terraform/internal/configs/config.go

558 lines
21 KiB
Go
Raw Permalink Normal View History

package configs
import (
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
"fmt"
backend/local: Check dependency lock consistency before any operations In historical versions of Terraform the responsibility to check this was inside the terraform.NewContext function, along with various other assorted concerns that made that function particularly complicated. More recently, we reduced the responsibility of the "terraform" package only to instantiating particular named plugins, assuming that its caller is responsible for selecting appropriate versions of any providers that _are_ external. However, until this commit we were just assuming that "terraform init" had correctly selected appropriate plugins and recorded them in the lock file, and so nothing was dealing with the problem of ensuring that there haven't been any changes to the lock file or config since the most recent "terraform init" which would cause us to need to re-evaluate those decisions. Part of the game here is to slightly extend the role of the dependency locks object to also carry information about a subset of provider addresses whose lock entries we're intentionally disregarding as part of the various little edge-case features we have for overridding providers: dev_overrides, "unmanaged providers", and the testing overrides in our own unit tests. This is an in-memory-only annotation, never included in the serialized plan files on disk. I had originally intended to create a new package to encapsulate all of this plugin-selection logic, including both the version constraint checking here and also the handling of the provider factory functions, but as an interim step I've just made version constraint consistency checks the responsibility of the backend/local package, which means that we'll always catch problems as part of preparing for local operations, while not imposing these additional checks on commands that _don't_ run local operations, such as "terraform apply" when in remote operations mode.
2021-09-30 02:31:43 +02:00
"log"
various: helpers for collecting necessary provider types Since schemas are required to interpret provider, resource, and provisioner attributes in configs, states, and plans, these helpers intend to make it easier to gather up the the necessary provider types in order to preload all of the needed schemas before beginning further processing. Config.ProviderTypes returns directly the list of provider types, since at this level further detail is not useful: we've not yet run the provider allocation algorithm, and so the only thing we can reliably extract here is provider types themselves. State.ProviderAddrs and Plan.ProviderAddrs each return a list of absolute provider addresses, which can then be turned into a list of provider types using the new helper providers.AddressedTypesAbs. Since we're already using configs.Config throughout core, this also updates the terraform.LoadSchemas helper to use Config.ProviderTypes to find the necessary providers, rather than implementing its own discovery logic. states.State is not yet plumbed in, so we cannot yet use State.ProviderAddrs to deal with the state but there's a TODO comment to remind us to update that in a later commit when we swap out terraform.State for states.State. A later commit will probably refactor this further so that we can easily obtain schema for the providers needed to interpret a plan too, but that is deferred here because further work is required to make core work with the new plan types first. At that point, terraform.LoadSchemas may become providers.LoadSchemas with a different interface that just accepts lists of provider and provisioner names that have been gathered by the caller using these new helpers.
2018-06-22 02:39:27 +02:00
"sort"
version "github.com/hashicorp/go-version"
"github.com/hashicorp/hcl/v2"
"github.com/hashicorp/terraform/internal/addrs"
backend/local: Check dependency lock consistency before any operations In historical versions of Terraform the responsibility to check this was inside the terraform.NewContext function, along with various other assorted concerns that made that function particularly complicated. More recently, we reduced the responsibility of the "terraform" package only to instantiating particular named plugins, assuming that its caller is responsible for selecting appropriate versions of any providers that _are_ external. However, until this commit we were just assuming that "terraform init" had correctly selected appropriate plugins and recorded them in the lock file, and so nothing was dealing with the problem of ensuring that there haven't been any changes to the lock file or config since the most recent "terraform init" which would cause us to need to re-evaluate those decisions. Part of the game here is to slightly extend the role of the dependency locks object to also carry information about a subset of provider addresses whose lock entries we're intentionally disregarding as part of the various little edge-case features we have for overridding providers: dev_overrides, "unmanaged providers", and the testing overrides in our own unit tests. This is an in-memory-only annotation, never included in the serialized plan files on disk. I had originally intended to create a new package to encapsulate all of this plugin-selection logic, including both the version constraint checking here and also the handling of the provider factory functions, but as an interim step I've just made version constraint consistency checks the responsibility of the backend/local package, which means that we'll always catch problems as part of preparing for local operations, while not imposing these additional checks on commands that _don't_ run local operations, such as "terraform apply" when in remote operations mode.
2021-09-30 02:31:43 +02:00
"github.com/hashicorp/terraform/internal/depsfile"
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
"github.com/hashicorp/terraform/internal/getproviders"
)
// A Config is a node in the tree of modules within a configuration.
//
// The module tree is constructed by following ModuleCall instances recursively
// through the root module transitively into descendent modules.
//
// A module tree described in *this* package represents the static tree
// represented by configuration. During evaluation a static ModuleNode may
// expand into zero or more module instances depending on the use of count and
// for_each configuration attributes within each call.
type Config struct {
// RootModule points to the Config for the root module within the same
// module tree as this module. If this module _is_ the root module then
// this is self-referential.
