From: Andreas Ericsson <ae@op5.se>
Date: Thu, 08 Nov 2007 08:41:55 GMT
Subject: Re: [PATCH MISC 1/1] Make gcc warning about empty if body go away.
Message-ID: <4732CBD3.503@op5.se>
In-Reply-To: <7vode57awg.fsf@gitster.siamese.dyndns.org>

Junio C Hamano wrote:
> Pierre Habouzit <madcoder@debian.org> writes:
> 
>> diff --git a/builtin-diff.c b/builtin-diff.c
>> index f77352b..80392a8 100644
>> --- a/builtin-diff.c
>> +++ b/builtin-diff.c
>> @@ -204,7 +204,7 @@ static void refresh_index_quietly(void)
>>  		if (write_cache(fd, active_cache, active_nr) ||
>>  		    close(fd) ||
>>  		    commit_locked_index(lock_file))
>> -			; /*
>> +			(void)0; /*
>>  			   * silently ignore it -- we haven't mucked
>>  			   * with the real index.
>>  			   */
> 
> Wouldn't this be much easier to read, by the way?
> 
> The point is that if we touched the active_cache, we try to
> write it out and make it the index file for later users to use
> by calling "commit", but we do not really care about the failure
> from this sequence because it is done purely as an optimization.
> 
> The original code called three functions primarily for their
> side effects, which is admittedly a bad style.
> 
>  builtin-diff.c |   12 +++---------
>  1 files changed, 3 insertions(+), 9 deletions(-)
> 
> diff --git a/builtin-diff.c b/builtin-diff.c
> index f77352b..906c924 100644
> --- a/builtin-diff.c
> +++ b/builtin-diff.c
> @@ -200,15 +200,9 @@ static void refresh_index_quietly(void)
>  	discard_cache();
>  	read_cache();
>  	refresh_cache(REFRESH_QUIET|REFRESH_UNMERGED);
> -	if (active_cache_changed) {
> -		if (write_cache(fd, active_cache, active_nr) ||
> -		    close(fd) ||
> -		    commit_locked_index(lock_file))
> -			; /*
> -			   * silently ignore it -- we haven't mucked
> -			   * with the real index.
> -			   */
> -	}
> +	if (active_cache_changed &&
> +	    !write_cache(fd, active_cache, active_nr) && !close(fd))
> +		commit_locked_index(lock_file);
>  	rollback_lock_file(lock_file);
>  }
>  

Ack, obviously, as it no longer requires a comment to explain it, although
I'd prefer an empty line after commit_locked_index(lock_file); so as to not
confuse the rollback_lock_file() statement as being part of the conditional
path.

First I thought the rollback_lock_file() was the *only* statement to the
condition, and everyone who uses 4 for tabsize) will have double trouble
since commit_locked_index(lock_file) aligns with the second line of the
condition.

-- 
Andreas Ericsson                   andreas.ericsson@op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231