Hi Eric,
I've applied your patch and it solved the issue I was having with sorting in my mailboxes. I'm not sure why the server-side sorting isn't working for me, given that hMailserver supports it.. but I'll dig into that separately.
Thanks!!
Andrew
On Tue, 1 Aug 2006 13:14:26 -0600, Eric Stadtherr estadtherr@gmail.com wrote:
All,
I have a patch that restores the message sorting behavior in cases where the IMAP server does not support the "SORT" command:
http://stadtherr.bounceme.net/files/imap_no_sort.patch [1] It was created against a working copy of Subversion rev288.
Sorry for the goofup - I didn't have a non-sorting IMAP server to test with!
-Eric On Tue, 1 Aug 2006 11:07:00 -0600, Eric Stadtherr wrote: I am partially responsible for the sorting "fixes" of late, so I'd like to take part in the resolution of this one. Rev288 contains a fix I made to the code that manages the results of the server-side sorting. Although the fix is working, it seems that it negatively impacted the behavior in the absence of "SORT" capability on the IMAP server.
I believe the behavior before my fix was to simply display the messages in IMAP Message Sequence Number order (basically the order the messages arrived in the folder). I think I can come up with an easy patch to restore this behavior when "SORT" capability is absent. How does that sound?
-EricOn Tue, 01 Aug 2006 12:34:04 -0400, Jim Pingle wrote: I updated from SVN version 285 to 288, and in doing so I now have problems with the message sort order on BincIMAP. It's working fine on my Courier server.
What happens is that the messages in the inbox are in random order, even though the "date" column header is highlighted. If it refreshes to check new mail and finds something, it does go at the top. However, once you reload the page the messages return to their previous order. The order doesn't correspond to any discernible pattern.
If I move back and forth between pages, the messages stay in the same order.
BincIMAP does not support server-side sorting, so the usual sort functions in Roundcube are disabled (column headers are not clickable.) Perhaps there needs to be an additional capability check that can default back to client-side sorting in this case?
I really, really wish I didn't have to deal with BincIMAP, but I guess it does help to hunt down bugs in lesser-used code paths...
Jim
Links: