[6349] in Athena Bugs
pmax 7.1: Xmfbpmax memory allocation; xlogin
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Tue Nov 6 13:59:23 1990
Date: Tue, 6 Nov 90 13:59:06 EST
From: Ken Raeburn <Raeburn@MIT.Edu>
To: bugs@ATHENA.MIT.EDU
When the server runs out of memory, it logs:
start Tue Nov 6 08:38:40 1990
Fatal server bug!
Out of memory
and crashes, leaving a core file of, oh, say, 2105344 bytes, which
can't be debugged because there's no symbol table nor a decent stack
trace.
It would be much nicer if it would pass errors back to clients, and
continue to run itself. If it can't, just exiting would probably be
nice.
Of course, it would be better still if it didn't run out of memory...
is there more work that can be done on optimizing the memory use? Bit
fields for small numbers and booleans, custom allocators for
fixed-size structures, et cetera?
----------------
On a related note: My workstation display hung when this happened.
The deactivate script wouldn't run, and reboot messages didn't show up
until the "Autoload..." message appeared. I assume this means that
when xlogin and friends died (which I know they did, since I did a
`ps' listing while logged in remotely) the console output didn't
properly revert to the real console. This is a bug in the xlogin
program, or the console, or both.
