MySQL TranQL connector (fwd)

[Heikki Linnakangas]

Hi,

I sent the below message to the geronimo-devel list in December, but got
no replies. Maybe someone on this list would like to comment.

- Heikki

---------- Forwarded message ----------
Date: Sat, 10 Dec 2005 20:34:21 +0200 (EET)
From: Heikki Linnakangas <hlinnaka-X3B1VOXEql0@...>
Reply-To: dev-nmThH5wfGryY1pf4d8vP8EB+6BGkLq7r@...
Newsgroups: gmane.comp.java.geronimo.devel
Subject: MySQL TranQL connector

Hi,

I've been looking at the MySQL TranQL connector, trying to figure out how to
write one for PostgreSQL. Is this the right list to discuss TranQL?

This caught my eye at the org.tranql.connector.mysql.LocalMCF constructor:

     public LocalMCF() {
-->     super(new MysqlConnectionPoolDataSource(), new MysqlExceptionSorter(),
false);
         ds = (MysqlConnectionPoolDataSource) super.dataSource;
     }

The corresponding superclass constructor signature is:

protected AbstractLocalDataSourceMCF(javax.sql.DataSource xaDataSource,
     ExceptionSorter exceptionSorter, boolean commitBeforeAutocommit)

Now why does LocalMCF give the superclass constructor a
ConnectionPoolDataSource instance while it's expecting a DataSource? It seems
to work because MysqlConnectionPoolDataSource extends MysqlDataSource, but
that's just accidental, isn't it?

Also, the variable name in the superclass constructor is a bit confusing. I
would think that xaDataSource is an instance of XADataSource, but it's not.

- Heikki

[Jeremy Boynes]

This is probably the better list to discuss TranQL itself rather than
its Geronimo integration.

Agreed the name is confusing - I would assume someone (probably me)
was being lazy when they created the LocalMCF and just copied the XA
one.

Things have probably worked so far because most vendors have an
implementation hierarchy where things extend from a common base class.
On the "if it ain't broke don't fix it" principle I don't think we'd
change this unless someone provided a patch or we had an issue with a
new connector.

Having a PostgreSQL connector would be useful - how is it coming along?

--
Jeremy

On 1/19/06, Heikki Linnakangas <hlinnaka@...> wrote:

> Hi,
>
> I sent the below message to the geronimo-devel list in December, but got
> no replies. Maybe someone on this list would like to comment.
>
> - Heikki
>
> ---------- Forwarded message ----------
> Date: Sat, 10 Dec 2005 20:34:21 +0200 (EET)
> From: Heikki Linnakangas <hlinnaka-X3B1VOXEql0@...>
> Reply-To: dev-nmThH5wfGryY1pf4d8vP8EB+6BGkLq7r@...
> Newsgroups: gmane.comp.java.geronimo.devel
> Subject: MySQL TranQL connector
>
> Hi,
>
> I've been looking at the MySQL TranQL connector, trying to figure out how to
> write one for PostgreSQL. Is this the right list to discuss TranQL?
>
> This caught my eye at the org.tranql.connector.mysql.LocalMCF constructor:
>
>      public LocalMCF() {
> -->     super(new MysqlConnectionPoolDataSource(), new MysqlExceptionSorter(),
> false);
>          ds = (MysqlConnectionPoolDataSource) super.dataSource;
>      }
>
> The corresponding superclass constructor signature is:
>
> protected AbstractLocalDataSourceMCF(javax.sql.DataSource xaDataSource,
>      ExceptionSorter exceptionSorter, boolean commitBeforeAutocommit)
>
> Now why does LocalMCF give the superclass constructor a
> ConnectionPoolDataSource instance while it's expecting a DataSource? It seems
> to work because MysqlConnectionPoolDataSource extends MysqlDataSource, but
> that's just accidental, isn't it?
>
> Also, the variable name in the superclass constructor is a bit confusing. I
> would think that xaDataSource is an instance of XADataSource, but it's not.
>
> - Heikki
>
>

[djencks]

On Jan 19, 2006, at 9:50 AM, Heikki Linnakangas wrote:

> Hi,
>
> I sent the below message to the geronimo-devel list in December,  
> but got no replies. Maybe someone on this list would like to comment.

This is a better list :-) but I don't have very intelligible comments  
yet.

>
> - Heikki
>
> ---------- Forwarded message ----------
> Date: Sat, 10 Dec 2005 20:34:21 +0200 (EET)
> From: Heikki Linnakangas <hlinnaka-X3B1VOXEql0@...>
> Reply-To: dev-nmThH5wfGryY1pf4d8vP8EB+6BGkLq7r@...
> Newsgroups: gmane.comp.java.geronimo.devel
> Subject: MySQL TranQL connector
>
> Hi,
>
> I've been looking at the MySQL TranQL connector, trying to figure  
> out how to write one for PostgreSQL. Is this the right list to  
> discuss TranQL?
>
> This caught my eye at the org.tranql.connector.mysql.LocalMCF  
> constructor:
>
>     public LocalMCF() {
> -->     super(new MysqlConnectionPoolDataSource(), new  
> MysqlExceptionSorter(), false);
>         ds = (MysqlConnectionPoolDataSource) super.dataSource;
>     }
>
> The corresponding superclass constructor signature is:
>
> protected AbstractLocalDataSourceMCF(javax.sql.DataSource  
> xaDataSource,
>     ExceptionSorter exceptionSorter, boolean commitBeforeAutocommit)
>
> Now why does LocalMCF give the superclass constructor a  
> ConnectionPoolDataSource instance while it's expecting a  
> DataSource? It seems to work because MysqlConnectionPoolDataSource  
> extends MysqlDataSource, but that's just accidental, isn't it?

I think so.  I thought Jeremy had written a MCF based on  
ConnectionPoolDataSource, but I don't see it.  If he doesn't have it  
hidden at home :-) and no one beats me to it I will see about coming  
up with one: IMO we should use a MCF based on CPDS if the db provides  
one. You are welcome to write it also :-)

>
> Also, the variable name in the superclass constructor is a bit  
> confusing. I would think that xaDataSource is an instance of  
> XADataSource, but it's not.

I think that is a lack of updating after copying from XAMCF :-)

thanks
david jencks

>
> - Heikki
>