RE: [pcp] pcp updates - pmlc <---> pmlogger access controls

["Ken McDonell" <kenj@xxxxxxxxxxxxxxxx>]

> -----Original Message-----
> From: Nathan Scott [mailto:nathans@xxxxxxxxxx]
> Sent: Monday, 5 May 2014 5:33 PM
> ...
> I'm seeing a QA failure on 381 - looks like the pmlc flush command is able to
> get EPERM under some conditions.  Could be a primary logger using the (pre-
> existing) system logger config with a new pmlc?
> 
> Since flush is now a no-op, should this just always succeed perhaps,
> independent of any permission settings?  .bad file attached - everything else
> is passing for me here.

Thank for this.  And I agree.

With the next round of changes, pmlc sync/flush is a no-op at the new pmlogger  
and always returns success.

So this means ...

old pmlc <--> new pmlogger; does not show a regression
new pmlc <--> old pmlogger; historical behaviour preserved
new pmlc <--> new pmlogger; flush/sync is a no-op independent of access controls