Root *Config
// ParentModule points to the Config for the module that directly calls
// this module. If this is the root module then this field is nil.
Parent *Config
// Path is a sequence of module logical names that traverse from the root
// module to this config. Path is empty for the root module.
//
// This should only be used to display paths to the end-user in rare cases
// where we are talking about the static module tree, before module calls
// have been resolved. In most cases, an addrs.ModuleInstance describing
// a node in the dynamic module tree is better, since it will then include
// any keys resulting from evaluating "count" and "for_each" arguments.
Path addrs.Module
// ChildModules points to the Config for each of the direct child modules
// called from this module. The keys in this map match the keys in
// Module.ModuleCalls.
Children map[string]*Config
// Module points to the object describing the configuration for the
// various elements (variables, resources, etc) defined by this module.
Module *Module
// CallRange is the source range for the header of the module block that
// requested this module.
//
// This field is meaningless for the root module, where its contents are undefined.
CallRange hcl.Range
// SourceAddr is the source address that the referenced module was requested
Refactoring of module source addresses and module installation It's been a long while since we gave close attention to the codepaths for module source address parsing and external module package installation. Due to their age, these codepaths often diverged from our modern practices such as representing address types in the addrs package, and encapsulating package installation details only in a particular location. In particular, this refactor makes source address parsing a separate step from module installation, which therefore makes the result of that parsing available to other Terraform subsystems which work with the configuration representation objects. This also presented the opportunity to better encapsulate our use of go-getter into a new package "getmodules" (echoing "getproviders"), which is intended to be the only part of Terraform that directly interacts with go-getter. This is largely just a refactor of the existing functionality into a new code organization, but there is one notable change in behavior here: the source address parsing now happens during configuration loading rather than module installation, which may cause errors about invalid addresses to be returned in different situations than before. That counts as backward compatible because we only promise to remain compatible with configurations that are _valid_, which means that they can be initialized, planned, and applied without any errors. This doesn't introduce any new error cases, and instead just makes a pre-existing error case be detected earlier. Our module registry client is still using its own special module address type from registry/regsrc for now, with a small shim from the new addrs.ModuleSourceRegistry type. Hopefully in a later commit we'll also rework the registry client to work with the new address type, but this commit is already big enough as it is.
2021-05-28 04:24:59 +02:00
// from, as specified in configuration. SourceAddrRaw is the same
// information, but as the raw string the user originally entered.
//
Refactoring of module source addresses and module installation It's been a long while since we gave close attention to the codepaths for module source address parsing and external module package installation. Due to their age, these codepaths often diverged from our modern practices such as representing address types in the addrs package, and encapsulating package installation details only in a particular location. In particular, this refactor makes source address parsing a separate step from module installation, which therefore makes the result of that parsing available to other Terraform subsystems which work with the configuration representation objects. This also presented the opportunity to better encapsulate our use of go-getter into a new package "getmodules" (echoing "getproviders"), which is intended to be the only part of Terraform that directly interacts with go-getter. This is largely just a refactor of the existing functionality into a new code organization, but there is one notable change in behavior here: the source address parsing now happens during configuration loading rather than module installation, which may cause errors about invalid addresses to be returned in different situations than before. That counts as backward compatible because we only promise to remain compatible with configurations that are _valid_, which means that they can be initialized, planned, and applied without any errors. This doesn't introduce any new error cases, and instead just makes a pre-existing error case be detected earlier. Our module registry client is still using its own special module address type from registry/regsrc for now, with a small shim from the new addrs.ModuleSourceRegistry type. Hopefully in a later commit we'll also rework the registry client to work with the new address type, but this commit is already big enough as it is.
2021-05-28 04:24:59 +02:00
// These fields are meaningless for the root module, where their contents are undefined.
SourceAddr addrs.ModuleSource
SourceAddrRaw string
// SourceAddrRange is the location in the configuration source where the
// SourceAddr value was set, for use in diagnostic messages.
//
// This field is meaningless for the root module, where its contents are undefined.
SourceAddrRange hcl.Range
// Version is the specific version that was selected for this module,
// based on version constraints given in configuration.
//
// This field is nil if the module was loaded from a non-registry source,
// since versions are not supported for other sources.
//
// This field is meaningless for the root module, where it will always
// be nil.
