Re: [PATCH v2] index: add trace2 region for clear skip worktree
- From
- Jeff Hostetler <git@jeffhostetler.com>
- Date
- Oct 28, 2022, 16:50 UTC
- Message-ID
- <ff5d79c9-35f8-8881-8349-63977732c422@jeffhostetler.com>
- In-Reply-To
- <pull.1368.v2.git.git.1666917961644.gitgitgadget@gmail.com>
On 10/27/22 8:46 PM, Anh Le via GitGitGadget wrote: [...]
Show 47 quoted lines
> diff --git a/sparse-index.c b/sparse-index.c
> index e4a54ce1943..dbf647949c1 100644
> --- a/sparse-index.c
> +++ b/sparse-index.c
> @@ -493,24 +493,38 @@ void clear_skip_worktree_from_present_files(struct index_state *istate)
> int dir_found = 1;
>
> int i;
> + int path_counts[2] = {0, 0};
> + int restarted = 0;
>
> if (!core_apply_sparse_checkout ||
> sparse_expect_files_outside_of_patterns)
> return;
>
> + trace2_region_enter("index", "clear_skip_worktree_from_present_files", istate->repo);
> restart:
> for (i = 0; i < istate->cache_nr; i++) {
> struct cache_entry *ce = istate->cache[i];
>
> - if (ce_skip_worktree(ce) &&
> - path_found(ce->name, &last_dirname, &dir_len, &dir_found)) {
> - if (S_ISSPARSEDIR(ce->ce_mode)) {
> - ensure_full_index(istate);
> - goto restart;
> + if (ce_skip_worktree(ce)) {
> + path_counts[restarted]++;
> + if (path_found(ce->name, &last_dirname, &dir_len, &dir_found)) {
> + if (S_ISSPARSEDIR(ce->ce_mode)) {
> + ensure_full_index(istate);
> + restarted = 1;
> + goto restart;
> + }
> + ce->ce_flags &= ~CE_SKIP_WORKTREE;
> }
> - ce->ce_flags &= ~CE_SKIP_WORKTREE;
> }
> }
> +
> + if (path_counts[0] > 0) {
> + trace2_data_intmax("index", istate->repo, "clear_skip_worktree_from_present_files/path_count", path_counts[0]);
> + }
> + if (restarted) {
> + trace2_data_intmax("index", istate->repo, "clear_skip_worktree_from_present_files/full_index/path_count", path_counts[1]);
> + }
> + trace2_region_leave("index", "clear_skip_worktree_from_present_files", istate->repo);
> }This looks good. And I think it sets us up to later answer some of our earlier perf questions after you get some data.
For example, if the path_count[0] phase is expensive, does the restart always need to restart the loop at i=0? Granted, the `ensure_full_index()` may create a new cache-entry array, so the value of `i` prior to the call may not correspond to the same cell after the call at the `goto`, but we could cache the strdup the pathname of cache[i] before the call and do a find on the new cache[] afterwards to get the corresponding i-prime value and let the restart start the loop on i-prime. That's just an idea, but let's wait for the data to see if that would be helpful.
Thanks! Jeff