Closed Bug 78409 Opened 25 years ago Closed 25 years ago

Hierarchy menus detach at second level folders.

Categories

(SeaMonkey :: MailNews: Message Display, defect, P2)

x86
Linux
defect

Tracking

(Not tracked)

VERIFIED FIXED
mozilla0.9.1

People

(Reporter: laurel, Assigned: paulkchen)

References

Details

(Whiteboard: Checked in on Trunk, need verification before Branch)

Attachments

(1 file)

Using apr30-may01 commercial builds linux rh6.2 and Win98 Mac appears fine. This has been around awhile but seems to be getting worse -- at least it's more in my face than before. Very hard to use sometimes... When you try to access the second sub-level (or deeper) of a folder hierarchy menu such as File/Move/Copy, the submenu detaches at the second level folders level. When that next level detaches from the rest of the menu, it REALLY detaches and often displays at the opposite end of the screen making it quite difficult to secure a selection. This affects the File/Move/Copy menus, search UI scope dropdown and File button menu, and probably New Folder dialog, too. This really needs to be fixed.
Keywords: nsbeta1
QA Contact: esther → laurel
Summary: Hierarchy menus detach at second level folders. → Hierarchy menus detach at second level folders.
I think this is an xptoolkit bug. Reassigning to pinkerton. I see this a bunch too.
Assignee: sspitzer → pinkerton
ben? are you ever going to check in your menu fixes?
Assignee: pinkerton → ben
Target Milestone: --- → mozilla0.9.1
nav triage team: 1. So did Ben just go and fix these issues many moons ago? 2. Why haven't these fixes been checked in?
Ok, talked to Ben and Pink about this. The patch has possibly bitrotted on Ben's Mac, and Ben said that he sent the patch to Pink. Pink has just showed me this patch which is in bug 66848. We're not sure that the patch will fix this bug until we apply it and test it out. If I ever feel safe about pulling a tree something this week, I might get around to testing the patch out.
nav triage team; Reassigning to pchen to test patch for 66848
Assignee: ben → pchen
Keywords: nsbeta1nsbeta1+
Priority: -- → P2
I applied the patch to bug 66848 at home under linux, and it doesn't seem to make a difference. However, I don't have access to an imap server with nested folders. Creating a nested local folder hierarchy doesn't seem to cause any problems. Will try again tomorrow at work.
Maybe it's just me, but I can't see this problem. I guess my imap mailbox just isn't complicated enough. Can anyone else reproduce with a recent build? (I just tried 2001051604, but I don't think it's the build, more likely my setup)
I'm using 5/15 and I'm not seeing it right now. It didn't happen to me all of the time unfortunately, but right now I can't reproduce it.
On win98 and linux with may15 commercial build, I can easily see it when selecting the scope for Search Mail/News Messages or in the MoveToFolder action section of the Filter Rules dialog or if using the dropdown in New Folder from main mail window... I don't, however, see it when using the File button or Move/Copy dropdowns.
Mac, too, may15 trunk build.
Esther's experiencing it, too, on her windows system today's build in main mail File button dropdown. I'm seeing weirder problems in today's build with the file menu, will log separately.
New problems in today's build logged as bug 81252
OK, fairly easy to reproduce on main mail window (win98): 1. Have a hierarchy with at least one account having lots (20?) of top level folders and some of those which have several sublevel folders. 2. Login, select a message in a folder. 3. Put window in restored state, move window down so it is significantly off the bottom of the screen, say just a couple inched of the top of the window showing -- only a couple lines of the thread pane visible. 4. Click File button and select the account having lots of folders. The sublevel of the dropmenu will detach to the far left of the screen. Further sublevel menus will appear detached in somewhat random parts of the screen.
Ok, I've got this to reproduce on my linux build without the patch for 66848. On my win2k box with the patch, it looks fixed. So I think we've got a winner here. Pinkerton says that he also wants to make sure mac is fixed too. I'll apply the patch to my powerbook tomorrow. So far, I'd say it looks like we have a winner.
Whiteboard: fix in hand, but need to get it checked in asap
Whiteboard: fix in hand, but need to get it checked in asap → fix in hand, need testing, needs to be checked in by 5/23.
Not sure if we can keep this indefinitely as a beta stopper. We shd make sure we have a winner with the current patch, and if so, get it in asap.
Shouldn't the patch be attached to the bug? If we have the right patch, we should get this checked in when ready.
Just copied patch from bug 66848. Although fixing windows and linux, this patch trades one bug for another on mac. On the mac, submenus don't show up the first time you select them, and on subsequent tries, the submenu is empty. So I don't think the patch is ready. I'm thinking it will take on the order of a 2-3 days for me to fix the mac problem caused by the patch, mainly because I'm not that familiar with the code.
