Dear David,

well, 1D2F does cause the issue for me, or at least it _did_ until I found the 
real issue.

In fact, I could not reproduce the problem on a second machine, so I started 
testing for problems with my user environment (by cleaning up my shell start-up 
file to the bare minimum, deleting .coot-preferences etc.), but to no avail.
What finally turned out to be responsible was a tiny modification to the 
system-wide environment: I had recently added /usr/lib64/gtk-2.0/modules to the 
ld.so.conf path in order to expose these (default) libraries to third-party 
programs that would not otherwise find them. After removing this path, geometry 
analysis in coot started working again. But it's really hard to understand how 
these libraries should be harmful to the operation of coot, and specifically 
prevent it from properly handling OXT atoms.
And all this did not affect Paul's version...

Anyway, thanks a lot for looking into this
Oliver

================================================
  PD Dr. Oliver H. Weiergräber
  Institute of Complex Systems
  ICS-6: Structural Biochemistry
  Tel.: +49 2461 61-2028
  Fax: +49 2461 61-9540
================================================



________________________________________
From: David Waterman [[email protected]]
Sent: Friday, December 09, 2016 4:38 PM
To: Weiergräber, Oliver H.
Cc: [email protected]
Subject: Re: [ccp4bb] Bug in latest COOT

Dear Oliver,

I just tried this on 1D2F using computers at Diamond (Red Hat Enterprise Linux 
Workstation release 6.8 (Santiago)) with coot in CCP4 7.0.025 and all was well.

Could you send me, or point me towards, a PDB that causes the issue - and exact 
instructions for how to trigger it?

Cheers

-- David

On 9 December 2016 at 14:59, "Weiergräber, Oliver H." 
<[email protected]<mailto:[email protected]>> wrote:
Hello Paul and ccp4 devs,

as I mentioned, the crashing coot version is the one provided with the latest 
ccp4 update(s), i.e.

---------------------------------------
0.8.7 (revision-count 6498)
[with guile 1.8.8 embedded]
[with python 2.7.10 embedded]
Binary type: Linux-x86_64-centos-5-python-gtk2
Builder_info: CCP4, Oxfordshire
---------------------------------------

Running CentOS 6.8 here, but any EL6 incarnation should behave in the same way.

This is the console output after invoking geometry analysis:
---------------------------------------
 Unknown element ::
/usr/local/ccp4-7.0/bin/coot: line 284: 17416 Segmentation fault      (core 
dumped) $coot_bin "$@"
coot-exe: "/usr/local/ccp4-7.0/libexec/coot-bin"
coot-version:
/usr/local/ccp4-7.0/libexec/coot-bin
platform:
/bin/uname
core: #f
No core file found.  No debugging
---------------------------------------

Line 284 is just the final fi statement, without obvious (to me at least) 
relation to geometry analysis ...

Hope this helps
Oliver

================================================
  PD Dr. Oliver H. Weiergräber
  Institute of Complex Systems
  ICS-6: Structural Biochemistry
  Tel.: +49 2461 61-2028<tel:%2B49%202461%2061-2028>
  Fax: +49 2461 61-9540<tel:%2B49%202461%2061-9540>
================================================



________________________________________
From: CCP4 bulletin board [[email protected]<mailto:[email protected]>] 
on behalf of Paul Emsley 
[[email protected]<mailto:[email protected]>]
Sent: Friday, December 09, 2016 3:31 PM
To: [email protected]<mailto:[email protected]>
Subject: Re: [ccp4bb] Bug in latest COOT

On 09/12/16 14:13, "Weiergräber, Oliver H." wrote:
> Hello,
>
> after installing the latest ccp4 updates today, I found that coot will 
> consistently crash (core dump) if geometry analysis is run on _any_ structure 
> containing an OXT atom.
>
>

Hello Oliver,

When reporting a bug you should try to include as much information as
possible so that the developer can reproduce it.

Ideally include a backtrace, but at the minimum you should include the
operating system and version of coot that you are using:

$ coot --version-full

Having said that, I don't really follow what CCP4 do to Coot before they
distribute the binaries.  Official coot binaries are tested before
publication - so I'd recommend you use those.

It goes without saying that I can't reproduce the problem that you
mentioned - and even CCP4's build of Coot 0.8.7 (on Macintosh 10.10)
seems not to have this problem.

http://www.chiark.greenend.org.uk/~sgtatham/bugs.html

Paul.


------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher
Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender),
Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------

Reply via email to