Version *version.Version
}
// ModuleRequirements represents the provider requirements for an individual
// module, along with references to any child modules. This is used to
// determine which modules require which providers.
type ModuleRequirements struct {
Name string
Refactoring of module source addresses and module installation It's been a long while since we gave close attention to the codepaths for module source address parsing and external module package installation. Due to their age, these codepaths often diverged from our modern practices such as representing address types in the addrs package, and encapsulating package installation details only in a particular location. In particular, this refactor makes source address parsing a separate step from module installation, which therefore makes the result of that parsing available to other Terraform subsystems which work with the configuration representation objects. This also presented the opportunity to better encapsulate our use of go-getter into a new package "getmodules" (echoing "getproviders"), which is intended to be the only part of Terraform that directly interacts with go-getter. This is largely just a refactor of the existing functionality into a new code organization, but there is one notable change in behavior here: the source address parsing now happens during configuration loading rather than module installation, which may cause errors about invalid addresses to be returned in different situations than before. That counts as backward compatible because we only promise to remain compatible with configurations that are _valid_, which means that they can be initialized, planned, and applied without any errors. This doesn't introduce any new error cases, and instead just makes a pre-existing error case be detected earlier. Our module registry client is still using its own special module address type from registry/regsrc for now, with a small shim from the new addrs.ModuleSourceRegistry type. Hopefully in a later commit we'll also rework the registry client to work with the new address type, but this commit is already big enough as it is.
2021-05-28 04:24:59 +02:00
SourceAddr addrs.ModuleSource
SourceDir string
Requirements getproviders.Requirements
Children map[string]*ModuleRequirements
}
// NewEmptyConfig constructs a single-node configuration tree with an empty
// root module. This is generally a pretty useless thing to do, so most callers
// should instead use BuildConfig.
func NewEmptyConfig() *Config {
ret := &Config{}
ret.Root = ret
ret.Children = make(map[string]*Config)
ret.Module = &Module{}
return ret
}
// Depth returns the number of "hops" the receiver is from the root of its
// module tree, with the root module having a depth of zero.
func (c *Config) Depth() int {
ret := 0
this := c
for this.Parent != nil {
ret++
this = this.Parent
}
return ret
}
// DeepEach calls the given function once for each module in the tree, starting
// with the receiver.
//
// A parent is always called before its children and children of a particular
// node are visited in lexicographic order by their names.
func (c *Config) DeepEach(cb func(c *Config)) {
cb(c)
names := make([]string, 0, len(c.Children))
for name := range c.Children {
names = append(names, name)
}
for _, name := range names {
c.Children[name].DeepEach(cb)
}
}
// AllModules returns a slice of all the receiver and all of its descendent
// nodes in the module tree, in the same order they would be visited by
// DeepEach.
func (c *Config) AllModules() []*Config {
var ret []*Config
c.DeepEach(func(c *Config) {
ret = append(ret, c)
})
return ret
}
// Descendent returns the descendent config that has the given path beneath
// the receiver, or nil if there is no such module.
//
// The path traverses the static module tree, prior to any expansion to handle
// count and for_each arguments.
//
// An empty path will just return the receiver, and is therefore pointless.
func (c *Config) Descendent(path addrs.Module) *Config {
current := c
for _, name := range path {
current = current.Children[name]
if current == nil {
return nil
}
}
return current
}
// DescendentForInstance is like Descendent except that it accepts a path
// to a particular module instance in the dynamic module graph, returning
// the node from the static module graph that corresponds to it.
//
// All instances created by a particular module call share the same
// configuration, so the keys within the given path are disregarded.
func (c *Config) DescendentForInstance(path addrs.ModuleInstance) *Config {
current := c
for _, step := range path {
current = current.Children[step.Name]
if current == nil {
return nil
}
}
return current
}
various: helpers for collecting necessary provider types Since schemas are required to interpret provider, resource, and provisioner attributes in configs, states, and plans, these helpers intend to make it easier to gather up the the necessary provider types in order to preload all of the needed schemas before beginning further processing. Config.ProviderTypes returns directly the list of provider types, since at this level further detail is not useful: we've not yet run the provider allocation algorithm, and so the only thing we can reliably extract here is provider types themselves. State.ProviderAddrs and Plan.ProviderAddrs each return a list of absolute provider addresses, which can then be turned into a list of provider types using the new helper providers.AddressedTypesAbs. Since we're already using configs.Config throughout core, this also updates the terraform.LoadSchemas helper to use Config.ProviderTypes to find the necessary providers, rather than implementing its own discovery logic. states.State is not yet plumbed in, so we cannot yet use State.ProviderAddrs to deal with the state but there's a TODO comment to remind us to update that in a later commit when we swap out terraform.State for states.State. A later commit will probably refactor this further so that we can easily obtain schema for the providers needed to interpret a plan too, but that is deferred here because further work is required to make core work with the new plan types first. At that point, terraform.LoadSchemas may become providers.LoadSchemas with a different interface that just accepts lists of provider and provisioner names that have been gathered by the caller using these new helpers.
2018-06-22 02:39:27 +02:00
// EntersNewPackage returns true if this call is to an external module, either
// directly via a remote source address or indirectly via a registry source
// address.
//
// Other behaviors in Terraform may treat package crossings as a special
// situation, because that indicates that the caller and callee can change
// independently of one another and thus we should disallow using any features
// where the caller assumes anything about the callee other than its input
// variables, required provider configurations, and output values.
//
// It's not meaningful to ask if the Config representing the root module enters
// a new package because the root module is always outside of all module
// packages, and so this function will arbitrarily return false in that case.
func (c *Config) EntersNewPackage() bool {
return moduleSourceAddrEntersNewPackage(c.SourceAddr)
}
backend/local: Check dependency lock consistency before any operations In historical versions of Terraform the responsibility to check this was inside the terraform.NewContext function, along with various other assorted concerns that made that function particularly complicated. More recently, we reduced the responsibility of the "terraform" package only to instantiating particular named plugins, assuming that its caller is responsible for selecting appropriate versions of any providers that _are_ external. However, until this commit we were just assuming that "terraform init" had correctly selected appropriate plugins and recorded them in the lock file, and so nothing was dealing with the problem of ensuring that there haven't been any changes to the lock file or config since the most recent "terraform init" which would cause us to need to re-evaluate those decisions. Part of the game here is to slightly extend the role of the dependency locks object to also carry information about a subset of provider addresses whose lock entries we're intentionally disregarding as part of the various little edge-case features we have for overridding providers: dev_overrides, "unmanaged providers", and the testing overrides in our own unit tests. This is an in-memory-only annotation, never included in the serialized plan files on disk. I had originally intended to create a new package to encapsulate all of this plugin-selection logic, including both the version constraint checking here and also the handling of the provider factory functions, but as an interim step I've just made version constraint consistency checks the responsibility of the backend/local package, which means that we'll always catch problems as part of preparing for local operations, while not imposing these additional checks on commands that _don't_ run local operations, such as "terraform apply" when in remote operations mode.
2021-09-30 02:31:43 +02:00
// VerifyDependencySelections checks whether the given locked dependencies
// are acceptable for all of the version constraints reported in the
// configuration tree represented by the reciever.
//
// This function will errors only if any of the locked dependencies are out of
// range for corresponding constraints in the configuration. If there are
// multiple inconsistencies then it will attempt to describe as many of them
// as possible, rather than stopping at the first problem.
//
// It's typically the responsibility of "terraform init" to change the locked
// dependencies to conform with the configuration, and so
// VerifyDependencySelections is intended for other commands to check whether
// it did so correctly and to catch if anything has changed in configuration
// since the last "terraform init" which requires re-initialization. However,
// it's up to the caller to decide how to advise users recover from these
// errors, because the advise can vary depending on what operation the user
// is attempting.
func (c *Config) VerifyDependencySelections(depLocks *depsfile.Locks) []error {
var errs []error
reqs, diags := c.ProviderRequirements()
if diags.HasErrors() {
// It should be very unusual to get here, but unfortunately we can
// end up here in some edge cases where the config loader doesn't
// process version constraint strings in exactly the same way as
// the requirements resolver. (See the addProviderRequirements method
// for more information.)
errs = append(errs, fmt.Errorf("failed to determine the configuration's provider requirements: %s", diags.Error()))
}
for providerAddr, constraints := range reqs {
if !depsfile.ProviderIsLockable(providerAddr) {
continue // disregard builtin providers, and such
}
if depLocks != nil && depLocks.ProviderIsOverridden(providerAddr) {
// The "overridden" case is for unusual special situations like
// dev overrides, so we'll explicitly note it in the logs just in
// case we see bug reports with these active and it helps us
// understand why we ended up using the "wrong" plugin.
log.Printf("[DEBUG] Config.VerifyDependencySelections: skipping %s because it's overridden by a special configuration setting", providerAddr)
continue
}
var lock *depsfile.ProviderLock
if depLocks != nil { // Should always be true in main code, but unfortunately sometimes not true in old tests that don't fill out arguments completely
lock = depLocks.Provider(providerAddr)
}
if lock == nil {
log.Printf("[TRACE] Config.VerifyDependencySelections: provider %s has no lock file entry to satisfy %q", providerAddr, getproviders.VersionConstraintsString(constraints))
errs = append(errs, fmt.Errorf("provider %s: required by this configuration but no version is selected", providerAddr))
continue
}
selectedVersion := lock.Version()
allowedVersions := getproviders.MeetingConstraints(constraints)
log.Printf("[TRACE] Config.VerifyDependencySelections: provider %s has %s to satisfy %q", providerAddr, selectedVersion.String(), getproviders.VersionConstraintsString(constraints))
if !allowedVersions.Has(selectedVersion) {
// The most likely cause of this is that the author of a module
// has changed its constraints, but this could also happen in
// some other unusual situations, such as the user directly
// editing the lock file to record something invalid. We'll
// distinguish those cases here in order to avoid the more
// specific error message potentially being a red herring in
// the edge-cases.
