2015-02-11 10:48:45 -06:00
package terraform
2017-07-21 11:41:29 -05:00
import (
"fmt"
2018-08-31 12:41:42 -05:00
"log"
2019-10-28 20:16:45 -05:00
"sort"
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 12:33:53 -05:00
"github.com/hashicorp/terraform/addrs"
2018-11-30 12:56:50 -06:00
"github.com/hashicorp/terraform/configs"
2019-12-18 18:55:46 -06:00
"github.com/hashicorp/terraform/plans"
2018-08-16 10:40:08 -05:00
"github.com/hashicorp/terraform/providers"
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
"github.com/hashicorp/terraform/states"
2018-11-30 12:56:50 -06:00
"github.com/hashicorp/terraform/tfdiags"
2017-07-21 11:41:29 -05:00
)
2015-02-11 10:48:45 -06:00
// EvalReadState is an EvalNode implementation that reads the
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// current object for a specific instance in the state.
2015-02-11 10:48:45 -06:00
type EvalReadState struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Addr is the address of the instance to read state for.
Addr addrs . ResourceInstance
// ProviderSchema is the schema for the provider given in Provider.
ProviderSchema * * ProviderSchema
// Provider is the provider that will subsequently perform actions on
// the the state object. This is used to perform any schema upgrades
// that might be required to prepare the stored data for use.
2018-08-16 10:40:08 -05:00
Provider * providers . Interface
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Output will be written with a pointer to the retrieved object.
Output * * states . ResourceInstanceObject
2015-02-11 10:48:45 -06:00
}
2015-02-14 00:58:41 -06:00
func ( n * EvalReadState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if n . Provider == nil || * n . Provider == nil {
panic ( "EvalReadState used with no Provider object" )
}
if n . ProviderSchema == nil || * n . ProviderSchema == nil {
panic ( "EvalReadState used with no ProviderSchema object" )
}
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
2018-08-31 12:41:42 -05:00
log . Printf ( "[TRACE] EvalReadState: reading state for %s" , absAddr )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
src := ctx . State ( ) . ResourceInstanceObject ( absAddr , states . CurrentGen )
if src == nil {
// Presumably we only have deposed objects, then.
2018-08-31 12:41:42 -05:00
log . Printf ( "[TRACE] EvalReadState: no state present for %s" , absAddr )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
return nil , nil
}
2018-11-27 17:30:18 -06:00
schema , currentVersion := ( * n . ProviderSchema ) . SchemaForResourceAddr ( n . Addr . ContainingResource ( ) )
2018-11-30 12:56:50 -06:00
if schema == nil {
// Shouldn't happen since we should've failed long ago if no schema is present
return nil , fmt . Errorf ( "no schema available for %s while reading state; this is a bug in Terraform and should be reported" , absAddr )
}
var diags tfdiags . Diagnostics
src , diags = UpgradeResourceState ( absAddr , * n . Provider , src , schema , currentVersion )
if diags . HasErrors ( ) {
// Note that we don't have any channel to return warnings here. We'll
// accept that for now since warnings during a schema upgrade would
// be pretty weird anyway, since this operation is supposed to seem
// invisible to the user.
return nil , diags . Err ( )
2018-11-27 17:30:18 -06:00
}
2018-09-07 16:45:51 -05:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
obj , err := src . Decode ( schema . ImpliedType ( ) )
if err != nil {
return nil , err
}
2018-09-07 16:45:51 -05:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if n . Output != nil {
* n . Output = obj
}
return obj , nil
2015-02-27 11:38:17 -06:00
}
// EvalReadStateDeposed is an EvalNode implementation that reads the
2015-03-02 13:05:38 -06:00
// deposed InstanceState for a specific resource out of the state
2015-02-27 11:38:17 -06:00
type EvalReadStateDeposed struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Addr is the address of the instance to read state for.
Addr addrs . ResourceInstance
2015-02-27 11:38:17 -06:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Key identifies which deposed object we will read.
Key states . DeposedKey
// ProviderSchema is the schema for the provider given in Provider.
ProviderSchema * * ProviderSchema
// Provider is the provider that will subsequently perform actions on
// the the state object. This is used to perform any schema upgrades
// that might be required to prepare the stored data for use.
2018-08-16 10:40:08 -05:00
Provider * providers . Interface
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Output will be written with a pointer to the retrieved object.
