Meissner examples - the last one to fall

[Dominique Dhumieres]

> This, I believe, is a regression for which I do not think that I am
> responsible.  Are you in a position to track down an approximate
> date on which it occurred? It is very recent.

I have included the Meissner's examples in my tests only very recently
(Sep 28), this is why I was not in a position to conclude that it was
a regression. In addition I compiled the code with g95, found out that
I did not have the right data file, so I did not try gfortran (too bad!).

Although technically possible (based on weekly snapshots), I am not
very fond to do the tracing back since I am not keeping the previous
builds (I am using a hack of the Fink script to do so and it takes
~ 4 hours to do the build, though I certainly can trim it down to less
than one hour by removing the java builds). What I can propose, if you
don't find a better alternative, is to revert to the 20050924 snapshot
next weekend.

Dominique