Re: pgAdmin IV : Unittest modular patch(database child nodes) and trigger function

On Tue, Aug 23, 2016 at 1:56 PM, Priyanka Shendge <
priyanka(dot)shendge(at)enterprisedb(dot)com> wrote:

> I noticed whilst testing that I've started seeing the error below
> (with or without the patches I just applied). Any idea what's causing
> it?
>
> The user may already present in database (i.e. abc(at)EXAMPLE(dot)COM) which we
> creating as dummy user.
> When we run test case it tries to create dummy user. As user is already
> present it throws an error.
> It does not get an id for same user.
>

Ahh, yes - can you add a check for that so it throws a more helpful error
please?

Also; related to the issue I've been seeing with databases not being
removed, I recently ran into problems where tablespaces weren't removed,
following a ctrl+c, which then required manual cleanup once I'd figured out
what was wrong. We need to fix this; I'd suggest:

- Keep track of objects that are created during tests. This probably just
needs to be objects to the database level, e.g. pgAdmin users, tablespaces,
databases and roles (because, if we drop the test database then everything
in it will go anyway).

- Remove objects from the list again if a test drops them.

- Call a function on exit (or abort/ctrl+c/whatever) that iterates through
the list of objects that have been created but not dropped, and clean them
up manually before exiting.

Thanks!

--
Dave Page
VP, Chief Architect, Tools & Installers
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake