Has anyone tried building the AVL program from...
http://web.mit.edu/drela/Public/web/avl/
with gfortran? The fink packager of this program is
noticing regressions compared to using g95. On my
ppc Darwin8 box, I noticed that the session1.txt
tests seem to go astray. In particular, for the
section...
C1 enter horizontal-flight trim setup menu
G 9.8 set gravitational acceleration
D 1.2 set air density
M 100.0 set mass
B 30.0 set 30 deg bank angle
<Return> return to OPER menu
X execute calculation
T show Trefftz Plane plot
The calculation shows nothing but NaN results. This
is with all of the gfortran code compiled at -O3
and -fdefault-real-8 while the c code is compiled
at -O3. It might be worth looking at this for
some more corner cases where gfortran is failing.