terraform/internal/terraform/eval_context_builtin.go

385 lines
11 KiB
Go
Raw Normal View History

2015-02-04 02:46:11 +01:00
package terraform
import (
"context"
2015-02-04 02:46:11 +01:00
"fmt"
2015-02-13 18:05:09 +01:00
"log"
2015-02-04 02:46:11 +01:00
"sync"
"github.com/hashicorp/terraform/internal/instances"
"github.com/hashicorp/terraform/internal/plans"
"github.com/hashicorp/terraform/internal/providers"
"github.com/hashicorp/terraform/internal/provisioners"
"github.com/hashicorp/terraform/internal/refactoring"
"github.com/hashicorp/terraform/version"
"github.com/hashicorp/terraform/internal/states"
"github.com/hashicorp/hcl/v2"
"github.com/hashicorp/terraform/internal/configs/configschema"
"github.com/hashicorp/terraform/internal/lang"
"github.com/hashicorp/terraform/internal/tfdiags"
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
"github.com/hashicorp/terraform/internal/addrs"
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
"github.com/zclconf/go-cty/cty"
2015-02-04 02:46:11 +01:00
)
// BuiltinEvalContext is an EvalContext implementation that is used by
// Terraform by default.
type BuiltinEvalContext struct {
// StopContext is the context used to track whether we're complete
StopContext context.Context
// PathValue is the Path that this context is operating within.
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
PathValue addrs.ModuleInstance
// pathSet indicates that this context was explicitly created for a
// specific path, and can be safely used for evaluation. This lets us
// differentiate between PathValue being unset, and the zero value which is
// equivalent to RootModuleInstance. Path and Evaluation methods will
// panic if this is not set.
pathSet bool
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
// Evaluator is used for evaluating expressions within the scope of this
// eval context.
Evaluator *Evaluator
// VariableValues contains the variable values across all modules. This
// structure is shared across the entire containing context, and so it
// may be accessed only when holding VariableValuesLock.
// The keys of the first level of VariableValues are the string
// representations of addrs.ModuleInstance values. The second-level keys
// are variable names within each module instance.
VariableValues map[string]map[string]cty.Value
VariableValuesLock *sync.Mutex
// Plugins is a library of plugin components (providers and provisioners)
// available for use during a graph walk.
Plugins *contextPlugins
Hooks []Hook
InputValue UIInput
ProviderCache map[string]providers.Interface
ProviderInputConfig map[string]map[string]cty.Value
ProviderLock *sync.Mutex
ProvisionerCache map[string]provisioners.Interface
ProvisionerLock *sync.Mutex
ChangesValue *plans.ChangesSync
StateValue *states.SyncState
RefreshStateValue *states.SyncState
PrevRunStateValue *states.SyncState
InstanceExpanderValue *instances.Expander
MoveResultsValue refactoring.MoveResults
2015-02-04 02:46:11 +01:00
}
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
// BuiltinEvalContext implements EvalContext
var _ EvalContext = (*BuiltinEvalContext)(nil)
func (ctx *BuiltinEvalContext) WithPath(path addrs.ModuleInstance) EvalContext {
newCtx := *ctx
2020-09-30 16:45:55 +02:00
newCtx.pathSet = true
newCtx.PathValue = path
return &newCtx
}
func (ctx *BuiltinEvalContext) Stopped() <-chan struct{} {
// This can happen during tests. During tests, we just block forever.
if ctx.StopContext == nil {
return nil
}
return ctx.StopContext.Done()
}
2015-02-11 22:43:07 +01:00
func (ctx *BuiltinEvalContext) Hook(fn func(Hook) (HookAction, error)) error {
for _, h := range ctx.Hooks {
action, err := fn(h)
if err != nil {
return err
}
switch action {
case HookActionContinue:
continue
case HookActionHalt:
// Return an early exit error to trigger an early exit
2015-02-13 18:05:09 +01:00
log.Printf("[WARN] Early exit triggered by hook: %T", h)
return nil
2015-02-11 22:43:07 +01:00
}
}
return nil
}
2015-02-14 02:59:54 +01:00
func (ctx *BuiltinEvalContext) Input() UIInput {
return ctx.InputValue
}
func (ctx *BuiltinEvalContext) InitProvider(addr addrs.AbsProviderConfig) (providers.Interface, error) {
// If we already initialized, it is an error
if p := ctx.Provider(addr); p != nil {
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
return nil, fmt.Errorf("%s is already initialized", addr)
2015-02-04 02:46:11 +01:00
}
// Warning: make sure to acquire these locks AFTER the call to Provider
// above, since it also acquires locks.
