On Thu, Feb 9, 2017 at 9:40 PM, Michael Paquier
<michael.paquier@gmail.com> wrote:
> On Thu, Feb 9, 2017 at 8:17 PM, Amit Kapila <amit.kapila16@gmail.com> wrote:
>> +1. Sounds sensible thing to do.
>
> Hm. It seems to me that PGPASSFILE still needs to be treated as an
> exception because it is set to $HOME/.pgpass without any value set in
> PQconninfoOption->compiled and it depends on the environment. Similar
> rules apply to fallback_application_name, dbname and replication as
> well, so they would need to be kept as checked on an individual basis.
>
> Now it is true that pg_basebackup -R enforces the value set for a
> parameter in the created string if its environment variable is set.
> Bypassing those values would potentially break applications that rely
> on the existing behavior.
Hmm, I didn't think about environment variables.
> In short, I'd like to think that we should just filter out those two
> parameters by name and call it a day. Or introduce an idea of value
> set for the environment by adding some kind of tracking flag in
> PQconninfoOption? Though I am not sure that it is worth complicating
> libpq to just generate recovery.conf in pg_basebackup.
Yeah, I'm not sure what the best solution is. I just thought it was strange.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company