2016-04-28 04:46:46 -05:00
package terraform
import (
"fmt"
2020-09-14 15:53:37 -05:00
"log"
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"
2020-03-20 07:15:29 -05:00
"github.com/hashicorp/terraform/configs"
2018-08-24 18:13:50 -05:00
"github.com/hashicorp/terraform/providers"
2018-05-29 14:17:59 -05:00
"github.com/hashicorp/terraform/tfdiags"
2016-04-28 04:46:46 -05:00
)
// ImportStateTransformer is a GraphTransformer that adds nodes to the
// graph to represent the imports we want to do for resources.
type ImportStateTransformer struct {
Targets [ ] * ImportTarget
2020-03-20 07:15:29 -05:00
Config * configs . Config
2016-04-28 04:46:46 -05:00
}
func ( t * ImportStateTransformer ) Transform ( g * Graph ) error {
for _ , target := range t . Targets {
2020-03-20 07:15:29 -05:00
// This is only likely to happen in misconfigured tests
if t . Config == nil {
return fmt . Errorf ( "Cannot import into an empty configuration." )
}
// Get the module config
modCfg := t . Config . Descendent ( target . Addr . Module . Module ( ) )
if modCfg == nil {
return fmt . Errorf ( "Module %s not found." , target . Addr . Module . Module ( ) )
}
2020-06-23 13:08:27 -05:00
providerAddr := addrs . AbsProviderConfig {
Module : target . Addr . Module . Module ( ) ,
2020-03-20 07:15:29 -05:00
}
2020-06-23 13:08:27 -05:00
// Try to find the resource config
rsCfg := modCfg . Module . ResourceByAddr ( target . Addr . Resource . Resource )
if rsCfg != nil {
// Get the provider FQN for the resource from the resource configuration
providerAddr . Provider = rsCfg . Provider
2020-03-20 07:15:29 -05:00
2020-06-23 13:08:27 -05:00
// Get the alias from the resource's provider local config
providerAddr . Alias = rsCfg . ProviderConfigAddr ( ) . Alias
} else {
// Resource has no matching config, so use an implied provider
// based on the resource type
rsProviderType := target . Addr . Resource . Resource . ImpliedProvider ( )
providerAddr . Provider = modCfg . Module . ImpliedProviderForUnqualifiedType ( rsProviderType )
2018-12-04 14:39:46 -06:00
}
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
node := & graphNodeImportState {
Addr : target . Addr ,
2017-11-01 17:34:18 -05:00
ID : target . ID ,
2018-12-04 14:39:46 -06:00
ProviderAddr : providerAddr ,
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
}
g . Add ( node )
2016-04-28 04:46:46 -05:00
}
return nil
}
type graphNodeImportState struct {
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
Addr addrs . AbsResourceInstance // Addr is the resource address to import into
ID string // ID is the ID to import as
2018-05-08 13:36:07 -05:00
ProviderAddr addrs . AbsProviderConfig // Provider address given by the user, or implied by the resource type
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
ResolvedProvider addrs . AbsProviderConfig // provider node address after resolution
2016-04-28 12:18:19 -05:00
2018-08-24 18:13:50 -05:00
states [ ] providers . ImportedResource
2016-04-28 04:46:46 -05:00
}
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
var (
2020-03-10 20:21:19 -05:00
_ GraphNodeModulePath = ( * graphNodeImportState ) ( nil )
2020-09-14 15:53:37 -05:00
_ GraphNodeExecutable = ( * graphNodeImportState ) ( nil )
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
_ GraphNodeProviderConsumer = ( * graphNodeImportState ) ( nil )
_ GraphNodeDynamicExpandable = ( * graphNodeImportState ) ( nil )
)
2016-04-28 04:46:46 -05:00
func ( n * graphNodeImportState ) Name ( ) string {
2018-05-08 13:36:07 -05:00
return fmt . Sprintf ( "%s (import id %q)" , n . Addr , n . ID )
2016-04-28 04:46:46 -05:00
}
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
// GraphNodeProviderConsumer
2020-03-06 07:33:44 -06:00
func ( n * graphNodeImportState ) ProvidedBy ( ) ( addrs . ProviderConfig , bool ) {
2018-05-08 13:36:07 -05:00
// We assume that n.ProviderAddr has been properly populated here.
