Re: svn commit: r1493066 - /tomcat/trunk/java/org/apache/catalina/startup/ContextConfig.java

2013-06-14 Thread Rainer Jung
On 14.06.2013 18:36, Christopher Schultz wrote:
> Mark,
> 
> On 6/14/13 12:22 PM, Mark Thomas wrote:
>> On 14/06/2013 17:15, Christopher Schultz wrote:
>>> I was actually hoping that certain JDBC drivers would start packaging a
>>> listener that could do things like stop driver-launched threads, etc.
>>> that the user doesn't really know have been started. Will this then
>>> always be a requirement of the webapp to provide a listener that
>>> essentially does nothing on startup but then terminates things on
>>> shutdown?
>>
>> They'd have to include an SCI (which is always processed) and that SCI
>> can register a ServletContextListener.
> 
> Sorry... I'm not yet sensitive to the different language being used
> here. I read the original commit comment as "don't scan container JARs
> at all" instead of what you said which was to "not scan for
> annotations". Big difference ;)

Yup, later commit by mark http://svn.apache.org/r1493080 clarified SCI
use for container Jars.

Rainer


-
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org



Re: svn commit: r1493066 - /tomcat/trunk/java/org/apache/catalina/startup/ContextConfig.java

2013-06-14 Thread Christopher Schultz
Mark,

On 6/14/13 12:22 PM, Mark Thomas wrote:
> On 14/06/2013 17:15, Christopher Schultz wrote:
>> I was actually hoping that certain JDBC drivers would start packaging a
>> listener that could do things like stop driver-launched threads, etc.
>> that the user doesn't really know have been started. Will this then
>> always be a requirement of the webapp to provide a listener that
>> essentially does nothing on startup but then terminates things on
>> shutdown?
> 
> They'd have to include an SCI (which is always processed) and that SCI
> can register a ServletContextListener.

Sorry... I'm not yet sensitive to the different language being used
here. I read the original commit comment as "don't scan container JARs
at all" instead of what you said which was to "not scan for
annotations". Big difference ;)

-chris



signature.asc
Description: OpenPGP digital signature


Re: svn commit: r1493066 - /tomcat/trunk/java/org/apache/catalina/startup/ContextConfig.java

2013-06-14 Thread Mark Thomas

On 14/06/2013 17:15, Christopher Schultz wrote:

I was actually hoping that certain JDBC drivers would start packaging a
listener that could do things like stop driver-launched threads, etc.
that the user doesn't really know have been started. Will this then
always be a requirement of the webapp to provide a listener that
essentially does nothing on startup but then terminates things on shutdown?


They'd have to include an SCI (which is always processed) and that SCI 
can register a ServletContextListener.


Mark


-
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org



Re: svn commit: r1493066 - /tomcat/trunk/java/org/apache/catalina/startup/ContextConfig.java

2013-06-14 Thread Christopher Schultz
Mark,

On 6/14/13 11:54 AM, Mark Thomas wrote:
> On 14/06/2013 16:38, Rainer Jung wrote:
>> On 14.06.2013 15:43, ma...@apache.org wrote:
>>> Author: markt
>>> Date: Fri Jun 14 13:43:56 2013
>>> New Revision: 1493066
>>>
>>> URL: http://svn.apache.org/r1493066
>>> Log:
>>> Container provided JARs should not be scanned for deployment
>>> annotations nor should they be checked for web-fragment.xml files.
>>
>> Sorry I haven't really investigated this, but:
>>
>> - I think I have seen you committing a change that labels common loader
>> loaded Jars as container provided
> 
> Correct.
> 
>> - we usually suggest to pack JDBC drivers into the common loader, right?
> 
> Correct.
> 
>> So as a general consistency check to the above: we don't expect JDBC
>> drivers to contain relevant annotations, web fragments etc. in the near
>> future?
> 
> I certainly don't.

I was actually hoping that certain JDBC drivers would start packaging a
listener that could do things like stop driver-launched threads, etc.
that the user doesn't really know have been started. Will this then
always be a requirement of the webapp to provide a listener that
essentially does nothing on startup but then terminates things on shutdown?

Or is the idea that these issues are moot unless the driver is actually
in WEB-INF/lib and in that case the listeners will actually be
discovered and fire?

-chris



signature.asc
Description: OpenPGP digital signature


Re: svn commit: r1493066 - /tomcat/trunk/java/org/apache/catalina/startup/ContextConfig.java

2013-06-14 Thread Mark Thomas

On 14/06/2013 16:38, Rainer Jung wrote:

On 14.06.2013 15:43, ma...@apache.org wrote:

Author: markt
Date: Fri Jun 14 13:43:56 2013
New Revision: 1493066

URL: http://svn.apache.org/r1493066
Log:
Container provided JARs should not be scanned for deployment annotations nor 
should they be checked for web-fragment.xml files.


Sorry I haven't really investigated this, but:

- I think I have seen you committing a change that labels common loader
loaded Jars as container provided


Correct.


- we usually suggest to pack JDBC drivers into the common loader, right?


Correct.


So as a general consistency check to the above: we don't expect JDBC
drivers to contain relevant annotations, web fragments etc. in the near
future?


I certainly don't.

Note that any framework JARs in the common loader will always have SCIs 
processed.


Mark


-
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org



Re: svn commit: r1493066 - /tomcat/trunk/java/org/apache/catalina/startup/ContextConfig.java

2013-06-14 Thread Rainer Jung
On 14.06.2013 15:43, ma...@apache.org wrote:
> Author: markt
> Date: Fri Jun 14 13:43:56 2013
> New Revision: 1493066
> 
> URL: http://svn.apache.org/r1493066
> Log:
> Container provided JARs should not be scanned for deployment annotations nor 
> should they be checked for web-fragment.xml files.

Sorry I haven't really investigated this, but:

- I think I have seen you committing a change that labels common loader
loaded Jars as container provided

- we usually suggest to pack JDBC drivers into the common loader, right?

So as a general consistency check to the above: we don't expect JDBC
drivers to contain relevant annotations, web fragments etc. in the near
future?

Regards,

Rainer

-
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org