[Debichem-devel] Bug#848781: Bug#848781: tiledarray: FTBFS: E: Build killed with signal TERM after 150 minutes of inactivity

[Michael Banck]

Hi,

On Mon, Dec 19, 2016 at 10:26:11PM +0100, Lucas Nussbaum wrote:
> > unknown location(0): fatal error: in "distributed_storage_suite/array_operator": memory access violation at address: 0x7f53709e95d0: no mapping at fault address
> > /<<PKGBUILDDIR>>/tests/distributed_storage.cpp(102): last checkpoint: "array_operator" entry.
> > unknown location(0): fatal error: in "distributed_storage_suite/array_operator": memory access violation at address: 0x7fc69c66f5d0: no mapping at fault address
> > /<<PKGBUILDDIR>>/tests/distributed_storage.cpp(102): last checkpoint: "array_operator" entry.
> > E: Build killed with signal TERM after 150 minutes of inactivity

Upstream is now on this and could reproduce the issue, see
https://github.com/ValeevGroup/tiledarray/issues/100 

I've had a look at the new upstream version (0.6.0), and I've also
switch libmadness-dev and (locally) tiledarray to MPICH, but that did
only help in so far as the build is no longer hung after the fatal
error, but just errors out, proceeding to the binary-arch stage and
succeeding eventually (as we ignore test suite failures).

So it would no longer FTBFS, but there's still something fishy going on
here.


Michael