ctx.ProviderLock.Lock()
defer ctx.ProviderLock.Unlock()
key := addr.String()
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
p, err := ctx.Plugins.NewProviderInstance(addr.Provider)
2015-02-05 02:02:18 +01:00
if err != nil {
return nil, err
}
log.Printf("[TRACE] BuiltinEvalContext: Initialized %q provider for %s", addr.String(), addr)
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
ctx.ProviderCache[key] = p
2015-02-05 02:02:18 +01:00
return p, nil
2015-02-04 02:46:11 +01:00
}
func (ctx *BuiltinEvalContext) Provider(addr addrs.AbsProviderConfig) providers.Interface {
ctx.ProviderLock.Lock()
defer ctx.ProviderLock.Unlock()
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
return ctx.ProviderCache[addr.String()]
}
func (ctx *BuiltinEvalContext) ProviderSchema(addr addrs.AbsProviderConfig) (*ProviderSchema, error) {
return ctx.Plugins.ProviderSchema(addr.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 (ctx *BuiltinEvalContext) CloseProvider(addr addrs.AbsProviderConfig) error {
ctx.ProviderLock.Lock()
defer ctx.ProviderLock.Unlock()
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
key := addr.String()
provider := ctx.ProviderCache[key]
if provider != nil {
delete(ctx.ProviderCache, key)
return provider.Close()
}
return nil
}
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 (ctx *BuiltinEvalContext) ConfigureProvider(addr addrs.AbsProviderConfig, cfg cty.Value) tfdiags.Diagnostics {
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
var diags tfdiags.Diagnostics
if !addr.Module.Equal(ctx.Path().Module()) {
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
// This indicates incorrect use of ConfigureProvider: it should be used
// only from the module that the provider configuration belongs to.
panic(fmt.Sprintf("%s configured by wrong module %s", addr, ctx.Path()))
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
}
p := ctx.Provider(addr)
if p == nil {
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
diags = diags.Append(fmt.Errorf("%s not initialized", addr))
return diags
}
providerSchema, err := ctx.ProviderSchema(addr)
if err != nil {
diags = diags.Append(fmt.Errorf("failed to read schema for %s: %s", addr, err))
return diags
}
if providerSchema == nil {
diags = diags.Append(fmt.Errorf("schema for %s is not available", addr))
return diags
}
req := providers.ConfigureProviderRequest{
TerraformVersion: version.String(),
Config: cfg,
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
}
resp := p.ConfigureProvider(req)
return resp.Diagnostics
}
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 (ctx *BuiltinEvalContext) ProviderInput(pc addrs.AbsProviderConfig) map[string]cty.Value {
2015-02-14 02:59:54 +01:00
ctx.ProviderLock.Lock()
defer ctx.ProviderLock.Unlock()
if !pc.Module.Equal(ctx.Path().Module()) {
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
// This indicates incorrect use of InitProvider: it should be used
// only from the module that the provider configuration belongs to.
panic(fmt.Sprintf("%s initialized by wrong module %s", pc, ctx.Path()))
}
if !ctx.Path().IsRoot() {
// Only root module provider configurations can have input.
return nil
}
return ctx.ProviderInputConfig[pc.String()]
2015-02-14 02:59:54 +01:00
}
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 (ctx *BuiltinEvalContext) SetProviderInput(pc addrs.AbsProviderConfig, c map[string]cty.Value) {
absProvider := pc
if !pc.Module.IsRoot() {
// Only root module provider configurations can have input.
