i was having a similar problem with Server.Transfer. having written my own MVC framework under .net i know how frustrating Server.Transfer can be.
the problem, i believe, is related to recursive calls to the page handler.
i have a solution that seemed to work for the most simple solution of WebForm1.aspx navigating to WebForm2.aspx with Server.Transfer. i doubt it is a great solution, and with apologies to the spring guys, i had to hack up their code a bit to make it work. perhaps, the solution will be of some use in finding a final answer, especially if no one has any other ideas (not much activity on this post recently).
my solution works on the concept of aliasing the page you request and then using Server.Transfer to go to the actual page. it was the only way i could get my personal mvc framework to operate and it is how microsoft specifies the building of a front controller as part of their best practices.
to get this to work i made a few changes. there are probably more elegant ways to do the things i did and perhaps some are unnecessary, but i don't know the spring code base well enough to make those kind of design decisions.
the first change is in the web.config file:
<httpHandlers>
<add verb="*" path="Pg*.aspx" type="Spring.Web.Support.PageHandlerFactory, Spring.Web"/>
</httpHandlers>
note the addition of the prefix "Pg" in the path attribute. to request a page you simply request the .aspx page prefixed with the "Pg" value. for example, a page called WebForm1.aspx would be requested by PgWebForm1.aspx.
the object definitions change slightly:
<object type="WebForm1.aspx" name="PgWebForm1" id="PgWebForm1">
<property name="Results">
<dictionary>
<entry key="go">
<value>transfer:WebForm2.aspx</value>
</entry>
</dictionary>
</property>
</object>
<object type="WebForm2.aspx" name="PgWebForm2" id="PgWebForm2"></object>
spring did not require you to specify a name or id for the object (not really sure which one i needed here so i specified both to be sure). the name and id will be the name of the page prefixed with the "Pg" value. the idea here is to provide a mapping of the page request "PgWebForm1.aspx" to the physical "WebForm1.aspx". the rest of the object definition is the same as the spring examples.
to make all this work, i then just made some minor adjustments in the Spring.Web code to make it accept the "Pg" prefix for object lookups.
in Spring.Objects.Factory.Support.WebObjectDefinition Factory:
public override IConfigurableObjectDefinition CreateObjectDefinition(string typeName, string parent, ConstructorArgumentValues arguments, MutablePropertyValues properties, AppDomain domain)
{
Type objectType = null;
bool isPage = StringUtils.HasText(typeName) && typeName.ToLower().EndsWith(".aspx");
if (!isPage && StringUtils.HasText(typeName) && domain != null)
{
objectType = ObjectUtils.ResolveType(typeName);
}
if (!StringUtils.HasText(parent))
{
if (objectType != null)
{
return new RootWebObjectDefinition(objectType, arguments, properties);
}
else if (isPage)
{
// SPMVA: change the name of the object
IConfigurableObjectDefinition def = new RootWebObjectDefinition(typeName, properties);
def.FactoryObjectName = "Pg" + typeName;
return def;
}
else
{
return new RootWebObjectDefinition(typeName, arguments, properties);
}
}
else
{
if (objectType != null)
{
return new ChildWebObjectDefinition(parent, objectType, arguments, properties);
}
else if (isPage)
{
return new ChildWebObjectDefinition(parent, typeName, properties);
}
else
{
return new ChildWebObjectDefinition(parent, typeName, arguments, properties);
}
}
}
in Spring.Web.Support.PageHandlerFactory:
public IHttpHandler GetHandler(HttpContext context, string requestType, string url, string path)
{
IHttpHandler handler = (IHttpHandler) handlers[url];
if (handler != null)
{
return handler;
}
InternalSecurityPermissions.UnmanagedCode.Assert() ;
string pageName = WebUtils.GetPageName(url);
IApplicationContext appContext;
try
{
appContext = (IApplicationContext) context.GetConfig("spring/context");
}
catch (ConfigurationException ex)
{
if (ex.InnerException != null)
{
throw ex.InnerException;
}
else
{
throw ex;
}
}
if (appContext.ContainsObjectDefinition(pageName))
{
Type pageType = appContext.GetType(pageName);
// SPMVA: change the object name
string realName = pageName.Substring(2, pageName.Length -2);
if (typeof(IRequiresSessionState).IsAssignableFrom(pa geType))
{
handler = new SessionAwarePageHandler(appContext, realName); //SPMVA: pageName);
}
else
{
handler = new PageHandler(appContext, realName); //SPMVA: pageName);
}
}
else
{
handler = PageParser.GetCompiledPageInstance(url, path, context);
}
if (handler.IsReusable)
{
handlers[url] = handler;
}
return handler;
}
and also in that internal class PageHandler:
public void ProcessRequest(HttpContext context)
{
// SPMVA: change the object name
IHttpHandler page = (IHttpHandler) appContext.GetObject("Pg" + pageName);
page.ProcessRequest(context);
}
i think that was it. in any case i don't know what the down sides are to making things happen this way, but i do like that things work! <smiley_face>