Ross Walker wrote:
> As it seems that this message mysteriously vanished into the ether without a
> trace I am resending it...
>
> Hi Mark,
>
> I fixed this bug in the amber 10 cvs tree about a month ago and produced a
> patch file to update amebr 9 but then completely forgot to post it. I'll
> post the bugfix that makes it match the amber 10 tree when I get back in the
> office on Thursday. Thanks for bringing this to my attention and your
> attempts at debugging it. I was very dissapointed myself in the ifort error
> message and can't remember now how I tracked it down. Probably by using
> print statements in the code.
Hi David and Ross,
Thanks for replies and the information about g95. I had not really
understood fully the forking between g95/gfortran. As an aside/follow
up, I managed to track down the issue in the noesy test. It was problem
with a namelisted READ function with the NXP.8May95 file and a certain
formating element (?) within this file. I managed to produce a testcase
which explains it better:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30193
From the resulting discussion on this bug, it would appear it's fixed
in gfortran 4.3. I've not tested and 4.2pres, but there was an
interesting comment (
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30193#c6 ) from someone who
was testing with SUSE.
So, hopefully.. one is one step closer to having AMBER compile with
gfortran.
regards,
Mark
Received on Thu Dec 14 2006 - 04:59:59 PST