Output * * states . ResourceInstanceObject
2015-03-02 13:05:38 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
func ( n * EvalReadStateDeposed ) Eval ( ctx EvalContext ) ( interface { } , error ) {
2018-11-29 18:47:18 -06:00
if n . Provider == nil || * n . Provider == nil {
panic ( "EvalReadStateDeposed used with no Provider object" )
}
if n . ProviderSchema == nil || * n . ProviderSchema == nil {
panic ( "EvalReadStateDeposed used with no ProviderSchema object" )
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
key := n . Key
if key == states . NotDeposed {
return nil , fmt . Errorf ( "EvalReadStateDeposed used with no instance key; this is a bug in Terraform and should be reported" )
2015-02-27 11:38:17 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
2018-08-31 12:41:42 -05:00
log . Printf ( "[TRACE] EvalReadStateDeposed: reading state for %s deposed object %s" , absAddr , n . Key )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
src := ctx . State ( ) . ResourceInstanceObject ( absAddr , key )
if src == nil {
// Presumably we only have deposed objects, then.
2018-08-31 12:41:42 -05:00
log . Printf ( "[TRACE] EvalReadStateDeposed: no state present for %s deposed object %s" , absAddr , n . Key )
2015-02-27 11:38:17 -06:00
return nil , nil
}
2018-11-27 17:30:18 -06:00
schema , currentVersion := ( * n . ProviderSchema ) . SchemaForResourceAddr ( n . Addr . ContainingResource ( ) )
2018-11-30 12:56:50 -06:00
if schema == nil {
// Shouldn't happen since we should've failed long ago if no schema is present
return nil , fmt . Errorf ( "no schema available for %s while reading state; this is a bug in Terraform and should be reported" , absAddr )
}
var diags tfdiags . Diagnostics
src , diags = UpgradeResourceState ( absAddr , * n . Provider , src , schema , currentVersion )
if diags . HasErrors ( ) {
// Note that we don't have any channel to return warnings here. We'll
// accept that for now since warnings during a schema upgrade would
// be pretty weird anyway, since this operation is supposed to seem
// invisible to the user.
return nil , diags . Err ( )
2018-11-27 17:30:18 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
obj , err := src . Decode ( schema . ImpliedType ( ) )
2015-03-02 13:05:38 -06:00
if err != nil {
return nil , err
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if n . Output != nil {
* n . Output = obj
2015-02-27 11:38:17 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
return obj , nil
2015-02-27 11:38:17 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// EvalRequireState is an EvalNode implementation that exits early if the given
// object is null.
2015-02-25 00:45:47 -06:00
type EvalRequireState struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
State * * states . ResourceInstanceObject
2015-02-25 00:45:47 -06:00
}
func ( n * EvalRequireState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
if n . State == nil {
return nil , EvalEarlyExitError { }
}
state := * n . State
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if state == nil || state . Value . IsNull ( ) {
2015-02-25 00:45:47 -06:00
return nil , EvalEarlyExitError { }
}
return nil , nil
}
2015-02-23 21:09:48 -06:00
// EvalUpdateStateHook is an EvalNode implementation that calls the
// PostStateUpdate hook with the current state.
type EvalUpdateStateHook struct { }
func ( n * EvalUpdateStateHook ) Eval ( ctx EvalContext ) ( interface { } , error ) {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// In principle we could grab the lock here just long enough to take a
// deep copy and then pass that to our hooks below, but we'll instead
// hold the hook for the duration to avoid the potential confusing
// situation of us racing to call PostStateUpdate concurrently with
// different state snapshots.
stateSync := ctx . State ( )
state := stateSync . Lock ( ) . DeepCopy ( )
defer stateSync . Unlock ( )
2015-02-23 21:09:48 -06:00
// Call the hook
err := ctx . Hook ( func ( h Hook ) ( HookAction , error ) {
return h . PostStateUpdate ( state )
} )
if err != nil {
return nil , err
}
return nil , nil
}
2020-05-13 08:40:13 -05:00
// evalWriteEmptyState wraps EvalWriteState to specifically record an empty
// state for a particular object.
type evalWriteEmptyState struct {
EvalWriteState
}
func ( n * evalWriteEmptyState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
var state * states . ResourceInstanceObject
n . State = & state
return n . EvalWriteState . Eval ( ctx )
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// EvalWriteState is an EvalNode implementation that saves the given object
// as the current object for the selected resource instance.
2015-02-11 10:48:45 -06:00
type EvalWriteState struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Addr is the address of the instance to read state for.
