When running in server mode, make the Environment validate a client
option is provided on instantiation. If no client option is provided,
raise an exception since it is required in server mode.
Since an initial Environment is required during start, include a
stub value for the client so the initial Environment can be instantiated
successfully and the GRPC service can be started.
printf the special character for "things" is % + STUFF, so if you want just a % you have to use the special character % and then combine it with % to get a single %
Previously the clear line function would insert extra lines making
output look like
"""
==> default: Importing base box 'hashicorp/bionic64'...
Progress: 10%!(NOVERB)
Progress: 20%!(NOVERB)
Progress: 30%!(NOVERB)
Progress: 40%!(NOVERB)
"""
now, output for progress looks like
"""
==> default: Importing base box 'hashicorp/bionic64'...
Progress: 10%!(NOVERB)
Progress: 20%!(NOVERB)
Progress: 30%!(NOVERB)
Progress: 40%!(NOVERB)
"""
When computing the solution set, if a gem is already loaded, make sure
to use the specification of the loaded one instead of the first
available as otherwise there is a risk that when multiple matches are
available the specification for the wrong version may be picked.
When this happens an error message will be triggered that looks like
can't activate json-2.3.0, already activated json-2.5.1
This can occur for distribution packaged vagrants as well as installs
for development purposes where the ruby install may contain a default
gem spec of an older version than is needed.
Fixes: #12521Fixes: vagrant-libvirt/vagrant-libvirt#1390
VirtualBox introduced a restriction on the valid range for hostonly
networks. When using a version of VirtualBox which includes this
restriction a check is performed on the defined IP address to validate
it is within either the default range (as defined in the VirtualBox
documentation) or the values defined in the network configuration
file.
Set flag on RSA keys of deprecated RSA SHA1 support when loading
keys based on server version of the transport. This ensures keys
are properly flagged. Flag name has been updated to provide context
on usage.
Version matching on the OpenSSH server version has also been updated
to handle customized naming in the version string (as seen in the
Windows port) and to properly handle when no match is found.
Fixes#12344#12408#12381
Keep the constraint on net-ssh tight so we can be confident that the
patching will be successful and that a net-ssh release won't inadvertently
cause our local updates to become non-functional.
Fix patch to only update the behavior for RSA based keys when the server
is recent enough to include the signature changes
Modifies `OpenSSL::PKey::RSA` to provide a `#signature_algorithm`
method which provides the signature algorithm value expected by
OpenSSH. The `#ssh_do_sign` method is updated to use the set
algorithm (SHA256) and `#to_blob` is updated to include the
signature algorithm instead of the key type.
Parts of the stdlib which have been externalized but are still
included within Ruby introduce issues when pinning Vagrant's
dependencies to resolve plugin installs. When determining
Vagrant's dependency list prior to solution generation, check
the specification and ignore any default gems to prevent
pinning versions that are not actual dependencies.
When uploading box file, check if the size is greater than
5GB. If the size is larger and the direct to storage option
is enabled, disable the option due to current 5GB restriction
on direct uploads.
This also checks if the redirect notification has been displayed
before inspecting the source and location to prevent repeat checks
after the notification has been sent.
This sets the `authenticate_box_url` hook as deprecated and also
disables the cloud auth middleware from adding an access token
as a URL parameter by default. An environment variable has been
added which can be used for re-enabling the access token URL
parameter behavior if required for some legacy system which does
not support the authorization header.
When starting up, and before any loading, find our current
specification and activate all the internal dependencies
while also collecting the activated specifications. Store
these for later use when doing plugin resolutions. We bypass
the builtin list when running in bundler since they will
still show up as not activated, but we use the entire list
regardless.
Inspect the error message received when failing to install a
plugin. If it's something we can determine the cause, send
back a cleaner error message to the user on how to resolve.
To prevent resolution issues with the introduction
of a prerelease constraint, update the net-ssh constraint
to be a minimum at the latest release. Include monkey
patches to include support for wanted host key algos.
