Seg fault in resonance mapping

[Juergen Reuter]

The following SINDARIN script leads to a seg fault when run with -r:

model = SM
process e1 = e1, E1 => e2, E2
beams = e1, E1 => isr
sqrts = 500 GeV
integrate (e1)
integrate (e1)

The second integration of the same process leads to

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x0000000000000000
0x00000001001a2e70 in __sf_mappings_MOD_sf_channel_set_res_mapping (channel=<value temporarily unavailable, due to optimizations>, i_sf=<value temporarily unavailable, due to optimizations>, m=0.18237639999999999, w=0.0048859999999999997) at sf_mappings.f90:817
817	    channel%map_code(i_sf) = SFMAP_MULTI_RES
(gdb) bt
#0  0x00000001001a2e70 in __sf_mappings_MOD_sf_channel_set_res_mapping (channel=<value temporarily unavailable, due to optimizations>, i_sf=<value temporarily unavailable, due to optimizations>, m=0.18237639999999999, w=0.0048859999999999997) at sf_mappings.f90:817
#1  0x000000010057ad2a in __dispatch_MOD_dispatch_sf_channels (sf_channel=<incomplete type>, sf_string=<value temporarily unavailable, due to optimizations>, coll=Invalid F77 type code 3 in symbol table.
) at dispatch.f90:840

[kilian]

Depends on the ISR part of #500.

[kilian]

Was an artefact of the work-in-progress state of this setup.

Works with r5470.

Note: this process requires either a cut on sqrts_hat, or a large number of iterations in the first pass.