// It's the responsibility of the code creating a graphNodeImportState
// to populate this, possibly by calling DefaultProviderConfig() on the
// resource address to infer an implied provider from the resource type
// name.
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
return n . ProviderAddr , false
2017-11-01 17:34:18 -05:00
}
2020-03-10 13:43:57 -05:00
// GraphNodeProviderConsumer
2020-03-18 07:58:20 -05:00
func ( n * graphNodeImportState ) Provider ( ) addrs . Provider {
2020-03-10 13:43:57 -05:00
// We assume that n.ProviderAddr has been properly populated here.
// It's the responsibility of the code creating a graphNodeImportState
// to populate this, possibly by calling DefaultProviderConfig() on the
// resource address to infer an implied provider from the resource type
// name.
return n . ProviderAddr . Provider
}
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
// GraphNodeProviderConsumer
func ( n * graphNodeImportState ) SetProvider ( addr addrs . AbsProviderConfig ) {
n . ResolvedProvider = addr
2016-04-28 04:46:46 -05:00
}
2020-03-05 15:13:54 -06:00
// GraphNodeModuleInstance
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
func ( n * graphNodeImportState ) Path ( ) addrs . ModuleInstance {
return n . Addr . Module
2016-04-28 04:46:46 -05:00
}
2020-03-10 20:21:19 -05:00
// GraphNodeModulePath
func ( n * graphNodeImportState ) ModulePath ( ) addrs . Module {
return n . Addr . Module . Module ( )
}
2020-09-14 15:53:37 -05:00
// GraphNodeExecutable impl.
2020-10-28 12:47:04 -05:00
func ( n * graphNodeImportState ) Execute ( ctx EvalContext , op walkOperation ) ( diags tfdiags . Diagnostics ) {
2016-04-28 12:18:19 -05:00
// Reset our states
n . states = nil
2020-09-14 15:53:37 -05:00
provider , _ , err := GetProvider ( ctx , n . ResolvedProvider )
2020-10-28 12:47:04 -05:00
diags = diags . Append ( err )
if diags . HasErrors ( ) {
return diags
2020-09-14 15:53:37 -05:00
}
// import state
absAddr := n . Addr . Resource . Absolute ( ctx . Path ( ) )
// Call pre-import hook
2020-10-28 12:47:04 -05:00
diags = diags . Append ( ctx . Hook ( func ( h Hook ) ( HookAction , error ) {
2020-09-14 15:53:37 -05:00
return h . PreImportState ( absAddr , n . ID )
2020-10-28 12:47:04 -05:00
} ) )
if diags . HasErrors ( ) {
return diags
2016-04-28 12:18:19 -05:00
}
2020-09-14 15:53:37 -05:00
resp := provider . ImportResourceState ( providers . ImportResourceStateRequest {
TypeName : n . Addr . Resource . Resource . Type ,
ID : n . ID ,
} )
diags = diags . Append ( resp . Diagnostics )
if diags . HasErrors ( ) {
2020-10-28 12:47:04 -05:00
return diags
2020-09-14 15:53:37 -05:00
}
imported := resp . ImportedResources
for _ , obj := range imported {
log . Printf ( "[TRACE] graphNodeImportState: import %s %q produced instance object of type %s" , absAddr . String ( ) , n . ID , obj . TypeName )
}
n . states = imported
// Call post-import hook
2020-10-28 12:47:04 -05:00
diags = diags . Append ( ctx . Hook ( func ( h Hook ) ( HookAction , error ) {
2020-09-14 15:53:37 -05:00
return h . PostImportState ( absAddr , imported )
2020-10-28 12:47:04 -05:00
} ) )
return diags
2016-04-28 12:18:19 -05:00
}
// GraphNodeDynamicExpandable impl.
//
// We use DynamicExpand as a way to generate the subgraph of refreshes
// and state inserts we need to do for our import state. Since they're new
// resources they don't depend on anything else and refreshes are isolated
// so this is nearly a perfect use case for dynamic expand.
func ( n * graphNodeImportState ) DynamicExpand ( ctx EvalContext ) ( * Graph , error ) {
2018-05-29 14:17:59 -05:00
var diags tfdiags . Diagnostics
2016-04-28 12:18:19 -05:00
g := & Graph { Path : ctx . Path ( ) }
2016-04-30 17:29:03 -05:00
// nameCounter is used to de-dup names in the state.
nameCounter := make ( map [ string ] int )
// Compile the list of addresses that we'll be inserting into the state.