currentConstraints := getproviders.VersionConstraintsString(constraints)
lockedConstraints := getproviders.VersionConstraintsString(lock.VersionConstraints())
switch {
case currentConstraints != lockedConstraints:
errs = append(errs, fmt.Errorf("provider %s: locked version selection %s doesn't match the updated version constraints %q", providerAddr, selectedVersion.String(), currentConstraints))
default:
errs = append(errs, fmt.Errorf("provider %s: version constraints %q don't match the locked version selection %s", providerAddr, currentConstraints, selectedVersion.String()))
}
}
}
// Return multiple errors in an arbitrary-but-deterministic order.
sort.Slice(errs, func(i, j int) bool {
return errs[i].Error() < errs[j].Error()
})
return errs
}
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
// ProviderRequirements searches the full tree of modules under the receiver
// for both explicit and implicit dependencies on providers.
//
// The result is a full manifest of all of the providers that must be available
// in order to work with the receiving configuration.
//
// If the returned diagnostics includes errors then the resulting Requirements
// may be incomplete.
func (c *Config) ProviderRequirements() (getproviders.Requirements, hcl.Diagnostics) {
reqs := make(getproviders.Requirements)
diags := c.addProviderRequirements(reqs, true)
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
return reqs, diags
}
// ProviderRequirementsShallow searches only the direct receiver for explicit
// and implicit dependencies on providers. Descendant modules are ignored.
//
// If the returned diagnostics includes errors then the resulting Requirements
// may be incomplete.
func (c *Config) ProviderRequirementsShallow() (getproviders.Requirements, hcl.Diagnostics) {
reqs := make(getproviders.Requirements)
diags := c.addProviderRequirements(reqs, false)
return reqs, diags
}
// ProviderRequirementsByModule searches the full tree of modules under the
// receiver for both explicit and implicit dependencies on providers,
// constructing a tree where the requirements are broken out by module.
//
// If the returned diagnostics includes errors then the resulting Requirements
// may be incomplete.
func (c *Config) ProviderRequirementsByModule() (*ModuleRequirements, hcl.Diagnostics) {
reqs := make(getproviders.Requirements)
diags := c.addProviderRequirements(reqs, false)
children := make(map[string]*ModuleRequirements)
for name, child := range c.Children {
childReqs, childDiags := child.ProviderRequirementsByModule()
childReqs.Name = name
children[name] = childReqs
diags = append(diags, childDiags...)
}
ret := &ModuleRequirements{
SourceAddr: c.SourceAddr,
SourceDir: c.Module.SourceDir,
Requirements: reqs,
Children: children,
}
return ret, diags
}
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
// addProviderRequirements is the main part of the ProviderRequirements
// implementation, gradually mutating a shared requirements object to
// eventually return. If the recurse argument is true, the requirements will
// include all descendant modules; otherwise, only the specified module.
func (c *Config) addProviderRequirements(reqs getproviders.Requirements, recurse bool) hcl.Diagnostics {
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
var diags hcl.Diagnostics
// First we'll deal with the requirements directly in _our_ module...
if c.Module.ProviderRequirements != nil {
for _, providerReqs := range c.Module.ProviderRequirements.RequiredProviders {
fqn := providerReqs.Type
if _, ok := reqs[fqn]; !ok {
// We'll at least have an unconstrained dependency then, but might
// add to this in the loop below.
reqs[fqn] = nil
}
// The model of version constraints in this package is still the
// old one using a different upstream module to represent versions,
// so we'll need to shim that out here for now. The two parsers
// don't exactly agree in practice 🙄 so this might produce new errors.
// TODO: Use the new parser throughout this package so we can get the
// better error messages it produces in more situations.
constraints, err := getproviders.ParseVersionConstraints(providerReqs.Requirement.Required.String())
if err != nil {
diags = diags.Append(&hcl.Diagnostic{
Severity: hcl.DiagError,
Summary: "Invalid version constraint",
// The errors returned by ParseVersionConstraint already include
// the section of input that was incorrect, so we don't need to
// include that here.
Detail: fmt.Sprintf("Incorrect version constraint syntax: %s.", err.Error()),
Subject: providerReqs.Requirement.DeclRange.Ptr(),
})
}
reqs[fqn] = append(reqs[fqn], constraints...)
