I have a general question on how to wrap transaction boundaries around
requests
to a WebResource. As discussed in the thread
http://markmail.org/message/tbpg65xd5x26v7xf
there are two possible requests: an ordinary web request (with a
RequestCycle)
and a HEAD request (where no RequestCycle exists). If an application
needs to
wrap transaction boundaries around such requests, how should it do this?
The former kind of request can wrapped by listening to WebRequestCycle's
onBeginRequest and onEndRequest methods, but the latter needs *another*
way, it seems. Is there a way that handles both cases?
Thanks a lot for any hint!
Kaspar
On 07.03.2008, at 13:06, Kaspar Fischer wrote:
On 07.03.2008, at 13:01, lars vonk wrote:
I guess you can't. Since you are in the Application init method I
don't think there is a requestcycle available (request cycles
represents the processing of a request).
I am not trying to get the request cycle inside my application's
init(),
but in getResourceStream().
Still, I think your explanation applies again: getResourceStream()
is apparently called (why?) even when the web server is doing a
HEAD request, in which case there is (?) no request cycle either.
Could this be the case?
On Fri, Mar 7, 2008 at 12:19 PM, Kaspar Fischer
<[EMAIL PROTECTED]> wrote:
How can I get hold of the current request cycle in a subclass of
WebResource?
In
public final class RepositoryFileResource extends WebResource
{
/* ... */
public IResourceStream getResourceStream()
{
RequestCycle cycle = RequestCycle.get();
cycle is null.
P.S. I am registering my resource in my application's init()
getSharedResources().add("repo", new RepositoryFileResource());
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