Addr addrs . ResourceInstance
// State is the object state to save.
State * * states . ResourceInstanceObject
// ProviderSchema is the schema for the provider given in ProviderAddr.
ProviderSchema * * ProviderSchema
// ProviderAddr is the address of the provider configuration that
// produced the given object.
ProviderAddr addrs . AbsProviderConfig
2019-10-17 15:05:27 -05:00
// Dependencies are the inter-resource dependencies to be stored in the
// state.
2020-03-23 14:26:18 -05:00
Dependencies * [ ] addrs . ConfigResource
2015-02-11 10:48:45 -06:00
}
2015-02-14 00:58:41 -06:00
func ( n * EvalWriteState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if n . State == nil {
// Note that a pointer _to_ nil is valid here, indicating the total
// absense of an object as we'd see during destroy.
panic ( "EvalWriteState used with no ResourceInstanceObject" )
}
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
state := ctx . State ( )
2020-02-13 14:32:58 -06:00
if n . ProviderAddr . Provider . Type == "" {
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 07:23:07 -06:00
return nil , fmt . Errorf ( "failed to write state for %s: missing provider type" , absAddr )
2018-12-18 12:09:40 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
obj := * n . State
2018-09-07 16:45:51 -05:00
if obj == nil || obj . Value . IsNull ( ) {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// No need to encode anything: we'll just write it directly.
state . SetResourceInstanceCurrent ( absAddr , nil , n . ProviderAddr )
2018-08-31 12:41:42 -05:00
log . Printf ( "[TRACE] EvalWriteState: removing state object for %s" , absAddr )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
return nil , nil
}
2019-10-17 15:05:27 -05:00
// store the new deps in the state
2019-10-28 20:16:45 -05:00
if n . Dependencies != nil {
log . Printf ( "[TRACE] EvalWriteState: recording %d dependencies for %s" , len ( * n . Dependencies ) , absAddr )
obj . Dependencies = * n . Dependencies
}
2019-10-17 15:05:27 -05:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if n . ProviderSchema == nil || * n . ProviderSchema == nil {
// Should never happen, unless our state object is nil
panic ( "EvalWriteState used with pointer to nil ProviderSchema object" )
}
2018-09-11 18:29:11 -05:00
if obj != nil {
log . Printf ( "[TRACE] EvalWriteState: writing current state object for %s" , absAddr )
} else {
log . Printf ( "[TRACE] EvalWriteState: removing current state object for %s" , absAddr )
}
2018-08-31 12:41:42 -05:00
2018-11-27 17:30:18 -06:00
schema , currentVersion := ( * n . ProviderSchema ) . SchemaForResourceAddr ( n . Addr . ContainingResource ( ) )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if schema == nil {
// It shouldn't be possible to get this far in any real scenario
// without a schema, but we might end up here in contrived tests that
// fail to set up their world properly.
return nil , fmt . Errorf ( "failed to encode %s in state: no resource type schema available" , absAddr )
}
2018-11-27 17:30:18 -06:00
src , err := obj . Encode ( schema . ImpliedType ( ) , currentVersion )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if err != nil {
return nil , fmt . Errorf ( "failed to encode %s in state: %s" , absAddr , err )
}
state . SetResourceInstanceCurrent ( absAddr , src , n . ProviderAddr )
return nil , nil
2015-03-02 15:34:05 -06:00
}
2015-03-04 18:01:46 -06:00
// EvalWriteStateDeposed is an EvalNode implementation that writes
// an InstanceState out to the Deposed list of a resource in the state.
2015-03-02 15:34:05 -06:00
type EvalWriteStateDeposed struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Addr is the address of the instance to read state for.
Addr addrs . ResourceInstance
2015-03-02 15:34:05 -06:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Key indicates which deposed object to write to.
Key states . DeposedKey
// State is the object state to save.
State * * states . ResourceInstanceObject
// ProviderSchema is the schema for the provider given in ProviderAddr.
ProviderSchema * * ProviderSchema
// ProviderAddr is the address of the provider configuration that
// produced the given object.
