84ded8d67e
Fix machine ssh_info tests
2013-02-05 21:27:26 -08:00
864f1b7506
Pass a proper Pathname for SSHExec to check key perms
2013-02-04 21:17:12 -08:00
434cc79a83
Properly expand config.ssh.private_key_path
2013-02-04 21:13:29 -08:00
c57ba9de58
Give a nice human-friendly error message when problems loading Vfile
2013-01-30 20:01:41 -08:00
7eec127704
Support refreshing the machine cache
2013-01-30 10:37:40 -08:00
64042a5d35
Downloading the box_url component works again
2013-01-30 10:24:47 -08:00
3399b74163
Fix invalid box test by not relying on Archive;:Tar::Minitar
2013-01-29 11:10:44 -08:00
948a31d7b3
Merge pull request #1326 from lolindrath/master
...
Add test for invalid box file
2013-01-29 11:09:05 -08:00
2073ce697f
Move cli test to proper path
2013-01-29 11:00:56 -08:00
a21744a12d
Merge pull request #1336 from ithinkihaveacat/return-non-zero-if-invalid-command
...
Return exit status of 1 on invalid command
2013-01-29 10:58:55 -08:00
7531c94dbf
Allow path and exception for Lock middleware to be procs
2013-01-28 20:46:39 -08:00
9ae3a373c8
Add new middleware builtin: Lock
...
This will do a process lock by flocking a file. If it fails, it will
raise an exception of choice.
2013-01-28 17:08:37 -08:00
64af19c783
Remove archive::tar::minitar from tests
2013-01-28 13:28:40 -08:00
7941748db7
Return exit status of 1 on invalid command
...
Makes "vagrant destroyjj" and similar return an exit code of 1, so that
"vagrant destroyjj && vagrant up" works as expected.
2013-01-27 19:52:43 +00:00
6a3c0bd425
Call built-in supports args for imddleware
2013-01-22 12:20:02 -08:00
51540496a9
GracefulHalt builtin
2013-01-22 11:56:49 -08:00
8ad8f73846
Raise an exception if a non-MachineState is returned
2013-01-21 11:31:20 -06:00
7bdf54923a
Create the MachineState class
2013-01-21 11:25:28 -06:00
2d57afbbda
Support warnings/errors when upgrading Vagrantfiles internally
2013-01-20 22:04:50 -05:00
8a8b771e0a
Add test for invalid box file
...
Filled in pending test. It creates a valid tar file, over-writes the
checksum field in the tar's header and then checks that the
BoxUnpackageFailure exception is thrown.
2013-01-20 14:28:48 -05:00
e651eb3aa1
Add a V2 config helper to merge errors since that seems common
2013-01-18 13:03:07 -08:00
a8b57ba13f
Ignore empty error groups
2013-01-18 12:27:29 -08:00
e6f9586d83
New validation method on the root that returns errors
2013-01-18 12:14:40 -08:00
23668d4ed0
Fix some newline weirdness with StringBlockEditor
2013-01-15 21:53:11 -08:00
adec64baa4
Add the #get method to StringBlockEditor
2013-01-15 20:34:41 -08:00
335503a688
StringBlockEditor to help plugins with owning parts of files
2013-01-15 17:49:46 -08:00
25fcb59e38
Use registries for the V2 plugin manager
2013-01-13 13:05:31 -08:00
f3b340aae0
Use registry merging for provider configs
2013-01-13 13:00:06 -08:00
45879132a3
Registries support merging
2013-01-13 12:58:48 -08:00
e66c5066e4
Plugin configuration can have scopes now, ex. provider
2013-01-13 12:38:17 -08:00
0c612f695f
Util::ScopedHashOverride
2013-01-11 15:44:35 -08:00
cf2c5a10c1
One more final fix for 1.8.7
2012-12-30 18:16:51 -10:00
037dbf24db
Fix some failing tests for 1.8.7.
2012-12-30 18:13:27 -10:00
a30a92bb4f
active_machines should return name as a symbol
2012-12-30 11:51:22 -10:00
4c46091746
Environment#primary_machine_name and use it for with_target_vms
...
This makes the single-provider and default provider semantics work with
primary VMs.
2012-12-30 11:12:56 -10:00
24c3c9a7ae
Better test coverage on with_target_vms. Added pending tests to work on.
2012-12-30 10:59:59 -10:00
07157b47ae
Only allow one provider active machine at a time.
...
Temporary limitation of Vagrant to only allow one active machine with a
provider at a time. That means you cant `up` a machine with both vmware
and virtualbox at the same time. In the future you will be able to but
to avoid various edge cases for now we're disallowing it.
2012-12-30 10:52:01 -10:00
9257fe3d98
Environment#active_machines
...
This returns a list of active machines for the environment. An active
machine is a machine that at one point was created by Vagrant.
2012-12-27 19:28:05 -10:00
4e649cc987
Upgrade V1-style dotfile to V2
...
See the code and comments for details on how this is done. As usual, we
are very careful about this so as not to inadvertently destruct real
user data.
