Re: Hot Standby, release candidate?

On Mon, 2009-12-14 at 11:07 +0000, Simon Riggs wrote:
> Thanks for the further review, much appreciated.
>
> On Mon, 2009-12-14 at 11:54 +0200, Heikki Linnakangas wrote:
> > Simon Riggs wrote:
> > > Enclose latest version of Hot Standby.
>
> > * LockAcquireExtended needs a function comment. Or at least something
> > explaining what report_memory_error does. And perhaps the argument
> > should be named "reportMemoryError" to be in line with the other arguments.
>
> OK

Done

> > * We tend to not add remarks about authors in code (comments in standby.c).
>
> OK

Done

> > * This optimization in GetConflictingVirtualXIDs():
> >
> > > + /*
> > > + * If we don't know the TransactionId that created the conflict, set
> > > + * it to latestCompletedXid which is the latest possible value.
> > > + */
> > > + if (!TransactionIdIsValid(limitXmin))
> > > + limitXmin = ShmemVariableCache->latestCompletedXid;
> > > +
> >
> > needs a lot more explanation. It took me a very long time to figure out
> > why using latest completed xid is safe.
>
> OK. Took me a long time as well.

Done

> > * Can you merge with CVS HEAD, please? There's some merge conflicts.
>
> Huh? I did. And tested patch against a CVS checkout before submitting.
> Can you explain further?

Still not sure what conflicts you see or where they might come from...

I am now unable to push these changes to the shared repo. What is
happening?

--
Simon Riggs www.2ndQuadrant.com