}
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
}
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
// Each resource in the configuration creates an *implicit* provider
// dependency, though we'll only record it if there isn't already
// an explicit dependency on the same provider.
for _, rc := range c.Module.ManagedResources {
fqn := rc.Provider
if _, exists := reqs[fqn]; exists {
// Explicit dependency already present
continue
}
reqs[fqn] = nil
}
for _, rc := range c.Module.DataResources {
fqn := rc.Provider
if _, exists := reqs[fqn]; exists {
// Explicit dependency already present
continue
}
reqs[fqn] = nil
}
// "provider" block can also contain version constraints
for _, provider := range c.Module.ProviderConfigs {
fqn := c.Module.ProviderForLocalConfig(addrs.LocalProviderConfig{LocalName: provider.Name})
if _, ok := reqs[fqn]; !ok {
// We'll at least have an unconstrained dependency then, but might
// add to this in the loop below.
reqs[fqn] = nil
}
if provider.Version.Required != nil {
// The model of version constraints in this package is still the
// old one using a different upstream module to represent versions,
// so we'll need to shim that out here for now. The two parsers
// don't exactly agree in practice 🙄 so this might produce new errors.
// TODO: Use the new parser throughout this package so we can get the
// better error messages it produces in more situations.
constraints, err := getproviders.ParseVersionConstraints(provider.Version.Required.String())
if err != nil {
diags = diags.Append(&hcl.Diagnostic{
Severity: hcl.DiagError,
Summary: "Invalid version constraint",
// The errors returned by ParseVersionConstraint already include
// the section of input that was incorrect, so we don't need to
// include that here.
Detail: fmt.Sprintf("Incorrect version constraint syntax: %s.", err.Error()),
Subject: provider.Version.DeclRange.Ptr(),
})
}
reqs[fqn] = append(reqs[fqn], constraints...)
}
}
if recurse {
for _, childConfig := range c.Children {
moreDiags := childConfig.addProviderRequirements(reqs, true)
diags = append(diags, moreDiags...)
}
}
internal/getproviders: A new shared model for provider requirements We've been using the models from the "moduledeps" package to represent our provider dependencies everywhere since the idea of provider dependencies was introduced in Terraform 0.10, but that model is not convenient to use for any use-case other than the "terraform providers" command that needs individual-module-level detail. To make things easier for new codepaths working with the new-style provider installer, here we introduce a new model type getproviders.Requirements which is based on the type the new installer was already taking as its input. We have new methods in the states, configs, and earlyconfig packages to produce values of this type, and a helper to merge Requirements together so we can combine config-derived and state-derived requirements together during installation. The advantage of this new model over the moduledeps one is that all of recursive module walking is done up front and we produce a simple, flat structure that is more convenient for the main use-cases of selecting providers for installation and then finding providers in the local cache to use them for other operations. This new model is _not_ suitable for implementing "terraform providers" because it does not retain module-specific requirement details. Therefore we will likely keep using moduledeps for "terraform providers" for now, and then possibly at a later time consider specializing the moduledeps logic for only what "terraform providers" needs, because it seems to be the only use-case that needs to retain that level of detail.
2020-03-26 20:04:48 +01:00
return diags
}
// resolveProviderTypes walks through the providers in the module and ensures
// the true types are assigned based on the provider requirements for the
// module.
func (c *Config) resolveProviderTypes() {
for _, child := range c.Children {
child.resolveProviderTypes()
}
// collect the required_providers, and then add any missing default providers
providers := map[string]addrs.Provider{}
for name, p := range c.Module.ProviderRequirements.RequiredProviders {
providers[name] = p.Type
}
// ensure all provider configs know their correct type
for _, p := range c.Module.ProviderConfigs {
addr, required := providers[p.Name]
if required {
p.providerType = addr
} else {
addr := addrs.NewDefaultProvider(p.Name)
p.providerType = addr
providers[p.Name] = addr
}
}
// connect module call providers to the correct type
for _, mod := range c.Module.ModuleCalls {
for _, p := range mod.Providers {
if addr, known := providers[p.InParent.Name]; known {
p.InParent.providerType = addr
}
}
}
// fill in parent module calls too
if c.Parent != nil {
for _, mod := range c.Parent.Module.ModuleCalls {
for _, p := range mod.Providers {
if addr, known := providers[p.InChild.Name]; known {
p.InChild.providerType = addr
}
}
}
}
}
// ProviderTypes returns the FQNs of each distinct provider type referenced
various: helpers for collecting necessary provider types Since schemas are required to interpret provider, resource, and provisioner attributes in configs, states, and plans, these helpers intend to make it easier to gather up the the necessary provider types in order to preload all of the needed schemas before beginning further processing. Config.ProviderTypes returns directly the list of provider types, since at this level further detail is not useful: we've not yet run the provider allocation algorithm, and so the only thing we can reliably extract here is provider types themselves. State.ProviderAddrs and Plan.ProviderAddrs each return a list of absolute provider addresses, which can then be turned into a list of provider types using the new helper providers.AddressedTypesAbs. Since we're already using configs.Config throughout core, this also updates the terraform.LoadSchemas helper to use Config.ProviderTypes to find the necessary providers, rather than implementing its own discovery logic. states.State is not yet plumbed in, so we cannot yet use State.ProviderAddrs to deal with the state but there's a TODO comment to remind us to update that in a later commit when we swap out terraform.State for states.State. A later commit will probably refactor this further so that we can easily obtain schema for the providers needed to interpret a plan too, but that is deferred here because further work is required to make core work with the new plan types first. At that point, terraform.LoadSchemas may become providers.LoadSchemas with a different interface that just accepts lists of provider and provisioner names that have been gathered by the caller using these new helpers.
