Request Hets's version outside of mutex-block

[eugenk]

When I call HetsInstance.choose!
And the instance of Hets is not reachable (a timeout occurs)
And there are other processes that call HetsInstance.choose!
Then the other instances must wait for the timeout to expire as well.

This is because choose! requests the version of Hets in a mutex-block.

Change the choosing mechanics in a way that the timeout does not make other processes wait.

[tillmo]

is this a possible explanation for ontologies being indefinitely pending on develop.ontohub.org?

[eugenk]

I'm not 100% (but some 80%) sure that it is.