Feedback-ID: if26b431b:Fastmail
From: Junio C Hamano <gitster@pobox.com>
To: "D. Ben Knoble" <ben.knoble@gmail.com>
Cc: git@vger.kernel.org,  Todd Zullinger <tmz@pobox.com>,  Tian Yuchen
 <cat@malon.dev>,  Patrick Steinhardt <ps@pks.im>,  Olamide Caleb Bello
 <belkid98@gmail.com>
Subject: Re: [PATCH v3 3/3] core: convert build-time USE_NSEC into runtime
 core.useNanosec
In-Reply-To: <48fceb4b575ca39346cf2f59f621584a19049008.1787065125.git.ben.knoble@gmail.com>
	(D. Ben Knoble's message of "Tue, 18 Aug 2026 10:59:47 -0400")
References: <cover.1786103607.git.ben.knoble@gmail.com>
	<cover.1787065125.git.ben.knoble@gmail.com>
	<48fceb4b575ca39346cf2f59f621584a19049008.1787065125.git.ben.knoble@gmail.com>
Date: Tue, 18 Aug 2026 11:51:46 -0700
Message-ID: <xmqqh5krxnwd.fsf@gitster.g>
User-Agent: Gnus/5.13 (Gnus v5.13)
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain

"D. Ben Knoble" <ben.knoble@gmail.com> writes:

> diff --git a/environment.c b/environment.c
> index 6676e6f5ae..c7f6b801f4 100644
> --- a/environment.c
> +++ b/environment.c
> @@ -571,6 +571,13 @@ int git_default_core_config(const char *var, const char *value,
>  		return 0;
>  	}
>  
> +#ifndef NO_NSEC
> +	if (!strcmp(var, "core.usenanosec")) {
> +		cfg->use_nanosec = git_config_bool(var, value);
> +		return 0;
> +	}
> +#endif

What this hunk tells us: At build time, you could choose to ignore
core.usenanosec configuration variable, preventing cfg->use_nanosec
from getting flipped to true by the configured value.

> @@ -769,6 +776,9 @@ void repo_config_values_init(struct repo_config_values *cfg)
>  	cfg->ignore_case = 0;
>  	cfg->trust_executable_bit = 1;
>  	cfg->has_symlinks = platform_has_symlinks();
> +#ifndef NO_NSEC
> +	cfg->use_nanosec = 0;
> +#endif

I think we want to unconditionally initialize it to 0, unless the
definition of the .use_nanosec member itself in the structure is
conditional on NO_NSEC.  And ...

>  
>  	/* section "sparse" config values */
>  	cfg->sparse_expect_files_outside_of_patterns = 0;
> diff --git a/environment.h b/environment.h
> index e7ec5b0437..a35534afe5 100644
> --- a/environment.h
> +++ b/environment.h
> @@ -139,6 +139,7 @@ struct repo_config_values {
>  	int ignore_case;
>  	int trust_executable_bit;
>  	int has_symlinks;
> +	int use_nanosec;

... that is not the case.  

Which means that git_default_core_config() does keep the initial
value of the member without getting affected by the configuration,
but it does not necessarily be keeping "false".  It may be keeping
the uninitialized state instead ;-).

> diff --git a/read-cache.c b/read-cache.c
> index 6c449f393d..31888f77ee 100644
> --- a/read-cache.c
> +++ b/read-cache.c
> @@ -353,12 +353,18 @@ static int ce_match_stat_basic(const struct cache_entry *ce, struct stat *st)
>  static int is_racy_stat(const struct index_state *istate,
>  			const struct stat_data *sd)
>  {
> +#ifndef NO_NSEC
> +	int use_nsec = repo_config_values(istate->repo)->use_nanosec;
> +#endif
> +
>  	return (istate->timestamp.sec &&
> -#ifdef USE_NSEC
> -		 /* nanosecond timestamped files can also be racy! */
> -		(istate->timestamp.sec < sd->sd_mtime.sec ||
> -		 (istate->timestamp.sec == sd->sd_mtime.sec &&
> -		  istate->timestamp.nsec <= sd->sd_mtime.nsec))
> +#ifndef NO_NSEC
> +		/* nanosecond timestamped files can also be racy! */
> +		use_nsec
> +		? (istate->timestamp.sec < sd->sd_mtime.sec ||
> +		   (istate->timestamp.sec == sd->sd_mtime.sec &&
> +		    istate->timestamp.nsec <= sd->sd_mtime.nsec))
> +		: istate->timestamp.sec <= sd->sd_mtime.sec
>  #else
>  		istate->timestamp.sec <= sd->sd_mtime.sec
>  #endif

Ugly.  How about getting rid of the latter #ifndef/#else/#endif and
instead keeping the "if use_nsec, pay attention to nsec, otherwise
only the seconds part" ternary?  As to the early part, as you can
arrange cfg's '.use_nanosec' to always hold a sensible value, the
function can become

        return (istate->timestamp.sec &&
                (repo_config_values(istate->repo)->use_nanosec
                 ? (istate->timestamp.sec < sd->sd_mtime.sec ||
                   (istate->timestamp.sec == sd->sd_mtime.sec &&
                    istate->timestamp.nsec <= sd->sd_mtime.nsec))
                 : istate->timestamp.sec <= sd->sd_mtime.sec));

I think.

The code you presented here for is_racy_stat() sprinkled with
#ifndef/#else/#endif would be sensible if repo_config_values struct
defined the '.use_nanosec' member conditionally.  But that is not
what is happening here.