2012-12-26 22:41:42 -08:00
cd969e1e05
Remove the dotfile_name configuration option.
...
The dotfile is gone now so the configuration option is obselete
2012-12-26 21:45:24 -08:00
c0c3e7bf43
Remove Vagrant::DataStore
...
We just don't use it yet and the old implementation was sketchy. I was
not happy with it.
2012-12-26 21:45:24 -08:00
3baa31460f
Store machine ID in "id" file in data directory.
...
Instead of storing an "active" hash in the local_data of an Environment,
we now place the ID of a machine in the "id" file of the machine data
directory. This file is read upon re-instantiation in order to load the
proper state.
2012-12-26 21:45:23 -08:00
cc18492c7a
Local data path introduced
...
The local data path is set to the `ROOT_DIR/.vagrant` by default and is
a directory where Vagrant can store environment-local state. This can be
overriden on a per-Environment basis using the `local_data_path`
option.
2012-12-26 21:45:23 -08:00
8fe0f86dbd
The --provider flag for `up` now actually does something.
2012-12-23 21:23:08 -08:00
3808ea377f
Upgrade all other default configurations to V2
2012-12-23 16:29:26 -08:00
6478139cee
Move config.vm.customize to VirtualBox specific option
2012-12-23 16:29:25 -08:00
dffbf6eb09
Allow COnfirm calls to be forced by specifying a force_key
2012-12-23 16:29:25 -08:00
50d7b0aba4
Fix bug where Call didn't propagate recovery. Warden has no recovery.
...
The issue here is that when a middleware failed and a recovery sequence
started, it would halt at the "call" step because the "Call" didn't
properly recover the child sequence.
An additional issue was that a Warden had no "recover" method, meaning
embedded Wardens wouldn't recover their stacks properly.
2012-12-23 16:29:25 -08:00
a6e0d3908f
Create a "dummy" provider for tests
2012-12-23 16:29:24 -08:00
ac7958a43b
Environment#primary_machine takes a provider argument. Use it.
2012-12-23 16:29:24 -08:00
128c06e78d
Environment#default_provider
2012-12-23 16:29:24 -08:00
1ee470a551
Begin work on supporting provider-specific configuration
...
This works by registering a `config` with `:provider => true` with the
same name as your provider. Vagrant will then automatically configure
the provider when `config.vm.provider` is used.
2012-12-23 16:29:24 -08:00
99c5cddb7d
V2 config supports the concept of an "UNSET" value when doing merges
...
This value should be used as a default, and allows the parent `merge`
method to "just work" most of the time.
2012-12-23 16:29:24 -08:00
819140bbc2
Don't error if box doesn't exist on load
2012-12-23 16:29:23 -08:00
0ec639c8e1
BoxCollection#find should only find V1 box when searching for vbox
2012-12-23 16:29:23 -08:00
e2dd0b0cf2
Get machine tests passing again with new environment
2012-12-23 16:29:23 -08:00
f2b91d26fd
Load the configuration per machine, so that provider boxes work
...
Boxes are provider-specific, and we don't know the provider until
Environment#machine is called, so we need to build up the machine
configuration during this time.
2012-12-23 16:29:23 -08:00
0180ed849d
Move config loader order out from an ivar into a param for the #load
...
method
2012-12-23 16:29:23 -08:00
1559f7b7a7
Get rid of the old Environment#vms calls. Use #machine everywhere.
2012-12-23 16:29:23 -08:00
5002b8d9c2
Fix failing test on 1.8.7
2012-12-20 22:05:05 -08:00
2ef6ff10f0
Cache the machine objects created in Environment#machine.
...
Each machine and provider backing should be represented by only one
machine.
2012-11-07 22:08:06 -08:00
67855be77b
Add the Environment#machine method
...
This will eventually replace the Environment#vms method. Because of the
introduction of providers, the environment doesn't know what the backing
of the machines will be (and they're _machines_ now, not _vms_).
Instead, users of Environment will now call `#machine` on the
environment to retrieve a machine with the given backing provider as it
needs it.
2012-11-07 21:45:09 -08:00
f52f8ecc97
Get tests passing again, use V2 config in tests.
2012-11-07 20:06:55 -08:00
53860f90ab
V2 loader now properly upgrades V1 configuration.
...
This is done by calling the `upgrade` method on the _old_ configuration
classes. The old configuration classes are given the complete new
configuration and can set whatever settings they need to on it.
2012-11-07 20:01:39 -08:00
253b433864
Support configuring the version for test plugins.
2012-11-06 21:32:26 -08:00
02e499da47
Update tests to create V2 plugins
2012-11-06 21:19:27 -08:00
be294e002a
Build the v2 interface, which is just a copy of V1 for now.
2012-11-06 20:51:03 -08:00
efa0a6b8c2
Clean up tests a bit more
2012-11-04 08:10:48 -08:00
0ca783e8d1
V1 config loader loads upgradable config if not V1
...
This is to prepare for the upcoming V2 configuration. When we're in V2,
we only want to load _upgradable_ configuration keys.
