apycot-moved #19367 New proposal for code retrieving [open]

Cloning is expensive especially on large repositories. We could use a new function which will retrieve only the concerned revision:

  • mercurial : hg archive -t files -r $REVISION
  • subversion : svn export -r $REVISION
  • ...
priorityimportant
typeenhancement
done in<not specified>
load2.000
load left2.000
closed by<not specified>

Comments

  • 2010/01/20 10:48, written by sthenault

    the pb is that archive needs a local repository

  • 2010/02/07 18:46, written by pydavid

    We may cache all remote repositories in a dedicated space. It would fix the "local_repository relation issue" at the same time.

  • 2010/09/23 11:50, written by sthenault
    1. not sure it's quick than cloning a local repo
    2. needs apycot change to control/get used revision
    • 2010/09/23 12:01, written by pydavid
      1. It's only as fast if hardlinking are avaible. If the repository are on another partition than the apycot execution, cloning will be much slower
      2. hg archive contains a .hg_archival.txt file with the following data:
      • repo: b97547f5f1fa96cb25eab1b10d877db8ae6dedbf
      • node: b3fd91524132d0fc6b0bff042b7d8589ef946a05
      • branch: stable
      • latesttag: cubicweb-debian-version-3.9.7-1
      • latesttagdistance: 35
add comment

Ticket #19367 - latest update on 2010/09/23, created on 2009/12/15