AW: Can the backend return more than one error message per PQexec? - Mailing list pgsql-hackers

From Zeugswetter Andreas SB
Subject AW: Can the backend return more than one error message per PQexec?
Date
Msg-id 11C1E6749A55D411A9670001FA687963368312@sdexcsrv1.f000.d0188.sd.spardat.at
Whole thread Raw
List pgsql-hackers
>     ERROR: blah blah
>     CODE: 12345
>     LOCATION: some/file.c line NNN

It might be handy to have the LOCATION in the postmaster log,
or make it something that needs to be explicitly switched on.
I do not think it is of general interest to users (most errors 
will result from normal operation, and not bugs that need to be tracked).

Since access to SQLSTATE will become a hot path once savepoints 
are available I think that having SQLSTATE up front would be
more convenient.

exec sql insert into blabla values ....;
if (strncmp(sqlca.sqlstate, "23", 2) == 0)    // duplicate key valueexec sql update blabla set ... ;

Andreas


pgsql-hackers by date:

Previous
From: Zeugswetter Andreas SB
Date:
Subject: AW: Re: [PATCHES] Fw: Isn't pg_statistic a security hol e - Solution Proposal
Next
From: Tom Lane
Date:
Subject: Re: [SQL] behavior of ' = NULL' vs. MySQL vs. Standards