On Thu, Dec 4, 2014 at 10:41 AM, Junio C Hamano <gitster@xxxxxxxxx> wrote: > Stefan Beller <sbeller@xxxxxxxxxx> writes: > >> This is the whole refs-transactions-reflog series[1], >> which was in discussion for a bit already. It applies to origin/master. >> >> The idea is to have the reflog being part of the transactions, which >> the refs are already using, so the we're moving towards a database >> like API in the long run. This makes git easier to maintain as well >> as opening the possibility to replace the backend with a real database. >> >> If you've followed the topic a bit, start reading at patch >> "[PATCH 06/13] refs.c: add a transaction function to append a reflog" >> as the first 5 patches have been discussed a lot separately and >> can be found in origin/pu already[2]. > >> The first two patches are deduplicating code. >> The third patch is ripping some code out of log_ref_write and introduces >> log_ref_write_fd, which does the actual writing. >> The patches 4+5 are renaming variables for clarity. > > Thanks. It seems that we have a bit of hashing out the approaches > that is necessary between Michael's and this one. I'll refrain from > picking this up for today (even though I would read it through as > time permits). It might turn out to be necessary to drop those five > early patches from my tree when this series gets rerolled to take > input from the alternative Michael's working on, but that droppage > does not have to happen today. > Ok, let's see how Michaels series evolves and if I can base the transactions series on that afterwards. Michael has picked up 3 out the 5 patches to get his series started, so maybe we can resolve it nicely. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html