Re: error message from IOC software

[Jeff Hill <johill@lanl.gov>]

Oleg A. Makarov wrote:

I am not the author of nor a frequent user of the sequencer. Nevertheless since
there is some useful information here I attempted to track down the cause of the 
problem.

> info_ID: A call to "assert (textBuffer != NULL)" failed in ../tsSubr.c at
> 1045
> Please send a copy of the output from "tt (0x551008)" and a copy of this
> message
> to the author or "tech-talk@aps.anl.gov"
> task: 0X6709e8 taskwd
> task 551008 info_ID suspended
> iocid08> tt (0x551008)
>   b2ac _vxTaskEntry   +10 : _sequencer (563378, 2af8, 55991d, 0, 0, 0, 0, 0,
>   0, 0)
> 56181a _sequencer     +d8 : _ss_entry ([563378, 55b524, 0, 0, 0])
> 56193a _ss_entry      +114: 558568 (55b524, 0, 0)
> 5585a8 info_ID.o_text +3c8: _tsStampToText (&_tmc, 1)
> 6f9f24 _tsStampToText +22 : _epicsAssert (6f93e0, 415, 6f9e92)
> 6fcaa4 _epicsAssert   +50 : _taskSuspend ([551008, 1, 1, 0, 55b524])
> value = 0 = 0x0

The task trace above indicates that tsStampToText() is being called with 2 
instead of the required 3 arguments. Since assert is complaining that
the last argument should never be nill this appears to be a plausible 
explanation of the problem. However when I grep in your version of
the sequencer I dont see this occurring:

seq_main.c:     tsStampToText(&timeStamp, TS_TEXT_MMDDYY, pBfr);
seq_qry.c:              tsStampToText(&pDB->timeStamp, TS_TEXT_MMDDYY, tsBfr);

The WRS "tt()" utility is occasionally wrong about the number of arguments.
I dont expect either of the above two locations in the source because 
the entry points are not declared "static" and therefore the name
of the routine would be in the task trace. Also these particular calls
to tsStampToText() do not appear to be likely candidates for passing
a nill 3rd arg to tsStampToText().

With this limited information I am unable to
locate a cause for this problem. Is it possible that you
are calling tsStampToText() either directly or indirectly from
within your sequence program? The symbol "tmc" in the
stack trace should be a good clue. Try typing lkup "tmc".
This should provide an address for the symbol "tmc". Next
type lkAddr <symbol addr> which should indicate which source
code in your system is making this call. The symbol "tmc"
does not show up here on an mv167 running EPICS 3.12.2.

Rozelle indicated that many problems were recently fixed in the sequencer.

Looking in CVS at APS it does appear that some changes have been made 
since version 1.9.0 (the version that you are using). Here is the
entry in the log:

revision 1.14
date: 1996/05/09 19:30:16;  author: mrk;  state: Exp;  lines: +1 -1
I am committing changes made by LANL and LBL

These changes are in EPICS 3.13 (however it appears that nothing
has changed related to calls to tsStampToText()).

Jeff

-- 

______________________________________________________________________
Jeffrey O. Hill			Internet	johill@lanl.gov
LANL MS H820			Voice		505 665 1831
Los Alamos, NM 87545 USA 	FAX		505 665 5107