While running crest I got the following error (I compiled the current version with gcc):
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x14ae0605c70f in ???
#1 0x555e5a46a331 in write_cts_nci_
at crest/src/constraining.f90:526
#2 0x555e5a453ee9 in setmdrun2_
at crest/src/confscript2_misc.f90:777
#3 0x555e5a5e86a1 in trialmd_
at crest/src/trialmd.f90:120
#4 0x555e5a44c1b0 in confscript2i_
at crest/src/confscript2_main.f90:72
#5 0x555e5a6151c5 in crest
at crest/src/crest_main.f90:249
#6 0x555e5a615876 in main
at crest/src/crest_main.f90:25
Since I did not specify NCI mode, it was quite strange that it was in a function write_cts_NCI(ich,cts) in a place where it could never get if cts%NCI was false. So - I looked at where the variable is referenced in code and could not find any place where it is initalized to .false. which would be logical.
src/classes.f90
89: logical :: NCI
src/constraining.f90
524: if (cts%NCI) then
545: if (cts%NCI) then
src/ncigeo.f90
50: env%cts%NCI = .true.
src/qcg/solvtool.f90
964: env%cts%NCI = .true. !Activating to have wall pot. written in coord file for xtb
1123: if (env%cts%NCI) then
1287: env%cts%NCI = .false. !Dactivating the wall pot.
Threfore, I believe that line 89 should be logical :: NCI = .false. After that change it no longer crashes for me.
While running crest I got the following error (I compiled the current version with gcc):
Since I did not specify NCI mode, it was quite strange that it was in a function
write_cts_NCI(ich,cts)
in a place where it could never get ifcts%NCI
was false. So - I looked at where the variable is referenced in code and could not find any place where it is initalized to .false. which would be logical.Threfore, I believe that line 89 should be
logical :: NCI = .false.
After that change it no longer crashes for me.