Re: [PATCH] for-each-ref: add split message parts to %(contents:*).

> On Thu, Sep 01, 2011 at 06:50:01PM +0200, Michał Górny wrote:
> 
> > -	buf = strchr(buf, '\n');
> > -	if (!buf) {
> > -		*body = "";
> > -		return; /* no body */
> > -	}
> > -	while (*buf == '\n')
> > -		buf++; /* skip blank between subject and body */
> > -	*body = buf;
> > +	buf = format_subject(NULL, buf, NULL);
> > +
> > +	/* When having a signed tag without body, format_subject()
> > +	 * will start to eat the signature. */
> > +	if (buf > *signature)
> > +		*body = *signature;
> > +	else /* - 1 to get a trailing newline to strip */
> > +		*body = buf - 1;
> 
> This last line is wrong if there is no trailing newline, no? Running
> even the existing tests in t6300 against your new patch, I get:
> 
> expecting success:
>   git for-each-ref --format='%(subject)' refs/heads/master >actual &&
>   test_cmp expected actual
> 
> --- expected    2011-09-02 16:36:38.306058729 +0000
> +++ actual      2011-09-02 16:36:38.318058729 +0000
> @@ -1 +1 @@
> -Initial
> +Initia
> not ok - 28 basic atom: head subject