The monkey patches are only applied to the latest
net-ssh currently and will be ignored once the current
prerelease has been fully released.
Maintain the solution file persisting dependency information on
disk but update the runtime representation to
Gem::Resolver::DependencyRequest instances which are expected
by the sets when locating matches.
Properly abide by prerelease setting in customized sets and
force prerelease matching when in the builtin set. If a request
is matched on a prerelease, and the request itself is not set
to allow prereleases, update it to ensure successful resolution.
The raw action name tracking should be sufficient for preventing
multi-insertions where only a single modification should occur.
With action name, hooks _should_ be able to be applied in multiple
builder stacks as they are expanding.
Fixes#12035
The raw actions are used for applying the original trigger behavior
which can insert before and/or after the entire set of actions. When
processing the stack items, mark when the raw action has been applied
to prevent it from being applied again. Triggers around the raw actions
should only ever be applied _once_.
Fixes#12034
When initializing for internal plugin resolution inspect contraints
on all defined dependencies. If a prerelease constraint is detected,
automatically enable prerelease resolution.
When expanding stack track the origin action name and only apply
it once the stack has completed its expansion. The local env data
is marked with origin action to prevent it from being applied in
nested builders as they are expanded. The value of the stored action
name is checked and invalidated if another action is applied to the
same env in the future so hooks / triggers for that action are
applied as expected.
This PR is dependent on the 2.1.0 release of the vagrant_cloud
library. It updates the `cloud` command to use the new interface
for interacting with the Vagrant Cloud API. It also adds support
for direct to backend storage uploads, and defaults to this
method.
Also included is a bit of cleanup refactoring, addition of method
documentation, and fixing up some small issues around custom username
usage within the internal client for authentication.
Passwords are (and should) never stored be as part of folder data.
I think there is also a case to be made about desensitizes
information here can lead to leaking of credentials. For example if
an exported folder is named "vagrant" and the users password is
"vagrant", the synced_folder cache will show "****" in place of the
folder name, indicating that it is also password.
This includes updates for resolving all warnings provided by Ruby
for deprecations and/or removed methods. It also enables support
for Ruby 2.7 in the specification constraint as all 2.7 related
warnings are resolved with this changeset.
This commit adds a unique error message for an empty box value. It
requires modifications to vagrantfile.rb because some Vagrantfile config
is used before validation occurs.
This commit adds a new error message to be raised if a VM has no
supported storage controllers. This lets us differentiate between two
different "controller not found" scenarios:
1. If we are looking for a controller that we're expecting to find (i.e.
one that was recorded in the disk metadata file)
2. If we are poking around for the *best* controller to use in a
configuration task
This commit adds a new error type that can be raised whenever a storage
controller of the required type is not found. This indicates that a user
needs to either add the storage controller manually or change their disk
configuration.
It also removes the last hardcoded instance of "SATA Controller" as a
default argument.
This commit changes the SSH communicator to raise an error if Vagrant
doesn't receive an exit status from an SSH command, for example if the
command is terminated by the OOM-killer.
Prior to this commit, if a created but exited container bound a port,
and a new container grabed that same port (say for an ssh port forward),
when the initial container came back up it would fail because the port
also got bound to the second container. This commit fixes that behavior
by first looking at what containers are already bound prior to creating
a container.
During a plugin install, if the plugin is already installed and
activated, no specification will be returned as there was nothing
new installed. In this situation, look for the requested plugin
within the activated specifications. If it is found, then proceed
since the plugin is installed. If it is not found, return an error.
Sets prerelease on the request set based on Vagrant's version. This
allows installing plugins which may include a vagrant version constraint
when running on a development version without needing to a manual local
installation of the gem.
Retains the original default value of 15 seconds for SSH connect
timeout. Allows users to modify this timeout via SSH communicator
option. Enforces integer values for timeout and validates custom
values are greater than 0.
These updates allow the after trigger to behave the same as the
original with regards to the execution location of the trigger
within the execution stack.
