Toon Claes <toon@xxxxxxxxx> writes:
> OPT_STRING('b', "branch", &option_branch, N_("branch"),
> N_("checkout <branch> instead of the remote's HEAD")),
> + OPT_STRING(0, "revision", &option_rev, N_("rev"),
> + N_("clone single revision <rev> and check out")),
OK, this thing comes as a string; we'll parse it down to a commit
later, hopefully?
> - refspec_appendf(&remote->fetch, "+%s*:%s*", src_ref_prefix,
> - branch_top.buf);
> + if (!option_rev)
> + refspec_appendf(&remote->fetch, "+%s*:%s*", src_ref_prefix,
> + branch_top.buf);
> + die_for_incompatible_opt2(!!option_rev, "--revision",
> + !!option_branch, "--branch");
> + die_for_incompatible_opt2(!!option_rev, "--revision",
> + option_mirror, "--mirror");
So here is where we mark the new thing incompatible with these two,
and when either of them is given with "--revision", we would bail
out. OK.
> + // TODO --no-single-branch
Style.
> @@ -1381,7 +1396,15 @@ int cmd_clone(int argc,
> if (transport->smart_options && !deepen && !filter_options.choice)
> transport->smart_options->check_self_contained_and_connected = 1;
>
> - strvec_push(&transport_ls_refs_options.ref_prefixes, "HEAD");
> + if (option_rev) {
> + option_tags = 0;
> + option_branch = 0;
> + option_single_branch = 0;
> + opts.wants_head = 0;
> + opts.detach = 1;
option_branch is of type "char *" so sparse rightfully complains
that you are assigning an integer 0 to it, which follows stronger
rules than plain vanilla C standard to help us avoid mistakes.
But stepping back a bit, hasn't we already been ruled out earlier
that when option_rev is set, we cannot possibly be affected by the
"--branch" option that was given at the same time? I do not now
about other assignments we see in this block, but are there others
that are unnecessary? For example, you do not clear option_mirror
in this block. Is option_branch so special that it needs clearing,
and if so why?
Thanks.
