git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] git exproll: steps to tackle gc aggression

From
Junio C Hamano <gitster@pobox.com>
Date
Aug 8, 2013, 19:37 UTC
Message-ID
<7vwqnw7z47.fsf@alter.siamese.dyndns.org>
In-Reply-To
<CALkWK0kqE8azzxp_GkzhPNT41nD8NzeLqXSe1xi0jbVo=7Xz3A@mail.gmail.com>
Ramkumar Ramachandra <artagnon@gmail.com> writes:
> I asked you a very simple question: what happens when I do "git push"?

You asked "does push send unreachable cruft?" and I said No. Does that answer your question?

> 3. After a few days of working, the gc heuristics figure out that I
> have too much garbage and too many packs; a cleanup is required. The
> gc --auto which doesn't tolerate fragmentation: it tries to put
> everything into one large pack.

Today's "gc --auto" skims the history shallowly and packs loose objects into a single _new_ pack. If you start from a single pack and enough loose objects and run "gc --auto", you will have two packs, one intact from the original, one new.

> We're talking about tackling the gc aggression problem in 3.
> when I say "git push" (or when gc decides to repack), won't I need
> to explode the young pack into loose objects, do a reachability
> analysis, and repack anyway?

You do not explode anything. "push" will read objects from anywhere, either from loose or packed, and send a newly created pack over the wire.

And I see from "(or when ...)" part that you do not know what you are talking about. "push" will not stream existing pack to the other end without computation, and it never will. It will reuse existing individual pieces when it can, and having data in a pack (even without deltifying, or as a suboptimal delta) is much better for push performance than having the same data in a loose object if only for this reason, as you do not have to recompress and reencode it in a different way (loose objects and undelitifed object in pack are encoded differently).

> ... are you willing to explain it to me?
Hope that the above clarifies.

I've treated this thread as a design discussion, not an education session, but the above ended up having more education material than anything that would advance the design.

Previous: Ramkumar RamachandraNext: Ramkumar Ramachandra
Message 12 of 33 in “git exproll: steps to tackle gc aggression”
  1. git exproll: steps to tackle gc aggressionRamkumar Ramachandra, Aug 6, 2013
  2. Duy NguyenAug 6, 2013
  3. Junio C HamanoAug 6, 2013
  4. Ramkumar RamachandraAug 7, 2013
  5. Junio C HamanoAug 8, 2013
  6. Ramkumar RamachandraAug 8, 2013
  7. Junio C HamanoAug 8, 2013
  8. Martin FickAug 8, 2013
  9. Junio C HamanoAug 8, 2013
  10. Ramkumar RamachandraAug 8, 2013
  11. Ramkumar RamachandraAug 8, 2013
  12. Junio C HamanoAug 8, 2013
  13. Ramkumar RamachandraAug 8, 2013
  14. Martin LanghoffAug 8, 2013
  15. Jeff KingAug 9, 2013
  16. Ramkumar RamachandraAug 9, 2013
  17. Junio C HamanoAug 9, 2013
  18. Jeff KingAug 9, 2013
  19. Duy NguyenAug 10, 2013
  20. Jeff KingAug 10, 2013
  21. Junio C HamanoAug 10, 2013
  22. Ramkumar RamachandraAug 10, 2013
  23. Duy NguyenAug 10, 2013
  24. Duy NguyenAug 10, 2013
  25. Jeff KingAug 10, 2013
  26. Duy NguyenAug 10, 2013
  27. Ramkumar RamachandraAug 10, 2013
  28. Duy NguyenAug 10, 2013
  29. Jeff KingAug 10, 2013
  30. Martin FickAug 7, 2013
  31. Duy NguyenAug 8, 2013
  32. Martin FickAug 7, 2013
  33. Ramkumar RamachandraAug 7, 2013

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.