Permalink
Commits on Oct 30, 2016
  1. Version bump to 2.4.0

    nirvdrum committed Oct 30, 2016
  2. Preparing the 2.4.0 release.

    nirvdrum committed Oct 30, 2016
Commits on Oct 18, 2016
  1. Merge pull request #224 from pdf/multiple_branches_tags

    nirvdrum committed Oct 18, 2016
    Allow `tags` and `branches` arguments to be specified multiple times
Commits on Oct 12, 2016
  1. Allow `tags` and `branches` arguments to be specified multiple times

    pdf committed Jul 5, 2016
    This mirrors upstream, though I'm not certain of which git-svn version
    this functionality was added.
    
    Fixes #175
Commits on Jul 1, 2016
  1. Merge pull request #223 from pdf/fix_exclude

    nirvdrum committed Jul 1, 2016
    Fix exclude arg (incorrectly quoted string)
  2. Fix exclude arg (incorrectly quoted string)

    pdf committed Jul 1, 2016
    The quote in the appended `--ignore-paths` cmd argument is in the wrong
    place.  Also add the conventional space in case of additional future
    args.
Commits on Apr 17, 2015
  1. Merge pull request #190 from edpbx/passwordoption

    nirvdrum committed Apr 17, 2015
    Add support for --password option
  2. Add support for --password option

    edpbx committed Apr 15, 2015
Commits on Apr 16, 2015
  1. Merge pull request #191 from edpbx/readmeids

    nirvdrum committed Apr 16, 2015
    Stop README ID-extract snippet leaving spaces end of line
Commits on Apr 15, 2015
  1. Merge pull request #169 from psexton/master

    nirvdrum committed Apr 15, 2015
    Remove unnecessary package install in README
Commits on Aug 1, 2014
  1. Update README.markdown

    Paul Sexton committed Aug 1, 2014
    On Debian systems, the `rubygems` package has been rolled into the `ruby` package.
Commits on Jun 8, 2014
  1. Preparing the 2.3.2 release.

    nirvdrum committed Jun 8, 2014
  2. Preparing the 2.3.2 release.

    nirvdrum committed Jun 8, 2014
  3. Switched back from open4 to IO.popen.

    nirvdrum committed Jun 8, 2014
    open4 was added as a dependency when I was trying to parse STDERR and needed to work with STDIN. As it turns out, $stdin is good enough.  Redirecting STDERR to STDOUT seems to be good enough as well.  While open4 is probably still a cleaner approach, its reliance on fork() means it doesn't work on non-JRuby in Windows (echnically it doesn't work in JRuby either, but JRuby has an IO.popen4 special method that does virtually the same thing and we were using that instead).
Commits on May 15, 2014
  1. Version bump to 2.3.1

    nirvdrum committed May 15, 2014
  2. Preparing the 2.3.1 release.

    nirvdrum committed May 15, 2014
  3. Fixed issue attempting to track remote SVN branches in git >= 1.8.3.2.

    nirvdrum committed May 15, 2014
    In order not to break semantics with those using svn2git with git < 1.8.3.2, the current tracking behavior is preserved but deprecated.  Its functionality is superceded by our '--rebase' option for re-syncing with upstream.
  4. Merge pull request #160 from ktdreyer/minitest

    nirvdrum committed May 15, 2014
    tests: switch to minitest
  5. tests: switch to minitest

    ktdreyer committed Apr 8, 2014
    Ruby 1.9+ uses Minitest as the backend for Test::Unit. As of Minitest 5,
    the shim no longer supports Test::Unit::TestCase.
    
    Adjust the svn2git test suite to support Minitest 5's syntax.
    
    Minitest versions 4 and below do not support the newer Minitest::Test
    class that arrived in version 5. For that case, use the
    MiniTest::Unit::TestCase class as a fallback.
  6. Merge branch 'gemspec-license' of github.com:ktdreyer/svn2git into kt…

    nirvdrum committed May 15, 2014
    …dreyer-gemspec-license
    
    Conflicts:
    	Rakefile
Commits on May 14, 2014
  1. Version bump to 2.3.0

    nirvdrum committed May 14, 2014
  2. Preparing the 2.3.0 release.

    nirvdrum committed May 14, 2014
  3. Robust STDIN pass-through to sub-process.

    nirvdrum committed May 14, 2014
    My previous attempt at preventing explosive thread growth didn't work because I couldn't reliably interrupt a .gets call.  The timeout in the thread did cause the thread to begin aborting, but  would still be blocked.  Starting multiple such threads meant any user input could end up being consumed by any one of those threads.  Moving towards a single thread for getting user input and communicating to a new pass-through thread via a Queue avoids those problems and allows us to stop the pass-through threads when they're no longer needed.
  4. Don't join on the stdin pass-through thread.

    nirvdrum committed May 14, 2014
    Since the thread would otherwise block forever waiting for user input, we'll wrap it in a timeout and loop checking the status of the sub-process STDIN pipe. This should prevent runaway thread growth.