eneas77 commented 2 years ago

Hi,

I´m newbie with tinker. I´m unable to run properly pdbxyz routine with the peptide.pdb example. Following message appears:

 ######################################################################

##########################################################################

Tinker --- Software Tools for Molecular Design

Version 8.10.2 February 2022

Copyright (c) Jay William Ponder 1990-2022

All Rights Reserved

########################################################################## ######################################################################

Enter Protein Data Bank File Name : peptide.pdb

Enter Parameter File Name [=NONE] : test.key corrupted size vs. prev_size

Program received signal SIGABRT: Process abort signal.

Backtrace for this error:

0 0x7f015f52a700 in ???

1 0x7f015f5298a5 in ???

2 0x7f015f20b0bf in ???

3 0x7f015f20b03b in ???

4 0x7f015f1ea858 in ???

5 0x7f015f25529d in ???

6 0x7f015f25d32b in ???

7 0x7f015f25d99a in ???

8 0x7f015f25eeba in ???

9 0x55dfeb31a4c4 in ???

10 0x55dfeb316b45 in ???

11 0x55dfeb30351e in ???

12 0x7f015f1ec0b2 in ???

13 0x55dfeb30357d in ???

14 0xffffffffffffffff in ???

test.key file is:

parameters /apps/tinker/tinker/params/oplsaal

Any hints about what is going on?

Thank you kindly, E77

FHedin commented 2 years ago

Hi all,

I observed the same problem with a call to analyze when using poltype2.

I recompiled the last commit (d461cd9c3150f782e970631fd976d8bcfc0b0abc) in debug mode using:

cmake -D CMAKE_BUILD_TYPE="Debug" ../cmake

Here is the problem:

Calling: analyze /root/poltype/VersionFiles/water.xyz -k /root/poltype/VersionFiles/water.key e> version.out
At line 263 of file /root/tinker/source/initprm.f
Fortran runtime error: Index '501' of dimension 2 of array 'stbn' above upper bound of 500

Error termination. Backtrace:
#0  0x7f959e8d8bfa
#1  0x7f959e8d9685
#2  0x7f959e8d9c76
#3  0x56261dd0fb87
#4  0x56261dd0510a
#5  0x56261dd00e41
#6  0x56261da23ad4
#7  0x56261d8a3565
#8  0x56261d8a4667
#9  0x7f959e5d6d09
#10  0x56261d884639
#11  0xffffffffffffffff

I think the problem was induced by this commit :

d461cd9c3150f782e970631fd976d8bcfc0b0abc

The change of maxnsb from 2000 to 500 seems to cause the segfault:

allocate (stbn(2,maxnsb))

The water.key and water.xyz used as input are from the poltype2 repo: https://github.com/TinkerTools/poltype2/tree/master/VersionFiles

Regards,

F Hedin (@FHedin) & D Loco (@DanieleLoco) from Qubit Pharmaceuticals

jayponder commented 2 years ago

OK, thanks, I'll take a look at it. The change from 2000 to 500 should not break anything. In fact that's kind of the point of this earlier rewrite of some of the parameter parsing mechanism- to better use dynamic allocation for parameter storage. That first number in the "max" intrinsic is just a safeguard- the code should even work correctly if the 2000 is changed to 0. So the actual problem is somewhere else with the counting of the number of parameters. This is likely easy to fix, I'll just need to track it down.