log.Printf("[WARN] BuiltinEvalContext: attempt to SetProviderInput for non-root module")
return
}
// Save the configuration
ctx.ProviderLock.Lock()
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
ctx.ProviderInputConfig[absProvider.String()] = c
ctx.ProviderLock.Unlock()
2015-02-14 02:59:54 +01:00
}
func (ctx *BuiltinEvalContext) Provisioner(n string) (provisioners.Interface, error) {
2015-02-09 20:15:54 +01:00
ctx.ProvisionerLock.Lock()
defer ctx.ProvisionerLock.Unlock()
p, ok := ctx.ProvisionerCache[n]
if !ok {
var err error
p, err = ctx.Plugins.NewProvisionerInstance(n)
if err != nil {
return nil, err
}
2015-02-09 20:15:54 +01:00
ctx.ProvisionerCache[n] = p
}
2015-02-09 20:15:54 +01:00
return p, nil
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
}
func (ctx *BuiltinEvalContext) ProvisionerSchema(n string) (*configschema.Block, error) {
return ctx.Plugins.ProvisionerSchema(n)
2015-02-09 20:15:54 +01:00
}
func (ctx *BuiltinEvalContext) CloseProvisioners() error {
var diags tfdiags.Diagnostics
ctx.ProvisionerLock.Lock()
defer ctx.ProvisionerLock.Unlock()
for name, prov := range ctx.ProvisionerCache {
err := prov.Close()
if err != nil {
diags = diags.Append(fmt.Errorf("provisioner.Close %s: %s", name, err))
}
}
return diags.Err()
}
func (ctx *BuiltinEvalContext) EvaluateBlock(body hcl.Body, schema *configschema.Block, self addrs.Referenceable, keyData InstanceKeyEvalData) (cty.Value, hcl.Body, tfdiags.Diagnostics) {
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
var diags tfdiags.Diagnostics
scope := ctx.EvaluationScope(self, keyData)
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
body, evalDiags := scope.ExpandBlock(body, schema)
diags = diags.Append(evalDiags)
val, evalDiags := scope.EvalBlock(body, schema)
diags = diags.Append(evalDiags)
return val, body, diags
}
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
func (ctx *BuiltinEvalContext) EvaluateExpr(expr hcl.Expression, wantType cty.Type, self addrs.Referenceable) (cty.Value, tfdiags.Diagnostics) {
scope := ctx.EvaluationScope(self, EvalDataForNoInstanceKey)
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
return scope.EvalExpr(expr, wantType)
2015-02-04 02:46:11 +01:00
}
func (ctx *BuiltinEvalContext) EvaluationScope(self addrs.Referenceable, keyData InstanceKeyEvalData) *lang.Scope {
if !ctx.pathSet {
panic("context path not set")
}
data := &evaluationStateData{
Evaluator: ctx.Evaluator,
ModulePath: ctx.PathValue,
InstanceKeyData: keyData,
Operation: ctx.Evaluator.Operation,
}
scope := ctx.Evaluator.Scope(data, self)
// ctx.PathValue is the path of the module that contains whatever
// expression the caller will be trying to evaluate, so this will
// activate only the experiments from that particular module, to
// be consistent with how experiment checking in the "configs"
// package itself works. The nil check here is for robustness in
// incompletely-mocked testing situations; mc should never be nil in
// real situations.
if mc := ctx.Evaluator.Config.DescendentForInstance(ctx.PathValue); mc != nil {
scope.SetActiveExperiments(mc.Module.ActiveExperiments)
}
return scope
}
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
func (ctx *BuiltinEvalContext) Path() addrs.ModuleInstance {
if !ctx.pathSet {
panic("context path not set")
}
return ctx.PathValue
}
core: Handle root and child module input variables consistently Previously we had a significant discrepancy between these two situations: we wrote the raw root module variables directly into the EvalContext and then applied type conversions only at expression evaluation time, while for child modules we converted and validated the values while visiting the variable graph node and wrote only the _final_ value into the EvalContext. This confusion seems to have been the root cause for #29899, where validation rules for root module variables were being applied at the wrong point in the process, prior to type conversion. To fix that bug and also make similar mistakes less likely in the future, I've made the root module variable handling more like the child module variable handling in the following ways: - The "raw value" (exactly as given by the user) lives only in the graph node representing the variable, which mirrors how the _expression_ for a child module variable lives in its graph node. This means that the flow for the two is the same except that there's no expression evaluation step for root module variables, because they arrive as constant values from the caller. - The set of variable values in the EvalContext is always only "final" values, after type conversion is complete. That in turn means we no longer need to do "just in time" conversion in evaluationStateData.GetInputVariable, and can just return the value exactly as stored, which is consistent with how we handle all other references between objects. This diff is noisier than I'd like because of how much it takes to wire a new argument (the raw variable values) through to the plan graph builder, but those changes are pretty mechanical and the interesting logic lives inside the plan graph builder itself, in NodeRootVariable, and the shared helper functions in eval_variable.go. While here I also took the opportunity to fix a historical API wart in EvalContext, where SetModuleCallArguments was built to take a set of variable values all at once but our current caller always calls with only one at a time. That is now just SetModuleCallArgument singular, to match with the new SetRootModuleArgument to deal with root module variables.
