User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.7) Gecko/2009021910 Firefox/3.0.7 (.NET CLR 3.5.30729) Build Identifier: Thunderbird 3.0b2 A closed parent folder isn't highlighted when a subfolder receives new mail until Thunderbird repaints itself (e.g. when the mouse moves over the parent folder or when I use Alt-Tab to switch to another app then back to TB). This is bad because I can't see which folder branches contain new mail. I know there are several related bugs, but I believe this one is different because the situation is fixed not by clicking on any folder but simply when a repaint happens. Perhaps the problem is related to the new non-RDF folder tree. I use POP3. FWIW, I think the relevant CSS is in folderPane.css: treechildren::-moz-tree-cell-text(closed, subfoldersHaveUnreadMessages-true) { font-weight: bold; } Reproducible: Always

I can confirm this behavior in Thunderbird beta 3 too. It's absolutely dangerous because closed folders may contain new emails and the user may fail to read them if he misses the original arrival alert. Just imagine what a mess is when you have tens of folders and subfolders for different clients and/or businesses.

The deficiency would appear to be around: http://mxr.mozilla.org/comm-central/source/mail/base/content/folderPane.js#1331 The code should probably be modified to find the folder or its first visible ancestor. It should then invalidate the given folder and all of its (implicitly visible) ancestors. It is possible that no ancestor is visible in certain folder modes.

Confirmed same bahavior with TB 3.0.3 on Win Vista (all updates installes as of 13 March 2010).

Is this bug somehow connected to Bug 533928 ? I can see both behavior that at first glance may be connected.

sid, is the capability of Bug 536042 - Provide a way for folder tree modes to express parent relationships and tell us which folder a message belongs to - useful in fixing this issue?

I don't think so -- the only thing related to this that would really affect is the new mail notification.

However, the fix asuth's describing would certainly involve the code there. :)