I did that since my fix is rather kludgy and I wanted to leave it to
others to apply a proper fix upstream of sda_xmm55. Since Stepan's fix is
in sda_xmm55 also, I suspect that his is pretty kludgy as well (haven't
looked at it yet). We will get together an install a compromise kludgy fix
on the main branch.
Rob Feuerbach looked at this and concluded that a proper fix was not
completely obvious. He tells me that this error occurs for some good
tracks, and simply abandoning the track altogether would incur a
significant hit in efficiency.
Mark M. Ito, Thomas Jefferson National Accelerator Facility
12000 Jefferson Ave., Mail Stop 12H, Newport News, VA 23606
Email: marki@jlab.org, Office: (757)269-5295, Pager: (757)680-7175
WWW: http://claspc2.cebaf.gov/~marki/
On Thu, 12 Mar 1998, James A. Mueller wrote:
> > From list-error@jlab.org Wed Mar 11 20:55:40 1998
> > Return-Path: <list-error@jlab.org>
> > Date: Wed, 11 Mar 1998 20:50:22 -0500
> > From: Alexander Vlassov <Alexander.Vlassov@itep.ru>
> > Subject: bag report
> > Sender: vlassov@itep.ru
> > To: clas_offline@jlab.org
> > Errors-to: list-error@jlab.org
> > Warnings-to: list-error@jlab.org
> > Message-id: <35073F4F.74BC@itep.ru>
> > Organization: ITEP, Moscow
> > MIME-version: 1.0
> > X-Mailer: Mozilla 3.01 (X11; U; SunOS 5.5 sun4u)
> > Content-type: text/plain; charset=us-ascii
> > Content-transfer-encoding: 7bit
> > Content-Length: 748
> > Status: RO
> >
> Alexander,
> The problem with xmm55 was fixed by Mark Ito on January 29th,
> and is in release-1-13. It also appears to have been fixed by Stepan
> on March 5, but I like Mark's better, because is prints an annoying
> warning. The problem I see with these two fixes is that they are separate
> branches of the code (which is probably why it got fixed twice).
>
> Mark,
> You should probably merge you changes with Stepan's or copy yours
> back from a branch to the head.
>
> -Jim Mueller
>
> > Hello,
> >
> > I am using RECSIS version 1-12. I had an experience, that
> > my program crashes after 1500 - 2000 events on HP and and works
> > on Sun. In order to fix the problem, I ran the program under
> > debugger (ddd) and found that the error occurs from sda_xmm55
> > subroutine. Actually it could be wrong index number in calling this
> > routine from sda_trmat.
> > After that I update trk package and put some warning printing
> > statements into both program; make library and recompile the
> > executable. As a result, I have no any tarack. ( Nor HB nor TB)
> >
> > What's happend since version 1-12 with trk package , tcl -
> > variables etc which can cause this result ?
> >
> > Sincerely,
> > Alexander Vlassov.
> >
> > Home page : http://www.cebaf.gov/~vlassov/
>