Hi Junio,

On Sun, 24 Nov 2019, Junio C Hamano wrote:

"Nika Layzell via GitGitGadget" [off-list ref] writes:

quoted

The refresh command is called from git-add--interactive to ensure that git's
view of the worktree is up-to-date. This is necessary for most commands
which use git-add--interactive, as they are interacting with the worktree,
however it's not necessary for git-reset, which exclusively operates on the
index. This patch skips the refresh call when performing a git-reset -p,
which can improve performance on large repositories which have out-of-date,
or no, caches of the current worktree state.

I chose to use the existing FILTER property of the flavour to make this
decision. A separate REFRESH property could be added instead.

Hmph, I wonder why this was sent my way.  How does GGG determine
whom to send patches to?  I, like other reviewers, prefer not to see
earlier rounds of patches sent directly to me unless they are about
areas that I am mostly responsible for (other patches I'll see them
and review them on the copies sent to the mailing list anyway).

CC'ing Dscho who has stakes in keeping the scripted "add -i" frozen,
while he is rewriting it in C.

Thanks for pinging me!

As it were, I was aware of this effort, and I am quite happy about it (and
of course I will adjust my patch series accordingly).

Thank you,
Dscho

Thanks.

quoted

Nika Layzell (1):
  add--interactive: skip index refresh in reset patch mode

 git-add--interactive.perl | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)


base-commit: 5fa0f5238b0cd46cfe7f6fa76c3f526ea98148d9
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-475%2Fmystor%2Findex-only-refresh-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-475/mystor/index-only-refresh-v1
Pull-Request: https://github.com/gitgitgadget/git/pull/475