Hi, everybody!
I have a few comments about Simon's changes:
> I do not see a reason why Moller
> stuff should be contained in the blastmc directory, so I moved it
> to userGen/userMoller and make it there, where all other user
> generators are also made. For this reason, I commented out the
> mollerInter stuff from blast_proc.f since otherwise linking of blastmc
> will fail. Note that now even Moller is called with KINE 10, not 20,
> but if Adrian insists on his scheme, I give up.
I do not insist on anything, if this new scheme works. However, there have
been problems lately, so I had to work around them and that was my
solution; it might not the best, but since Moller is more about
studying background than physics and since the background subroutine is
also in the blastmc directory, I thought this would be acceptable. And if
you do not activate the Moller generator with IKINE 20, you do not really
care, pretty much like if you do not have a negative IKINE, you do not
care about background... So there was some symmetry in my solution too.
> I thought this
> was more symmetric. In this way each user-supplied event generator
> sits and gets compiled in its own location and once when it is
> --enabled and compiled, blast is linked to the individual libuserGen.so.
This was the problem, I was having a hard time using this method. Maybe
you already checked this, anyway, when linking with a user generator
compiled in another location and simply moved to the "run" directory, I
recently got error messages like: "relocation error" etc. I know this used
to work and I do not know what the problems are now, but I would rather
have a less elegant solution that works than a very elegant one that does
not.
> I hope these changes do not interfere with those who prefer to use
> ordinary Makefiles. Please let me know if something fails to compile
> or run. Best regards,
I will surely try your method, if it works, I will be very happy with
this, if not, I will have to get back to the other...
Cheers!
Adrian
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:28 EST