DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://issues.apache.org/bugzilla/show_bug.cgi?id=30424>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://issues.apache.org/bugzilla/show_bug.cgi?id=30424

possible enhancement on bootup (or proving how little I know)...

[EMAIL PROTECTED] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|WONTFIX                     |



------- Additional Comments From [EMAIL PROTECTED]  2004-07-31 19:47 -------
Ah, from that perspective it makes sense.

(as much as I was trying to avoid another included technolgoy for a 'minor' fix
from a specific webapp servlet instantiation perspective.  I've used JNDI for
ejbs - of course)

However, it appears that you aren't supposed to think in that perspective or
there is some confusion in the doc, which is probably why I did not consider the
"global copy of the codebase" problem in the first place.  Exhibit A:  in
http://struts.apache.org/userGuide/configuration.html#config_add, the Sidebar
about Sharing JAR Files Across Web Applications, it specifically states:  "...it
is possible that sharing the Struts and Commons JAR files might appear to work
for you. However, this is NOT a supported configuration."

So, as a result of reading that part of the doc, my focus for an app from day 1
has within this "walled garden" approach to where it is self-contained within
the webapp, including the duplication of codebases (jar files).  From this
perspective, bolstered by the doc, I would not have run into scenario you
described, which might indeed require a more "global namespace/content"
technology such as JNDI.

Actually, I still submit that setting up a global lookup technology such as JNDI
is still more gyrations to go through for an application-specific servlet-based
bootup-sequence when a clean and minor enhancement could keep at the servlet
lifecycle level where you would want it in some situations (and yes, I realize
you create the same end effect based on your marshalling namespace example).

Furthermore, there is a workaround to my original patch proposal:  Generate the
lookup and list each time from the specific context within the methods instead
of caching the results globally like I originally did.  Since this approach
would retrieve the context and therefore the underlying tokens at invocation
time, this implementation would always retrieve the correct results.

(sidebar:  If you're looking to re-factor to take Struts startup from containers
other than servlets down the road, yes could change the landscape.  But the idea
of Initialization Parameters and a "default context" is present from an
assembler viewpoint as well as from a servlet viewpoint, although other
containers have a more intrinsic perspective of namespaces via JNDI than
servlets traditionally have)

In conclusion, it might make sense to at least revise that part of the
documentation to avoid what I would still maintain is a reasonable conclusion
based on the published scope of the .jar files.  At least I hope you can see how
I got to the assumption(s) leading up to the original proposal.  And why I
wanted to bring up these points, even at the risk of 'beating a dead horse'.

Thanks for listening!

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to