ProviderAddr addrs . AbsProviderConfig
2015-03-04 18:01:46 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
func ( n * EvalWriteStateDeposed ) Eval ( ctx EvalContext ) ( interface { } , error ) {
if n . State == nil {
// Note that a pointer _to_ nil is valid here, indicating the total
// absense of an object as we'd see during destroy.
panic ( "EvalWriteStateDeposed used with no ResourceInstanceObject" )
2015-03-02 15:34:05 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
key := n . Key
state := ctx . State ( )
2015-03-02 15:34:05 -06:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if key == states . NotDeposed {
// should never happen
return nil , fmt . Errorf ( "can't save deposed object for %s without a deposed key; this is a bug in Terraform that should be reported" , absAddr )
2015-03-02 15:34:05 -06:00
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
obj := * n . State
if obj == nil {
// No need to encode anything: we'll just write it directly.
2018-08-31 12:41:42 -05:00
state . SetResourceInstanceDeposed ( absAddr , key , nil , n . ProviderAddr )
log . Printf ( "[TRACE] EvalWriteStateDeposed: removing state object for %s deposed %s" , absAddr , key )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
return nil , nil
}
if n . ProviderSchema == nil || * n . ProviderSchema == nil {
// Should never happen, unless our state object is nil
panic ( "EvalWriteStateDeposed used with no ProviderSchema object" )
2015-03-02 15:34:05 -06:00
}
2018-11-27 17:30:18 -06:00
schema , currentVersion := ( * n . ProviderSchema ) . SchemaForResourceAddr ( n . Addr . ContainingResource ( ) )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if schema == nil {
// It shouldn't be possible to get this far in any real scenario
// without a schema, but we might end up here in contrived tests that
// fail to set up their world properly.
return nil , fmt . Errorf ( "failed to encode %s in state: no resource type schema available" , absAddr )
}
2018-11-27 17:30:18 -06:00
src , err := obj . Encode ( schema . ImpliedType ( ) , currentVersion )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if err != nil {
return nil , fmt . Errorf ( "failed to encode %s in state: %s" , absAddr , err )
2015-03-04 12:15:53 -06:00
}
2015-03-02 15:34:05 -06:00
2018-08-31 12:41:42 -05:00
log . Printf ( "[TRACE] EvalWriteStateDeposed: writing state object for %s deposed %s" , absAddr , key )
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
state . SetResourceInstanceDeposed ( absAddr , key , src , n . ProviderAddr )
2015-03-02 15:34:05 -06:00
return nil , nil
}
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// EvalDeposeState is an EvalNode implementation that moves the current object
// for the given instance to instead be a deposed object, leaving the instance
// with no current object.
// This is used at the beginning of a create-before-destroy replace action so
// that the create can create while preserving the old state of the
// to-be-destroyed object.
2015-02-13 17:57:37 -06:00
type EvalDeposeState struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
Addr addrs . ResourceInstance
2018-09-20 14:30:52 -05:00
// ForceKey, if a value other than states.NotDeposed, will be used as the
// key for the newly-created deposed object that results from this action.
// If set to states.NotDeposed (the zero value), a new unique key will be
// allocated.
ForceKey states . DeposedKey
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// OutputKey, if non-nil, will be written with the deposed object key that
// was generated for the object. This can then be passed to
// EvalUndeposeState.Key so it knows which deposed instance to forget.
OutputKey * states . DeposedKey
2015-02-13 17:57:37 -06:00
}
// TODO: test
2015-02-14 00:58:41 -06:00
func ( n * EvalDeposeState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
state := ctx . State ( )
2015-02-13 17:57:37 -06:00
2018-09-20 14:30:52 -05:00
var key states . DeposedKey
if n . ForceKey == states . NotDeposed {
key = state . DeposeResourceInstanceObject ( absAddr )
} else {
key = n . ForceKey
state . DeposeResourceInstanceObjectForceKey ( absAddr , key )
}
2018-09-18 20:16:31 -05:00
log . Printf ( "[TRACE] EvalDeposeState: prior object for %s now deposed with key %s" , absAddr , key )
2015-02-13 17:57:37 -06:00
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
if n . OutputKey != nil {
* n . OutputKey = key
2015-02-13 17:57:37 -06:00
}
return nil , nil
}
2018-09-20 20:28:27 -05:00
// EvalMaybeRestoreDeposedObject is an EvalNode implementation that will
// restore a particular deposed object of the specified resource instance
// to be the "current" object if and only if the instance doesn't currently
// have a current object.
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
//
2018-09-20 20:28:27 -05:00
// This is intended for use when the create leg of a create before destroy
// fails with no partial new object: if we didn't take any action, the user
// would be left in the unfortunate situation of having no current object
// and the previously-workign object now deposed. This EvalNode causes a
// better outcome by restoring things to how they were before the replace
// operation began.
