Re: Patch: Implement failover on libpq connect level.

On 20 October 2016 at 01:15, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>
> On Wed, Oct 19, 2016 at 12:04 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> > On Thu, Oct 13, 2016 at 5:53 AM, Victor Wagner <vitus(at)wagner(dot)pp(dot)ru> wrote:
> >> On Thu, 13 Oct 2016 12:30:59 +0530
> >> Mithun Cy <mithun(dot)cy(at)enterprisedb(dot)com> wrote:
> >>> On Fri, Sep 30, 2016 at 2:14 PM, Victor Wagner <vitus(at)wagner(dot)pp(dot)ru>
> >>> wrote:
> >>> Okay but for me consistency is also important. Since we agree to
> >>> disagree on some of the comments and others have not expressed any
> >>> problem I am moving it to committer.
> >>
> >> Thank you for your efforts improving my patch
> >
> > I haven't deeply reviewed this patch, but on a quick read-through it
> > certainly seems to need a lot of cleanup work.
>
> Some more comments:
>
> - I am pretty doubtful that the changes to connectOptions2() work as
> intended. I think that's only going to be called before actualhost
> and actualport could possibly get set. I don't think we keep
> reinvoking this function for every new host we try.
>
> - It's pretty clear that this isn't going to work if the host list
> includes a mix of hostnames and UNIX socket addresses. The code that
> handles the UNIX socket address case is totally separate from the code
> that handles the multiple-hostname case.

So should it be the case that it disallows UNIX socket addresses
entirely? I can't imagine a list of UNIX socket addresses being that
useful.

Thom