Re: DataSource / OOM problem

Jakarta Commons consists of several small, useful and well defined 
components (BeanUtils, Chain, Collections, Lang, i18n, ...).
The larger projects, such as Tomcat, is often built by these 
components (or rather, the components have been found reusable and 
therefore extracted from Tomcat).

My only objection is that many of them use the notorious Commons Logging API.


At 2006-06-20 15:30, sks wrote:
>The couple of projects I've used are massively overbloated (James, Tomcat).
>
>
>
>>At 2006-06-20 13:15, sks wrote:
>>>I don't like to use jakarta stuff
>>
>>Why??? Because of the license?
>>There are lots of useful Apache projects, even if you don't fancy Tomcat.
>>
>>
>>>but I've tried this pool with limited success.
>>>http://homepages.nildram.co.uk/~slink/java/DBPool/
>>
>>There is also http://xapool.experlog.com/ which supports JTA.
>>
>>
>>>>At 2006-06-20 12:54, sks wrote:
>>>>>Yes Scott has acknowledged it as an issue. I can get around it 
>>>>>by stopping using a datasource and getting a connection to the 
>>>>>database using the good old fashioned 
>>>>>java.sql.DriverManager.getConnection but obviously that's quite inefficient.
>>>>
>>>>
>>>>Couldn't you just put a DBCP 
>>>>(http://jakarta.apache.org/commons/dbcp/) pool in JNDI?
>>