aeb0d1a480
commands/plugin: fix plugin existence middleware, add tests
2014-01-05 20:57:55 -08:00
5fe2994005
commands/plugin: convert all actions to use the new classes
2014-01-05 20:50:25 -08:00
8904319beb
commands/plugin: install version and entrypoints work
2014-01-05 20:47:02 -08:00
9dc1307b7c
commands/plugin: remove --plugin-prerelease
2014-01-05 17:46:20 -08:00
0c73a5ee05
remove clean on install... doesn't work right now
2014-01-05 17:40:38 -08:00
a571222431
core: tolerate errors in gemrc
2014-01-05 17:12:01 -08:00
d368b3cf62
core: Tolerate syntax errors in ~/.gemrc [GH-2760]
2014-01-05 17:11:24 -08:00
2fd144611e
Cleaning gems properly only removes them from the local dir
2014-01-05 17:04:50 -08:00
eabc0f04fa
friendly errors if a non-existent gem install is tried
2014-01-05 16:57:51 -08:00
fe8842c795
Fix failing tests
2014-01-05 16:54:10 -08:00
36f64db874
Always setup plugins, but just don't always require them
2014-01-05 16:49:25 -08:00
f7e4c4df6b
setup the Bundler path always
2014-01-05 16:43:59 -08:00
e69723b63c
core: Plugin::Manager.installed_specs doesn't use Bundler
2014-01-05 16:31:49 -08:00
f8b49afe8a
Nice error message on plugin install conflict
2014-01-05 16:28:05 -08:00
8861510520
Lower some branching logic
2014-01-05 16:17:39 -08:00
73c71dbcc6
Work some things around so that Bundler is not setup with no plugins
2014-01-05 16:16:04 -08:00
f51b6d0708
Lock to log4r < 1.1.11 because we got fucked
...
1.1.11 was released today and seriously broke backwards compatibility.
While they never officially made any promise to follow semver, it is
almost expected at this point, but log4r decided to just fuck that.
1.1.11 changed the arity of Log4r::Logger.initialize. That seriously
breaks _everything_.
Darwin awarddddddddd goes to...
2014-01-05 16:07:24 -08:00
76de267d1e
Don't load plugins again on `vagrant plugin` commands
2014-01-05 16:04:53 -08:00
8823f43f3a
commands/plugin: error if uninstall non-existent plugin
2014-01-05 16:03:00 -08:00
e1c9431085
commands/plugin: uninstall works again
2014-01-05 16:02:00 -08:00
8cfa24143d
Put plugins in their own Gemfile group so we can load on their own
2014-01-05 15:57:10 -08:00
e800743d06
fast-path vagrant --version
2014-01-05 15:32:19 -08:00
177bfc1d0c
Lock to log4r < 1.1.11 because we got fucked
...
1.1.11 was released today and seriously broke backwards compatibility.
While they never officially made any promise to follow semver, it is
almost expected at this point, but log4r decided to just fuck that.
1.1.11 changed the arity of Log4r::Logger.initialize. That seriously
breaks _everything_.
Darwin awarddddddddd goes to...
2014-01-05 15:27:20 -08:00
817f7f326e
fast path Vagrant --version
2014-01-05 15:27:03 -08:00
91751f6e41
load with bundler context
2014-01-05 15:26:44 -08:00
95aeb9443d
commands/plugin: Remove PruneGems
2014-01-05 14:55:44 -08:00
86610bf735
installing gems works
2014-01-05 14:54:50 -08:00
1eef75a715
commands/plugin: list uses PluginManager
2014-01-05 08:42:34 -08:00
279b171339
commands/plugin: remove unnecessary line
2014-01-04 17:22:07 -08:00
ae17dc09eb
commands/plugin: list plugins uses Bundler
2014-01-04 17:21:53 -08:00
8450f20e87
Move StateFile into Vagrant::Plugin core
2014-01-04 16:45:41 -08:00
d98868d150
core: Vagrant.require_plugin is gone
2014-01-04 16:35:28 -08:00
5387984e0f
commands/plugin: allow within Bundler env (cause we always are)
2014-01-04 16:32:46 -08:00
8adef9c15f
warn if running in a bundler env, don't load plugins
2014-01-04 16:29:22 -08:00
f2c6175d16
Use bundler to load plugins
2014-01-04 16:15:12 -08:00
7c8fc34a25
update CHANGELOG
2014-01-03 11:36:59 -08:00
0f989ce554
Merge pull request #2754 from mitchellh/f-insert-private-key
...
Support password-based SSH, automatically insert insecure key
2014-01-03 11:34:53 -08:00
a6f4f56ba7
website/docs: update docs for insert_key
2014-01-03 11:31:33 -08:00
4c6957b5cf
communicators/ssh: if insert_key is false, don't insert a key
2014-01-03 11:29:44 -08:00
92413d0393
core: error if vagrant ssh -c with password
2014-01-03 11:26:41 -08:00
54e640b0c9
communicators/ssh: output UI when inserting key
2014-01-03 11:22:05 -08:00
074bb2c7fb
core: fix potential exception case in SSHExec middleware
2014-01-03 11:19:17 -08:00
b3a9e6a088
insert_public_key cap, and use that if it exists
2014-01-03 11:13:21 -08:00
664aaa0088
core: raise exceptions if they happen in WaitForCommunicator threads
2014-01-03 10:57:04 -08:00
d1fdee7ae3
core: warn if password only on `vagrant ssh`
2014-01-03 10:45:42 -08:00
38e7166a21
communicators/ssh: support passwords
2014-01-03 09:55:34 -08:00
e115322e78
core: accept passwords in ssh_info
2014-01-03 09:48:35 -08:00
0ee22aa6d1
Merge pull request #2752 from sferik/patch-1
...
core: replace `exit 1` with `abort`
2014-01-03 08:33:05 -08:00
56175b8132
Replace `exit 1` with `abort`
...
According to the documentation, Kernel.abort calls Kernel.exit(false). The failure status 1 is an implementation detail on Unix-like systems. In theory, a future system could exist where 1 indicates a successful exit and 0 indicates a failure.
2014-01-03 17:32:45 +01:00
32d8b507c1
update changelog
2014-01-03 08:31:14 -08:00
Mitchell Hashimoto
Erik Michaels-Ober