Adds method to shared helpers for adding procs to be evaluated
which can add default modifications to the option parser used
by commands. Customized option parser class within Vagrant
handles processing defined procs to set options.
This adjusts how triggers are implemented during a normal run. Any
defined triggers which are applicable are located and injected into
the run stack as the stack is built, including hook type triggers.
Support is included for dynamic hook lookup.
The data type used when defining triggers has also been relaxed to
support symbols, strings, or constants.
This commit changes the behavior of the port check to check all possible
IPv4 network interfaces when the host IP is `nil` or `0.0.0.0`. This
means that if the desired port is available on any network interfaces, a
forward from 0.0.0.0 will use that interface.
If the port is open (in use) on all interfaces, then it's treated as a
collision and will either throw an error or auto-correct the port, based
on the Vagrantfile configuration.
Windows commands that run over SSH are wrapped in a script that writes a
special marker to the two output streams (stdout and stderr). This
allows Vagrant to consume the output streams.
Unfortunately, this leads to a sort of chicken-and-egg problem where no
commands can be run before a wrapper script exists. For example, you
can't make a destination directory to upload the wrapper script without
first creating a wrapper script to make the directory. :)
This commit changes the behavior of the WinSSH communicator to assume
that the destination directory already exists for provisioning scripts.
It also moves the default `upload_path` from the shell provisioner
config so we can have OS-specific defaults.
Finally, it introduces a Windows-specific #upload method which will
properly use a Windows path separator on a non-Windows host.
Removes dynamic calls of before/after hooks and replaces it with
proper lookups for hooks defined for the action to run. If hooks
are found for an action, the action is placed in a new builder
and the hooks are applied. The new stack is extracted, finalized,
and then executed.
This change allows the vagrant user to see the systemd process in the
event that the hidepid mount option is enabled.
Also adds sudo: true to other tests that use `systemd?`
This pull request adds an enhancement to the internal Bundler class
to cache solution sets. This prevents Vagrant from generating a
solution for configured plugins on every run. Modifications to
the configured plugin list (global or local) will result in the
cached solution being invalidatd and resolved again.
Also included is the removal of the GEMRC modifications required
for Windows.
Prior to this commit, the check used to look for the config option
`name` in a provisioner config would accidentally create a "DummyConfig"
based on how vagrant handles missing config options. This commit fixes
that by instead using the `respond_to?` method to check for the
existance of the config option name.
config
Prior to this commit, if a user had configured a provisioner that had a
config with a `name` option, it would not properly set the top level
provisioner classes name config option which would lead to some
understanibly confusing results when trying to `--provision-with`. This
commit fixes that by checking to see if the top level name isn't set,
look to see if that provisioners config defines a name, and use that
instead.
This commit changes the behavior of the builtin SSHRun action to use a Windows
shell if the WinSSH communicator is active. This allows for running one-off SSH
commands with Windows Command Prompt or PowerShell. By default, this will not
allocate a TTY for any SSH commands.
Example usage:
```
vagrant ssh -c 'dir "c:\program files"'
```
Updates docs to reflect a change made in
a55a53e6a4.
This reverts commit 81553263ab.
This fixes a regression with Windows port detection which led to port
collisions not being fixed on `vagrant up`.
PR #8517 changed `IsPortOpen#is_port_open?` to rescue
Errno::EADDRNOTAVAIL, but when we merged it into master, there was code
in `HandleForwardedPortCollisions#port_check` that depended on that
error bubbling up.
This commit catches the Errno::EPERM raised by the operating system if
the machine folder is inaccessible and displays it as a more friendly
error message.
This can be an issue on macOS Catalina if virtual machine files are kept
in a special directory (Documents/Downloads/Desktop) that Vagrant's
embedded Ruby is not allowed to access.
Prior to this commit, if a user had recently checked for updates, there
was no way to force Vagrant to re-check without manually deleting a
state file in the local `.vagrant` data dir. This commit fixes that by
giving users the ability to force check for updates for a given box with
a flag to the `vagrant box outdated` command.
