Home » PANDA » PandaRoot » Bugs, Fixes, Releases » Bug fix in PndEmcMakeBump.cxx
| Bug fix in PndEmcMakeBump.cxx [message #7772] |
Thu, 29 January 2009 17:34  |
Bertram Kopf
Messages: 110
Registered: March 2006
|
continuous participant |
From: *ep1.ruhr-uni-bochum.de
|
|
Hi,
by running some emc macros with the latest trunk revision I observed that no PndEmcBumps have not been stored in the root files anymore. This was caused by the following two commented lines in PndEmcMakeBump.cxx:
...
for (unsigned int i = 0; i < theBumps.size(); i++){
// Int_t size_ba = fBumpArray->GetEntriesFast();
// PndEmcBump* theNextBump = new((*fBumpArray)[size_ba]) PndEmcBump(*(theBumps[i]));
if ((fVerbose>=1)&&(theBumps.size()>1)){
...
I uncommented these two lines and it seems to me that it works fine again. The changes have been done in revision 4441. Please check whether everything is o.k. now.
Cheers,
Bertram.
|
|
|
|
|
|
|
|
|
|
| Re: Bug fix in PndEmcMakeBump.cxx [message #7779 is a reply to message #7777] |
Fri, 30 January 2009 10:25  |
StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino
|
first-grade participant |
From: *dip.t-dialin.net
|
|
Hi Bertram and Florian,
the svn comments are fine, but the normal developer/user does not read svn usually, only when he finds some troubles.
The "rule" -> 'announcement of code changes' was stated I think two years ago, but it is not always followed.
The EMC thread is done for this for the emc detector, considering that all the emc developers are sitting in different institutes, as well the tracking thread and so on.
In general even this thread "Bugs, Fixes, Releases" can be used for this purpose, one should just use it.
Even if for small and decorative changes the message can be skipped, when the code is mid-heavily changed these are mandatory, in particular for developers which are using the trunk and uploading the code almost day by day, and even for people working on the same code that find something that can change their results.
Regards
|
|
|
|
|
|
|
|
|
|
| Re: Changelog bookeeping [message #7786 is a reply to message #7782] |
Sun, 01 February 2009 13:19 |
Johan Messchendorp
Messages: 693
Registered: April 2007
Location: University of Groningen
|
first-grade participant |
From: *xs4all.nl
|
|
Dear all,
I would support the line of thoughts of Florian, e.g.
1) try to provide as accurate comments while committing your changes. I know this is not easy and asks for some discipline (...and, ouch!, I am certainly not the best example of this!). And with "trac", one even gets an effective overview as far as I understand......
2) The forum can also be used to inform about changes. Here, I would not submit all the changes, but primarily inform about really major changes or additions to the code, e.g. to tell that new features are added, etc.
3) Concerning the branches: Indeed, for the trunk it is not required to follow QA, but to be able to compile and link. If you require more, please use the "stable" branch, which - at the moment - is not really used and outdated.
4) The package manager should know about what is going on in the part of the code which he/she is managing... (Stefano's comment!)
Greetings,
Johan.
[Updated on: Sun, 01 February 2009 13:34] Report message to a moderator |
|
|
|
|
|
|
|
| Re: Changelog bookeeping [message #7789 is a reply to message #7788] |
Sun, 01 February 2009 16:00 |
Jens Sören Lange
Messages: 193
Registered: June 2005
|
first-grade participant |
From: *web.vodafone.de
|
|
Hi all,
I also think that Stefano and Johan are right. However, maybe the confusion also arises from the fact, that in the structure before Sep08 the package manager was also (sort of) the (sub)coordinator. As an example, in case of emc, Stefano did all the emc code restructuring, merging of all the different calorimeters, all the phone calls, tested all the macros etc. etc. At least 50% of this work is emc (sub)coordinator work, which we now have in addition. So now we have two persons who are being responsible if the code doesn't work and/or there is not sufficient documentation. So maybe we just have to clearly define, who is taking care of what. The developers and package coordinators know the code and maybe write more svn comments, but maybe the (sub)coordinators should test the macros and maybe write more the Wiki (e.g. as it was in my case, and that is sort of an implicit filter so that not all the comments reach the Wiki). What would you think?
cheers, Soeren
|
|
|
|
| Re: Bug fix in PndEmcMakeBump.cxx [message #7790 is a reply to message #7772] |
Sun, 01 February 2009 17:29 |
Mohammad Al-Turany
Messages: 518
Registered: April 2004
Location: GSI, Germany
|
first-grade participant |
From: *dip.t-dialin.net
|
|
Hi every body,
More comments in the SVN would surely help, this is the easiest way to follow the changes, and keep in mind that the trunk is assumed to compile, and the more tested code should be moved to the stable! that what we discussed few months ago, also I remember we had some discussion about responsibles! speaking about coordinators and sub-coordinator is simply confusing for me, because of the many ways how the people define the responsibilities of a coordinator!!
one can call them coordinators or sub-coordinators or what ever he likes, I have a list in the svn authority file of the people who can decide if a package goes from stable to release, these people for me are the responsibles (They all agreed at that time to take care of the the packages they work with!):
abiegun, acecchi, asanchez, awronska, johan, kgoetzen, pablo, ralfk, sneubert, soeren, spataro, tstockm
So I would suggest that these people may write something in the wiki when they move something to release! and every developer should try to write more comments in trunk and stable.
cheers, Mohammad
|
|
|
|
Goto Forum:
Current Time: Sat Apr 27 20:11:53 CEST 2024
Total time taken to generate the page: 0.00799 seconds
|
GSI Forum
Search
Help
Members
Login
Home
