Detecting a segfault in dmd itself

Segfault filtering

One way to do it is this, which has the advantage of providing dmd's output before segfaulting:

#!/bin/sh

# arguments to dmd
DMDARGS="-debug -g ..."

OUTPUT=$(gdb --batch -ex "run $DMDARGS" -ex 'bt 30' dmd 2>&1)
echo "$OUTPUT" | egrep -q "in dwarf_typidx .+ at backend/dwarf\.c" # adapt for your needs, use fgrep if you don't need regexps
return $?

OS X (lldb)

#!/bin/sh

# arguments to dmd
DMDARGS="-debug -g ..."

OUTPUT=$(lldb --batch -o "run $DMDARGS" -k 'bt 30' -k 'quit' -o 'quit' dmd 2>&1)
echo "$OUTPUT" | egrep -q "in dwarf_typidx .+ at backend/dwarf\.c" # adapt for your needs, use fgrep if you don't need regexps
exit $?

Detect just any segfault

A simpler, but less detailed way, is to do this:

#!/bin/bash
dmd ...args...
# Segfaulting programs will likely return an error code of 139.
if [ $? -eq 139 ]; then
        exit 0
else 
        exit 1
fi