// We do this ahead of time so we can verify that we aren't importing
// something that already exists.
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
addrs := make ( [ ] addrs . AbsResourceInstance , len ( n . states ) )
2016-04-30 17:29:03 -05:00
for i , state := range n . states {
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
addr := n . Addr
2018-08-24 18:13:50 -05:00
if t := state . TypeName ; t != "" {
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
addr . Resource . Resource . Type = t
2016-04-30 17:29:03 -05:00
}
// Determine if we need to suffix the name to de-dup
key := addr . String ( )
count , ok := nameCounter [ key ]
if ok {
count ++
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
addr . Resource . Resource . Name += fmt . Sprintf ( "-%d" , count )
2016-04-30 17:29:03 -05:00
}
nameCounter [ key ] = count
// Add it to our list
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
addrs [ i ] = addr
2016-04-30 17:29:03 -05:00
}
2016-04-30 17:53:48 -05:00
// Verify that all the addresses are clear
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 ( )
2016-04-30 17:53:48 -05:00
for _ , addr := range addrs {
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
existing := state . ResourceInstance ( addr )
if existing != nil {
diags = diags . Append ( tfdiags . Sourceless (
tfdiags . Error ,
"Resource already managed by Terraform" ,
fmt . Sprintf ( "Terraform is already managing a remote object for %s. To import to this address you must first remove the existing object from the state." , addr ) ,
) )
2018-05-29 14:17:59 -05:00
continue
2016-04-30 17:53:48 -05:00
}
}
2018-05-29 14:17:59 -05:00
if diags . HasErrors ( ) {
// Bail out early, then.
return nil , diags . Err ( )
}
2016-04-30 17:53:48 -05:00
2016-04-28 12:18:19 -05:00
// For each of the states, we add a node to handle the refresh/add to state.
2020-09-14 15:53:37 -05:00
// "n.states" is populated by our own Execute with the result of
// ImportState. Since DynamicExpand is always called after Execute, this is
// safe.
2016-04-30 17:29:03 -05:00
for i , state := range n . states {
2016-04-28 12:18:19 -05:00
g . Add ( & graphNodeImportStateSub {
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
TargetAddr : addrs [ i ] ,
2017-11-01 17:34:18 -05:00
State : state ,
ResolvedProvider : n . ResolvedProvider ,
2016-04-28 12:18:19 -05:00
} )
}
// Root transform for a single root
t := & RootTransformer { }
if err := t . Transform ( g ) ; err != nil {
return nil , err
}
// Done!
2018-05-29 14:17:59 -05:00
return g , diags . Err ( )
2016-04-28 12:18:19 -05:00
}
// graphNodeImportStateSub is the sub-node of graphNodeImportState
// and is part of the subgraph. This node is responsible for refreshing
// and adding a resource to the state once it is imported.
type graphNodeImportStateSub struct {
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
TargetAddr addrs . AbsResourceInstance
2018-08-24 18:13:50 -05:00
State providers . ImportedResource
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
ResolvedProvider addrs . AbsProviderConfig
2016-04-28 12:18:19 -05:00
}
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
var (
2020-03-05 15:13:54 -06:00
_ GraphNodeModuleInstance = ( * graphNodeImportStateSub ) ( nil )
2020-09-14 15:53:37 -05:00
_ GraphNodeExecutable = ( * graphNodeImportStateSub ) ( nil )
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
)
2016-04-28 12:18:19 -05:00
func ( n * graphNodeImportStateSub ) Name ( ) string {
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 fmt . Sprintf ( "import %s result" , n . TargetAddr )
2016-04-28 12:18:19 -05:00
}
terraform: ugly huge change to weave in new HCL2-oriented types
Due to how deeply the configuration types go into Terraform Core, there
isn't a great way to switch out to HCL2 gradually. As a consequence, this
huge commit gets us from the old state to a _compilable_ new state, but
does not yet attempt to fix any tests and has a number of known missing
parts and bugs. We will continue to iterate on this in forthcoming
commits, heading back towards passing tests and making Terraform
fully-functional again.
