Re: Problem with FDC hits in new savehits.F

[Richard Jones <richard.t.jones@uconn.edu>]

David,

As for policy on the update schedule for the hddsGeant3.F file in the 
HDGeant directory, you can exercise some executive powers on that 
decision.  In my opinion, that is not a source file but more like an 
object file.  The only reason I stuck a copy of it into the HDGeant 
directory was so that a newbie just checking out HDGeant for the first 
time could build it without errors.

Well, isn't it an error to make the wires sensitive?  That was done in 
former times as a kludge, but it is no longer appropriate, now that I 
have implemented a more correct algorithm for aggregation of separated 
hit segments for a single track within single active volume.  The reason 
it breaks the code to make the wires active is that it makes a situation 
where an active volume is embedded inside another active volume.  If you 
do that, then it breaks.  I can refine the algorithm to deal with that 
case if you want, but when I was writing it I could not think of a 
concrete case when I want active volumes inside other active volumes, so 
I assumed not -- think of it as a stack with a maximum depth of 1.  And 
of course, if I cannot think of a case where it might happen then it 
won't happen, right?

If you want a more general algorithm, I can create full stack semantics 
instead of just a single set of registers.  Let me know what people want 
to see.

-Richard J.


David Lawrence wrote:
>
> Hi Richard,
>
>    I believe I have found the error. After verifying that the problem 
> persisted using a fresh build on both OS X 10.5 and Linux F8, I 
> noticed that the hddsGeant3.F file in the HDGeant directory was not 
> consistent with the XML in the hdds directory. Specifically, the 
> change made on Jan. 3 (revision 4652) changed the FDC wires to be 
> insensitive. I'm not sure how this leads to such a dramatic change, 
> but it sure seems to. I tried remaking the hddsGeant3.F file from the 
> XML and now I see lots of hits in the first event. When I change that 
> FDAW volume's "sensitive" attribute to "true", I get only a couple of 
> hits.
>
>    I have now committed an updated version of the hddsGeant3.F file to 
> the release-2009-01-06 since that was broken as it was. For the future 
> we will change the Makefile in the hdds directory to automatically 
> place the hddsGeant3.F file in the HDGeant directory and we will 
> remove the hddsGeant3.F file from the repository altogether to force 
> regeneration.
>
> Regards,
> -David
>