Re: [PATCH v2 0/5] packfile: use oidset for bad objects

From: Derrick Stolee <stolee@gmail.com>
Date: Sep 11, 2021, 23:59 UTC
Message-ID: <9cdf54db-132e-d771-e177-6673f7ad02bf@gmail.com>
In-Reply-To: <e50c1465-59de-7fe1-de01-800404c7640e@web.de>

On 9/11/21 4:31 PM, René Scharfe wrote:

Show 13 quoted lines

> Replace the custom hash array for remembering corrupt pack entries with
> an oidset.  This shortens and simplifies the code.
> 
> Changes since v1:
> - inline oidset_size()
> - inline nth_midxed_pack_entry() early
> - use oidset_size() to avoid a function call if no bad objects exist
> 
>   oidset: make oidset_size() an inline function
>   midx: inline nth_midxed_pack_entry()
>   packfile: convert mark_bad_packed_object() to object_id
>   packfile: convert has_packed_and_bad() to object_id
>   packfile: use oidset for bad objects

These were easy reads, and I understand the value of them.

I initially hesitated to support the drop of nth_midxed_pack_entry(), since it was designed with things like midx bitmaps in mind (specifically, to also support lex-order-to-stable-order conversions). However, it seems that the midx bitmap series by Taylor is succeeding without needing such a translation.

Thanks, -Stolee