The three main goals here are:
- Use the configuration models from the "configs" package instead of the
older models in the "config" package, which is now deprecated and
preserved only to help us write our migration tool.
- Do expression inspection and evaluation using the functionality of the
new "lang" package, instead of the Interpolator type and related
functionality in the main "terraform" package.
- Represent addresses of various objects using types in the addrs package,
rather than hand-constructed strings. This is not critical to support
the above, but was a big help during the implementation of these other
points since it made it much more explicit what kind of address is
expected in each context.
Since our new packages are built to accommodate some future planned
features that are not yet implemented (e.g. the "for_each" argument on
resources, "count"/"for_each" on modules), and since there's still a fair
amount of functionality still using old-style APIs, there is a moderate
amount of shimming here to connect new assumptions with old, hopefully in
a way that makes it easier to find and eliminate these shims later.
I apologize in advance to the person who inevitably just found this huge
commit while spelunking through the commit history.
2018-04-30 12:33:53 -05:00
func ( n * graphNodeImportStateSub ) Path ( ) addrs . ModuleInstance {
return n . TargetAddr . Module
2016-04-30 01:28:58 -05:00
}
2020-09-14 15:53:37 -05:00
// GraphNodeExecutable impl.
2020-10-28 12:47:04 -05:00
func ( n * graphNodeImportStateSub ) Execute ( ctx EvalContext , op walkOperation ) ( diags tfdiags . Diagnostics ) {
2016-04-28 12:18:19 -05:00
// If the Ephemeral type isn't set, then it is an error
2018-08-24 18:13:50 -05:00
if n . State . TypeName == "" {
2020-10-28 12:47:04 -05:00
diags = diags . Append ( fmt . Errorf ( "import of %s didn't set type" , n . TargetAddr . String ( ) ) )
return diags
2016-04-28 12:18:19 -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
state := n . State . AsInstanceObject ( )
2020-09-14 15:53:37 -05:00
provider , providerSchema , err := GetProvider ( ctx , n . ResolvedProvider )
2020-10-28 12:47:04 -05:00
diags = diags . Append ( err )
if diags . HasErrors ( ) {
return diags
2020-09-14 15:53:37 -05:00
}
// EvalRefresh
evalRefresh := & EvalRefresh {
Addr : n . TargetAddr . Resource ,
ProviderAddr : n . ResolvedProvider ,
Provider : & provider ,
ProviderSchema : & providerSchema ,
State : & state ,
Output : & state ,
}
2020-10-28 12:47:04 -05:00
diags = diags . Append ( evalRefresh . Eval ( ctx ) )
2020-10-28 11:03:00 -05:00
if diags . HasErrors ( ) {
2020-10-28 12:47:04 -05:00
return diags
2020-09-14 15:53:37 -05:00
}
// Verify the existance of the imported resource
if state . Value . IsNull ( ) {
diags = diags . Append ( tfdiags . Sourceless (
tfdiags . Error ,
"Cannot import non-existent remote object" ,
fmt . Sprintf (
"While attempting to import an existing object to %s, the provider detected that no object exists with the given id. Only pre-existing objects can be imported; check that the id is correct and that it is associated with the provider's configured region or endpoint, or use \"terraform apply\" to create a new remote object for this resource." ,
n . TargetAddr . Resource . String ( ) ,
) ,
) )
2020-10-28 12:47:04 -05:00
return diags
2020-09-14 15:53:37 -05:00
}
2016-04-28 12:18:19 -05:00
2020-09-14 15:53:37 -05:00
//EvalWriteState
evalWriteState := & EvalWriteState {
Addr : n . TargetAddr . Resource ,
ProviderAddr : n . ResolvedProvider ,
ProviderSchema : & providerSchema ,
State : & state ,
2016-04-28 04:46:46 -05:00
}
2020-10-28 11:23:03 -05:00
diags = diags . Append ( evalWriteState . Eval ( ctx ) )
2020-10-28 12:47:04 -05:00
return diags
2016-04-28 04:46:46 -05:00
}