74674 - bad rows when switching from large to small (or empty) folder

Status: VERIFIED FIXED

(SeaMonkey :: MailNews: Message Display, defect, P1)

Description

[(not reading, please use [e-mail] instead)]

try this:

load a large folder, scroll to the bottom.
switch to a folder with a 5 messages.  
sometimes I see 10, with 5 blank bogus messages at the top

load a large folder, scroll to the bottom.
switch to a folder with a 0 messages.  
sometimes I crash

fix for the crasher on the way, but I need to debug to figure out why this is
happening.

Comment 1

[(not reading, please use [e-mail] instead)]

we crash when we ask for the cell text at row "-8"

I've got a patch for the crash, but I need to figure out why we ask for row -8

my guess is when I tell the outliner that the # of rows has changes, I'm giving
the wrong value.

Comment 2

[(not reading, please use [e-mail] instead)]

crash = p1, 0.9

Comment 3

[(not reading, please use [e-mail] instead)]

crasher is gone, bad row problem remains.

Comment 4

[(not reading, please use [e-mail] instead)]

doesn't happen when I'm scrolled to the top of the large folder.

I'll tackle this tomorrow.

Comment 5

[[e-mail] (gone - use [e-mail] instead)]

*** Bug 74817 has been marked as a duplicate of this bug. ***

Comment 6

[(not reading, please use [e-mail] instead)]

I'm also seeing a bad row count if on startup if my inbox is small.

I'll work on this today.

Comment 7

[scottputterman]

yeah, this doesn't look good! marking nsbeta1+

Comment 8

[(not reading, please use [e-mail] instead)]

fix in hand.

I'll need a review from hyatt before I can check it in.

Comment 9

[(not reading, please use [e-mail] instead)]

Attachment: fix, please review.

Comment 10

[David Hyatt]

I don't buy this patch.  Can you come by on Monday and explain the problem?

Comment 11

[(not reading, please use [e-mail] instead)]

that patch makes more sense when you see it applied, than in patch form.

the current code does this:

if (mView) {
  // some stuff
  mView = nsnull;
}

if (mView) {
  // some other stuff.
}

"some other stuff" will never get reached.  

with the patch, we do this:

if (mView) {
  // some stuff
  mView = nsnull;
  // some other stuff
}

if you need more explanation, I'll come down and talk to you about it on monday.

Comment 12

[(not reading, please use [e-mail] instead)]

talked to hyatt, he's given the sr=.

Comment 13

[(not reading, please use [e-mail] instead)]

fixed.

Comment 14

[laurel]

OK using apr17 commercial trunk build: linux rh6.2, mac OS 9.0, win98
no crash
no bogus headers seen