I'm wondering if I should go the other direction. When running inside an application server, the initial context should already be configured. So just creating the InitialContext() instance would be enough (with no environment). At least I know this is true for the application servers I've used. I'm not familiar with JBoss. Does it work the same way?
Even if I don't support any JNDI properties directly in the code, they can still be set in various other ways (as you mentioned). At the time I wrote the original code I wasn't sure how it would work in an application server context.
Thoughts?
Hi Steve,
JBoss works the same way - you shouldn't need any JNDI context settings if you're connecting within the JVM (and I don't think it makes datasources available outside the JVM, but could be wrong).
Maybe a separate option to pass the DataSource programatically to JdbcxxxFactory and have it pass it directly to JdbcStore/Log would be useful? This way the user can have complete control over how the data source is found (JNDI, commons-dbcp, proxool etc).
Disclaimer: I don't use JDBC with quickfix/j 
Cheers,
Brad.
Thanks, Brad. I don't often use the JDBC plugins either.
Jörg, what do you think about these suggestions? I'd be removing the existing settings-based properties and adding the capability to set the datasource explicitly on the factory. Or you could make the changes, if you'd like.
I would like to keep JdbcSetting.SETTING_JDBC_DS_NAME.
In addition, if also JdbcSetting.SETTING_JNDI_PROVIDER_URL is set, the InitialContext should be creating using the Hashtable
containing all JNDI properties mentioned in the Context and InitialContext classes. Check every property if it is set and put it
in the Hashtable. I was worried about the incomplete support: If you cannot set any JNDI properties and have to rely on the
InitialContext as it is provided (via application server or jndi.properties), this would be fine.
The ability to set the data source programmatically would be an extra option which makes much sense. But in general, I would
like to see more things to be configurable without extra programming.
My background to use the data source that I have code that should run both in the application server and in a standalone JVM.
Accidentally resolved the wrong issue.
JdbcUtil: Removed empty HashMap argument from InitialContext constructor.