On 21-7-2016 10:59, Tone Zhang wrote: > Hi, > > Could you please share me that after reclone the ceph code you can > fetch the latest code of ceph/rocksdb or not? I mean the same as the > upstream facebook/rocksdb. > I have tried, the HEAD of ceph/rocksdb is still > 5647fa427c4eab87c0caf8bcbb91f5846a4623f9. In my side, the code cloned > to local is different from the files saw in github filebrowser. Right. That is exactly my problem as well. https://github.com/ceph/rocksdb claims 6797e6f Which is 17 commits behind on the master version So what I get from the other mails, is to actually get that forward, a commit for the new tree needs to get into ceph/ceph. Which is what Josh and Greg are writting about. And there is a lingering PULL from Kefu that is actually doing something like that. --WjW > Thanks. > > Br, > Tone > > On 21 July 2016 at 16:15, Willem Jan Withagen <wjw@xxxxxxxxxxx> wrote: >> On 21-7-2016 09:15, Tone Zhang wrote: >>> Hi, >>> >>> I met the same issue, and I run "git submodule update --init >>> --recursive --remote" to fetch the latest commit from >>> facebook/rocksdb. >>> In fact, there is one similar PR has been raised, it is >>> https://github.com/ceph/rocksdb/pull/10, it is in review stage. >>> In fact, the PR is raised to fix the rocksdb compiling error in >>> AArm64, the reason the ceph/rocksdb is out of date. >>> Sure we should fetch all the delta from facebook/rocksdb to ceph/rocksdb. >>> Thanks. >>> >>> Br, >>> Tone >>> >>> On 20 July 2016 at 22:23, Josh Durgin <jdurgin@xxxxxxxxxx> wrote: >>>> On 07/20/2016 06:52 AM, Willem Jan Withagen wrote: >>>>> >>>>> On 20-7-2016 15:32, Gregory Farnum wrote: >>>>>> >>>>>> On Wed, Jul 20, 2016 at 3:38 AM, Willem Jan Withagen <wjw@xxxxxxxxxxx> >>>>>> wrote: >>>>>>> >>>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> I submitted a small patch to rocksdb for FreeBSD >>>>>>> >>>>>>> https://github.com/facebook/rocksdb/commit/b726bf59619d43651838e03cce9f876b3560cbda >>>>>>> >>>>>>> And was hopping that it would be tracked by my Ceph stuff as well. >>>>>>> But refetching things, it does not seem to be in the tree ceph cmake >>>>>>> fetches. >>>>>>> >>>>>>> And I'm not fluent enough in git to actually check if my current >>>>>>> checkout with hash: 5647fa427c4eab87c0caf8bcbb91f5846a4623f9 >>>>>>> should contain the above commit. >>>>>> >>>>>> >>>>>> Use "git log" on both trees and search for the hashes? :) >>>>>> >>>>>> Anyway, the Ceph submodule does not auto-update (git submodules don't >>>>>> support that AFAIK). You can go into the rocksdb folder and try >>>>>> checking out the latest upstream master, then committing the new hash >>>>>> as a commit in the Ceph git repo (at least, once your commit is >>>>>> merged; not sure if it is yet). >>>>> >>>>> >>>>> Although I (sort of) understand what you are saying, I'm not able to >>>>> reproduce the actual commands to do this. Comparing the logs I >>>>> understand. First just have to clone rocksdb. >>>>> >>>>> And as far as I know the commit I'm looking for is in the current >>>>> rocksdb tree. I got commit about 22 days ago, and the current hash seems >>>>> te be from just a bit before that. >>>> >>>> >>>> To clarify, to update the rocksdb submodule, you'd need to first update >>>> the ceph fork. (cat .gitmodules shows src/rocksdb comes from >>>> https://github.com/ceph/rocksdb). >>>> >>>> You can make a PR to ceph/rocksdb with the latest upstream code from >>>> within your submodule checkout (src/rocksdb) which functions like a >>>> normal git repo itself. You could e.g. 'git pull >>>> https://github.com/facebook/rocksdb' to get the latest code there. >>>> >>>> Then once that's merged into ceph/rocksdb, make a PR to ceph (from the >>>> root of your ceph checkout 'git commit src/rocksdb' will update the >>>> sha1 the rocksdb submodule references). >> >> 'mmmm, >> >> If I look in ceph/rocksdb at github thru the filebrowser, the commit I >> made is actually in this tree. So I guess I should be getting it. >> So I guess I'm going to start recloning my sources... >> >> --WjW >> >> -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html