//
// The create operation may have produced a partial result even though it
// failed and it's important that we don't "forget" that state, so in that
// situation the prior object remains deposed and the partial new object
// remains the current object, allowing the situation to hopefully be
// improved in a subsequent run.
type EvalMaybeRestoreDeposedObject struct {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
Addr addrs . ResourceInstance
2019-12-18 18:55:46 -06:00
// PlannedChange might be the action we're performing that includes
// the possiblity of restoring a deposed object. However, it might also
// be nil. It's here only for use in error messages and must not be
// used for business logic.
PlannedChange * * plans . ResourceInstanceChange
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
// Key is a pointer to the deposed object key that should be forgotten
// from the state, which must be non-nil.
Key * states . DeposedKey
2015-02-13 17:57:37 -06:00
}
// TODO: test
2018-09-20 20:28:27 -05:00
func ( n * EvalMaybeRestoreDeposedObject ) Eval ( ctx EvalContext ) ( interface { } , error ) {
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
2018-09-20 20:28:27 -05:00
dk := * n . Key
terraform: Ugly huge change to weave in new State and Plan types
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
2018-08-14 16:24:45 -05:00
state := ctx . State ( )
2015-02-13 17:57:37 -06:00
2019-12-18 18:55:46 -06:00
if dk == states . NotDeposed {
// This should never happen, and so it always indicates a bug.
// We should evaluate this node only if we've previously deposed
// an object as part of the same operation.
var diags tfdiags . Diagnostics
if n . PlannedChange != nil && * n . PlannedChange != nil {
diags = diags . Append ( tfdiags . Sourceless (
tfdiags . Error ,
"Attempt to restore non-existent deposed object" ,
fmt . Sprintf (
"Terraform has encountered a bug where it would need to restore a deposed object for %s without knowing a deposed object key for that object. This occurred during a %s action. This is a bug in Terraform; please report it!" ,
absAddr , ( * n . PlannedChange ) . Action ,
) ,
) )
} else {
diags = diags . Append ( tfdiags . Sourceless (
tfdiags . Error ,
"Attempt to restore non-existent deposed object" ,
fmt . Sprintf (
"Terraform has encountered a bug where it would need to restore a deposed object for %s without knowing a deposed object key for that object. This is a bug in Terraform; please report it!" ,
absAddr ,
) ,
) )
}
return nil , diags . Err ( )
}
2018-09-20 20:28:27 -05:00
restored := state . MaybeRestoreResourceInstanceDeposed ( absAddr , dk )
if restored {
log . Printf ( "[TRACE] EvalMaybeRestoreDeposedObject: %s deposed object %s was restored as the current object" , absAddr , dk )
} else {
log . Printf ( "[TRACE] EvalMaybeRestoreDeposedObject: %s deposed object %s remains deposed" , absAddr , dk )
}
2015-02-13 17:57:37 -06:00
return nil , nil
}
2018-09-11 16:22:25 -05:00
// EvalWriteResourceState is an EvalNode implementation that ensures that
// a suitable resource-level state record is present in the state, if that's
// required for the "each mode" of that resource.
//
// This is important primarily for the situation where count = 0, since this
// eval is the only change we get to set the resource "each mode" to list
// in that case, allowing expression evaluation to see it as a zero-element
// list rather than as not set at all.
type EvalWriteResourceState struct {
2020-03-20 14:19:01 -05:00
Addr addrs . AbsResource
2018-09-11 16:22:25 -05:00
Config * configs . Resource
ProviderAddr addrs . AbsProviderConfig
}
func ( n * EvalWriteResourceState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
var diags tfdiags . Diagnostics
state := ctx . State ( )
2020-03-20 14:19:01 -05:00
// We'll record our expansion decision in the shared "expander" object
// so that later operations (i.e. DynamicExpand and expression evaluation)
// can refer to it. Since this node represents the abstract module, we need
// to expand the module here to create all resources.
