Localization of the program crash point

Hello everybody, pls HELP!
I ran natural program on SUN Solaris in interactive mode. The program is crashed.
I can see:
1., on the terminal:
Encore Thread (1) <> (Signal 16) Callstack: Stack Backtrace ©: 24027: natural ----------------- lwp# 1 / thread# 1 -------------------- ffffffff7e7a5c78 poll (ffffffff7fffbe30,
0, 1f4) ffffffff7e755550 select (ffffffff7e8bc790, ffffffff7fffbe30, ffffffff7e8bc790, ffffffff7e8bc……………………………………

2., in a Natural log file:
NATURAL V 6.1.1 Pl 14 Software AG 2005
^[[62;1"p^[P1;1|23/1B^[^[[?4l^[[?7h^[[?8h^[[2J^[)0^O^[[m^O^[[?25l^[=^[>^[[?1h^[[2J^G^[[1;2H2010-06-28^[[1;37HNATURAL^[[1;61HLibrary: SYSTEM^[[2;2H18:21:49^[[2;25HV 6.1.1 Pl 14 Software AG 2005^[[2;61HMode : REPORT^[[3;2HUser: BATCH^[[3;58HWor
k Area : empty^[[23;1H^[[7m NAT0082 Invalid command, or Object TMA50C1 does not exist in library ^[[1;1H^[[?25h^^^A

Actually the program was working something because some output report was come out.
My main question is how I can figure out that in which program line was the crash?

I cannot quite recognise the terminal input - but then again unix is not my primary turf…
But what I notice is this info:

Library: SYSTEM
NAT0082 Invalid command, or Object TMA50C1 does not exist in library

Are you sure you want to run the program in library SYSTEM ?
A qualified guess is that you should run your program in the library where TMA50C1 is located :wink:

Finn

Hi Finn - thanks for the reply!

Actually I was in the Natural editor session and just put PF4 to run my program.

The current library and program name were totally different. SYSTEM lib and TMA50C1 are never used by me.

Based on “NAT0082 Invalid command” I can assume that It was a kind of Natural session crash caused my program.

So I’m looking for method to identify the guilty line in my program.

Thanks in advance!

Given that the log shows “Work Area: empty”, either the program you are RUNing (or something it calls - check your maps, CALLNATs, copycode…use XREF if you have it to check) is clearing itself (using terminal commands like %Z) or there is a software error.

Does the issue happen if you STOW the program then execute (not RUN) it? If not, it is more likely a software issue.

Assuming the program isn’t obviously clobbering itself, you might try upgrading to a more recent version of Natural (6.3.8 is available; 6.1 is out of support) and seeing if the problem is still present. If it is, then you can open a support request with Software AG to look at the issue.

Are you “playing” with the Stack?

Is that perhaps where TMA50C1 came from?

steve

Thanks for ideas, next findings:

  • software error possible but I assume config/setup problem because I have (carefully!) run the program on production system and gives NAT1316 with line number

    Question: if it is then how I can identify the config/setup problem on the test system, let say what files/parameters need to compare?

  • STOW works well, no Stack “playing”, no sub program calling

  • could I have link to related SAG page to see supported Natural versions

Thanks again!

See http://empower.softwareag.com for supported versions. You will need a user id for empower.

Thank you Douglas,

I look after the version upgrade.

Any idea that where I can find Natural system/user environment parameters? (This Natural is running on Unix)

Thanks: Lajos Pasti

http://techcommunity.softwareag.com/ecosystem/documentation/natural/nat638unx/config/cfg-over.htm

I have compared all “sagenv” scripts. No difference between production and test system.

Only strange I have found that the two systems do not contain “sag” group definition in /etc/group file. But might be it is OK because production system works well.