From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 4.0) BuildID: 20010813 [maybe - browser titlebar is all zeros!] Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:0.9.3) Gecko/20010813 (as noted, "BuildID" is all zeros in the browser title bar). While running just mailnews, an ever increasing amount of CLOSE_WAIT connections [NNTP to the server] are left lying around. These seem to be created once every 20-30 mins., even with 'auto-fetch' [I use 10 mins] disabled for that account. IIUC, this is mozilla not issuing 'close()' calls on connections. May possibly be related to PC bug 67957. Compiled for Solaris 2.6. Reproducible: Always Steps to Reproduce: 1. mozilla -mail 2. wait 3. netstat -f inet | grep nntp Actual Results: Lots of nntp CLOSE_WAIT entries displayed. Expected Results: Little or none of the above! I'm not convinced that the rate of creation of these is related to the auto- fetch time (whether on or off) as I'd then expect far more than I get. I'm putting this as normal not critical, even though it's a noticable resource leak. I don't *think* there're any real side effects.

I later filed bug 70605 for the problem I originally reported on bug 67957.

I believe this is either the same bug as bug 67957, or dependant upon it, since if you don't have lots of connections left open, you can't have lots of close_waits, either ;-)

*** This bug has been marked as a duplicate of 67597 ***

.

*** This bug has been marked as a duplicate of 67957 ***

Verified dup. But I haven't seen this or bug 67957 in a while. I let my connection sit there for a while, and it only opened up a max of 3 NNTP sockets when I was reading multiple news messages.

For what it's worth, I can no longer reproduce the fault with the released 0.9.4; I raised it on 0.9.3. Esp. since this was marked as a dup., I reckon this can now be closed.