Skip to content
Commits on Mar 12, 2012
  1. @tmm1

    Merge pull request #13 from benburkert/elasticsearch

    Build of the elasticsearch 0.19.0 release.
    tmm1 committed Mar 12, 2012
  2. @benburkert
  3. @jnewland

    s3cmd

    jnewland committed Mar 12, 2012
  4. @jnewland

    basic kestrel formula

    jnewland committed Feb 20, 2012
Commits on Mar 6, 2012
  1. @benburkert
  2. @benburkert

    Update lockrun.

    benburkert committed Mar 5, 2012
Commits on Mar 5, 2012
  1. @tmm1

    nethogs

    tmm1 committed Mar 4, 2012
Commits on Mar 3, 2012
  1. @tmm1

    bprobe

    tmm1 committed Mar 2, 2012
  2. @tmm1

    google-coredumper

    tmm1 committed Mar 2, 2012
  3. @tmm1

    megacli homepage

    tmm1 committed Mar 2, 2012
  4. @tmm1

    smarter shlibs detection

    tmm1 committed Mar 2, 2012
Commits on Mar 2, 2012
  1. @tmm1

    add init.d, example cfg, etc

    tmm1 committed Mar 1, 2012
  2. @tmm1

    haproxy

    tmm1 committed Mar 1, 2012
Commits on Mar 1, 2012
  1. @tmm1
  2. @tmm1

    geoip-city formula

    tmm1 committed Feb 29, 2012
Commits on Feb 22, 2012
  1. @tmm1

    filtergen

    tmm1 committed Feb 21, 2012
Commits on Feb 18, 2012
  1. @tmm1
  2. @tmm1
  3. @tmm1
  4. @tmm1

    Merge pull request #12 from github/git-detect-clones

    git: detect clones more accurately in upload-pack patch
    tmm1 committed Feb 17, 2012
  5. @tmm1
  6. @tmm1

    bump nginx

    tmm1 committed Feb 17, 2012
Commits on Feb 17, 2012
  1. @tmm1

    bump zeromq

    tmm1 committed Feb 17, 2012
  2. @tmm1
  3. @peff

    git: detect clones more accurately in upload-pack patch

    We keep statistics on clones versus fetches by asking
    upload-pack to run a hook for us. It tells us which it saw
    by checking git's existing create_full_pack flag. However,
    that flag is intended for use by pack-objects and is defined
    like this:
    
       int create_full_pack = (nr_our_refs == want_obj.nr && !have_obj.nr);
    
    I.e., the other side asked for everything we had, and
    claimed to have nothing else. But the first half of the
    condition is not true for most GitHub repositories, as we
    have refs outside of refs/heads and refs/tags that are not
    fetched by normal, non-mirror clones (e.g., our pull request
    refs).
    
    A better heuristic for clone vs fetch is to just take the
    second half of that condition: check that the remote side
    claimed not to have any objects in its repository. That will
    catch clones and fetches into empty repositories, both of
    which are identical from the server's perspective.
    
    It will miss a clone with "--reference" to an alternates
    repository. But that is a minority case, and arguably not
    really a clone (it's indistinguishable on the server end
    from a fetch from a related repository cloned from somewhere
    else).
    peff committed Feb 17, 2012
  4. @tmm1
  5. @tmm1

    add certmaster and func packages

    tmm1 committed Feb 16, 2012
Commits on Feb 16, 2012
  1. @tmm1

    latest freight

    tmm1 committed Feb 16, 2012
Commits on Feb 15, 2012
  1. @tmm1

    fix debhelper requirement

    tmm1 committed Feb 15, 2012
  2. @tmm1

    new lsof

    tmm1 committed Feb 15, 2012
  3. @tmm1
Commits on Feb 11, 2012
  1. @tmm1
  2. @tmm1

    handle errors during dh_shlibdeps

    tmm1 committed Feb 11, 2012
  3. @tmm1

    nginx: add syslog patch

    tmm1 committed Feb 10, 2012
  4. @tmm1

    nginx: use @nickh's chunkin patches

    tmm1 committed Feb 10, 2012
Something went wrong with that request. Please try again.