Re: [pgScript patch] Output + bug fix

Sorry - I meant some pgScript code that would create a message so i
can figure out where it gets lost.

On 3/6/09, Mickael Deloison <mdeloison@gmail.com> wrote:
> 2009/3/6 Dave Page <dpage@pgadmin.org>:
>> There were some cases where error messages would be lost, and part of
>> the fix was to try to make errors all go through SetLastResultError()
>> for consistency in output. I modified SetLastResultError() so that if
>> it's passed NULL, it gets the message via: lastResultError.msg_primary
>> = GetLastError(); instead, so it /should/ get the same message.
>>
>> Can you supply an example of how to trigger the warning please? I
>> assume it's a pgScript thing, as I get notices and errors from the
>> backend just fine.
>>
>
> Here is the code you requested, I removed the useless parts:
>
> pgQueryThread thread(m_app->connection(), stmt);
>
> if (thread.Create() == wxTHREAD_NO_ERROR)
> {
>     if (thread.Run() == wxTHREAD_NO_ERROR)
>     {
>         // ...
>     }
>
>     if (thread.ReturnCode() != PGRES_COMMAND_OK
>             && thread.ReturnCode() != PGRES_TUPLES_OK)
>     {
>         // ...
>         wxString message(stmt + wxT("\n") +
> -->            thread.GetMessagesAndClear().Strip(wxString::both));
>         // ...
>         (*m_cout) << message << wxT("\n");
>         // ...
>     }
>
>
> Now, I always have nothing when I call GetMessagesAndClear().
>


--
Dave Page
EnterpriseDB UK:   http://www.enterprisedb.com