Re: Git Test Coverage Report (Thurs. June 27)

[Christian Couder <christian.couder@gmail.com>]

On Thu, Jun 27, 2019 at 7:35 PM Derrick Stolee <stolee@gmail.com> wrote:
>
> Here are some interesting sections I found when examining the test coverage
> report. I am only highlighting these sections because they seem to include
> non-trivial logic. In some cases, maybe the code isn't needed.
>
> On 6/27/2019 1:05 PM, Derrick Stolee wrote:
> > promisor-remote.c
> > db27dca5 25) die(_("Remote with no URL"));
> > 48de3158 61) warning(_("promisor remote name cannot begin with '/': %s"),
> > 48de3158 63) return NULL;
> > faf2abf4 93) previous->next = r->next;
> > 4ca9474e 108) return git_config_string(&core_partial_clone_filter_default,
> > fa3d1b63 139) return 0;
> > 9e27beaa 202) static int remove_fetched_oids(struct repository *repo,
> > 9e27beaa 206) int i, remaining_nr = 0;
> > 9e27beaa 207) int *remaining = xcalloc(oid_nr, sizeof(*remaining));
> > 9e27beaa 208) struct object_id *old_oids = *oids;
> > 9e27beaa 211) for (i = 0; i < oid_nr; i++)
> > 9e27beaa 212) if (oid_object_info_extended(repo, &old_oids[i], NULL,
> > 9e27beaa 214) remaining[i] = 1;
> > 9e27beaa 215) remaining_nr++;
> > 9e27beaa 218) if (remaining_nr) {
> > 9e27beaa 219) int j = 0;
> > 9e27beaa 220) new_oids = xcalloc(remaining_nr, sizeof(*new_oids));
> > 9e27beaa 221) for (i = 0; i < oid_nr; i++)
> > 9e27beaa 222) if (remaining[i])
> > 9e27beaa 223) oidcpy(&new_oids[j++], &old_oids[i]);
> > 9e27beaa 224) *oids = new_oids;
> > 9e27beaa 225) if (to_free)
> > 9e27beaa 226) free(old_oids);
> > 9e27beaa 229) free(remaining);
> > 9e27beaa 231) return remaining_nr;
> > 9e27beaa 248) if (remaining_nr == 1)
> > 9e27beaa 249) continue;
> > 9e27beaa 250) remaining_nr = remove_fetched_oids(repo, &remaining_oids,
> > 9e27beaa 252) if (remaining_nr) {
> > 9e27beaa 253) to_free = 1;
> > 9e27beaa 254) continue;
> > 9e27beaa 262) free(remaining_oids);
>
> Christian: this section continues to be untested, but I think you were
> working on creating tests for this.

Yeah, I am planning to work on this soon.

> > t/helper/test-oidmap.c
> > 11510dec 52) if (get_oid(p1, &oid)) {
> > 11510dec 53) printf("Unknown oid: %s\n", p1);
> > 11510dec 54) continue;
> > 11510dec 58) FLEX_ALLOC_STR(entry, name, p2);
> > 11510dec 59) oidcpy(&entry->entry.oid, &oid);
> > 11510dec 62) oidmap_put(&map, entry);
>
> Christian: this block looks like the test-oidmap helper never uses the "add"
> subcommand. Is that correct?

Yeah, I initially copied it from hashmap, but then I realized that it
was nearly identical as the "put" subcommand, so not worth testing
separately. I should have removed it and will do it soon.

> > 11510dec 97) if (get_oid(p1, &oid)) {
> > 11510dec 98) printf("Unknown oid: %s\n", p1);
> > 11510dec 99) continue;
> > 11510dec 103) entry = oidmap_remove(&map, &oid);
> > 11510dec 106) puts(entry ? entry->name : "NULL");
> > 11510dec 107) free(entry);
>
> Similarly, this block means we are not using the "remove" subcommand.

Yeah, it looks like I forgot to implement a test for that subcommand.
Will add it soon.

Thanks,
Christian.