Re: debug a mess - Mailing list pgsql-interfaces

From Tom Lane
Subject Re: debug a mess
Date
Msg-id 13032.1046208448@sss.pgh.pa.us
Whole thread Raw
In response to Re: debug a mess  (Wei Weng <wweng@kencast.com>)
Responses Re: debug a mess  (Wei Weng <wweng@kencast.com>)
List pgsql-interfaces
Wei Weng <wweng@kencast.com> writes:
> The entire code from PQconnect until PQfinish goes like this:

That sure looks harmless.  I wonder if there is something failing in the
PQconnect() step (like the queries that libpq sends on its own)?

I'd suggest enabling query logging --- see postgresql.conf, or just doexport PGOPTIONS="--log_statement=1"
before starting your program.  (Check Admin Guide first, as this
parameter got renamed recently.)  Then look in the postmaster log to get
a better idea of what's really going on.

BTW, I'm pretty sure that PQclear, and most other libpq functions,
behave sanely when handed a null pointer.  So the test in
>     if (res)
>         PQclear(res);
is unnecessary.
        regards, tom lane


pgsql-interfaces by date:

Previous
From: "Jeroen T. Vermeulen"
Date:
Subject: Re: debug a mess
Next
From: "D'Arcy J.M. Cain"
Date:
Subject: Re: Python interface and Money?