Does anyone else care to comment on this before I submit more patches?
Sean Legassick wrote:
>
> On Tue, Jun 27, 2000 at 08:04:50PM -0400, Nissim wrote:
>
> > OK, Here are the issues that aren't resolved:
>
> Good summary of the questions, here are my proposed answers.
>
> > 1) Should the connection for the transaction be stored within BasePeer
> > (through some sort of static Hashtable keyed on a transactionID) or
> > through a new non abstract class?
>
> Or we use the Connection object itself as the cookie that represents the
> transaction. This seems the simplest, and John's changes to BasePeer are
> a clear step towards this.
>
> Then we can add some convenience methods to BasePeer like
> beginTransaction(), commitTransaction() and rollbackTransaction().
>
> As to the MySQL issue I would say a single supportsTransactions() in DB
> should do the job. After all transactions are required in SQL-92
> (allegedly) and MySQL 3.23 (now GPL :-) ) has them (again allegedly, I
> haven't tried).
>
> > 2) Should the check for whether transactions are needed (i.e. for
> > postgres) be done within BasePeer, or in a higher level peer?
>
> Okay, so when all the above is in place we can look at this Postgres
> issue. It seems reasonable to have code to force a transaction if it is
> actually *needed* as you've proposed in your patches. But this should
> use a more general transaction mechanism like John's patches + the
> above.
>
> > 3) Should the check be:
> > a: Is it postgres
> > b: Check some method in the DB implementation (objectDataNeedsTrans)
>
> Definitely (b) - I've given my reasons.
>
> > 4) should we also check if there's a large object in the table, and only
> > set up a transaction then , or always do transactions on postgresql?
>
> I dunno - How bad is the overhead on just always doing the transaction -
> it would certainly be easier and less intrusive on the code.
>
> Otherwise some work is needed to restructure executeQuery so it has
> knows whether any of the tables contain binary data. This should be
> possible; although we'll need to change the signature of executeQuery -
> does anyone actually use this method directly.
>
> --
> Sean Legassick
> [EMAIL PROTECTED]
>
> ------------------------------------------------------------
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
> Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
> Problems?: [EMAIL PROTECTED]
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
Problems?: [EMAIL PROTECTED]
