As far as I know, Postgres is smart enough that the SQL statement
CREATE EXTENSION IF NOT EXISTS postgis
actually fails on permissions, only if the extension doesn't exist (we
ran into a very similar issue with the 'hstore' extension just the
other day).
So long as that is true -- I don't really see a problem here. If the
extension is installed, there will be no error; if it isn't, and the
user doesn't have permissions to create it, I see little difference
between an error when setting up the database and an error encountered
in migrations.
Shai.
On Thu, 11 Jan 2018 07:59:05 -0300
Ramiro Morales <[email protected]> wrote:
> On Wed, Jan 10, 2018 at 3:12 AM, Niclas Kron <[email protected]>
> wrote:
>
> > Hello,
> > I am writing to you to establish some sort of consensus on how to
> > proceed with this bug/PR:
> >
> > * PR https://github.com/django/django/pull/9518
> > * BUG https://code.djangoproject.com/ticket/28975
> >
> > With regard to the problem - the current implementation is that a
> > query is ran to determine
> > if the extension exists or not in the current database. That query
> > unlike `CREATE EXTENSION ...`
> > does not require superuser status on the executing database user.
> >
> > The pros and cons for this implementation:
> >
> > ### Pros
> > * Operates as normal with the only addition that if it already
> > exists we dont
> > run the CREATE EXTENSION thus the application wont fail if the user
> > is a non-superuser.
> >
> > ### Cons:
> > * Hard to test due to the fragility of the tests
> >
> > By hard to test, I mean that I can't easily test it because current
> > implementation leads some sort of race condition
> > and sometimes the tests fail. At least if you look at my current
> > tests (yes I am a new to this).
> >
> > I did talk about this on the irc channel #django and #django-dev to
> > try to solve it - however
> > the recommendation was to write on the mailing list to try to
> > establish a consensus on a proper
> > way to go forward. This especially since another create suggestion
> > was put forward by Ramiro Morales
> > as per: https://code.djangoproject.com/ticket/28975#comment:5
> >
> >
> Please disregard that comment as I posted it before or in parallel
> with the discussion you and Tim were
> having via the GitHub PR about a non-admin Postgres user being able to
> introspect at runtime if the
> postgis extension is installed or not.
>
> Had considered the idea myself bud assumed it woudn't be possible or
> that it would introuduce too
> much complexityto the code. Was wrong on both counts.
>
>
> Regards,
>
--
You received this message because you are subscribed to the Google Groups
"Django developers (Contributions to Django itself)" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/django-developers.
To view this discussion on the web visit
https://groups.google.com/d/msgid/django-developers/20180112191150.779cd7a1.shai%40platonix.com.
For more options, visit https://groups.google.com/d/optout.
