Last import time is updated even if previous import failed

[jar398]

In Publishing / Repository classes, rake sync task:
On a failed import (e.g. if the URL in secrets.yml is incorrect), a new ImportRun record is created, even if no resources/terms got imported. So, because the last-import time gets updated, a subsequent retry with a repair (e.g. a corrected secrets.yml) does nothing.
I think the correct behavior would be to keep track of whether an error occurred (e.g. in the rescue around the HTTP GET) and, if so, don't update the import time. Alternatively, keep track of whether anything got updated, and don't update the import time if nothing got updated - that option may be more robust.

[jar398]

I see that this is a known issue. Here is a quote from gitter in May 2018:

Jeremy Rice @JRice May 14 16:51
Well... yesssss... bu-u-u-u-u-u-ut... I manually ran it and manually
"faked" it to roll back the time from which it syncs. 
It's a long, obnoxious story. One of those "oh, this will be a cool
feature!" features that ... wasn't cool. 
It ATTEMPTS to remember when you ran sync last, and then uses that as
the date from which to look for new terms. 
...The problem is that if the last run failed, it doesn't know that,
so it ... doesn't get all terms. 
Manually bypassing this is actually really obnoxious.
The next time it happens I will clean up the code a bit to at least
make it easier to override. :|