2021-11-11 02:29:45 +01:00
func (ctx *BuiltinEvalContext) SetRootModuleArgument(addr addrs.InputVariable, v cty.Value) {
ctx.VariableValuesLock.Lock()
defer ctx.VariableValuesLock.Unlock()
log.Printf("[TRACE] BuiltinEvalContext: Storing final value for variable %s", addr.Absolute(addrs.RootModuleInstance))
key := addrs.RootModuleInstance.String()
args := ctx.VariableValues[key]
if args == nil {
args = make(map[string]cty.Value)
ctx.VariableValues[key] = args
}
args[addr.Name] = v
}
func (ctx *BuiltinEvalContext) SetModuleCallArgument(callAddr addrs.ModuleCallInstance, varAddr addrs.InputVariable, v cty.Value) {
ctx.VariableValuesLock.Lock()
defer ctx.VariableValuesLock.Unlock()
if !ctx.pathSet {
panic("context path not set")
}
core: Handle root and child module input variables consistently Previously we had a significant discrepancy between these two situations: we wrote the raw root module variables directly into the EvalContext and then applied type conversions only at expression evaluation time, while for child modules we converted and validated the values while visiting the variable graph node and wrote only the _final_ value into the EvalContext. This confusion seems to have been the root cause for #29899, where validation rules for root module variables were being applied at the wrong point in the process, prior to type conversion. To fix that bug and also make similar mistakes less likely in the future, I've made the root module variable handling more like the child module variable handling in the following ways: - The "raw value" (exactly as given by the user) lives only in the graph node representing the variable, which mirrors how the _expression_ for a child module variable lives in its graph node. This means that the flow for the two is the same except that there's no expression evaluation step for root module variables, because they arrive as constant values from the caller. - The set of variable values in the EvalContext is always only "final" values, after type conversion is complete. That in turn means we no longer need to do "just in time" conversion in evaluationStateData.GetInputVariable, and can just return the value exactly as stored, which is consistent with how we handle all other references between objects. This diff is noisier than I'd like because of how much it takes to wire a new argument (the raw variable values) through to the plan graph builder, but those changes are pretty mechanical and the interesting logic lives inside the plan graph builder itself, in NodeRootVariable, and the shared helper functions in eval_variable.go. While here I also took the opportunity to fix a historical API wart in EvalContext, where SetModuleCallArguments was built to take a set of variable values all at once but our current caller always calls with only one at a time. That is now just SetModuleCallArgument singular, to match with the new SetRootModuleArgument to deal with root module variables.
