Re: [PATCH v4 05/19] ref-filter: add parse_opt_merge_filter()

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

On Sun, Jun 21, 2015 at 1:48 PM, Karthik Nayak <karthik.188@gmail.com> wrote:
> Add 'parse_opt_merge_filter()' to parse '--merged' and '--no-merged'
> options and write MACROS for the same.

Why SHOUT here?

>
> This is copied from 'builtin/branch.c' which will eventually be removed
> when we port 'branch.c' to use ref-filter APIs.

Hmph. I somehow thought Matthieu's instruction was to finish tag.c
side first and then
do branch (i.e. with 3 and 4 you brought things from tag to
for-each-ref, now it is a time
to rewrite tag by using what you wrote for for-each-ref with 3 and 4,
before moving to
this patch)? Was that plan scrapped or found inappropriate or something?

> +int parse_opt_merge_filter(const struct option *opt, const char *arg, int unset)
> +{
> +       struct ref_filter *rf = opt->value;
> +       unsigned char sha1[20];
> +
> +       rf->merge = starts_with(opt->long_name, "no")
> +               ? REF_FILTER_MERGED_OMIT
> +               : REF_FILTER_MERGED_INCLUDE;
> +
> +       if (!arg)
> +               arg = "HEAD";

When does this trigger? You have lastarg-default with "HEAD", and I am
having trouble guessing when "arg" upon entry to this function can ever
be NULL.

> +       if (get_sha1(arg, sha1))
> +               die(_("malformed object name %s"), arg);
> +
> +       rf->merge_commit = lookup_commit_reference_gently(sha1, 0);

Can --merged (or --no-merged) be given more than once? Is the rule
"the last one wins"?
Does the old value of rf->merge_commit leak (no, it does not, but I am
just asking for
completeness)?