2012-11-03 21:54:32 -07:00
59d5c5ed92
Add support for "upgrade safe" config classes to plugins.
...
These are classes that use NO core classes of Vagrant, and are therefore
safe to load for upgrades. i.e. a V2 core can load a V1 config class
that is deemed upgrade safe without crashing Vagrant.
2012-11-03 21:39:06 -07:00
68923ff556
Additional tests for the plugin manager.
2012-11-03 21:30:22 -07:00
c803b0508a
Much cleaner plugin part querying syntax.
...
Before we were manually going over every plugin and getting each piece,
all over the place. Now we have a central manager that will give us all
the pieces we want. There is still some cleanup to do here but this is
much better overall.
2012-11-03 21:25:28 -07:00
6ca671e2b8
Don't allow registry to be given value in non-block form
2012-11-03 20:46:24 -07:00
6df6f6764f
Remove plugin activation. It really isn't necessary.
...
It was only used in a couple places and it isn't necessary since you can
do the loading within the actual blocks themselves.
2012-11-03 20:29:34 -07:00
5cf737d4fc
Remove the buildbot stuff, we'll move away from it.
2012-11-03 13:47:57 -07:00
bad38ca935
fixes an exception issue when a non-directory appears in the boxes dir
2012-09-25 21:13:38 -05:00
50c04ac927
`box.url` can use `file://` URI
2012-08-24 09:53:50 -04:00
1cdd69bc91
All tests passing
2012-08-19 19:17:19 -07:00
47fe278667
`vagrant box add` works again. Box verification remove temporarily.
...
The built-in middleware sequences will now be hardcoded onto
Vagrant::Action. Other plugins can hook into these sequences to provide
verification and so on. So the VirtualBox plugin will hook into that
action sequence and add verification.
2012-08-18 16:13:14 -07:00
b659191a02
`vagrant up`!
2012-08-14 22:38:41 -07:00
85a499ffb8
Properly handle the case that VM doesn't exist for the VB driver.
2012-08-14 20:27:28 -07:00
2fc18f7207
`destroy` gets a little farther, and properly halts the VM
2012-08-13 23:18:50 -07:00
296878cff5
Add basic loop detection for distro_dispatch
2012-08-12 19:03:22 -07:00
f9752d78d8
Properly resolve and load the guest class impl for Machines
2012-08-12 18:35:19 -07:00
28f341ec75
Begin working on the #guest method for Machine instances
2012-08-12 16:46:00 -07:00
de0865b9e2
Test to ensure that the environment hash is returned from Machine#action
...
This will allow methods that call into the action to use any results
that might be set onto the environment hash by any of the middleware.
2012-08-11 20:33:09 -07:00
64afd578b3
Just always return the SSH communicator for machines for now.
...
In the future we'll actually find a matching communicator but for now
since we're just focusing on machine abstraction, we just return SSH.
2012-08-08 21:57:08 -07:00
5ae3e0e80c
Allow the definition of communicators in plugins
2012-08-08 21:52:25 -07:00
a1cef830e3
Add the Communicator plugin API.
...
This allows communication protocols to be defined for the machine. This
is how things like SSH will be implemented.
2012-08-08 21:28:28 -07:00
b1ced46d7c
SSHExec raises proper exception if SSH is not yet ready
2012-08-06 10:29:25 -07:00
8d50c4774e
Tests for SSHExec, and forward in the ssh options
2012-08-05 18:37:41 -07:00
e0ec679838
`vagrant ssh` with full console works with new provider.
...
This works by now calling the `:ssh` action on the provider. This action
is allowed to do whatever it pleases, but should at some point probably
call the `SSHExec` built-in middleware.
The `SSHExec` built-in middleware was added. This uses the information
returned by `Machine#ssh_info` and uses the `Vagrant::Util::SSH` helper
to exec into the remote machine. The provider should do any work upfront
in verifying that the machine is ready to be SSHed into.
2012-08-05 13:45:24 -07:00
a1145615d0
Machine#action supports passing in extra env vars for action runner
2012-08-05 13:16:08 -07:00
984c4f4025
Add Util::SSH which has methods for checking key permissions and exec
2012-08-05 13:12:53 -07:00
1a2a8b49c0
Provider API to return SSH info, must implement `ssh_info`.
...
Since SSH is such a critical part of Vagrant, each provider must
implement a method that returns the proper way to SSH into the machine.
2012-08-05 12:41:05 -07:00
9db982f7a4
Expose the provider via the machine object.
2012-08-04 11:16:31 -07:00
d8cb02d55d
Load the machine ID prior to initializing the provider
2012-08-03 14:22:17 -07:00
e5f250121a
Call now only yields the environment
2012-07-27 19:34:46 -07:00
118377e6f0
Destroy sequence asks the user for confirmation.
2012-07-27 19:29:40 -07:00
e20326d577
Added the Confirm built-in middleware
...
This middleware asks the user a question and expects a Y/N answer. This
middleware can be used with the Call middleware.
2012-07-27 19:16:44 -07:00
Mitchell Hashimoto
Michael Stillwell
Andy Williams
skim
Nate Smith