Обсуждение: Double feature error messages

Hi developers!

Testing the pgAdmin III 1.8.0  Beta 5 (Sep 10 2007, rev: 6631). Client
Win XP, host: Debian Etch / PG 8.2.4 and Debian Sarge / PG 8.1.8


Error messages are displayed twice. Just try any faulty statement to see.
One line seems redundant?


SELECT foo;

=== quote ===

FEHLER:  Spalte »foo« existiert nicht

********** Error **********

FEHLER: Spalte »foo« existiert nicht
SQL state: 42703

=== unquote ===


Maybe helpful:
The first instance of the error message has two spaces after the colon,
the second has only one. Otherwise they are identical.


Regards
Erwin

Erwin Brandstetter wrote:
> Error messages are displayed twice. Just try any faulty statement to see.
Forgot to mention: concerns SQL query dialog window.

/E

I have checked on pgAdmin v1.6.3 and v1.7. Both print only one instance
of the message.
So this was introduced with v1.8 at some point, probably together with
the line "********** Error **********".

/E

Erwin Brandstetter wrote:
> Hi developers!
>
> Testing the pgAdmin III 1.8.0  Beta 5 (Sep 10 2007, rev: 6631). Client
> Win XP, host: Debian Etch / PG 8.2.4 and Debian Sarge / PG 8.1.8
>
>
> Error messages are displayed twice. Just try any faulty statement to see.
> One line seems redundant?

That behavious makes more sense when you understand why it does it that
way. Whilst the queries are running, messages from the backend are
periodically collected and displayed by the query thread. This is the
first error you see:

ERROR:  column "gerp" does not exist
LINE 1: select gerp
               ^

When the query completes the final status is displayed with the *****
Error ***** title:

********** Error **********

ERROR: column "gerp" does not exist
SQL state: 42703
Character: 13

This largely comes about from some refactoring of the query tool when we
discovered that you couldn't run more than 2 queries concurrently. I've
yet to figure out a better way of handling this.

BTW; I've committed a change to use the SQL Font (normally fixed width)
for the messages and history panes - makes the carat error markers work
more meaningfully!!

Thanks, Dave

dpage@postgresql.org wrote:
> (...)
> This largely comes about from some refactoring of the query tool when we
> discovered that you couldn't run more than 2 queries concurrently. I've
> yet to figure out a better way of handling this.
>

Well, it doesn't break anything. It's just a bit confusing.


> BTW; I've committed a change to use the SQL Font (normally fixed width)
> for the messages and history panes - makes the carat error markers work
> more meaningfully!!
>

Sounds like a nice by-product. :)


Regards
Erwin