[Repronim-trd3] rapid development: can't make today's meetup :-/

[Kyle Meyer]

Yaroslav O Halchenko <yoh at onerussian.com> writes:

> In my case I did try another "reproman run" but ran into another issue
> I am yet to troubleshoot and possibly report.  A glimpse of it is
> below if Kyle has an idea right away (e.g. me again forgetting about
> some known issue again):

[...]

> 2019-08-21 15:27:28,017 [ERROR ] 'datalad publish' failed. Try running
> 'datalad update -s smaug --merge --recursive' first
> [orchestrators.py:prepare_remote:792] (OrchestratorError)

Assuming that is being executed after another run, that should be fixed
by 2cbeda75f (ENH: datalad-pair-run orchestrator: Reset to original
commit, 2019-08-20), which is currently in kyleam/run-more-ssh-and-dirt
(gh-452).

A large part of that PR is switching the tests over to SSH (localhost)
rather than using a localshell resource so that we can test the
create-sibling/publish more thoroughly.  Doing so caught the case above
and should make it easier to test more complicated scenarios.