Re: [PATCH 2/2] [GSOC] ref-filter: add %(header) atom

[Junio C Hamano <gitster@pobox.com>]

"ZheNing Hu via GitGitGadget" <gitgitgadget@gmail.com> writes:

>  		struct {
>  			enum { RAW_BARE, RAW_LENGTH } option;
>  		} raw_data;
> +		struct {
> +			enum { H_BARE, H_LENGTH } option;
> +		} header;

Raw does not use R_{BARE,LENGTH} and uses raw_data member.  Header
should follow suit unless there is a compelling reason not to, no?

		struct {
			enum { HEADER_BARE, HEADER_LENGTH } option;
		} header_data;

perhaps?

> @@ -1372,6 +1389,15 @@ static void grab_raw_data(struct atom_value *val, int deref, void *buf, unsigned
>  				    &bodypos, &bodylen, &nonsiglen,
>  				    &sigpos, &siglen);
>  
> +		if (starts_with(name, "header")) {
> +			size_t header_len = subpos - (const char *)buf - 1;

Hmph, is this correct?  I would expect that the "header" part of a
commit or a tag object excludes the blank line after the header
fields.  In other words, the "header" would be separated by a blank
line from the "body", and that separating blank line is not part of
"header" or "body".

Otherwise, if there is a user of %(header), it needs to be coded to
ignore the last blank line but has to diagnose it as an error if
there is a blank line before that.

> +			if (atom->u.header.option == H_BARE) {
> +				v->s = xmemdupz(buf, header_len);
> +			} else if (atom->u.header.option == H_LENGTH)
> +				v->s = xstrfmt("%"PRIuMAX, (uintmax_t)header_len);
> +			continue;
> +		}
> +
>  		if (atom->u.contents.option == C_SUB)
>  			v->s = copy_subject(subpos, sublen);
>  		else if (atom->u.contents.option == C_SUB_SANITIZE) {