On Saturday, July 9, 2011, Guillaume Lelarge <guillaume@lelarge.info> wrote:
> On Sat, 2011-07-09 at 13:39 +0100, Dave Page wrote:
>> On Saturday, July 9, 2011, Guillaume Lelarge <guillaume@lelarge.info> wrote:
>>
>> > I would prefer to only refresh the old parent node, and the new parent
>> > node. But we have no way to know the new parent node. So I refresh all
>> > schemas. Which will bring other performance issues for users with big
>> > schemas.
>>
>> I wonder if we need to think about a way of passing more info down to
>> the lower guts of the execution mechanism. We kinda did that already
>> when we added the 2 part SQL execution, maybe now we should think
>> about a more extensible technique (maybe as simple as passing a simple
>> struct of stuff to deal with).
>>
>
> Do you have an example? I'm not sure I really understand what you mean.
(I'm not looking at any code right now, so please bear that in mind...)
Currently the code that executes the SQL that's generated by a dialog
when you click the ok button, just does a relatively simple check to
see if there's something to execute in either of the SQL boxes iirc.
I'm suggesting that we might have a more complex data structure that
we populate with info about "stuff to do" when the OK button is
clicked - for example, a schema name to refresh, an array of SQL
statements to run in separate transactions etc.
--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake
EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company