On Tue, Feb 26 2019, Thomas Gummerer wrote:

As I was advocating for this series to go into 'next' without a large
refactor of this series, I'll put my money were my mouth is and try to
make the cleanups and fixes required, though without trying to avoid
further external process calls, or changing the series around too
much.

One thing to consider here is that we have a GSoC project planned
based on 'git stash'.  If we can't get this to 'next' soon, I'd vote
for taking that project out of this years GSoC, and maybe try again
next year, if nobody implemented the feature in the meantime.

FWIW I'd like to +1 getting it into "next" so this can be given thorough
testing in the 2.22

If there's still bugs or other regressions I think it's better sorted
out without the cognitive load of reviewing it all again.

Worst case we can always add something on top to flip the default of
stash.useBuiltin.