Updating status whiteboard. In debugging, I got as far as finding out that submenus of the bookmarks popup menu on the personal toolbar are coming up blank because the titles of all the menu items are blank (well, the empty string). Pinkerton, Ben, and myself are all puzzled as to why this patch cause that effect, especially only on the mac. Further investigation will require me to traipse through the creation of the popup menu in the xul template builder. It's looking like another couple days of working on this only to fix.
Whiteboard: fix in hand, need testing, needs to be checked in by 5/23. → patch hoses mac, still investigating, needs to be checked in by 5/23.
The further I dig, the stranger it gets. So as far as I can tell, layout is building the submenu a couple times, but when it goes to draw the submenu, it draws using a completely DIFFERENT set nsTextBoxFrame objects that we so painstakingly create. It just so happens that those nsTextBoxFrame objects have empty string titles. Sigh. Adding pinkerton and ben to cc list.
Ok, I'm going to go off into a corner and weep profusely. The problem I'm seeing is bug 80512. Open a second window, and the bookmarks menu works fine on mac. Updating status whiteboard. Waiting for r= and sr=.
Whiteboard: patch hoses mac, still investigating, needs to be checked in by 5/23. → I'm an idiot, awaiting r= and sr=, needs to be checked in by 5/23.
Got an r=pinkerton. Adding alecf to cc for possible sr
Whiteboard: I'm an idiot, awaiting r= and sr=, needs to be checked in by 5/23. → I'm an idiot, awaiting sr=, needs to be checked in by 5/23.
wow. I really need to defer this to hyatt - but I have to add that the tabs are screwed up :) pchen is now in Taiwan - I really don't have a mac handy, is there someone who can take this?
Whiteboard: I'm an idiot, awaiting sr=, needs to be checked in by 5/23. → wanted for 0.9.1 awaiting sr= and a=
sr=hyatt
a= asa@mozilla.org for checkin to 0.9.1 and the trunk.
ok, I'll check this into the branch and trunk, since paul is away
Whiteboard: wanted for 0.9.1 awaiting sr= and a= → have r=,sr=,a=, alecf to checkin to branch and trunk
i must admit i'm a bit wary of landing this on the branch. it's a major change and it's been around for so long, with so many false-starts, that i'm just plain scared especially so close to beta.
alecf, pchen, can you just check this into the trunk. If everything looks good in a day or two and we haven't released yet we can look at it for the branch. Sorry for the premature approval and any confusion.
This has just been checked onto the trunk
Whiteboard: have r=,sr=,a=, alecf to checkin to branch and trunk → Checked in on Trunk, need verification before Branch
I have filed a bug introduced by this patch as bug 84121. It was commented on bug 78409 (2001-03-31 06:46).
I just tried the jun05 commercial trunk build on win98, mac OS 9.0, and linux rh6.2 and cannot get the separation problem anymore. Looks okay to me.
how could we build a set of test cases that would convince us that we didn't regress something else with this checkin... do we just need to pound the heck out of menus to confirm the fix? is there anything else we could do to shake out side effects?
If 84121 results in menus with unselectable items then I think we need that fix to come along with this one. Any other opinions?
Depends on: 84121
No longer depends on: 84121
So, I haven't seen an a= for the branch, if I don't see anything soon, I assume this is a no go for the branch.
a=chofmann on the branch
checked into the branch, marking fixed
Status: NEW → RESOLVED
Closed: 25 years ago
Resolution: --- → FIXED
OK with commercial branch 2001-06-06-12-0.9.1 win98
Looks ok to me, no detached menus on the commercial branch builds: OK with 2001-06-06-12-0.9.1 on mac OS 9.0 OK with 2001-06-06-13-0.9.1 on linux rh6.2 Note: On mac I do see lag time in drawing the submenus and will sometimes flash a blank white submenu before correctly drawing the correct contents... but I've seen this same effect previously when I'm in the condition spelled out in bug 80512. I don't see any separation, no detaching. So I'm marking this one verified.
Status: RESOLVED → VERIFIED
FYI, the flashing submenus should only happen the first time the submenu is created. It's RDF streaming in bookmarks asynchronously, so the window gets an initial size then resizes to the correct size when the content is built, causing the flashing.
It seems that this bug fix solved a related bookmark bug 62525 at the rame time intrducing a bookmark regression 85180. This regression (no arrow at the bottom of the screen if the sub menu is not fully visible) should probably be seen with long list of mail sub-folders, too.
The regression bug mentioned above is bug 84121. The newer 85180 is only its dup.
Product: Browser → Seamonkey
*** Bug 157780 has been marked as a duplicate of this bug. ***
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: