terraform/plugin/convert/schema.go

186 lines
4.9 KiB
Go
Raw Normal View History

package convert
import (
"encoding/json"
"reflect"
"sort"
"github.com/hashicorp/terraform/configs/configschema"
proto "github.com/hashicorp/terraform/internal/tfplugin5"
"github.com/hashicorp/terraform/providers"
)
// ConfigSchemaToProto takes a *configschema.Block and converts it to a
// proto.Schema_Block for a grpc response.
func ConfigSchemaToProto(b *configschema.Block) *proto.Schema_Block {
block := &proto.Schema_Block{
Description: b.Description,
DescriptionKind: protoStringKind(b.DescriptionKind),
Deprecated: b.Deprecated,
}
for _, name := range sortedKeys(b.Attributes) {
a := b.Attributes[name]
attr := &proto.Schema_Attribute{
Name: name,
Description: a.Description,
DescriptionKind: protoStringKind(a.DescriptionKind),
Optional: a.Optional,
Computed: a.Computed,
Required: a.Required,
Sensitive: a.Sensitive,
Deprecated: a.Deprecated,
}
ty, err := json.Marshal(a.Type)
if err != nil {
panic(err)
}
attr.Type = ty
block.Attributes = append(block.Attributes, attr)
}
for _, name := range sortedKeys(b.BlockTypes) {
b := b.BlockTypes[name]
block.BlockTypes = append(block.BlockTypes, protoSchemaNestedBlock(name, b))
}
return block
}
func protoStringKind(k configschema.StringKind) proto.StringKind {
switch k {
default:
return proto.StringKind_PLAIN
case configschema.StringMarkdown:
return proto.StringKind_MARKDOWN
}
}
func protoSchemaNestedBlock(name string, b *configschema.NestedBlock) *proto.Schema_NestedBlock {
configs/configschema: Introduce the NestingGroup mode for blocks In study of existing providers we've found a pattern we werent previously accounting for of using a nested block type to represent a group of arguments that relate to a particular feature that is always enabled but where it improves configuration readability to group all of its settings together in a nested block. The existing NestingSingle was not a good fit for this because it is designed under the assumption that the presence or absence of the block has some significance in enabling or disabling the relevant feature, and so for these always-active cases we'd generate a misleading plan where the settings for the feature appear totally absent, rather than showing the default values that will be selected. NestingGroup is, therefore, a slight variation of NestingSingle where presence vs. absence of the block is not distinguishable (it's never null) and instead its contents are treated as unset when the block is absent. This then in turn causes any default values associated with the nested arguments to be honored and displayed in the plan whenever the block is not explicitly configured. The current SDK cannot activate this mode, but that's okay because its "legacy type system" opt-out flag allows it to force a block to be processed in this way anyway. We're adding this now so that we can introduce the feature in a future SDK without causing a breaking change to the protocol, since the set of possible block nesting modes is not extensible.
2019-04-09 00:32:53 +02:00
var nesting proto.Schema_NestedBlock_NestingMode
switch b.Nesting {
case configschema.NestingSingle:
nesting = proto.Schema_NestedBlock_SINGLE
case configschema.NestingGroup:
nesting = proto.Schema_NestedBlock_GROUP
case configschema.NestingList:
nesting = proto.Schema_NestedBlock_LIST
case configschema.NestingSet:
nesting = proto.Schema_NestedBlock_SET
case configschema.NestingMap:
nesting = proto.Schema_NestedBlock_MAP
default:
nesting = proto.Schema_NestedBlock_INVALID
}
return &proto.Schema_NestedBlock{
TypeName: name,
Block: ConfigSchemaToProto(&b.Block),
configs/configschema: Introduce the NestingGroup mode for blocks In study of existing providers we've found a pattern we werent previously accounting for of using a nested block type to represent a group of arguments that relate to a particular feature that is always enabled but where it improves configuration readability to group all of its settings together in a nested block. The existing NestingSingle was not a good fit for this because it is designed under the assumption that the presence or absence of the block has some significance in enabling or disabling the relevant feature, and so for these always-active cases we'd generate a misleading plan where the settings for the feature appear totally absent, rather than showing the default values that will be selected. NestingGroup is, therefore, a slight variation of NestingSingle where presence vs. absence of the block is not distinguishable (it's never null) and instead its contents are treated as unset when the block is absent. This then in turn causes any default values associated with the nested arguments to be honored and displayed in the plan whenever the block is not explicitly configured. The current SDK cannot activate this mode, but that's okay because its "legacy type system" opt-out flag allows it to force a block to be processed in this way anyway. We're adding this now so that we can introduce the feature in a future SDK without causing a breaking change to the protocol, since the set of possible block nesting modes is not extensible.
2019-04-09 00:32:53 +02:00
Nesting: nesting,
MinItems: int64(b.MinItems),
MaxItems: int64(b.MaxItems),
}
}
// ProtoToProviderSchema takes a proto.Schema and converts it to a providers.Schema.
func ProtoToProviderSchema(s *proto.Schema) providers.Schema {
return providers.Schema{
Version: s.Version,
Block: ProtoToConfigSchema(s.Block),
}
}
// ProtoToConfigSchema takes the GetSchcema_Block from a grpc response and converts it
// to a terraform *configschema.Block.
func ProtoToConfigSchema(b *proto.Schema_Block) *configschema.Block {
block := &configschema.Block{
Attributes: make(map[string]*configschema.Attribute),
BlockTypes: make(map[string]*configschema.NestedBlock),
Description: b.Description,
DescriptionKind: schemaStringKind(b.DescriptionKind),
Deprecated: b.Deprecated,
}
for _, a := range b.Attributes {
attr := &configschema.Attribute{
Description: a.Description,
DescriptionKind: schemaStringKind(a.DescriptionKind),
Required: a.Required,
Optional: a.Optional,
Computed: a.Computed,
Sensitive: a.Sensitive,
Deprecated: a.Deprecated,
}
if err := json.Unmarshal(a.Type, &attr.Type); err != nil {
panic(err)
}
block.Attributes[a.Name] = attr
}
for _, b := range b.BlockTypes {
block.BlockTypes[b.TypeName] = schemaNestedBlock(b)
}
return block
}
func schemaStringKind(k proto.StringKind) configschema.StringKind {
switch k {
default:
return configschema.StringPlain
case proto.StringKind_MARKDOWN:
return configschema.StringMarkdown
}
}
func schemaNestedBlock(b *proto.Schema_NestedBlock) *configschema.NestedBlock {
configs/configschema: Introduce the NestingGroup mode for blocks In study of existing providers we've found a pattern we werent previously accounting for of using a nested block type to represent a group of arguments that relate to a particular feature that is always enabled but where it improves configuration readability to group all of its settings together in a nested block. The existing NestingSingle was not a good fit for this because it is designed under the assumption that the presence or absence of the block has some significance in enabling or disabling the relevant feature, and so for these always-active cases we'd generate a misleading plan where the settings for the feature appear totally absent, rather than showing the default values that will be selected. NestingGroup is, therefore, a slight variation of NestingSingle where presence vs. absence of the block is not distinguishable (it's never null) and instead its contents are treated as unset when the block is absent. This then in turn causes any default values associated with the nested arguments to be honored and displayed in the plan whenever the block is not explicitly configured. The current SDK cannot activate this mode, but that's okay because its "legacy type system" opt-out flag allows it to force a block to be processed in this way anyway. We're adding this now so that we can introduce the feature in a future SDK without causing a breaking change to the protocol, since the set of possible block nesting modes is not extensible.
2019-04-09 00:32:53 +02:00
var nesting configschema.NestingMode
switch b.Nesting {
case proto.Schema_NestedBlock_SINGLE:
nesting = configschema.NestingSingle
case proto.Schema_NestedBlock_GROUP:
nesting = configschema.NestingGroup
case proto.Schema_NestedBlock_LIST:
nesting = configschema.NestingList
case proto.Schema_NestedBlock_MAP:
nesting = configschema.NestingMap
case proto.Schema_NestedBlock_SET:
nesting = configschema.NestingSet
default:
// In all other cases we'll leave it as the zero value (invalid) and
// let the caller validate it and deal with this.
}
nb := &configschema.NestedBlock{
configs/configschema: Introduce the NestingGroup mode for blocks In study of existing providers we've found a pattern we werent previously accounting for of using a nested block type to represent a group of arguments that relate to a particular feature that is always enabled but where it improves configuration readability to group all of its settings together in a nested block. The existing NestingSingle was not a good fit for this because it is designed under the assumption that the presence or absence of the block has some significance in enabling or disabling the relevant feature, and so for these always-active cases we'd generate a misleading plan where the settings for the feature appear totally absent, rather than showing the default values that will be selected. NestingGroup is, therefore, a slight variation of NestingSingle where presence vs. absence of the block is not distinguishable (it's never null) and instead its contents are treated as unset when the block is absent. This then in turn causes any default values associated with the nested arguments to be honored and displayed in the plan whenever the block is not explicitly configured. The current SDK cannot activate this mode, but that's okay because its "legacy type system" opt-out flag allows it to force a block to be processed in this way anyway. We're adding this now so that we can introduce the feature in a future SDK without causing a breaking change to the protocol, since the set of possible block nesting modes is not extensible.
2019-04-09 00:32:53 +02:00
Nesting: nesting,
MinItems: int(b.MinItems),
MaxItems: int(b.MaxItems),
}
nested := ProtoToConfigSchema(b.Block)
nb.Block = *nested
return nb
}
// sortedKeys returns the lexically sorted keys from the given map. This is
// used to make schema conversions are deterministic. This panics if map keys
// are not a string.
func sortedKeys(m interface{}) []string {
v := reflect.ValueOf(m)
keys := make([]string, v.Len())
mapKeys := v.MapKeys()
for i, k := range mapKeys {
keys[i] = k.Interface().(string)
}
sort.Strings(keys)
return keys
}