I think I'm onto something here... Sorry for not supplying a patch this time
but I swamped the source with some homebrew trace macros to be able to follow
the execution path a little better...
I noticed that TreeItemMsgList::_numberFetchingStatus is set to LOADING in:
ObtainSynchronizedMailboxTask::syncFlags
but then it is not set to DONE although a reply for fetching flags is received.
I noticed that ObtainSynchronizedMailboxTask::handleStateHelper is called and
has an if-branch for "resp->tag == flagsCmd" which now looks like this in my
source: (added lines start at //TG)
} else if ( resp->tag == flagsCmd ) {
if ( resp->kind == Responses::OK ) {
log("Flags synchronized", LOG_MAILBOX_SYNC);
//qDebug() << "received OK for flagsCmd";
Q_ASSERT( status == STATE_SYNCING_FLAGS );
Q_ASSERT( mailboxIndex.isValid() ); // FIXME
TreeItemMailbox* mailbox = dynamic_cast<TreeItemMailbox*>(
static_cast<TreeItem*>( mailboxIndex.internalPointer() ));
Q_ASSERT( mailbox );
//TG set in syncFlags to LOADING but not set to DONE here...
TreeItemMsgList *list =
dynamic_cast<Imap::Mailbox::TreeItemMsgList*>( mailbox->_children[0] );
Q_ASSERT(list);
list->_numberFetchingStatus = TreeItem::DONE;
notifyInterestingMessages( mailbox );
model->emitMessageCountChanged(mailbox);
} else {
log("Flags synchronization failed", LOG_MAILBOX_SYNC);
// FIXME: error handling
}
status = STATE_DONE;
emit model->mailboxSyncingProgress( mailboxIndex, status );
_completed();
return true;
} else {
This fixes the issue for me: when a mailbox is selected and a message is marked
as read, the mailbox view reacts accordingly.
But1: At first I supposed the method
ObtainSynchronizedMailboxTask::handleFlags
is responsible for handling that response, but it doesn't seem to be called -
at least in my testcase - should the _numberFetchingStatus flag be unset there
as well?
But2: this manual unsetting of _numberFetchingStatus from several different
places seems cumbersome and error-prone - is it OK for me to have a go and try
to improve this situation? (accessor method or some RAII-based reference-
counting like magic, I'll have to dig into the source a little further to
think of something specific...)
Cheers,
Thomas
