disappearing atom mappings in MarvinSketch 5.3.8

Something we've run across and can't reproduce consistently is that when we provide students with a template molecule that includes atom mapping numbers and they copy the structure sometimes (but not always) the atom mapping numbers disappear from the copied structure.

For instance, when teaching students about resonance structures we'll give them a scaffold/template structure (see the attached template.mrv) to work with so they know how the atoms are connected, but it is up to them to decide the bond order and placement of lone pairs (see the attached solution.mrv).

Without atom mapping the structures are indistinguishable and therefore not gradeable, so we need the atom mapping numbers to persist. Up until 5.2.4 this was the case and worked well, but since we upgraded to 5.3.8 this error occurs once in a while. It seems to happen more often using the ctrl-C/ctrl-V (1 in 5 maybe) method than using the clipboard icon or the ctrl-drag method for duplicating structures.

Is there any reason why it would work as expected some times, but fail other times?

 

Thanks,

   Michiel

Hi Michiel,

truly sorry about the late answer, the new release causes some pressure in the last few days, and to be honest this question was a hard one.

First of all i could only reproduce it only when i run a program that fully utilize my hardware capabilities besides MarvinSketch, and I could reproduce it very rare. I have only a windows system, but i suspect that it is happening only on slow or on heavily loaded windows systems. In this case the java part is responsible for the image formats, and for the skc and cdx format, and OLE is responsible for the mrv, mol, ole data, and textual representations. The java part runs first, and the OLE part runs second. On a loaded system the OLE part can be slow a bit, and if the paste happens before OLE finished its work, then marvin will paste from CDX format and in that format this mapping is not supported.

There is no immidate fix which can be applied, all you can do for now is to try pasting a bit later again.

We will discuss this issue, and i think we will move back the mrv clipboard export into the java part of the running code. That can solve this issue. This fix can be the part of the 5.4.1 since the 5.4.0 release has completed all test and going to be released very soon.

I hope 5.4.1 will be suitable for you. Sorry for the inconvenience and sorry again for the late answer.

Regards,
Istvan

Hi,

i would like to inform you, that the possible solution for the problem you mentioned, is released in 5.4.1, could you please confirm that the bug is also disappeared on the mashines produces the issue with older versions? As i mentioned, i could reproduce it only under extreme circumstaces, and after i modified the code, i could not reproduce it at all. But i can not be absolutely sure, so any feedback is appreciated.

Regards,
Istvan

ifajth wrote:

Hi, i would like to inform you, that the possible solution for the problem you mentioned, is released in 5.4.1, could you please confirm that the bug is also disappeared on the mashines produces the issue with older versions? As i mentioned, i could reproduce it only under extreme circumstaces, and after i modified the code, i could not reproduce it at all. But i can not be absolutely sure, so any feedback is appreciated. Regards, Istvan

 

 

Sorry, but due to the isotopes.data issue (gzipped version with magic and checksum problems) introduced in 5.4 that affects Linux servers (discussed elsewhere) I am unable to run 5.4.1 at the moment.

   Michiel

Hi Michiel,

that issue has also been solved in 5.4.1.

Regards,
Istvan