
------------------------------------------------------------------------------

	                                lll      ii    bbb     
        	                        ll             bb     
	     rr rrr  nnnnn    aaaa     ll     iii     bb     
	     rrr rr nn  nn      aa    ll      ii     bbbbb  
	    rr  rr nn  nn   aaaaa    ll      ii     bb  bb 
	   rr     nn  nn  aa  aa    ll      ii     bb  bb 
	 rrrr    nn  nn   aaa aa  llll    iiii   bb bbb    (c)

 			     RNALIB 2.0  
 
            Copyright (c) 1989,1990, by Paolo Ventafridda

               Operator's Guide: Bugs & Problems Hunting
               =========================================

This document contains reports from beta-testers and rna operators;
common problems and possible solutions for various hardware/software
configurations. Please submit your reports to  venta@i2ack.sublink.org.



---

Due to the size of RNA, you might get some problems with the shell.
This is very likely to happen on 16bit-based xenix/unix systems,
since their shell have a fixed stack . Basically, here's what you can get.

(1)
DIAGNOSTIC:    sh: RNA: no space

CAUSE:
	your shell has not enough core space for keeping in memory RNA
	and its mail. The program does NOT start, so you don't loose
	rnalib mail.

SOLUTIONS:

	Check that it's really a problem of space. If you have more than
	one mail inside rnalib' mailbox, remove all but the first and
	retry executing RNA. If RNA can deal with a single message but
	crashes with others at the same time, you may try one of the
	following solutions:

	- 	Edit RNA. Remove the very first line, containing ':'.
		Or, if you own a version with the first line as '#/bin/sh'
		then remove this one. Then execute RNA with 
		' sh RNA ' (also from cron).

	-	Try lowering stack space (ugly!). 
		Do a copy of your /bin/sh. Do 'hdr /bin/sh and take note
		of the 'stack size' number. Then execute 'fixhdr -F yyyy sh'
		where yyyy is a number slightly lower than the old one.

	
(2)
DIAGNOSTIC:     RNA: no stack space

CAUSE:
		Your shell can't stand RNA internal environments.
		It starts but fails after boot; you (may) loose mail, beware.
		I am assuming you did not add or remove any ' { } '
		or brackets; if you modify bracket's position you may
		get serious problems and need to debug from scratch.
		Do not modify RNA unless you are 100% sure of what
		you are doing.  

SOLUTIONS:
		
	-	Run RNA with 'sh RNA' also from cron.

	-	Enlarge shell's stack size. Get the current one with
		' hdr /bin/sh ', then fix it with 'fixhdr -F yyyy sh'
		where yyyy is the new size. No warranty! Caution!

	
(3)
DIAGNOSTIC:	cannot unlink /usr/spool/mail/rnalib    (sort-of)

CAUSE:
	
		You are suffering of strange mail access permissions.
		Try removing your own mailbox with 'rm' (do a copy
		before!). Do 'id' and see if you have some 'egid'
		permissions.

SOLUTION:

		This is a problem of yours, not really an RNA one.
		Anyway try fixing the 'egid' thing; check that inside
		/etc/passwd rnalib has a /bin/sh or /bin/csh or nothing
		at all (defaults to /bin/sh). 
	
(4)
DIAGNOSTIC:
		RNA fails a 'find' command; everything looks like Ok, but
		a 'find' causes a 'Memory Fault'.

CAUSE:		It might really be a problem of 'grep'. 
		At least, on SCO Unix System V 3.2 this was the problem.

SOLUTION:	Search for the 'FindFile' procedure inside RNA source.
		Replace the line:

                ls -got $FFpath | StripPermAndLink | grep -y $FFchars && {

		with:
		
		ls -got $FFpath | StripPermAndLink | fgrep -y $FFchars && {
						     ^^^^

		(This was included inside source as a %%FIX%% ).


(5)  
SENT BY: palmer@univaq.sublink.org
DATE: 17/9/1990
SYSTEM: UNISYS 7000/40 running UNIX sysV R.2.21.02
---
E.R. Palmerini - Centro di Calcolo Facolta' di Ingegneria
Universita' Degli Studi De L'Aquila (0862-432511)
---

* RNA could not remove its own mailbox: it was fixed giving to rnalib
  group 'mail'.

	On UNISYS 7000 under old Unix 2.x problems appear
	with the StripPermAndLink() function.
* There was some trouble with 'head': executed by cron, RNA could not
  find this command. Fixed by putting full pathname to 'head'.

* Along with 'head', some other commands could not be found.
  Always executed by cron (not by shell), strange 'cannot shift' were
  appearing, generating ping-pong mails between rna and root (cron owner).
  Fixed by changing 'shift' with 'shift 2>/dev/null' 

AUTHOR'S NOTE:
	It could be just a stupid problem of 'PATH'. In this case a simple
        PATH=yourpaths/as/usual: could solve the whole situation.
	However i am not sure, since have no 7000/40 at home ;-)
	So i included PATH declaration inside RNA few hours before shipping!

There was a patch for the 7000, but basically all you have to do
is changing 'shift' with 'shift 2>/dev/null' inside the program,
and a major workout on the followings:

359c359
< 	echo "$4\t$5,$6\t$7\t$8"
---
> 	echo "$3\t$4,$5\t$6\t$7"
(*** inside StripPermAndLink ***)

402c402
< 	ls -ot $LLpath | StripPermAndLink
---
> 	ls -got $LLpath | StripPermAndLink
427c427
< 	ls -ot $FFpath | StripPermAndLink | fgrep -i $FFchars && {
---
> 	ls -got $FFpath | StripPermAndLink | fgrep -i $FFchars && {

There's a patch for the 7000/40 which applies to release 2.0 rev.A of RNA.
Does nothing but changing shift to shift 2>/dev/null and above things.
If you are lazy, this patch is named 'ptc.7000.a'. Use at your own risk!
		

=== your suggestions will be put here in the next release ===
Please give a description of your problems and a solution, if any.
Don't send big 'diffs' since they are related to versions of RNA
which are going to be obsoleted.
In any case, report:
1) hardware in use
2) operating system/release
3) rna release referring to

Thanks a lot, any contributions are welcome.
Remember not to post too much stuff cause i pay for the traffic ;-(

-- Paolo
