[Halld-offline] usage of HALLD_HOME in Makefiles with new sim-recon root

[Mark M. Ito]

Richard and David,

After this exchange, I'm thinking that suggestion 2 is more or less 
unavoidable.

On cost/benefit, the short term benefit is clearly zilch, but the 
overall idea of making the switch was to make the structure clearer and 
easier to describe, knowing that there would be some pain in having to 
switch. Having the legacy environment variable hanging around is kind of 
like not finishing the job. Long-term, it is easier to understand.

We could somehow grandfather in the use of HALLD_HOME in the makefiles 
by checking for the case where HALLD_HOME is defined and SIM_RECON_HOME 
is not and making the definition of the latter on the fly. We could warn 
that such usage is deprecated.

Typing? How about SR_HOME, SRSYS, SR,...?

   -- Mark

On 04/21/2010 10:46 PM, David Lawrence wrote:
>
> Hi Richard,
>
>     I'll let Mark give his take on this, but  since you asked on a 
> public mailing list, I can't help but chime in :)
>
>
>     We did discuss this at the March 9th meeting. I believe Mark 
> advocated something close to your suggestion 2., but was shouted down 
> by the majority in attendance. One major reason being that HALLD_HOME 
> is pervasive not just in lots of Makefiles, but in peoples' login 
> scripts etc. Changing it would have some ripple that would likely 
> cause more pain (or at least annoyance) than any gain from simply 
> changing the name of the variable. It may not have been well 
> advertised, but the standard install is now supposed to have the 
> HALLD_HOME environment variable point to the sim-recon directory and 
> everything else should work as before.
>
>     Your suggestion 1. has the problem that HALLD_HOME could no longer 
> be used to point to a tagged release. Because our tagged releases are 
> now named e.g.  sim-recon-2010-03-29 you'd want your environment 
> variable (whatever it is named) to point to that directory, not its 
> parent. Also, it seems more standard to have the "HOME" directory 
> contain src, include, bin, and lib directories as in your suggestion 2.
>
>     IMHO, I can see possibly going with suggestion 2., but I would 
> still advocate keeping it as it is for two reasons:
>
> A: Cost/Benefit doesn't strongly motivate a change
> B: "HALLD_HOME" is quicker to type than "SIM_RECON_HOME" (and some of 
> us type it a lot!)
>
> My $0.02
>
> Regards,
> -David
>
>
>
> Richard Jones wrote:
>> Dear Mark,
>>
>> In trying to move over to the new structure of the Gluex software 
>> tree, I encountered the following issue.  First, I created a new 
>> directory called $(HALLD_HOME)/sim-recon.  Then I cd'ed into 
>> sim-recon and did a fresh checkout of the trunk from subversion into 
>> the directory $(HALLD_HOME)/sim-recon/src.  After this, I did cd src, 
>> followed by make.  This fails in HDGeant.
>>
>> The reason for this is that in HDGeant/Makefile.bms there is the 
>> assumption that $(HALLD_HOME)/src is the absolute path to the 
>> sim-recon/src directory.  I see a couple of ways forward.
>>
>>    1. go through a modify all instances of $(HALLD_HOME)/src to
>>       become $(HALLD_HOME)/sim-recon/src in the makefiles.  You might
>>       then also consider whether $(HALLD_HOME)/lib and
>>       $(HALLD_HOME)/bin should also descend into the sim-recon folder
>>       or not.
>>    2. invent a new environment variable called SIM_RECON_HOME and
>>       then use $(SIM_RECON_HOME)/src in place of $(HALLD_HOME)/src in
>>       the makefiles.  This is more consistent with other cases, such
>>       as $(JANA_HOME).
>>
>> For the time being, I have implemented option 1 above (src only).  I 
>> have only modified my local copy, and have not checked my changes 
>> into the trunk.
>>
>> -Richard J.
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Halld-offline mailing list
>> Halld-offline at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/halld-offline
>
> -- 
>
> ------------------------------------------------------------------------
>   David Lawrence Ph.D.
>   Staff Scientist                 Office: (757)269-5567   [[[  [   [ [
>   Jefferson Lab                   Pager:  (757)584-5567   [  [ [ [ [ [
>   http://www.jlab.org/~davidl      davidl at jlab.org          [[[  [[ [[ [[[
> ------------------------------------------------------------------------
>
>    
>
>
> _______________________________________________
> Halld-offline mailing list
> Halld-offline at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-offline

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-offline/attachments/20100422/d731491e/attachment-0002.html>