Maybe OT... what not to have in iBATIS

[Sven Boden <li...@pandora.be>]

Do we ever vote on stuff we don't want to have in iBATIS?... Below is a 
list of JIRA requests/bugs I would not implement.

Regards,
Sven

- IBATIS-22: Specify Query Timeout
The only projects I can think of that would need it would be kind of 
reporting systems. Most systems I know of either want to have the or 
data or not. Timeout is also not supported by JDBC and implementing a 
hack for it (2 threads and then cancelling after the timeout) causes 
problems in DB2 for sure (hanging connections). And a workaround exists 
by using it via a Timeout aware datasource.

- IBATIS-73: EHCache for iBATIS
I would rather have JCS (http://jakarta.apache.org/jcs/index.html) as 
cache then EHCache. JCS was written as a EHCache replacement. But maybe 
we already have enough dependencies,

- IBATIS-119: use label from sql for mapping (XML stuff)
Doesn't seem very useful.

- IBATIS-120: Use of java.util.List for input parameter
Too error prone.

- IBATIS-206: Support for JDBC 3.0 Disconnected Resultsets such as 
WebRowSet/CachedRowSet
iBATIS is into returning domain objects, not returning ResultSets

- IBATIS-211: Cache should follow structure of hierarchical resultmaps
As stated doesn't make sense as iBATIS doesn't use object identity... 
something as flushOnFlush could possibly be implemented.

- IBATIS-222: Evict from cache
Could even be closed now... it doesn't make sense as iBATIS doesn't use 
object identity

- IBATIS-226: Support Commons DBUtils ResultHandler
Not as in the original request. Could be implemented as some kind of 
extension point

- IBATIS-240: Add support to generate either non-prepared statements or 
sequence of simple SQL statements
including bad practices is not on my wishlist

- IBATIS-259: SqlMapTransactionManager.setUserConnection(null) does not 
complete restore previous transaction state
Not the intended usage

Re: Maybe OT... what not to have in iBATIS

[Clinton Begin <cl...@gmail.com>]

+1 on all accounts.

The only comment I'd have is about EHCache and JCS....iBATIS has a pluggable
cache model, that should be compatible with either.  Although Cameron Purdy
once suggested we open up the CacheModel class for extension, to make
integrating JCS (Coherence in his case).  I think Cameron's idea is good,
and I believe it's already logged in JIRA.  We could add support for JCS or
EHCache, but I'd suggest the people who want them should implement them and
then contribute them to the project.  I don't want to be an ivory tower of
generating plugins for everything under the sun.  By waiting for someone
else to implement it, we're ensuring there's an audience for the feature.

In any case.....I say close 'em!

BTW:  Great job triaging the issue tracker entries Sven!

Cheers,
Clinton


On 3/18/06, Sven Boden <li...@pandora.be> wrote:
>
>
> Do we ever vote on stuff we don't want to have in iBATIS?... Below is a
> list of JIRA requests/bugs I would not implement.
>
> Regards,
> Sven
>
> - IBATIS-22: Specify Query Timeout
> The only projects I can think of that would need it would be kind of
> reporting systems. Most systems I know of either want to have the or
> data or not. Timeout is also not supported by JDBC and implementing a
> hack for it (2 threads and then cancelling after the timeout) causes
> problems in DB2 for sure (hanging connections). And a workaround exists
> by using it via a Timeout aware datasource.
>
> - IBATIS-73: EHCache for iBATIS
> I would rather have JCS (http://jakarta.apache.org/jcs/index.html) as
> cache then EHCache. JCS was written as a EHCache replacement. But maybe
> we already have enough dependencies,
>
> - IBATIS-119: use label from sql for mapping (XML stuff)
> Doesn't seem very useful.
>
> - IBATIS-120: Use of java.util.List for input parameter
> Too error prone.
>
> - IBATIS-206: Support for JDBC 3.0 Disconnected Resultsets such as
> WebRowSet/CachedRowSet
> iBATIS is into returning domain objects, not returning ResultSets
>
> - IBATIS-211: Cache should follow structure of hierarchical resultmaps
> As stated doesn't make sense as iBATIS doesn't use object identity...
> something as flushOnFlush could possibly be implemented.
>
> - IBATIS-222: Evict from cache
> Could even be closed now... it doesn't make sense as iBATIS doesn't use
> object identity
>
> - IBATIS-226: Support Commons DBUtils ResultHandler
> Not as in the original request. Could be implemented as some kind of
> extension point
>
> - IBATIS-240: Add support to generate either non-prepared statements or
> sequence of simple SQL statements
> including bad practices is not on my wishlist
>
> - IBATIS-259: SqlMapTransactionManager.setUserConnection(null) does not
> complete restore previous transaction state
> Not the intended usage
>
>
>