On 5/11/2018 2:00 PM, Jeff King wrote:
This is a follow-up to the discussion from February: https://public-inbox.org/git/1519522496-73090-1-git-send-email-dstolee@xxxxxxxxxxxxx/ There I theorized that some of these extra has_object_file() checks were unnecessary. After poking around a bit, I've convinced myself that this is the case, so here are some patches. After Stolee's fix in ebbed3ba04 (revision.c: reduce object database queries, 2018-02-24), I doubt this will provide any noticeable speedup. IMHO the value is just in simplifying the logic. The first two patches are the actual has_object_file simplifications. The second two are an attempt to fix some head-scratching I had while reading mark_parents_uninteresting(). I hope the result is easier to follow, but I may have just shuffled one confusion for another. :) [1/4]: mark_tree_contents_uninteresting(): drop missing object check [2/4]: mark_parents_uninteresting(): drop missing object check [3/4]: mark_parents_uninteresting(): replace list with stack [4/4]: mark_parents_uninteresting(): avoid most allocation revision.c | 90 ++++++++++++++++++++++++++++++------------------------ 1 file changed, 50 insertions(+), 40 deletions(-) -Peff
This series looks good to me. I found Patch 3 hard to read in the diff, so I just looked at the final result and the new arrangement is very clear about how it should behave.
Reviewed-by: Derrick Stolee <dstolee@xxxxxxxxxxxxx>
