Hi Dave,
----- Original Message -----
> This is good news Nathan. It looks to me like most of the failures
> could
> potentially be due to the same ipv6 socket leak. I may be able to
> track
> it down before you wake up. It's Saturday for you now anyway, so no
> need
> for you to delve into it unnecessarily.
>
Thanks. I don't think its a leak, I think its just there now where
it wasn't before (an extra fd, and open port) - the extra log file
diagnostics, netstat output is showing another port, etc.
Oh, except for 197, not sure what that is yet.
> I assume that this is all with a NSS-enabled build. I'm still working
Yep.
> on a few missing IPv6 bits for native sockets.
>
> One other area that I need to address for IPv6 is the code which
> creates/processes ip address wilcards/masks for access control
> purposes.
Ah yes - I bucketed those into "net address related changes".
> Dave
>
> On 02/21/2013 09:04 PM, Nathan Scott wrote:
> > Hi Dave,
> >
> > ----- Original Message -----
> >> ...
> >> It's my expectation that Dave and I will continue knocking off the
> >> remaining QA issues, finishing up the last of the IPv6 and network
> >> API tweaks, then merge to dev, then master shortly thereafter, and
> >> release pcp-3.6.11 with these changes.
> >>
> > These are the remaining failures I have after fixing up everything
> > else after a full pcpqa run today.
> >
> > 023 - pmcd.log format, net address related changes, ipv6 port open
> > 051 - pmcd.log format, net address related changes, ipv6 port open
> > 062 - pmcd.log format, net address related changes
> > 067 - pmcd.log format, net address related changes, ipv6 port open
> > 172 - ipv6 port open? (changes netstat report)
> > 197 - suggests several file descriptors leaked? (expect 4 -> got
> > 9)
> > 243 - bind output on failure now reports for two open ports (ipv6)
> > 244 - pmcd.log format, net address related changes, ipv6 port open
> > 255 - pmcd.log format, ipv6 port open
> > 449 - pmcd.log format, net address related changes
> >
> > I'll take a deeper look into 197, and also look into
> > backward-compat
> > in the QA sources which we've not maintained for other folks at
> > this
> > stage (i.e. people like Ken - testing older PCP versions with
> > latest
> > pcpqa sources).
> >
> > cheers.
> >
> > --
> > Nathan
>
> _______________________________________________
> pcp mailing list
> pcp@xxxxxxxxxxx
> http://oss.sgi.com/mailman/listinfo/pcp
>
|