opentofu/terraform/graph_builder_eval.go
Martin Atkins 88b5607a7a core: Fetch schemas during context construction
Previously we fetched schemas during the AttachSchemaTransformer,
potentially multiple times as that was re-run for each graph built. Now
we fetch the schemas just once during context construction, passing that
result into each of the graph builders.

This only addresses the schema accesses during graph construction. We're
still separately loading schemas during the main walk for evaluation
purposes. This will be addressed in a later commit.
2018-10-16 18:49:20 -07:00

108 lines
3.2 KiB
Go

package terraform
import (
"github.com/hashicorp/terraform/addrs"
"github.com/hashicorp/terraform/configs"
"github.com/hashicorp/terraform/dag"
"github.com/hashicorp/terraform/tfdiags"
)
// EvalGraphBuilder implements GraphBuilder and constructs a graph suitable
// for evaluating in-memory values (input variables, local values, output
// values) in the state without any other side-effects.
//
// This graph is used only in weird cases, such as the "terraform console"
// CLI command, where we need to evaluate expressions against the state
// without taking any other actions.
//
// The generated graph will include nodes for providers, resources, etc
// just to allow indirect dependencies to be resolved, but these nodes will
// not take any actions themselves since we assume that their parts of the
// state, if any, are already complete.
//
// Although the providers are never configured, they must still be available
// in order to obtain schema information used for type checking, etc.
type EvalGraphBuilder struct {
// Config is the configuration tree.
Config *configs.Config
// State is the current state
State *State
// Components is a factory for the plug-in components (providers and
// provisioners) available for use.
Components contextComponentFactory
// Schemas is the repository of schemas we will draw from to analyse
// the configuration.
Schemas *Schemas
}
// See GraphBuilder
func (b *EvalGraphBuilder) Build(path addrs.ModuleInstance) (*Graph, tfdiags.Diagnostics) {
return (&BasicGraphBuilder{
Steps: b.Steps(),
Validate: true,
Name: "EvalGraphBuilder",
}).Build(path)
}
// See GraphBuilder
func (b *EvalGraphBuilder) Steps() []GraphTransformer {
concreteProvider := func(a *NodeAbstractProvider) dag.Vertex {
return &NodeEvalableProvider{
NodeAbstractProvider: a,
}
}
steps := []GraphTransformer{
// Creates all the data resources that aren't in the state. This will also
// add any orphans from scaling in as destroy nodes.
&ConfigTransformer{
Concrete: nil, // just use the abstract type
Config: b.Config,
Unique: true,
},
// Attach the state
&AttachStateTransformer{State: b.State},
// Attach the configuration to any resources
&AttachResourceConfigTransformer{Config: b.Config},
// Add root variables
&RootVariableTransformer{Config: b.Config},
// Must be before TransformProviders and ReferenceTransformer, since
// schema is required to extract references from config.
&AttachSchemaTransformer{Schemas: b.Schemas},
TransformProviders(b.Components.ResourceProviders(), concreteProvider, b.Config),
// Add the local values
&LocalTransformer{Config: b.Config},
// Add the outputs
&OutputTransformer{Config: b.Config},
// Add module variables
&ModuleVariableTransformer{Config: b.Config},
// Connect so that the references are ready for targeting. We'll
// have to connect again later for providers and so on.
&ReferenceTransformer{},
// Although we don't configure providers, we do still start them up
// to get their schemas, and so we must shut them down again here.
&CloseProviderTransformer{},
// Single root
&RootTransformer{},
// Remove redundant edges to simplify the graph.
&TransitiveReductionTransformer{},
}
return steps
}