Title: | DECmcc user notes file. Does not replace IPMT. |
Notice: | Use IPMT for problems. Newsletter location in note 6187 |
Moderator: | TAEC::BEROUD |
Created: | Mon Aug 21 1989 |
Last Modified: | Wed Jun 04 1997 |
Last Successful Update: | Fri Jun 06 1997 |
Number of topics: | 6497 |
Total number of notes: | 27359 |
I am developing an application on OpenVMS that does mcc_event_gets on wildcard classes with infinite scope of interest time (to get all events forever). After a few crashes or QUITting out of the debugger without cancelling the scope of interest, I start getting strange behavior from the event kernel. For example, mcc_event_puts start hanging in a compute loop, or mcc_event_gets with wildcarded class returns only certain classes and ignores others. The same code works as expected on a "clean" system. Is there an easy way to clear up/reset/restart things after a crash or a debugging session? Does this behavior sound like a bug in the Framework that I should be documenting and QARing? Thanks, John
T.R | Title | User | Personal Name | Date | Lines |
---|---|---|---|---|---|
5268.1 | Stop ALL MCC processes | TOOK::T_HUPPER | The rest, as they say, is history. | Wed Jul 14 1993 13:52 | 10 |
The cleanest way to get things straight is to stop all MCC processes on the system. This includes the MCC_* sink processes. The event pool will be fully recreated when new MCC processes start up. Having processes stop without cleaning up does not make the event pool happy. We have not had much use of wildcarded classes in the event manager. There may be a bug lurking in this area. Most of what we have seen has been wildcarded instances for specific classes. Ted | |||||
5268.2 | Please do not QUIT out of the VMS Debugger when running MCC | TOOK::GUERTIN | MCC: Legend or Nightmare? | Wed Jul 14 1993 14:45 | 4 |
QUITing out of the VMS debugger can cause serious problems because the Event Manager on VMS does cleanup through an Exit Handling routine. -Matt. |