[Gluster-devel] Reduce regression runs wait time - New gerrit/review work flow

[Jeff Darcy]

> I've discussed this idea with different people before. But the major concern
> was how do we identify that minimal set of basic tests that would do a good
> job of identifying most regressions. Considering that regression suite will
> keep growing and will take longer to complete in the future, nightly full
> regression runs would be nice to have.

I generally do the following on a test machine before I submit:

    rm tests/basic/rpm.t
    ./run-tests.sh tests/basic/*.t

It's a pretty good compromise between time to run and accuracy (compared to
a full set of regression tests).  IMO the smoke tests do too little to find
anything beyond "nobody ever tried to build this in a clean tree" errors,
and I think we need a little bit more than that before we move a patch to
the next stage of the process.