Re: [PATCH] mktag: don't check SHA-1 object length under SHA-256

On 2020-11-23 at 21:34:21, Jeff King wrote:

This is how fsck_tag() does it. In fact, it seems like we could just
reuse fsck_tag() here. It wants to pass an oid to report(), but it would
be OK to use null_oid here; ultimately it just ends up back in our
callback error_func().

I would very much appreciate getting rid of this custom code.  It was a
pain during the SHA-256 transition, and even with significant effort, it
appears I missed a few parts.  Also, while I'm for casual language, its
use here is going to be a little hard to understand for non-native
speakers.

Moreover, when we start mapping objects to write the SHA-1 values into
the loose object index (which is a series I'm slowly working on), most
of this custom code is going to have to disappear or be coalesced
anyway.  If we can use an existing, more robust function now, so much
the better.
-- 
brian m. carlson (he/him or they/them)
Houston, Texas, US