site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Hello Terry, We're being bitten again by this problem on OSX 10.5.8 and 10.6.1. On 5 Sep 2009, at 20:27, Terry Lambert wrote: Is this in the wrapper (dejagnu) or in the underlying expect? have you any pointer to a bug report that I can examine? google didn't give any meaningful return to my queries. thanks, Iain _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... More specifically, what's happening is that the current session is being hosed. FWIW: whether this is causing launchd to die .. or whether launchd dying is the cause of the hose remains to be seen. The tests that cause the issue are ones where gdb is attached. The tests pass, and do NOT cause the session to be trashed, if run by a non-root user from the command line. Not really. There is a bug in the "expect" they ship that eats a process slot per run because it doesn't actually check the test output fully. we see this with OSX/XCode released expect 5.42 (10.5) and 5.44.1.11 (10.6) and with fink 5.43 (on 10.5) since these things are Open source.. at least we could look at fixing the fault you identify - if it still present. This email sent to site_archiver@lists.apple.com