At least one person on the thread (@anthony) raised concerns but has not
replied since.
Also since this thread was started, the bug that miscounted files has been
fixed, which is sufficient that stress tests would have been run for about
95% of PRs that previously (erroneously) were given a free p
> On Mar 17, 2020, at 9:03 AM, John Blum wrote:
>
> Additionally, it'd be ideal if the deprecated method were then adapted to
> delegate to the new approach. This will cut down on the number of required
> tests since then you only need a Unit Tests asserting the method performs
> the translat
Additionally, it'd be ideal if the deprecated method were then adapted to
delegate to the new approach. This will cut down on the number of required
tests since then you only need a Unit Tests asserting the method performs
the translation/delegating appropriately, unless of course the behavior is
I think we are also missing the other side of the coin.
Once we deprecate something and we now need a equivalent test that tests
the same behavior using the new method/approach. i.e now we have to
double up on the testing of said deprecated method/feature/class. First
we have to keep the tests
Seems like I'm the only one on this thread who even quibbled about removing
the limit entirely. Let's go ahead and remove the limit.
-Dan
On Fri, Mar 13, 2020 at 8:26 AM Robert Houghton
wrote:
> I want the check to stay required for PR merges to be allowed. I also want
> it to do real work in a
