🚀 go-pugleaf

In article <2745930303-scc@uk.ac.cam.cl.fowler> scc@computer-lab.cambridge.ac.UK (Stephen Crawley) writes:
>>     I had a problem with 0915 a few times. The solution was to
>> select the Profile Tool, select "Apply" and Quit. For some strange
>> reason that I have not determined, it is sometimes necessary to
>> enter the Profile Tool before doing anything else.
>
>This sounds like an entirely different problem to me!
>
>What an 0915 code really means is "I've crashed ... debug me!"  The 
>trouble is that hardly anyone in the Interlisp community knows to use 
>the Mesa debugger.  So most people have to rely on guesswork.
>
>But why does the SystemTool crash so often anyway?  My theory (for what
>it is worth) is that the root cause of the problem is in the Mesa 
>programming language.  In the Mesa language, a procedure's interface 
>does not specify the exceptions that the procedure can raise.  Therefore
>it is not possible to spot at compile time cases where an exception is 
>raised but never caught.  At runtime, when an exception is raised but
>not caught you get an 0915.

I have had similar experiences.  A while ago in Koto I got a lot of problems
which resulted in 0915 or 0940.  Most of the time this occured after I had
problems with my Lisp VM, e.g. running out of VM backup space or the garbage
collector quitting because of too many non-single references to cells.  When
I tried to reboot and CopyVM after this I freqently got a 0915 and on the
second attempt a 0940.

My theory was that Lisp left my disk in a slightly wrong state.  SystemTools
is able to boot but crashes as soon as it detects (trips over) a "wrongness". 
The crash often occurred when i tried to select the last used Lisp volume or
LispFiles.  The crash often messes up the disk totally, making the disk
unbootable: 0940.

My solution was to make a script on the installation floppy to do a Physical
Scavenge of the disk followed by a Scavenge of EVERY logical volume.  This
process takes about an hour on my 80MB disk, this is a lot of time but after
this the system booted and worked without problems each time.  I preferred
this option over the Xerox suggestion to reformat the disk :-)  It took me a
fair amount of time to figure out how to make the script because of the
"overwhelming" amount of documentation and the lack of the "dangerous"
password. 


							The Fons

P.S.
Is it correct that the code 0950 is not a progess code (i.e. not an error)
indicating some scavenging is in progress??