDAP code suggestions (it's been years since
> I've run javac) addressed that in this instance.
>
> (1) This behavior is incorrect; in fact, we've already dealt with it. This
> obviously requires an upgrade for resolution. Hypothesize that
> situation--what's going to cause "No such object" as a lie in this
> situation? Most likely would be some bdb/hdb bug. RE23 CHANGES "egrep
> bdb|hdb | wc -l" shows 17 fixes since 2.3.27. Wouldn't you want those ALL
> ruled out?
>
> (2) This behavior is incorrect; we need to work on a solution. This is
> made a LOT easier when a motivated reporter is on the same page as the
> developers, which pretty much means running the latest code. The only proper
> alternative is writing up a clean reproduction, and if a user has that,
> they'd be posting to the ITS instead of openldap-software. So to this
> audience, an upgrade is near-required if it hasn't already been undertaken.
>
>
> Underlying all of this, there's the general Good Citizen maneuver. I mean,
> 2.3.27 has remote crasher vulnerabilities. Obviously that's something that
> sites can choose to consciously enter into, but if you're not clearly
> indicating that you're aware of that situation, we'd be amiss to NOT
> recommend an upgrade. Plus, if you're correct in the contention that this is
> improper behavior (#1 or #2), your next maneuver needs to be to upgrade
> anyway. Bottom line...we're suggesting it because it's the right answer.
>

Reply via email to