On Thu, Dec 16 2021, Junio C Hamano wrote:
> Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes:
>
>> This series refactors various small bits of builtin/reflog.c
>> (e.g. using a "struct string_list" now), and finally makes it handle
>> the "--verbose" output, instead of telling the files backend to emit
>> that same verbose output.
>
> I was quite confused even after spending quite some time until I
> realized that you are primarily talking about the "-v" option of
> "git reflog expire" (and "delete", perhaps) and not about "git
> reflog -v" having different output from the same command without
> "-v" option.
Will try to clear that up in an eventual re-roll, but considering this a
"look at only if re-rolling for other reasons" for now.
> I'll refrain from commenting on, or picking up, the series until
> Han-Wen has a chance to take a look and give his input as it is
> unclear if it helps or crashes with his plan.
He hasn't reviewed this series.
But I think I/he have looked at it sufficiently that you might want to
consider picking it up, particularly since he doesn't work on Friday's &
with the holidays etc. it might be a while.
He commented on this helping with his plans recently here:
https://lore.kernel.org/git/CAFQ2z_PSS9zOzR6nGYZ8DBK+6oOQzkJsEy_7y+NprwJ1OHNs7w@xxxxxxxxxxxxxx/
The snippet he's referring to can (last seen on-list, AFAICT) be seen
here:
https://lore.kernel.org/git/3d57f7c443082fd2a7f01aee003a9cd3ca2dd910.1629207607.git.gitgitgadget@xxxxxxxxx/
Which from my reading of the code (I haven't actually tried to combine
this with reftable/* for real) would Just Work when combined with this
series.
I.e. he had "dry_run" implemented, but not "verbose". Now he'll get
"verbose" for free. The "XXX" comment there about dry_run appears to be
a TODO about not doing needless work within the reflog backend.
