Number: 1001

Date:  7-May-84  9':19':41

Submitter: Sannella.PA

Source: Jellinek.pa

Subject: Inside Tedit': "compiler/microcode error': unknown UFN."

Assigned To: 

Attn: 

Status: Incomplete

In/By: 

Problem Type: Bug

Impact: Moderate

Difficulty: Hard

Frequency: Once

Priority: Unlikely

System: Language Support

Subsystem: Microcode

Machine: 1100

Disk: 

Lisp Version: 2-May-84 00':11':18

Source Files: 

Microcode Version: 

Memory Size: 

File Server: 

Server Software Version: 

Disposition: [this kind of stuff just cruds up the AR system with non-reproducable results. I''m marking it Incomplete to get it out of the Open-ARs list.]'
["Masinter" " 5-Sep-84 17':31':50" Attn': Status':(Open->Incomplete) Disposition':]

Description: '
Date':  4 May 84 16':46 PDT'
From': Jellinek.pa'
Subject': Lisp': compiler/microcode error'
To': LispSupport.pa'
cc': Jellinek.pa'
'
Lisp System Date':  2-May-84 00':11':18'
Machine': Dolphin (Duchamp)'
Microcode version': 24,1'
Memory size': 10000'
Frequency': Once'
Impact': Serious'
'
Susi asked me to take a look at her machine, which had dropped into Raid.  Raid said "compiler/microcode error': unknown UFN."  The stack looked like this': (forgive my punctuation/spelling)'
'
	raid'
	\mp.error'
	\unknown.ufn'
	\sysbufp'
	\waitforsysbufp'
	errorset'
	errorset'
	\tedit.command.loop'
	\tedit2'
	\evalform'
	errorset'
	\make.process0'
	T'
'
The sysout was that of 2-May-84 00':11':18; Susi was browsing mail when it happened.  I eventually ^N-ed back into Lisp and looked at the code of \SYSBUFP, which looked ok to my untrained eye.  Everything ran fine after that, and Susi was able to continue browsing.'
'
Blame it on cosmic rays.'
		Herb'
'
'
Date': 10 May 84 11':02 PDT'
From': vanMelle.pa'
'
On the evidence presented in the report, I have reclassified this from Tedit to Dolphin microcode, priority unlikely.  I.e., it looks like either the microcode took a wild leap in \SYSBUFP (not likely), it returned to the wrong place from a function call (also unlikely), or there is a bug in the stack code (Lisp or microcode), and the \SYSBUFP frame had just undergone a stack overflow or hard return (conceivable).'
'
Without a much more detailed examination of the stack, there isn''t enough to go on.  The one piece of evidence that I would have loved to have seen in the report was the pc of \SYSBUFP.'
'


Workaround: 

Test Case: 

Edit-By: Masinter

Edit-Date:  5-Sep-84 17':31':51