[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
- To: HIC at MIT-AI, (BUG LISPM) at MIT-AI
- From: RMS@MIT-AI
- Date: Mon ,13 Feb 79 07:32:19 EDT
HIC@MIT-MC 02/13/79 17:54:02
RMS@MIT-AI (Sent by RMS0@MIT-AI) 02/13/79 02:17:42
Prolonged SUPDUP'ing just used up all virtual memory in MAKE-PKT!
Before that, I reloaded the machine because it was out of virtual memory.
It had been used a lot for SUPDUP'ing prior to that.
Next time something like this happens, please tell me the pack and the band so that
I may look at the correct version of the software to track down the lossage.
As it is, I do not know which version (new or old) was running at the time.
It is not normal practice for users who find bugs in the current band
to spell out explicitly which band was current, so why the sudden
demand for it? Nor do I think there is reason to ask them to.
What's the chance that someone will install a new band with any change
that is relevant to the problem and you won't know?
If I had reloaded a band other than the current one I would have said
so. If you had just looked in the most obvious possible place and
tried the current band on whatever pack was loaded you would have
found the problem.
Since you wouldn't look into it, I just looked into it myself. It was
harder for me than it would have been for you, because I wasn't sure
how and where packets were linked, and because describe has a bug
whereby it doesn't tell about all of the components that the packet
has, which made it harder for me to find out where to look, but I
eventually managed to determine that there were 840. packets on
MADE-PKTS. A minute or so of slow typing at SUPDUP was enough to
gobble a whole region of working-storage. This is as much as I know
how to find out. I hope you will be willing to look further.
The problem occurred on CADR 2. I didn't mention this since I was
under the impression that the new software was on all machines (you
said "I have just installed it", period). However, this information
also should not have been necessary for you to investigate.
Finally, when I left the machine last night I left it in the state of
manifesting this bug (not completely full, but most of the space used
up from SUPDUPing) with a comment on the screen as to what was wrong
and that someone should look into this. If anyone wiped it out
without investigating it sufficiently, that was very irresponsible.