[Rivet] AGILe svn: r551 - in trunk: . src src/CCHerwig

[James Monk]

I'm a bit nervous of updating the AGILe on the Debrecen system now  
that I have one that works (albeit with some weird cout).

I also noted some warning on the getenv description page that the  
contents of the char* env pointer returned by getenv should not ever  
be modified (even tho' it is not a const pointer) - I wonder if that  
may have been happening somwhere in the string chopping, but I  
couldn't see where.

On 11 Aug 2008, at 17:42, Andy Buckley wrote:

> Holger Schulz wrote:
>> rivet at projects.hepforge.org wrote:
>>> Author: buckley
>>> Date: 2008-08-11 16:18:48 +0100 (Mon, 11 Aug 2008)
>>> New Revision: 551
>>>
>>> Modified:
>>>   trunk/ChangeLog
>>>   trunk/src/CCHerwig/CCHerwig.cc
>>>   trunk/src/Loader.cc
>>> Log:
>>> Factorising loader functions for more fine-grained valgrind bug  
>>> tracing on problem systems, and added special Herwig++ 'INFILE'  
>>> parameter handling via ThePEG repo 'read' command.
>
>> Now I don't need a cout anymore, great, thanks!
>
> Really?!? I thought I had more or less just translated what we  
> already had into
> "bite size" chunks which would be better for memory analysis! Good!
>
> The only "sort-of uninitialised" variable I saw was the "char* env =  
> 0"
> pointer, which was used repeatedly for checking whether environment  
> variables
> were set. It wasn't being reset in-between calls to getenv, since I  
> didn't
> think it needed to be, but on rewriting I was paranoid and inserted  
> lots of
> "env = 0". So maybe this, or the more localised scoping introduced  
> by the
> refactoring sorted the problem out. Either way, I'm glad the problem  
> is gone!
>
> Andy
> _______________________________________________
> Rivet mailing list
> Rivet at projects.hepforge.org
> http://www.hepforge.org/lists/listinfo/rivet