Re: [PATCH 2/2] Actually use retval

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Mar 22, 2011, 18:20 UTC
Message-ID: <alpine.DEB.1.00.1103221918540.1561@bonsai2>
In-Reply-To: <7v7hbrkovz.fsf@alter.siamese.dyndns.org>

Hi,

On Tue, 22 Mar 2011, Junio C Hamano wrote:

Show 6 quoted lines

> Johannes Schindelin <johannes.schindelin@gmx.de> writes:
> 
> > This is most likely a bug. Nocited by gcc 4.6.0.
> 
> While I don't doubt gcc 4.6.0 found the retval assigned is not used, I 
> think you misunderstood the value returned from this function.

Nope. I did not misunderstand. I did not understand in the first place. That is what I described pretty explicitly in the cover letter.

Show 9 quoted lines

> The caller uses the return value to decide if an entry from t1 (and not 
> from t2) was consumed, if an entry each from both t1 and t2 were 
> consumed, or an entry from t2 (and not from t1) was consumed.  It 
> doesn't change the fact that the entry at the beginning of each tree we 
> looked at in this function at that point shared the same name and we 
> consumed them, whatever the call to diff_tree_sha1() to run a recursive 
> comparison between the trees found.
> 
> The likely fix would be to remove assignment to retval instead.

Thanks, that is the alternative I suggested after the dashdashdash.

Ciao, Johannes