steinbrt.rci.rutgers.edu wrote:
> Hi,
>
> in my ongoing quest of failing to compile a fresh Amber checkout every
> single time I try, I came across this problem:
>
Hi Thomas,
As a side note and an tool in debugging, it's useful to look at:
http://git.ambermd.org:8080/dashboard/tab/build/detail/amber
and see if the robot build of the git amber tree is building as well.
If not, you can got back in time with ref to the checkins and narrow
down the issue. For example here, build 351 (numbering is relative to
cruisecontrol and has no meaning outside of it) was the last good build.
You can then go to the next build (i.e. the first failing one) and have
a look at the "Errors and warnings" tabs. Here towards the bottom, we see:
Topology.c:1728: warning: ‘nh2’ may be used uninitialized in this function
Trajectory.c:19:25: error: AmberNetcdf.h: No such file or directory
make[1]: *** [Trajectory.o] Error 1
make: *** [serial] Error 2
which is what you are seeing as well.
The real use of this now comes into play, you can look at the
modifications tab to give you a clue as to what changed to cause the
breakage. This tool is very useful for ruling out a local issue specific
to your compile setup, versus a real issue, that is a result of a tree
breakage.
Regards,
Mark
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Thu Feb 24 2011 - 09:00:06 PST
