Greenplum patch for pgAdmin

From: Dave Page <dpage(at)pgadmin(dot)org>
To: Chuck McDevitt <cmcdevitt(at)greenplum(dot)com>
Cc: pgadmin-hackers <pgadmin-hackers(at)postgresql(dot)org>
Subject: Greenplum patch for pgAdmin
Date: 2009-03-06 15:08:00
Message-ID: 937d27e10903060707u2d8ad26cle2b97f64c08e8523@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgadmin-hackers

Hi Chuck,

I've spent the last few hours working on your patches to pgAdmin. I've
fixed/cleaned up some issues, and had to leave others for you to work
on due to time constraints. Please see the attached patch which
includes all changes and new files, and should apply directly against
a clean source tree from SVN trunk.

Fixed items
=========

* Tablespace creation is disabled for Greenplum servers (do we need to
do the same for any other object types?).

* gpPartition objects were being created such that their factory
appeared to be tableFactory. This prevented the list of partitions
being displayed and possibly other problems. I've added a new
constructor to pgTable which allows us to pass partitionFactory from
gpPartition::gpPartition.

* There was an unused additional constructor in gpResourceQueue.
Unfortunately, only that one set the metaType correctly.

* We no longer offer an option to create a new resource queue.

Unfixed items
==========

* PDF helpfiles are unsupported. You can probably add support for them
fairly easily by hacking up DisplayExternalHelp() (in misc.cpp) to
have it open PDF's in the default browser.

* The behaviour of gpPartition objects seems bogus to me. There are a
number of issues:
- If I select New Partition from the context menu on the Partitions
node, I get 'Not implemented'. That message box should never be seen
in a release version of pgAdmin - instead, the New Partition menu
option should be removed. That also applies to the 'create an object
of this type' button, which is also enabled when a partition is
selected.
- I'm able to modify partitions by adding or dropping sub objects
such as columns or constraints (albeit, with the occasional
assertion). Shouldn't such options be disabled?
- If I should be able to modify partitions, then the refresh
mechanism is broken, as modified partitions will vanish from the
treeview until the Partitions node is refreshed.

The following items must also be fixed before release, but needn't
hold up the beta (per previous discussion):

* isGpApp() should check the output of pg_dump --help for
"--gp-syntax" to determine whether it is a Greenplum utility.

* gpPartition::GetSQL() should retrun an ALTER TABLE statement.

Can you let me have an eta for these fixes ASAP please, so I can plan
the beta schedule? The intention is still to release beta 1 early next
week after which no new features will be added until we go into
development mode again.

Thanks.

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

Attachment Content-Type Size
gp.zip application/zip 102.6 KB

Responses

Browse pgadmin-hackers by date

  From Date Subject
Next Message Dave Page 2009-03-06 16:45:22 pgAdmin - notice of intent to change licence
Previous Message svn 2009-03-05 20:35:33 SVN Commit by dpage: r7649 - trunk/www/development