Switch branches/tags
Commits on Sep 19, 2012
  1. minor cleanup

    indirect committed Sep 19, 2012
  2. update changelog

    hone committed Sep 19, 2012
Commits on Sep 18, 2012
  1. Version 1.2.1

    hone committed Sep 18, 2012
  2. use the API_TIMEOUT constant

    hone committed Sep 18, 2012
  3. shut the server down in the test

    hone committed Sep 18, 2012
  4. use Net::HTTP's read_timeout setting instead

    talked with @drbrain about net/http timeouts, and we decided it was
    better than having to do a manual Timeout.
    hone committed Sep 18, 2012
Commits on Sep 14, 2012
  1. update changelog for 1.2.1

    hone committed Sep 14, 2012
  2. have a timeout around the gemcutter API endpoint

    The API endpoint can sometimes hang and not return a response but open a
    connection. Bundler before would just hang for a long time. This adds a
    reasonable timeout, since the point of the API is to speed up the
    install process. If the API takes too long, we should fall back on
    cloudfront and the modern index, since that will likely be faster.
    hone committed Sep 14, 2012
  3. Revert "`clean` should not remove gems in the lockfile"

    This reverts commit aefa530.
    bundle clean should remove BUNDLE_WITHOUT gems. BUNDLE_WITHOUT settings
    should be standardized across commands as well if you are using
    BUNDLE_WITHOUT you probably don't want the gems to be around when
    hone committed Sep 14, 2012
Commits on Sep 12, 2012
  1. Fix #2096. Definition#new use nil for ruby_version default value

    Definition#validate_ruby! checks for nil and not an empty string
    hone committed Sep 12, 2012
Commits on Sep 3, 2012
  1. point to

    hone committed Sep 3, 2012
Commits on Aug 30, 2012
  1. Revert "Ruby 2.0 (trunk ruby) is not Ruby 1.9"

    As discussed in #1539 and #2056, Bundler uses platforms to distinguish Ruby ABI versions. That means the known (and ATM supported) use cases for the ruby_19 platform include Ruby 2.0.
    This reverts commit dcfb640.
    indirect committed with hone Aug 16, 2012
  2. fix whitespace test

    hone committed Aug 30, 2012
  3. Version 1.2.0

    hone committed Aug 30, 2012
  4. update changelog for 1.2.0

    hone committed Aug 30, 2012
  5. bundle platform man pages

    hone committed Aug 30, 2012
Commits on Aug 25, 2012
  1. Fix #2075. Return to old namespace behavior.

    $1 gets overidden by gsub.
    hone committed Aug 25, 2012
Commits on Aug 14, 2012
  1. fix test case names

    hone committed Aug 14, 2012
Commits on Aug 9, 2012
  1. Version 1.2.0.rc.2 and changelog

    hone committed Aug 9, 2012
Commits on Aug 7, 2012
  1. Merge pull request #1931 from coding46/patch-1

    Simple Typo
    radar committed Aug 7, 2012
Commits on Jul 18, 2012
  1. Fix 'LICENSCE' typo in CHANGELOG

    radar committed Jul 18, 2012
Commits on Jul 17, 2012
  1. Version 1.2.0.rc with changelog

    hone committed Jul 17, 2012
  2. don't need these tests in 1.2

    hone committed Jul 17, 2012
  3. Merge branch '1-1-stable'

    hone committed Jul 17, 2012
  4. Version 1.1.5 with changelog

    hone committed Jul 17, 2012
  5. special case `ruby` directive

    This is light of the upcoming 1.2.0 release which adds this feature.
    This way people still using 1.1.x can run 1.2.0 Gemfiles that employ
    this feature and get a notification about upgrading.
    hone committed Jul 17, 2012
  6. rename --no-lock to --dry-run

    hone committed Jul 17, 2012
  7. Make locking the Gemfile optional on $ bundle check

    Currently `bundle check` locks the Gemfile. According to @indirect this
    behaviour is intentional (see!/indirect/status/186493694013210624).
    I am working on a CLI monitoring tool that makes it easier to work with
    cross-repository changes on projects that are split up into several
    This tool is supposed to display status information such as out-of-date
    bundler Gemfiles etc. To do so it watches the filesystem and queries for
    information on demand. Using `bundle check` to do so currently leads to
    highly confusing side-effects because suddenly the git working directory
    might be dirty - having changed the Gemfile.lock - because `bundle
    check` has been run in the background.
    This patch adds a boolean method option `--[no-]lock` to `bundle check`
    that defaults to `true` in order to not change the current behaviour but
    make it possible for users to turn it off.
    Sidenote: apparently the same behaviour also applies to `bundle show`.
    Am I the only one who finds this highly confusing?
    Sven Fuchs committed with hone Apr 1, 2012