[BUG] - CREST exits with 0 code when optimization fails

[coltonbh]

Describe the bug
When the optimization fails to converge, CREST still exits with process code 0. This makes it hard to manage the process using scripts/other packages (e.g., python wrappers) since the process exit code is 0 (success) even when a calculation fails. I'd expect failed calculations to exit with a non-0 code.

To Reproduce

input.toml:

threads = 16
input = "structure.xyz"
runtype = "optimize"

[calculation]
level = [
    { alpb = "ch2cl2", method = "gfn2", charge = 0, uhf = 0 },
]

structure.xyz:

57
qcio_charge=0 qcio_multiplicity=1 qcio__identifiers_smiles=[Na+].[S-]/C(=N\c1cc(C(F)(F)F)cc(C(F)(F)F)c1)NC(C)C.CO.C[C@@H]1OC(=O)[C@H](C)OC1=O qcio__identifiers_canonical_smiles=C[C@@H]1OC(=O)[C@@H](OC1=O)C.CC(N/C(=N/c1cc(cc(c1)C(F)(F)F)C(F)(F)F)/[S-])C.CO.[Na+] qcio__identifiers_canonical_smiles_program=openbabel
Na     1.51736   -0.16911    2.65810
S      3.52532    0.92320    1.78029
C      4.07773    0.14584    0.33944
N      3.91815   -1.14422    0.21587
C      3.24544   -2.05996    1.06409
C      3.62308   -2.41609    2.37298
C      2.92629   -3.41746    3.10053
C      3.41955   -3.92439    4.44309
F      4.72702   -4.29030    4.42197
F      3.28317   -3.00620    5.43653
F      2.73666   -5.02789    4.87131
C      1.79207   -4.01406    2.52780
C      1.41823   -3.70956    1.21683
C      0.22723   -4.39297    0.59032
F      0.45586   -5.67914    0.23386
F     -0.84396   -4.40881    1.43493
F     -0.22038   -3.74272   -0.51999
C      2.16916   -2.75870    0.49929
N      4.67774    0.91581   -0.64964
C      3.98995    1.82517   -1.57192
C      4.43717    3.29283   -1.54276
C      3.95847    1.39803   -3.04027
C     -1.68650   -2.50932   -5.10934
O     -0.85052   -3.49558   -4.53205
C      7.43570   -5.22583   -7.89088
C      6.77752   -3.91669   -8.36706
O      7.72786   -2.86142   -8.40719
C      7.93435   -2.25242   -7.21042
O      9.01543   -2.04942   -6.72126
C      6.68452   -1.70860   -6.61192
C      7.08475   -1.13259   -5.24694
O      5.65878   -2.69177   -6.53143
C      5.50511   -3.52543   -7.60577
O      4.41540   -4.02771   -7.85468
H      4.49147   -1.93138    2.82144
H      1.22067   -4.75851    3.08935
H      1.92461   -2.55111   -0.54210
H      5.66587    0.77873   -0.79112
H      2.96855    1.76412   -1.26320
H      5.40860    3.45324   -2.01181
H      3.71197    3.90324   -2.11041
H      4.49493    3.71479   -0.53708
H      3.62021    0.36872   -3.17243
H      3.24621    2.03815   -3.58813
H      4.92911    1.52596   -3.52727
H     -2.62165   -2.38496   -4.56660
H     -1.90042   -2.77907   -6.14854
H     -1.11480   -1.58822   -5.05378
H      0.37707   -3.25647   -4.71933
H      6.79242   -6.10022   -8.02186
H      7.72632   -5.17523   -6.83850
H      8.34950   -5.45223   -8.44963
H      6.47751   -4.08937   -9.40601
H      6.34426   -0.88796   -7.25137
H      7.89090   -0.39105   -5.36734
H      6.25399   -0.63963   -4.74331
H      7.47456   -1.91234   -4.57971

stdout (Note that the calculation fails yet the program still "terminated normally":

┌────────────────────────────────────────────────────────────────────────────┐
│                                 CYCLE  219                                 │
└────────────────────────────────────────────────────────────────────────────┘
 * total energy  :  -110.6353976 Eh     change ΔE    -0.2350447E-03 Eh
   gradient norm :     0.0055371 Eh/α   predicted    -0.1975044E-03 ( -15.97%)
   displ. norm   :     0.6142322 α      lambda       -0.3150866E-03
   maximum displ.:     0.3844039 α      in ANC's #3, #7, #6, ...
   converged δE/grad : False / False
 * RMSD in coord.:     1.9023198 α      energy gain  -0.1566916E-01 Eh

   *** FAILED TO CONVERGE GEOMETRY OPTIMIZATION IN 219 ITERATIONS ***

 geometry optimization FAILED!

==========================================
 TOTAL ENERGY        -110.6353976330 Eh
 GRADIENT NORM          0.0055380583 Eh/a0
==========================================
 WARNING: geometry optimization FAILED!

 -----------------
 Wall Time Summary
 -----------------
 CREST runtime (total)               0 d,  0 h,  8 min, 28.221 sec
 ------------------------------------------------------------------
 Geometry optimization      ...        8 min, 28.135 sec ( 99.983%)
 I/O and setup              ...        0 min,  0.086 sec (  0.017%)
 ------------------------------------------------------------------
 * wall-time:     0 d,  0 h,  8 min, 28.221 sec
 *  cpu-time:     0 d,  1 h, 28 min, 18.349 sec
 * ratio c/w:    10.425 speedup
 ------------------------------------------------------------------
 * Total number of energy+grad calls: 220

 CREST terminated normally.

Expected behavior
I'd expect failed calculations to exit with a non-0 process code.

Thanks so much!

[pprcht]

ok, will do.