950528 - empty folder pane at startup with many remotely mounted "local" folders

Status: RESOLVED WORKSFORME

(Thunderbird :: Folder and Message Lists, defect)

Description

[Thomas Skrobala]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_6_8) AppleWebKit/534.59.10 (KHTML, like Gecko) Version/5.1.9 Safari/534.59.10

Steps to reproduce:

Thunderbird v24.2.0 on Windows XP service pack 3 on a VM. "Local Folders" is a remote mount to a (Solaris) server; it has about 576 subfolders (and sub-subfolders). I started Thunderbird up.




Actual results:

As the number of subfolders has grown, the startup from home (VPN over cable internet) has gotten slower & slower, sometimes with warnings about scripts running slowly; startup takes several minutes. This is annoying by itself, but  the big problem is that now, when TB finally comes up, the folders pane is empty, so I can see neither new messages nor local folders.

Some "Messages" in the Error Console, but none seem relevant. And no "Errors".

Bug reports 912782 and 923169 also complained about empty folder panes, but they didn't appear to have the same scenario.


Expected results:

I expect to see a nice listing of my mail server's folders and my local folders, as I used to see (when I had fewer subfolders). I still do see the expected result when I start Thunderbird from my work office (fast LAN connection to the local folder server), and I see the expected result if I change the Local Folders to somewhere on the home computer's local disk.

Comment 1

[:aceman]

Would you be able to test out patch in bug 939462 ?

Comment 2

[Thomas Skrobala]

Thank you for your response. I tried the "daily" Thunderbird version that I found in the web page for bug 939462 (the 2013-11-23 version), and I found that, while it still takes a long time for Thunderbird to start up to a point where it is usable (about 11 minutes), it does display the folders pane at the end. I tried this several times, and it worked successfully every time with the daily build, and it failed every time with the official version. Interestingly, though, the daily build had the slow-but-successful result both with the new "Accumulate folder statistics from subfolders in the folder pane" option checked and with it unchecked. Is this what you would expect?

Comment 3

[:aceman]

If the "Accumulate" option did not make any difference then it would mean some other change in TB since TB24 has improved things for you. Still good :)
The patch in bug 939462 is experimental and wants to see if the recursive accumulation of folder stats (which is O(n^2)) is actually slowing anything down. Without accumulation, the folder stats are only retrieved in O(n) complexity. I think there is no evidence (which you confirm) that this really is making any difference speed wise. Thanks for the test.

Comment 4

[Wayne Mery (:wsmwk)]

Please test TB23beta https://ftp.mozilla.org/pub/mozilla.org/thunderbird/releases/23.0b1/mac/ and confirm that it works better, so we know if this is bug 926181

Comment 5

[Thomas Skrobala]

Right now I can't say whether the 926181 fix fixes this.  The beta .exe started up okay and didn't exhibit the problem.  But then I started up v24.2.0 as a control, and that worked okay, too. Maybe my internet connection is faster than usual because it's New Year's Day?  I'll try again in a day or two, will see how things are then.

(In reply to Wayne Mery (:wsmwk) from comment #4)

Comment 6

[Thomas Skrobala]

I just tried starting v24.2.0, then your v23-based beta, then v24.2.0, etc. (each one three times).  The result: v24.2.0 failed (i.e. exhibited the problem) each time, and the beta succeeded (did not exhibit the problem) each time. So maybe the 926181 fix does indeed fix this.  (Unless my problem was introduced in v24, and it doesn't appear yet in a beta based on v23???)

Thank you.

(In reply to Wayne Mery (:wsmwk) from comment #4)
> Please test TB23beta
> https://ftp.mozilla.org/pub/mozilla.org/thunderbird/releases/23.0b1/mac/ and
> confirm that it works better, so we know if this is bug 926181

Comment 7

[Wayne Mery (:wsmwk)]

Thomas, better with version 31?
(bug 926181 is gone in 31)

Comment 8

[Thomas Skrobala]

So far so good with version 31.