2014-09-02 15:51 GMT+02:00 Violeta Georgieva <miles...@gmail.com>: > Hi, > > > 2014-09-02 16:37 GMT+03:00 Rémy Maucherat <r...@apache.org>: > > > > 2014-09-02 15:11 GMT+02:00 <violet...@apache.org>: > > > > > Author: violetagg > > > Date: Tue Sep 2 13:11:20 2014 > > > New Revision: 1621975 > > > > > > URL: http://svn.apache.org/r1621975 > > > Log: > > > According to JavaDoc javax.servlet.jsp.PageContext.include(String) and > > > javax.servlet.jsp.PageContext.include(String, boolean) must throw > > > IOException if I/O error occur during the operation. > > > When JSPServlet init parameter "development" is "false" then > IOException > > > is re-thrown, but when the parameter is "true", JasperException is > thrown > > > instead of IOException. > > > This change guarantees that IOException is always re-thrown. > > > > > > > The purpose of the JSPException is to give debug information to be able > to > > print out the details of the error. I am sorry Jasper will not pass the > TCK > > in dev mode, but likely this must remain as is. Unless you took care of > > that somehow, this will likely have to be -1ed. > > I kept the debug information: > > + if (options.getDevelopment()) { > + throw new IOException(handleJspException(ex).getMessage(), > ex); > > and it is not about passing TCK but it is about the inconsistency in the > behavior: > > in the production mode everything is OK but in debug mode the exception is > not re-thrown. > > Wdyt? > > I checked the exception report looked ok in the test with the usual details from the included page, so I'll assume it will be fine in the error page generated for the browser.
Rémy