2018-06-22 02:39:27 +02:00
// in the receiving configuration.
//
// This is a helper for easily determining which provider types are required
// to fully interpret the configuration, though it does not include version
// information and so callers are expected to have already dealt with
// provider version selection in an earlier step and have identified suitable
// versions for each provider.
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
func (c *Config) ProviderTypes() []addrs.Provider {
// Ignore diagnostics here because they relate to version constraints
reqs, _ := c.ProviderRequirements()
various: helpers for collecting necessary provider types Since schemas are required to interpret provider, resource, and provisioner attributes in configs, states, and plans, these helpers intend to make it easier to gather up the the necessary provider types in order to preload all of the needed schemas before beginning further processing. Config.ProviderTypes returns directly the list of provider types, since at this level further detail is not useful: we've not yet run the provider allocation algorithm, and so the only thing we can reliably extract here is provider types themselves. State.ProviderAddrs and Plan.ProviderAddrs each return a list of absolute provider addresses, which can then be turned into a list of provider types using the new helper providers.AddressedTypesAbs. Since we're already using configs.Config throughout core, this also updates the terraform.LoadSchemas helper to use Config.ProviderTypes to find the necessary providers, rather than implementing its own discovery logic. states.State is not yet plumbed in, so we cannot yet use State.ProviderAddrs to deal with the state but there's a TODO comment to remind us to update that in a later commit when we swap out terraform.State for states.State. A later commit will probably refactor this further so that we can easily obtain schema for the providers needed to interpret a plan too, but that is deferred here because further work is required to make core work with the new plan types first. At that point, terraform.LoadSchemas may become providers.LoadSchemas with a different interface that just accepts lists of provider and provisioner names that have been gathered by the caller using these new helpers.
2018-06-22 02:39:27 +02:00
ret := make([]addrs.Provider, 0, len(reqs))
for k := range reqs {
various: helpers for collecting necessary provider types Since schemas are required to interpret provider, resource, and provisioner attributes in configs, states, and plans, these helpers intend to make it easier to gather up the the necessary provider types in order to preload all of the needed schemas before beginning further processing. Config.ProviderTypes returns directly the list of provider types, since at this level further detail is not useful: we've not yet run the provider allocation algorithm, and so the only thing we can reliably extract here is provider types themselves. State.ProviderAddrs and Plan.ProviderAddrs each return a list of absolute provider addresses, which can then be turned into a list of provider types using the new helper providers.AddressedTypesAbs. Since we're already using configs.Config throughout core, this also updates the terraform.LoadSchemas helper to use Config.ProviderTypes to find the necessary providers, rather than implementing its own discovery logic. states.State is not yet plumbed in, so we cannot yet use State.ProviderAddrs to deal with the state but there's a TODO comment to remind us to update that in a later commit when we swap out terraform.State for states.State. A later commit will probably refactor this further so that we can easily obtain schema for the providers needed to interpret a plan too, but that is deferred here because further work is required to make core work with the new plan types first. At that point, terraform.LoadSchemas may become providers.LoadSchemas with a different interface that just accepts lists of provider and provisioner names that have been gathered by the caller using these new helpers.
2018-06-22 02:39:27 +02:00
ret = append(ret, k)
}
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
sort.Slice(ret, func(i, j int) bool {
return ret[i].String() < ret[j].String()
})
various: helpers for collecting necessary provider types Since schemas are required to interpret provider, resource, and provisioner attributes in configs, states, and plans, these helpers intend to make it easier to gather up the the necessary provider types in order to preload all of the needed schemas before beginning further processing. Config.ProviderTypes returns directly the list of provider types, since at this level further detail is not useful: we've not yet run the provider allocation algorithm, and so the only thing we can reliably extract here is provider types themselves. State.ProviderAddrs and Plan.ProviderAddrs each return a list of absolute provider addresses, which can then be turned into a list of provider types using the new helper providers.AddressedTypesAbs. Since we're already using configs.Config throughout core, this also updates the terraform.LoadSchemas helper to use Config.ProviderTypes to find the necessary providers, rather than implementing its own discovery logic. states.State is not yet plumbed in, so we cannot yet use State.ProviderAddrs to deal with the state but there's a TODO comment to remind us to update that in a later commit when we swap out terraform.State for states.State. A later commit will probably refactor this further so that we can easily obtain schema for the providers needed to interpret a plan too, but that is deferred here because further work is required to make core work with the new plan types first. At that point, terraform.LoadSchemas may become providers.LoadSchemas with a different interface that just accepts lists of provider and provisioner names that have been gathered by the caller using these new helpers.
