Sorry for the delay in responding to you.
On Mon, Feb 11, 2013 at 6:03 AM, Phil Sorber <phil@omniti.com> wrote:
> On Fri, Feb 8, 2013 at 1:07 PM, Phil Sorber <phil@omniti.com> wrote:
>> On Fri, Feb 8, 2013 at 12:46 PM, Fujii Masao <masao.fujii@gmail.com> wrote:
>>> No maybe. But I think that all the client commands should follow the
>>> same rule. Otherwise a user would get confused when specifying
>>> options.
>>
>> I would consider the rest of the apps using it as a consensus. I will
>> make sure it aligns in behavior.
>>
>
> I've done as you suggested, and made sure they align with other
> command line utils. What I have found is that dbname is passed
> (almost) last in the param array so that it clobbers all previous
> values. I have made this patch as minimal as possible basing it off of
> master and not off of my previous attempt. For the record I still like
> the overall design of my previous attempt better, but I have not
> included a new version based on that here so as not to confuse the
> issue, however I would gladly do so upon request.
>
> Updated patch attached.
Thanks! The patch basically looks good to me.
I updated the patch against current master and fixed some problems:
- Handle the hostaddr in the connection string properly.
- Remove the character 'V' from the third argument of getopt_long().
- Handle the error cases of PQconninfoParse() and PQconndefaults().
- etc...
Please see the attached patch.
Regards,
--
Fujii Masao