[Clas12_software] Coatjava 2.0 reconstruction crash

Mon Jul 6 15:03:32 EDT 2015

Hi Andrey,

On Mon, Jul 06, 2015 at 02:32:30PM -0400, Andrey Kim wrote:
>I think the policy we are trying to enforce is to limit the number of 
>people able to push any changes to the code. 

I think this is a good idea. Just to be clear, you want to use the pull request work flow.
There will be a master repository (i.e. Gagik's github repo) which only a few people can push to. 

>To my knowledge, only 
>Gagik and Veronique can push them at the moment. Obviously, two people 
>are not enough for that amount of work. But we want to prevent 
>wrong/harmful changes to the code that are hard to debug.

This is also a good idea to have some individuals repsonsible for managing  and overseeing new contributions.
If someone has changes to contribute, they simply make a "pull request". This can be done via the github webpage or a simple email (on this list).  

>Probably we have to come up with some kind of review procedure 
>(similar to what Harut suggested during the collaboration meeting's 
>presentation). May be appoint some "review committee" that should 
>approve the code changes before we push it to the final release.

I think this is probably a bit much considering the flexability of git.
A review committee is always a idea, but it shouldn't be required before any changes are merged since they can always be undone.
Of course this is also why there should be onlyh a few "managers" of the master git repo that can merge changes.

In fact, a lack of shared changes is what caused my initial problem. Changes were made to  magfield  without the latest updates. 

>I understand that sometimes little bug can be annoying and stop one's 
>work completely. But if we all will be able to push changes (similar 
>to CLAS6) we'll have code that is very hard to manage.

I was not involved with CLAS6 but my guess would be that git will solve many of these problems.  :)
Once the code is setup on git with straightforward build, the code managment will be much easier than in the past.

Cheers,
Whitney

>I think it is best to discuss this issue during the Thursday meeting.
>Meanwhile, you can use plotting from coatjava-2.0 and everything else 
>from coatjava-1.0.
>Best,
>Andrey.
>
>
>
>On 07/06/2015 01:15 PM, Whitney R. Armstrong wrote:
>>Gagik,
>>
>>Thanks for your reply. However, I needed to run 2.0 in order to make 
>>plots (see my earlier email). This now makes me run in a circle and 
>>constantly have to switch versions.
>>
>>To the point: where can I find the source code that needs fixed?
>>
>>Cheers,
>>Whitney
>>
>>On Mon, Jul 06, 2015 at 08:27:01AM -0400, Gagik Gavalian wrote:
>>>
>>>run reconstruction from version 1.0.
>>>
>>>
>>>>On Jul 3, 2015, at 5:04 PM, Armstrong, Whitney R. 
>>>><warmstrong at anl.gov> wrote:
>>>>
>>>>Thanks David.
>>>>
>>>>Since this is a bug, do you (or anybody else) know where I can 
>>>>checkout the latest source code for coatjava? This way I could 
>>>>maybe find the bug, fix it, push the fix upstream, and maybe 
>>>>learn something in the process.
>>>>
>>>>I seem to only be able to find jar files ...
>>>>
>>>>Cheers,
>>>>Whitney
>>>>
>>>>From: David Heddle [david.heddle at cnu.edu]
>>>>Sent: Friday, July 03, 2015 12:33 AM
>>>>To: Armstrong, Whitney R.
>>>>Cc: clas12_software at jlab.org
>>>>Subject: Re: [Clas12_software] Coatjava 2.0 reconstruction crash
>>>>
>>>>This is an inconsistency between the latest magfield package and 
>>>>the reconstruction you are using. The lastest magfield removed 
>>>>methods to invert the field and to test for inversion, instead 
>>>>(to be more consistent and simplify usage) it uses only a 
>>>>setScale method, which can also invert, e.g. by setting th scale 
>>>>factor to -1.
>>>>
>>>>I know that the reconstruction software was modified to comply 
>>>>with the new magfield interface--but somehow some old 
>>>>reconstruction code is still around.
>>>>
>>>>Hope that helps,
>>>>dph
>>>>
>>>>On Thu, Jul 2, 2015 at 11:53 PM, Armstrong, Whitney R. 
>>>><warmstrong at anl.gov <mailto:warmstrong at anl.gov>> wrote:
>>>>Hello Everyone,
>>>>
>>>>I switched over to coatjava 2.0 and the following error occurs 
>>>>when reconstructing simulation.
>>>>...
>>>>test location (XYZ): (-31.70070, 31.14780, 28.15800) magnitude: 
>>>>52.61172
>>>>test Field Vector (XYZ): (-2.01643,  1.98126, 64.05748) 
>>>>magnitude: 64.11983
>>>>test location (CYL): (135.50404, 44.44232, 28.15800) magnitude: 
>>>>52.61172
>>>>test Field Vector (CYL): ( 0.00000,  2.82690, 64.05748)
>>>>
>>>>Exception in thread "main" java.lang.NoSuchMethodError: 
>>>>cnuphys.magfield.Solenoid.isInvertField()Z
>>>>        at org.jlab.rec.bst.trajectory.BSTSwimmer.getMagneticFields(BSTSwimmer.java:239)
>>>>        at org.jlab.rec.bst.services.BSTReconstruction.init(BSTReconstruction.java:317)
>>>>        at org.jlab.clasrec.rec.CLASReconstruction.initDetectors(CLASReconstruction.java:95)
>>>>        at org.jlab.clasrec.rec.CLASReconstruction.run(CLASReconstruction.java:120)
>>>>
>>>>        at org.jlab.clasrec.rec.CLASReconstruction.main(CLASReconstruction.java:236)
>>>>
>>>>The version 1.0 runs without crashing.
>>>>
>>>>Cheers,
>>>>Whitney
>>>>_______________________________________________
>>>>Clas12_software mailing list
>>>>Clas12_software at jlab.org <mailto:Clas12_software at jlab.org>
>>>>https://mailman.jlab.org/mailman/listinfo/clas12_software 
>>>><https://mailman.jlab.org/mailman/listinfo/clas12_software>
>>>>
>>>>
>>>>
>>>>-- 
>>>>David P. Heddle, Ph.D.
>>>>Professor of Physics
>>>>Christopher Newport University
>>>>Newport News, VA 23606
>>>>
>>>>757.594.8434 (CNU)
>>>>_______________________________________________
>>>>Clas12_software mailing list
>>>>Clas12_software at jlab.org
>>>>https://mailman.jlab.org/mailman/listinfo/clas12_software
>>>
>>
>
>_______________________________________________
>Clas12_software mailing list
>Clas12_software at jlab.org
>https://mailman.jlab.org/mailman/listinfo/clas12_software

-- 
Whitney R. Armstrong