2018-06-22 02:39:27 +02:00
return ret
}
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
// ResolveAbsProviderAddr returns the AbsProviderConfig represented by the given
// ProviderConfig address, which must not be nil or this method will panic.
//
// If the given address is already an AbsProviderConfig then this method returns
// it verbatim, and will always succeed. If it's a LocalProviderConfig then
// it will consult the local-to-FQN mapping table for the given module
// to find the absolute address corresponding to the given local one.
//
// The module address to resolve local addresses in must be given in the second
// argument, and must refer to a module that exists under the receiver or
// else this method will panic.
func (c *Config) ResolveAbsProviderAddr(addr addrs.ProviderConfig, inModule addrs.Module) addrs.AbsProviderConfig {
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
switch addr := addr.(type) {
case addrs.AbsProviderConfig:
return addr
case addrs.LocalProviderConfig:
// Find the descendent Config that contains the module that this
// local config belongs to.
mc := c.Descendent(inModule)
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
if mc == nil {
panic(fmt.Sprintf("ResolveAbsProviderAddr with non-existent module %s", inModule.String()))
}
var provider addrs.Provider
if providerReq, exists := c.Module.ProviderRequirements.RequiredProviders[addr.LocalName]; exists {
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
provider = providerReq.Type
} else {
provider = addrs.ImpliedProviderForUnqualifiedType(addr.LocalName)
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
}
return addrs.AbsProviderConfig{
Module: inModule,
Provider: provider,
Alias: addr.Alias,
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
}
default:
panic(fmt.Sprintf("cannot ResolveAbsProviderAddr(%v, ...)", addr))
}
}
// ProviderForConfigAddr returns the FQN for a given addrs.ProviderConfig, first
// by checking for the provider in module.ProviderRequirements and falling
// back to addrs.NewDefaultProvider if it is not found.
Initial steps towards AbsProviderConfig/LocalProviderConfig separation (#23978) * Introduce "Local" terminology for non-absolute provider config addresses In a future change AbsProviderConfig and LocalProviderConfig are going to become two entirely distinct types, rather than Abs embedding Local as written here. This naming change is in preparation for that subsequent work, which will also include introducing a new "ProviderConfig" type that is an interface that AbsProviderConfig and LocalProviderConfig both implement. This is intended to be largely just a naming change to get started, so we can deal with all of the messy renaming. However, this did also require a slight change in modeling where the Resource.DefaultProviderConfig method has become Resource.DefaultProvider returning a Provider address directly, because this method doesn't have enough information to construct a true and accurate LocalProviderConfig -- it would need to refer to the configuration to know what this module is calling the provider it has selected. In order to leave a trail to follow for subsequent work, all of the changes here are intended to ensure that remaining work will become obvious via compile-time errors when all of the following changes happen: - The concept of "legacy" provider addresses is removed from the addrs package, including removing addrs.NewLegacyProvider and addrs.Provider.LegacyString. - addrs.AbsProviderConfig stops having addrs.LocalProviderConfig embedded in it and has an addrs.Provider and a string alias directly instead. - The provider-schema-handling parts of Terraform core are updated to work with addrs.Provider to identify providers, rather than legacy strings. In particular, there are still several codepaths here making legacy provider address assumptions (in order to limit the scope of this change) but I've made sure each one is doing something that relies on at least one of the above changes not having been made yet. * addrs: ProviderConfig interface In a (very) few special situations in the main "terraform" package we need to make runtime decisions about whether a provider config is absolute or local. We currently do that by exploiting the fact that AbsProviderConfig has LocalProviderConfig nested inside of it and so in the local case we can just ignore the wrapping AbsProviderConfig and use the embedded value. In a future change we'll be moving away from that embedding and making these two types distinct in order to represent that mapping between them requires consulting a lookup table in the configuration, and so here we introduce a new interface type ProviderConfig that can represent either AbsProviderConfig or LocalProviderConfig decided dynamically at runtime. This also includes the Config.ResolveAbsProviderAddr method that will eventually be responsible for that local-to-absolute translation, so that callers with access to the configuration can normalize to an addrs.AbsProviderConfig given a non-nil addrs.ProviderConfig. That's currently unused because existing callers are still relying on the simplistic structural transform, but we'll switch them over in a later commit. * rename LocalType to LocalName Co-authored-by: Kristin Laemmert <mildwonkey@users.noreply.github.com>
2020-01-31 14:23:07 +01:00
func (c *Config) ProviderForConfigAddr(addr addrs.LocalProviderConfig) addrs.Provider {
if provider, exists := c.Module.ProviderRequirements.RequiredProviders[addr.LocalName]; exists {
return provider.Type
}
return c.ResolveAbsProviderAddr(addr, addrs.RootModule).Provider
}