Race condition between repack and loose-objects maintenance task

[Gregory Szorc <gregory.szorc@gmail.com>]

`git repack -A` (such as invoked by `git gc`) loosens unreachable
objects from packfiles.

`git maintenance`'s `loose-objects` task consolidates loose objects
into a packfile.

As neither process takes a lock or is otherwise aware of the existence
of the other, there is a race condition:

1. `git repack -A` creates loose objects
2. `git maintenance`'s `loose-objects` task deletes those loose objects
3. `git repack -A` fails to find the loose objects it just created and
aborts with `fatal: unable to add recent objects`

I see this failure with regularity in a custom environment where high
mutation volume repositories invoke `git maintenance run` with
regularity (effectively in reaction to repo mutations). We have a Git
housekeeping strategy that frequently invokes `git maintenance` with
all but the `gc` task. On a ~daily frequency, we invoke `git gc` to
purge accumulated garbage. This `git gc` frequently fails due to the
aforementioned race condition since the long wall time of `git gc`
practically guarantees a `git maintenance` would have been triggered
since the repo is mutated with such high frequency.

In my scenario, I believe we have a workaround by omitting
`--task=loose-objects` from the frequent `git maintenance` invocation
if there exists a `gc.pid` file. However, this is only a partial
solution, as `git repack` and the `loose-objects` task aren't aware of
each other. (I suppose running the `gc` maintenance task and relying
on the maintenance lock file could also work. However, we're invoking
`git gc` explicitly since `git maintenance` doesn't allow passing
custom arguments through to `git gc`.)

Gregory