<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=UTF-8">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    <br>
    Hi All,<br>
    <br>
        I've just committed a fix to the seg. fault/hang problem based
    on Richard's analysis. This is slightly different than the fix
    Richard suggested. It avoids the 80bit/64bit comparison issue by
    pre-calculating the values to be compared rather than doing it in
    the sort algorithm itself. This should also speed things up a little
    since DVector3::Perp() is not being called repeatedly during the
    sort for the same object. <br>
    <br>
        I have been able to run through my one reliably-problematic
    event using the new code without any problem. If anyone notices an
    issue, please let me know.<br>
    <br>
        This problem has been marked as resolved in Mantis.<br>
    <br>
    Regards,<br>
    -David<br>
    <br>
    -------- Original Message --------
    <table class="moz-email-headers-table" border="0" cellpadding="0"
      cellspacing="0">
      <tbody>
        <tr>
          <th align="RIGHT" nowrap="nowrap" valign="BASELINE">Subject: </th>
          <td>r7587 - trunk/sim-recon/src/libraries/TRACKING</td>
        </tr>
        <tr>
          <th align="RIGHT" nowrap="nowrap" valign="BASELINE">Date: </th>
          <td>Thu, 17 Mar 2011 11:16:40 -0400</td>
        </tr>
        <tr>
          <th align="RIGHT" nowrap="nowrap" valign="BASELINE">From: </th>
          <td><a class="moz-txt-link-abbreviated" href="mailto:Hall-D.SVN.Repository@jlab.org">Hall-D.SVN.Repository@jlab.org</a></td>
        </tr>
        <tr>
          <th align="RIGHT" nowrap="nowrap" valign="BASELINE">To: </th>
          <td><a class="moz-txt-link-abbreviated" href="mailto:davidl@jlab.org">davidl@jlab.org</a>, <a class="moz-txt-link-abbreviated" href="mailto:brash@pcs.cnu.edu">brash@pcs.cnu.edu</a>, <a class="moz-txt-link-abbreviated" href="mailto:wolin@jlab.org">wolin@jlab.org</a>,
            <a class="moz-txt-link-abbreviated" href="mailto:zisis@uregina.ca">zisis@uregina.ca</a>, <a class="moz-txt-link-abbreviated" href="mailto:mashephe@indiana.edu">mashephe@indiana.edu</a>,
            <a class="moz-txt-link-abbreviated" href="mailto:remitche@indiana.edu">remitche@indiana.edu</a>, <a class="moz-txt-link-abbreviated" href="mailto:zihlmann@jlab.org">zihlmann@jlab.org</a>, <a class="moz-txt-link-abbreviated" href="mailto:somov@jlab.org">somov@jlab.org</a>,
            <a class="moz-txt-link-abbreviated" href="mailto:staylor@jlab.org">staylor@jlab.org</a></td>
        </tr>
      </tbody>
    </table>
    <br>
    <br>
    <pre>Author: davidl
Date: 2011-03-17 11:16:39 -0400 (Thu, 17 Mar 2011)
New Revision: 7587

Modified:
   trunk/sim-recon/src/libraries/TRACKING/DTrackCandidate_factory_CDC.cc
Log:
This is a fix for the seg. fault/hang problem that has been plaguing us 
for the last ~4 months. It precalculates the values used in the comparison
in the SortIntersections routine to avoid issues with values calculated
with 80bit precision being compared with values having been copied to and
from a 64bit register. See the report on the GlueX wiki here:

<a class="moz-txt-link-freetext" href="http://www.jlab.org/Hall-D/software/wiki/index.php/Diagnosing_segmentation_faults_in_reconstruction_software">http://www.jlab.org/Hall-D/software/wiki/index.php/Diagnosing_segmentation_faults_in_reconstruction_software</a>


</pre>
  </body>
</html>