Jeff King <> writes:

> On Wed, Aug 24, 2016 at 02:16:02PM -0700, Junio C Hamano wrote:
>> * jk/pack-objects-optim-mru (2016-08-11) 4 commits
>>   (merged to 'next' on 2016-08-11 at c0a7dae)
>>  + pack-objects: use mru list when iterating over packs
>>  + pack-objects: break delta cycles before delta-search phase
>>  + sha1_file: make packed_object_info public
>>  + provide an initializer for "struct object_info"
>>  "git pack-objects" in a repository with many packfiles used to
>>  spend a lot of time looking for/at objects in them; the accesses to
>>  the packfiles are now optimized by checking the most-recently-used
>>  packfile first.
>>  Will hold to see if people scream.
> Just a note that we've been running with this at GitHub on all of our
> servers for a bit over a week, and no screaming yet. That's not
> necessarily proof of anything, but it does make the audience of "people"
> a bit bigger than just "next", as we run quite a few invocations of
> pack-objects in a day.
> I don't think that changes anything in the near future, since this is
> obviously not for v2.10, but barring any complaints, it's probably a
> reasonable topic to consider for the version after (and of course I'll
> relay any issues we come across on our servers).

Yeah, that "Will hold" is primarily me being lazy [*1*] during the
-rc period when updating the "What's cooking".  After following the
"break delta cycles" patch carefully, I am no longer worried about
this topic.  It should be in 'master' early in the next cycle, among
other topics that are competently done.

> I'm planning to deploy the delta-cache topic soon, too, so that should
> give it some good exercise.

Good.  Thanks.


*1* ... and a bit more careful, as any "Will merge to 'next'" thing
gets marked as "Will merge to 'master'" by default and having any
entry under "Will merge to 'master'" in "Meta/cook -w" report tempts
me to merge them even during -rc period.
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to
More majordomo info at

Reply via email to