On 29/07/2013 20:34, Jeremy Boynes wrote: > On Mon, Jul 29, 2013 at 10:19 AM, Mark Thomas <ma...@apache.org> wrote: > >> >> Looking forward to the point where we use an SCI for JSPs, we don't want >> to / shouldn't have to do this for every test that uses JSPs. >> >> addWebapp() is the interface that provides the settings from the default >> web.xml and should be just like deploying a webapp in Tomcat. >> This should enable container SCIs (JSP & WebSocket). >> >> addContext() is the 100% programmatic interface that requires the user >> to explicitly enable everything. Currently, scanning for SCI >> @HandlesTypes matches isn't supported with addContext(). I think that is >> the way to go and we should just document it. >> > > This was an explicit test for the loading mechanism. addWebapp() does add > the SCI in the normal way
Then I don't get why you are testing the loading mechanism with a context created with addWebapp(). Mark - that was actually the problem that led to bug > 55319: by adding the SCI it triggered annotation scanning which failed > because the parent classloader contained test classes with invalid > annotations (which should not have been scanned because isWebapp was false > but that was not being captured in the fragment by the scanner). > > HTH > Jeremy > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
