regression in ns/tmp-objdir and ns/batched-fsync

From: Neeraj Singh <nksingh85@gmail.com>
Date: Oct 26, 2021, 22:21 UTC
Message-ID: <20211026222146.GA16928@neerajsi-x1.localdomain>
In-Reply-To: <xmqq5ytkzbt7.fsf@gitster.g>

On Mon, Oct 25, 2021 at 08:48:04PM -0700, Junio C Hamano wrote:

Show 11 quoted lines

> * ns/tmp-objdir (2021-10-04) 2 commits
>   (merged to 'next' on 2021-10-23 at 358d376f61)
>  + tmp-objdir: disable ref updates when replacing the primary odb
>  + tmp-objdir: new API for creating temporary writable databases
>  (this branch is used by ns/batched-fsync and ns/remerge-diff.)
> 
>  New interface into the tmp-objdir API to help in-core use of the
>  quarantine feature.
> 
>  Will merge to 'master'.
>

--snip--

Show 17 quoted lines

> * ns/batched-fsync (2021-10-08) 8 commits
>   (merged to 'next' on 2021-10-25 at e45c907d41)
>  + core.fsyncobjectfiles: performance tests for add and stash
>  + core.fsyncobjectfiles: tests for batch mode
>  + unpack-objects: use the bulk-checkin infrastructure
>  + update-index: use the bulk-checkin infrastructure
>  + core.fsyncobjectfiles: add windows support for batch mode
>  + core.fsyncobjectfiles: batched disk flushes
>  + bulk-checkin: rename 'state' variable and separate 'plugged' boolean
>  + Merge branch 'ns/tmp-objdir' into ns/batched-fsync
>  (this branch uses ns/tmp-objdir.)
> 
>  The "core.fsyncobjectfiles" configuration variable can now be set
>  to "batch" for improved performance.
> 
>  Will merge to 'master'.
>

Hi Junio, When DScho enabled batch fsync by default (for all platforms) in the git-for-windows tree, CI caught a regression. It came up through the heavy refactoring to use the tmp-objdir code.

The problem is related to setup_work_tree erasing the tmp_objdir "primary" odb.  Please expect a incremental patch from me shortly against ns/tmp-objdir. It will need to be merged into ns/batched-fsync.

Thanks, Neeraj