Hi Paul,
----- Original Message -----
> Hi Nathan,
>
> On 02/07/2014 10:57 PM, Nathan Scott wrote:
> > The attached patch is where I was at - it looks to see if the
> > root namespace file exists before attempting to add the new sub
> > tree, and if not, Rebuilds it and flags pmcd to be started later
> > on. I've not tested it though, and its something that needs a
> > few more eyes on it (affects every PMDA install). If you could
> > look over it and give it a whirl, that'd be much appreciated.
> >
> I've taken a quick looksee at the patch you provided and had
> to make some slight alterations in order to get it to allow an
> error free PMDA install with this scenario on my RHEL 7 machine.
> With the alterations the attached patch seems to work well and
> the PMDA installed works when PMCD is later started.
Thanks! Good to hear it - I'll take it for a full QA spin and
check all the wierd and wonderful corner cases, verifying lots
of PMDA install scripts in the process.
> After changing the logic of the test for Rebuild to continue
> if found I ran into an issue where PMCPP would error and bail
> whilst parsing a PMDA root (PMDADIR/root) if it included
> stdpmid. For lack of it existing yet.
I think that ones an unrelated bug (fixed already) - 1059004 and
tackled via commit 80601bca. Thus, I think that second step you
added is superfluous when that other fix is in place.
> I found that generating an initial copy of stdpmid using the
> available Make.stdpmid before calling Rebuild [...]
> Because it does not exist yet.
So, it should have, just some dope (me) managed to not install
it, in his eagerness to tackle some other unrelated issue. If
I could just stop introducing a bug for every bug I fix, well,
we'd all be better off. :)
cheers.
--
Nathan
|