Hi Steve,
It seems a bit strange that it is killed in the middle of this m4 generation unless, as you suggest, your machine has killed it as it had run out of some resource. The m4 does build a large file and to see if this is a problem, try using the config option --enable-instantiations="Polyhedron@BD_Shape<double>" and see if it builds ok. If it does, but you want other instantiations available, you can add here just what you need.
If this does not work, can you give us more information? How was the m4 command below called? Can you let us have the config.log file and also the generated file interfaces/ppl_interface_instantiations.m4
Pat
On Fri, 2 Jul 2010, Steve Murphy wrote:
I'm on a linux machine, a VM running Centos 5.5, just in case it makes any difference.
/usr/bin/m4 --prefix-builtin -I.. -I. -I./.. ./ppl_interface_generator_c_cc_files.m4
dies with "Killed".
The last thing it says before it dies, is:
int ppl_BD_Shape_double_add_recycled_constraints (ppl_BD_Shape_double_t ph, ppl_Constraint_System_t cs) try { BD_Shape<double>& pph = *to_nonconst(ph); Constraint_System& ccs = *to_nonconst(cs); pph.add_recycled_constraints(ccs); return 0; } CATCH_ALL
int ppl_BD_Shape_double_add_recycled_congruences (ppl_BD_Shape_double_t ph, ppl_Congruence_System_t cs) try { BD_Shape<double>& pph = *to_nonconst(ph); Congruence_System& ccs = *to_nonconst(cs); pph.add_recycled_congruences(ccs); return 0; } CATCH_ALL
DEFINE_OUTPUT_FUNCTIONS(BD_Shape_double) Killed
(which I see when it by hand.
What gives? It takes a few seconds after generating the "DEFINE" lines, and does this several times. But with *this* define, it gets Killed. Any way to avoid this? work around it? Is it running out of memory or swap or something?
murf
-- Steve Murphy ParseTree Corp