Timeout in instance stop

[pierlauro]

Added ability to specify stop timeout (default 10 - as for singularity instance stop specification)

[vsoch]

okay, looks like this goes back to at least 3.1 (and there are no docs for 3.0) so the check for version seems sufficient to honor the variable. Should we perhaps instead of doing a default, only add the variable given the case that the user has defined it and we have Singularity version 3? This would be akin to honoring the Singularity default (and how it might change over time) instead of hard coding it here. Other than that small change, if you can confirm this works for you, I'm okay with not adding a test for now because it seems like it would be challenging to easily do (unless you have an idea?)

[pierlauro]

okay, looks like this goes back to at least 3.1 (and there are no docs for 3.0) so the check for version seems sufficient to honor the variable. Should we perhaps instead of doing a default, only add the variable given the case that the user has defined it and we have Singularity version 3? This would be akin to honoring the Singularity default (and how it might change over time) instead of hard coding it here. Other than that small change, if you can confirm this works for you, I'm okay with not adding a test for now because it seems like it would be challenging to easily do (unless you have an idea?)

I totally agree on the proposed change, this way (1) older singularity versions would not be affected and (2) there will be no problem if singularity changes the default timeout.

[pierlauro]

Regarding the testing, as we're just making a call to singularity stop (that is already working) with an additional argument, I think it is not worth instrumenting some complex to setup time-based test

[vsoch]

Agree! And your changes look good.

[vsoch]

https://pypi.org/project/spython/0.0.77/