Re: [PATCH 1/6] refs/files: remove duplicate check in `split_symref_update()`

From: Patrick Steinhardt <ps@pks.im>
Date: Feb 7, 2025, 16:12 UTC
Message-ID: <Z6Yw9RK6JRKERmn9@pks.im>
In-Reply-To: <20250207-245-partially-atomic-ref-updates-v1-1-e6a3690ff23a@gmail.com>

On Fri, Feb 07, 2025 at 08:34:36AM +0100, Karthik Nayak wrote:

> In split_symref_update(), there were two redundant checks:
>    - At the start: checking if refname exists in `affected_refnames`.
>    - After adding refname: checking if the item added to
>      `affected_refnames` contains the util field.

Okay, it took me a bit longer to understand what's going on here. What you're saying is that we already use `string_list_has_string()` at the start of `split_symref_update()`, and if that returns true then we would bail out. Consequently, it is impossible for `string_list_insert()` to find a preexisting values.

Makes sense, but I think that could be explained a bit better.

Show 16 quoted lines

> diff --git a/refs/files-backend.c b/refs/files-backend.c
> index 29f08dced40418eb815072c6335e0c3d1a45c7d8..c6a3f6d6261a894e1c294bb1329fdf8079a39eb4 100644
> --- a/refs/files-backend.c
> +++ b/refs/files-backend.c
> @@ -2846,13 +2838,7 @@ static int files_transaction_prepare(struct ref_store *ref_store,
>  		if (update->flags & REF_LOG_ONLY)
>  			continue;
>  
> -		item = string_list_append(&affected_refnames, update->refname);
> -		/*
> -		 * We store a pointer to update in item->util, but at
> -		 * the moment we never use the value of this field
> -		 * except to check whether it is non-NULL.
> -		 */
> -		item->util = update;
> +		string_list_append(&affected_refnames, update->refname);

Nice to see this and other code removed.

Patrick