expander := ctx . InstanceExpander ( )
2018-09-11 16:22:25 -05:00
2020-04-08 16:09:17 -05:00
switch {
case n . Config . Count != nil :
count , countDiags := evaluateCountExpression ( n . Config . Count , ctx )
diags = diags . Append ( countDiags )
if countDiags . HasErrors ( ) {
return nil , diags . Err ( )
}
2018-09-11 16:22:25 -05:00
2020-04-24 09:43:53 -05:00
state . SetResourceProvider ( n . Addr , n . ProviderAddr )
2020-04-08 16:09:17 -05:00
expander . SetResourceCount ( n . Addr . Module , n . Addr . Resource , count )
2019-07-25 10:51:55 -05:00
2020-04-08 16:09:17 -05:00
case n . Config . ForEach != nil :
forEach , forEachDiags := evaluateForEachExpression ( n . Config . ForEach , ctx )
diags = diags . Append ( forEachDiags )
if forEachDiags . HasErrors ( ) {
return nil , diags . Err ( )
}
2020-03-20 14:19:01 -05:00
2020-04-08 16:09:17 -05:00
// This method takes care of all of the business logic of updating this
// while ensuring that any existing instances are preserved, etc.
2020-04-24 09:43:53 -05:00
state . SetResourceProvider ( n . Addr , n . ProviderAddr )
2020-03-20 14:19:01 -05:00
expander . SetResourceForEach ( n . Addr . Module , n . Addr . Resource , forEach )
2020-04-08 16:09:17 -05:00
2020-03-20 14:19:01 -05:00
default :
2020-04-24 09:43:53 -05:00
state . SetResourceProvider ( n . Addr , n . ProviderAddr )
2020-03-20 14:19:01 -05:00
expander . SetResourceSingle ( n . Addr . Module , n . Addr . Resource )
2019-11-21 20:45:43 -06:00
}
2018-09-11 16:22:25 -05:00
return nil , nil
}
2018-09-26 19:52:34 -05:00
// EvalForgetResourceState is an EvalNode implementation that prunes out an
// empty resource-level state for a given resource address, or produces an
// error if it isn't empty after all.
//
// This should be the last action taken for a resource that has been removed
// from the configuration altogether, to clean up the leftover husk of the
// resource in the state after other EvalNodes have destroyed and removed
// all of the instances and instance objects beneath it.
type EvalForgetResourceState struct {
Addr addrs . Resource
}
func ( n * EvalForgetResourceState ) Eval ( ctx EvalContext ) ( interface { } , error ) {
absAddr := n . Addr . Absolute ( ctx . Path ( ) )
state := ctx . State ( )
pruned := state . RemoveResourceIfEmpty ( absAddr )
if ! pruned {
// If this produces an error, it indicates a bug elsewhere in Terraform
// -- probably missing graph nodes, graph edges, or
// incorrectly-implemented evaluation steps.
return nil , fmt . Errorf ( "orphan resource %s still has a non-empty state after apply; this is a bug in Terraform" , absAddr )
}
log . Printf ( "[TRACE] EvalForgetResourceState: Pruned husk of %s from state" , absAddr )
return nil , nil
}
2019-10-28 20:16:45 -05:00
// EvalRefreshDependencies is an EvalNode implementation that appends any newly
// found dependencies to those saved in the state. The existing dependencies
// are retained, as they may be missing from the config, and will be required
// for the updates and destroys during the next apply.
type EvalRefreshDependencies struct {
// Prior State
State * * states . ResourceInstanceObject
// Dependencies to write to the new state
2020-03-23 14:26:18 -05:00
Dependencies * [ ] addrs . ConfigResource
2019-10-28 20:16:45 -05:00
}
func ( n * EvalRefreshDependencies ) Eval ( ctx EvalContext ) ( interface { } , error ) {
state := * n . State
if state == nil {
// no existing state to append
return nil , nil
}
2020-03-23 14:26:18 -05:00
depMap := make ( map [ string ] addrs . ConfigResource )
2019-10-28 20:16:45 -05:00
for _ , d := range * n . Dependencies {
depMap [ d . String ( ) ] = d
}
2019-11-15 15:00:13 -06:00
// We have already dependencies in state, so we need to trust those for
// refresh. We can't write out new dependencies until apply time in case
// the configuration has been changed in a manner the conflicts with the
// stored dependencies.
if len ( state . Dependencies ) > 0 {
* n . Dependencies = state . Dependencies
return nil , nil
2019-10-28 20:16:45 -05:00
}
2020-03-23 14:26:18 -05:00
deps := make ( [ ] addrs . ConfigResource , 0 , len ( depMap ) )
2019-10-28 20:16:45 -05:00
for _ , d := range depMap {
deps = append ( deps , d )
}
sort . Slice ( deps , func ( i , j int ) bool {
return deps [ i ] . String ( ) < deps [ j ] . String ( )
} )
* n . Dependencies = deps
return nil , nil
}