2021-11-11 02:29:45 +01:00
childPath := callAddr.ModuleInstance(ctx.PathValue)
log.Printf("[TRACE] BuiltinEvalContext: Storing final value for variable %s", varAddr.Absolute(childPath))
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
key := childPath.String()
args := ctx.VariableValues[key]
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
if args == nil {
core: Handle root and child module input variables consistently Previously we had a significant discrepancy between these two situations: we wrote the raw root module variables directly into the EvalContext and then applied type conversions only at expression evaluation time, while for child modules we converted and validated the values while visiting the variable graph node and wrote only the _final_ value into the EvalContext. This confusion seems to have been the root cause for #29899, where validation rules for root module variables were being applied at the wrong point in the process, prior to type conversion. To fix that bug and also make similar mistakes less likely in the future, I've made the root module variable handling more like the child module variable handling in the following ways: - The "raw value" (exactly as given by the user) lives only in the graph node representing the variable, which mirrors how the _expression_ for a child module variable lives in its graph node. This means that the flow for the two is the same except that there's no expression evaluation step for root module variables, because they arrive as constant values from the caller. - The set of variable values in the EvalContext is always only "final" values, after type conversion is complete. That in turn means we no longer need to do "just in time" conversion in evaluationStateData.GetInputVariable, and can just return the value exactly as stored, which is consistent with how we handle all other references between objects. This diff is noisier than I'd like because of how much it takes to wire a new argument (the raw variable values) through to the plan graph builder, but those changes are pretty mechanical and the interesting logic lives inside the plan graph builder itself, in NodeRootVariable, and the shared helper functions in eval_variable.go. While here I also took the opportunity to fix a historical API wart in EvalContext, where SetModuleCallArguments was built to take a set of variable values all at once but our current caller always calls with only one at a time. That is now just SetModuleCallArgument singular, to match with the new SetRootModuleArgument to deal with root module variables.
2021-11-11 02:29:45 +01:00
args = make(map[string]cty.Value)
ctx.VariableValues[key] = args
}
core: Handle root and child module input variables consistently Previously we had a significant discrepancy between these two situations: we wrote the raw root module variables directly into the EvalContext and then applied type conversions only at expression evaluation time, while for child modules we converted and validated the values while visiting the variable graph node and wrote only the _final_ value into the EvalContext. This confusion seems to have been the root cause for #29899, where validation rules for root module variables were being applied at the wrong point in the process, prior to type conversion. To fix that bug and also make similar mistakes less likely in the future, I've made the root module variable handling more like the child module variable handling in the following ways: - The "raw value" (exactly as given by the user) lives only in the graph node representing the variable, which mirrors how the _expression_ for a child module variable lives in its graph node. This means that the flow for the two is the same except that there's no expression evaluation step for root module variables, because they arrive as constant values from the caller. - The set of variable values in the EvalContext is always only "final" values, after type conversion is complete. That in turn means we no longer need to do "just in time" conversion in evaluationStateData.GetInputVariable, and can just return the value exactly as stored, which is consistent with how we handle all other references between objects. This diff is noisier than I'd like because of how much it takes to wire a new argument (the raw variable values) through to the plan graph builder, but those changes are pretty mechanical and the interesting logic lives inside the plan graph builder itself, in NodeRootVariable, and the shared helper functions in eval_variable.go. While here I also took the opportunity to fix a historical API wart in EvalContext, where SetModuleCallArguments was built to take a set of variable values all at once but our current caller always calls with only one at a time. That is now just SetModuleCallArgument singular, to match with the new SetRootModuleArgument to deal with root module variables.
2021-11-11 02:29:45 +01:00
args[varAddr.Name] = v
}
func (ctx *BuiltinEvalContext) GetVariableValue(addr addrs.AbsInputVariableInstance) cty.Value {
ctx.VariableValuesLock.Lock()
defer ctx.VariableValuesLock.Unlock()
modKey := addr.Module.String()
modVars := ctx.VariableValues[modKey]
val, ok := modVars[addr.Variable.Name]
if !ok {
return cty.DynamicVal
}
return val
}
func (ctx *BuiltinEvalContext) Changes() *plans.ChangesSync {
return ctx.ChangesValue
2015-02-12 00:22:03 +01:00
}
func (ctx *BuiltinEvalContext) State() *states.SyncState {
return ctx.StateValue
2015-02-11 17:48:45 +01:00
}
func (ctx *BuiltinEvalContext) RefreshState() *states.SyncState {
return ctx.RefreshStateValue
}
func (ctx *BuiltinEvalContext) PrevRunState() *states.SyncState {
return ctx.PrevRunStateValue
}
func (ctx *BuiltinEvalContext) InstanceExpander() *instances.Expander {
return ctx.InstanceExpanderValue
}
func (ctx *BuiltinEvalContext) MoveResults() refactoring.MoveResults {
return ctx.MoveResultsValue
}