[ https://issues.apache.org/jira/browse/MYFACES-4309?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16969125#comment-16969125 ]
Thomas Andraschko commented on MYFACES-4309: -------------------------------------------- Just to be sure: If you don't use ManagedBeans, your patch doesn't change the current behavior? Still the bigger buffer is required?! The ViewScoped/SessionScoped could be greated very late in the RENDER_RESPONSE phase, when the bean is used at the end of the view. And the response might already be commited here. > Session is broken in some cases due to MYFACES-4297 > --------------------------------------------------- > > Key: MYFACES-4309 > URL: https://issues.apache.org/jira/browse/MYFACES-4309 > Project: MyFaces Core > Issue Type: Bug > Affects Versions: 2.3.5 > Reporter: Bill Lucy > Assignee: Bill Lucy > Priority: Major > Time Spent: 10m > Remaining Estimate: 0h > > The changes made for MYFACES-4297 introduced a problem: in some cases a > session is not created early enough for a session cookie to be written out on > a response before it is committed, due to the small default response buffer > size. As mentioned in the comments for 4297, that behavior causes a problem > for the ViewScope (and I expect the session as well.) > As discussed, increasing the javax.faces.FACELETS_BUFFER_SIZE is a > workaround, but that's not ideal for a few reasons: > 1. apps using the default buffer size value will be broken by the new > behavior when updating to 2.3.5 > 2. there doesn't appear to be a way to update the buffer size for JSPs > We should consider revisiting the changes made in MYFACES-4297. Or, if > nothing else, we might want to add some way to change the default buffer size > on the JSP path. -- This message was sent by Atlassian Jira (v8.3.4#803005)