Prior to this commit, if Vagrant received checksum options from Vagrant
Cloud that were simply empty strings, it would try to validate its
checksum with those options. This commit fixes that by ignoring empty
string values.
This commit changes the behavior of `vagrant box outdated --global` so
it gets the latest version for the current machine's provider, rather
than the latest version for any provider.
This commit checks the local box collection to see if the latest box
update has already been downloaded. If it has, Vagrant will display the
destroy/recreate message instead of the "run vagrant update" message.
Prior to this commit, the file_checksum class used the `Vagrant::Errors`
class as if it were apart of the Vagrant module. However, since the
file_checksum class is an interface and not part of the Vagrant module,
it doesn't have access to that Error class like other Vagrant modules.
This commit fixes that by ensuring the `"vagrant/errors"` class is
loaded, and that the proper namespace is used.
Prior to this commit, vagrant was not grabbing all of the tokens on
Windows for showing the full drive because the ruby api for it behaves
differenly on windows compared to other platforms. This commit changes
that by ensuring the letter drive is attached to the path when showing
an exception.
Prior to this commit, `vagrant box prune --force` would not prompt a
user to prune Vagrant boxes, even if that box was in use. There was no
way to prune boxes, ignore the prompt, but keep in-use boxes. This
commit adds a new flag that can be combined with `--force`, that will
keep in-use boxes but prune older boxes without prompting the user.
Inserting at the right index places new item *before* the index, so no
need to decrement and then insert, since `insert` takes care of shifting
down elements of the array.
If the local data directory is unavailable, there will be no local
file to use within the plugin manager. Check for local file before
returning result, otherwise just return an empty hash.
Allow Vagrantfile#machine_config to load properly when the requested
provider may not be currently available. Update the Environment to
utilize this when searching for plugin information to properly allow
box provided Vagrantfiles to define required plugins.
When installed outside of the official installer and not running
within a bundler environment, properly activate core dependencies
and properly enforce constraints.
When the provider supports parallel actions and actions are being
run in parallel, do not immediately kill the process on failure.
Instead terminate the action thread and log the exit code. Once
all running actions have completed, the process will then exit
with the stored exit code.
Since plugin installation happens when the environment is first
initialized, attempt to determine the provider in use and load
any box provided Vagrantfiles to include any plugin configuration
they may include.
Prior to this commit, if a guest name was given as a symbol, the
filter_triggers method would fail to properly match it with the only_on
option, as it is not a valid type to the #String.match method. This
commit fixes that by converting the parameter to a string so that it can
be properly matched on the guest.
Prior to this commit, the `abort` option for triggers would just call
`exit`, which would end up raising a SystemExit exception, signaling
Vagrant to abort. This broke down however in a multithreaded context
like when running multiple guests at once on supported providers,
resulting in Vagrant failing to exit cleanly and instead raise an
exception. This commit changes that by instead using `Process.exit!` to
abort Vagrant.
Before writing synced folder configuration data to the local
data directory run content through the credential scrubber to
remove any sensitive content before write.
Prior to this commit, when Vagrant attempted to use the Gem library, it
would attempt to pass in a gemrc through an environment variable that
the rubygems library would try to split and parse. This is normally
fine, as the method in question would return empty if that file did not
exist. However if the user had a file that matches the drive that
Vagrant was installed on, rubygems would fail saying the folder was not
a file (or a gemrc, in this case).
This commit works around that by instead configuring the gemrc location
through ruby with `Gem.configuration`.
Related rubygems issue
[#2733](https://github.com/rubygems/rubygems/issues/2733)
If the provider does not include a metadata.json file prior to
compressing the box, determine current provider and write
metadata.json file before compressing.
Prior to this commit, if a box some how got on disk that had an
incorrect or invalid version number that did not match Gem::Version,
Vagrant would throw an exception when attempting to generate a list of
the boxes on disk. This commit fixes that by looking at the version from
the path generated, and shows a warning to the user about the box and
skips it from the list so they at least know about the problematic box
and can still get a list of boxes.
