[Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.0.2pre) Gecko/2008071303 Thunderbird/3.0a2pre] (nightly) (W2Ksp4) [Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.0.2pre) Gecko/2008071301 SeaMonkey/2.0a1pre] (nightly) (W2Ksp4) After checking bug 438647 comment 24, I noticed another error: 1. Start SM or TB. 2. Create a dumb profile and (POP3) account. 3. Open Account Settings. 4. Select "Junk Settings", on the (POP3) account or/then the "Local Folders". [Try this again, if the error doesn't show up "immediately"...] 4r. {{ Error: aFolder is null Source File: chrome://messenger/content/folderWidgets.xml Line: 616 }} A (SM) Venkman stack can look like: {{ #0: function selectFolder(aFolder=null:null) in <chrome://messenger/content/folderWidgets.xml> line 616 #1: function onInit(aServerId=string:"mailbox://nobody@Local%20Folders", aPageId=string:"am-junk.xul") in <chrome://messenger/content/am-junk.js> line 60 #2: function restorePage(serverId=string:"mailbox://nobody@Local%20Folders", pageId=string:"am-junk.xul") in <chrome://messenger/content/AccountManager.js> line 968 #3: function showPage(pageId=string:"am-junk.xul", serverId=string:"mailbox://nobody@Local%20Folders") in <chrome://messenger/content/AccountManager.js> line 766 #4: function onAccountClick(tree=XULElement:{7}) in <chrome://messenger/content/AccountManager.js> line 727 #5: function onselect(event=Event:{0}) in <chrome://messenger/content/AccountManager.xul> line 1 6: select / x-jsd:native-code #7: function onxblmousedown(event=MouseEvent:{0}) in <chrome://global/content/bindings/tree.xml> line 951 }} |selectFolder()| doesn't like |aFolder=null:null|; I don't know if we should blame the caller or the callee...

WFM over here on version 3.0a2pre (2008071303) /linux - not using global inbox

Thunderbird, that is.

Serge, in your preferences file, what entries do you have that contain "spamActionTargetAccount"? I suspect that your preference is not set for one of your server accounts.

Serge, please can you answer comment 3. Thanks.

[Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.1a2pre) Gecko/20080813031200 Shredder/3.0b1pre] (nightly) (W2Ksp4) Still there, per comment 0 steps (no global, no download). (In reply to comment #3) > Serge, in your preferences file, what entries do you have that contain "spamActionTargetAccount"? None on the disk. In <about:config>, {{ mail.server.default.spamActionTargetAccount / default / string / (empty) }}

I can reproduce this on SM and TB by setting up POP3 to not use the Global Inbox. The error occurs when I select junk settings on Local Folders, then again on the POP3 account (or vice versa). Since I've looked at this code as part of bug 444220, I'll take this as well.

Comment on attachment 333790 [details] [diff] [review] Stop nulling variables No, I don't think this is right.

This is actually a regression from bug 368303 - if you look carefully at the linked section of the patch you'll notice that only one of the references to pendingServerId was replaced :-(

So this is what you had in mind, right?

Comment on attachment 333894 [details] [diff] [review] Fix regression per Neil's suggestion I think we might want to fix this on branch too for both Thunderbird and SeaMonkey.

Checked in (to trunk), changeset id 103:adbc45171106

(In reply to comment #11) > (From update of attachment 333894 [details] [diff] [review]) > I think we might want to fix this on branch too for both Thunderbird and > SeaMonkey. > I cannot see any effects of this on branch. Can you produce an STR that shows an issue worth fixing?

[Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.1a2pre) Gecko/20080816031236 Shredder/3.0b1pre] (nightly) (W2Ksp4) V.Fixed

(In reply to comment #13) > I cannot see any effects of this on branch. Can you produce an STR that shows > an issue worth fixing? [Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8.1.17pre) Gecko/2008081603 Thunderbird/2.0.0.17pre] (nightly) (W2Ksp4) [Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8.1.17pre) Gecko/20080816 SeaMonkey/1.1.12pre] (nightly) (W2Ksp4) I can't reproduce this bug (with same steps) on 1.8(.1) branch either. Yet, as bug 368303 was "fixed1.8.1.2", it shouldn't hurt to port this fix there.

(In reply to comment #15) > > I can't reproduce this bug (with same steps) on 1.8(.1) branch either. > ... it shouldn't hurt to port this fix > there. > I thought the standard was much higher than "it shouldn't hurt" before we ported fixes to branch.

(In reply to comment #16) > I thought the standard was much higher than "it shouldn't hurt" before we > ported fixes to branch. It is ... That was just my 2 cts...

Comment on attachment 333894 [details] [diff] [review] Fix regression per Neil's suggestion If we can't reproduce on branch, then we shouldn't accept the risk of taking the fix. That said, the risk does look awfully small.