[0/1] ref: add object check for regular ref

Message

shejialuo Dec. 27, 2024, 1:40 p.m. UTC

Hi All:

When I implement the code for packed ref content checks, I somehow
notice that I ignore checks for the object. In the first glance, I think
I could make this patch in the first of my packed ref content check
series. However, this is not a good idea which may cause the reviewers
more overhead.

And this patch aims at checking whether the object exists and whether
the type of the object is correct.

Thanks,
Jialuo

shejialuo (1):
  ref: add object check for regular ref

 refs/files-backend.c     | 50 ++++++++++++++++++++++++++++--------
 t/t0602-reffiles-fsck.sh | 55 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 95 insertions(+), 10 deletions(-)

Comments

shejialuo Dec. 27, 2024, 1:58 p.m. UTC | #1

On Fri, Dec 27, 2024 at 09:40:47PM +0800, shejialuo wrote:
> Hi All:
> 
> When I implement the code for packed ref content checks, I somehow
> notice that I ignore checks for the object. In the first glance, I think
> I could make this patch in the first of my packed ref content check
> series. However, this is not a good idea which may cause the reviewers
> more overhead.
> 
> And this patch aims at checking whether the object exists and whether
> the type of the object is correct.
> 
> Thanks,
> Jialuo
> 
> shejialuo (1):
>   ref: add object check for regular ref
> 
>  refs/files-backend.c     | 50 ++++++++++++++++++++++++++++--------
>  t/t0602-reffiles-fsck.sh | 55 ++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 95 insertions(+), 10 deletions(-)
> 
> -- 
> 2.47.1
> 

Forget to tell Junio, this patch is based on the latest master branch.
I don't use the "sj/refs-content-check" due to there are some conflicts
when using "is_promisor_object" in the latest master branch.

shejialuo Jan. 5, 2025, 6:30 a.m. UTC | #2

On Fri, Dec 27, 2024 at 09:40:47PM +0800, shejialuo wrote:
> Hi All:
> 
> When I implement the code for packed ref content checks, I somehow
> notice that I ignore checks for the object. In the first glance, I think
> I could make this patch in the first of my packed ref content check
> series. However, this is not a good idea which may cause the reviewers
> more overhead.
> 
> And this patch aims at checking whether the object exists and whether
> the type of the object is correct.
> 
> Thanks,
> Jialuo
> 
> shejialuo (1):
>   ref: add object check for regular ref
> 
>  refs/files-backend.c     | 50 ++++++++++++++++++++++++++++--------
>  t/t0602-reffiles-fsck.sh | 55 ++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 95 insertions(+), 10 deletions(-)
> 
> -- 
> 2.47.1
> 

Please ignore this patch, I will send it together within the packed-ref
consistency implementation.

Thanks,