If a provider configuration is inherited from another module, any
interpolations in that config won't have variables declared locally. Let
the config only be validated in it's original location.
Registry modules can't be handled directly by the getter.Storage
implementation, which doesn't know how to handle versions. First see if
we have a matching module stored that satisfies our constraints. If
not, and we're getting or updating, we can look it up in the registry.
This essentially takes the place of a "registry detector" for go-getter,
but required the intermediate step of resolving the version dependency.
This also starts breaking up the huge Tree.Load method into more
manageable parts. It was sorely needed, as indicated by the difficulty
encountered in this refactor. There's still a lot that can be done to
improve this, but at least there are now a few easier to read methods
when we come back to it.
The detection of registry modules will have to happen in mutliple
phases. The go-getter interface requires that the detector return the
final URL, while we won't know that until we verify which version we
need. This leaves the regisry sources broken, to be re-integrated in a
following commit.
wire up HTTP so we can test the mock discovery service
test lookupModuleVersions
Add a versions endpoint to the mock registry, and use that to verify the
lookupModuleVersions behavior.
lookupModuleVersions takes a Disco as the argument so a custom Transport
can be injected, and uses that transport for its own client if it set.
test looking up modules with default registry
Add registry.terrform.io to the hostname that the mock registry resolves
to localhost.
ACC test looking up module versions
Lookup a basic module for the available version in the default registry.
This test highlights how changing an intermediate source path prevents
reloading of submodules. While this is somewhat of an edge case now, it
becomes quite common in the cacse where module versions are updated.
Adds basic detector for registry module source strings. While this isn't
a thorough validation, this will eliminate anything that is definitely
not a registry module, and split out our host and module id strings.
lookupModuleVersions interrogates the registry for the available
versions of a particular module and the tree of dependencies.
Submodules were located by using their module path as the storage key.
Now that modules may have versions, a submodule needs to know how to
locate the corect source depending on the versions of its ancestors in
the tree.
Add a version field to each Tree, and a pointer back to the parent Tree
to step back through the ancestors. The new versionedPathKey method uses
this information to build a unique key for each module, dependent on the
ancestor versions.
Not only do stored modules need to know their version if it exists, but
any relative source needs to know all the ancestor versions in order to
resolve correctly.
The getter.Storage abstraction is proving entirely inadequate here, but
we can't replace it wholesale at the moment.
The Tree loader needs to know the location of the manifest before it can
start loading any modules. Since the version will have to be part of the
hashed storage key, there is no way to know what version of each module
are stored. The storageDir function will extract the StorageDir field
from the underlying FolderStorage instance for the tree to locate the
manifest.
To add registry support, a workaround in the local module storage was
added to record the subdirectory containing the module source from
within the archive file. Here we replace that temporary implementation
with the full manifest needed to record the necessary module metadata
for module loading.
In order to support versioned modules, the actual stored version needs
to be recorded. This can't be derived from the configuration, because
the configuration only contains the constraints, and at load time we need
to be able to enumerate the stored modules and all versions in order to
resolve them.
While the local storage key will be derived from the source and version,
that information is lost once it's hashed. While the entire storage
layer could be replaced to encode the needed data in the path itself,
this provides a minimal change to work with the existing storage code.
Now that we can enforce local modules being relative or absolute paths,
we can be assured that any module source matching a registry pattern
must be found in the registry. This allows us to surface more useful
errors to the user, rather than simply stating that a source string
isn't valid.
Breaking change for 0.11.
Local files were checked first to avoid the possibility of breaking a
module with a local source that looked like a registry ID. Now we can
enfore that any source iwth the pattern "namespace/identifier/provider"
must be a registry module.
While merging the cached Input configs in the correct order prevents
overwriting existing config values, it doesn't prevent an earlier
provider from inserting unwanted values into later provider
configurations.
Diff the key-values returned by Input with the pre-input config, and
store only the "answers" that were added during the Input call.
Always call Input, even if we already have some values, since a
previously cached config may not be complete.
Previously when looking up cached provider input, the Input was taken in
its entirety, and only provider configuration fields that weren't in the
saved input were added. This would cause providers in modules to use the
entire configuration from parent modules, even if they themselves had
entirely different configs.
Note: this is only marginally beter than the old behavior. It may be
slightly more correct, but stil can't account for the user's intent, and
may be adding configured values from one provider into another.
Change the PathCacheKey to just join the path on a non-path character
(|), which makes for easier debugging.
Use the configured providers directly, rather than looking for inherited
provider configuration during graph evaluation.
First remove the provider config cache, and the associated
SetProviderConfig and ParentProviderConfig methods on the eval context.
Every provider must be configured, so there's no need to look for
configuration from other provider instances.
The config.ProviderConfig struct now has a Scope field which stores the
proper path for the interpolation scope. To get this metadata to the
interpolator, we add an EvalInterpolatProvider node which can carry the
ProviderConfig, and an InterpolateProvider context method to carry the
ProviderConfig.Scope into the InterplationScope.
Some of the tests could be adjusted to account for the new inheritance
behavior, and some were simply no longer valid and will be removed.
The remaining tests have questions on how they should work in practice.
This mostly concerns orphaned modules where there is no longer a way to
obtain a provider. In some cases we may require that a minimal provider
config be present to handle the destroy process, but we need further
testing.
All disabled code was commented out in this commit to record any
additional comments. The following commit will be a cleanup pass.
This implements provider inheritance during config loading, rather than
during graph evaluation. At this point it's much simpler to find the
desired configuration, and once all providers are declared, all the
inheritance code in the graph can be removed.
The inheritance is dome by simply copying the RawConfig from the parent
ProviderConfig into the module. Since this happens before any
evaluation, we record the original interpolation scope in the
ProviderConfig so that it can be properly resolved later on.
Add the Version and Providers fields to the module config.
Add ProviderConfig.Scope, which will be used to record the original
path of a ProviderConfig for interpolation.
For situations where the default network-based discovery is inappropriate
or inconvenient, this allows users to provide a hard-coded discovery
document for a particular hostname in the CLI config.
This is a new config block, rather than combined with the existing
"credentials" block, because credentials should ideally live in separate
files from other config so that they can be managed more carefully.
However, this new "host" block _is_ designed to have room for additional
host-specific configuration _other than_ credentials in future, which
might include TLS certificate overrides or other such things used during
the discovery step.
The default network-based discovery is not desirable for all situations,
so this mechanism allows callers to provide a services map for a given
hostname that was obtained some other way (caller-defined) which will then
cause network-based discovery to be skipped and the given map to be
returned verbatim.
We encourage users to share the "terraform version" output as part of
filing an issue, but previously it only printed the core Terraform version
and this left provider maintainers with no information about which
_provider_ version an issue relates to.
Here we make a best effort to show versions for providers, though we will
omit some or all of them if either "terraform init" hasn't been run (and
so no providers were selected yet) or if there are other inconsistencies
that would cause Terraform to object on startup and require a re-run of
"terraform init".
Two different errors here caused this test to pass even though it was
incorrect: the wanted version string was incorrect, but the test for it
was also inverted, and so together this made the test pass even though
it was actually not testing the output at all.
James Bardin