Files
opentf/internal/plans/plan.go
Martin Atkins 3785619f93 core: Use the new checks package for condition tracking
The "checks" package is an expansion what we previously called
plans.Conditions to accommodate a new requirement that we be able to track
which checks we're expecting to run even if we don't actually get around
to running them, which will be helpful when we start using checks as part
of our module testing story because test reporting tools appreciate there
being a relatively consistent set of test cases from one run to the next.

So far this should be essentially a no-op change from an external
functionality standpoint, aside from some minor adjustments to how we
report some of the error and warning cases from condition evaluation in
light of the fact that the "checks" package can now track errors as a
different outcome than a failure of a valid check.

As is often the case with anything which changes what we track
in the EvalContext and persist between plan and apply, Terraform Core is
pretty brittle and so this had knock-on effects elsewhere too. Again, the
goal is for these changes to not create any material externally-visible
difference, and just to accommodate the new assumption that there will
always be a "checks" object available for tracking during a graph walk.
2022-08-26 15:47:29 -07:00

181 lines
6.8 KiB
Go

package plans
import (
"sort"
"github.com/hashicorp/terraform/internal/addrs"
"github.com/hashicorp/terraform/internal/configs/configschema"
"github.com/hashicorp/terraform/internal/lang/globalref"
"github.com/hashicorp/terraform/internal/states"
"github.com/zclconf/go-cty/cty"
)
// Plan is the top-level type representing a planned set of changes.
//
// A plan is a summary of the set of changes required to move from a current
// state to a goal state derived from configuration. The described changes
// are not applied directly, but contain an approximation of the final
// result that will be completed during apply by resolving any values that
// cannot be predicted.
//
// A plan must always be accompanied by the configuration it was built from,
// since the plan does not itself include all of the information required to
// make the changes indicated.
type Plan struct {
// Mode is the mode under which this plan was created.
//
// This is only recorded to allow for UI differences when presenting plans
// to the end-user, and so it must not be used to influence apply-time
// behavior. The actions during apply must be described entirely by
// the Changes field, regardless of how the plan was created.
UIMode Mode
VariableValues map[string]DynamicValue
Changes *Changes
DriftedResources []*ResourceInstanceChangeSrc
TargetAddrs []addrs.Targetable
ForceReplaceAddrs []addrs.AbsResourceInstance
Backend Backend
// Checks captures a snapshot of the (probably-incomplete) check results
// at the end of the planning process.
//
// If this plan is applyable (that is, if the planning process completed
// without errors) then the set of checks here should be complete even
// though some of them will likely have StatusUnknown where the check
// condition depends on values we won't know until the apply step.
Checks *states.CheckResults
// RelevantAttributes is a set of resource instance addresses and
// attributes that are either directly affected by proposed changes or may
// have indirectly contributed to them via references in expressions.
//
// This is the result of a heuristic and is intended only as a hint to
// the UI layer in case it wants to emphasize or de-emphasize certain
// resources. Don't use this to drive any non-cosmetic behavior, especially
// including anything that would be subject to compatibility constraints.
RelevantAttributes []globalref.ResourceAttr
// PrevRunState and PriorState both describe the situation that the plan
// was derived from:
//
// PrevRunState is a representation of the outcome of the previous
// Terraform operation, without any updates from the remote system but
// potentially including some changes that resulted from state upgrade
// actions.
//
// PriorState is a representation of the current state of remote objects,
// which will differ from PrevRunState if the "refresh" step returned
// different data, which might reflect drift.
//
// PriorState is the main snapshot we use for actions during apply.
// PrevRunState is only here so that we can diff PriorState against it in
// order to report to the user any out-of-band changes we've detected.
PrevRunState *states.State
PriorState *states.State
}
// CanApply returns true if and only if the recieving plan includes content
// that would make sense to apply. If it returns false, the plan operation
// should indicate that there's nothing to do and Terraform should exit
// without prompting the user to confirm the changes.
//
// This function represents our main business logic for making the decision
// about whether a given plan represents meaningful "changes", and so its
// exact definition may change over time; the intent is just to centralize the
// rules for that rather than duplicating different versions of it at various
// locations in the UI code.
func (p *Plan) CanApply() bool {
switch {
case !p.Changes.Empty():
// "Empty" means that everything in the changes is a "NoOp", so if
// not empty then there's at least one non-NoOp change.
return true
case !p.PriorState.ManagedResourcesEqual(p.PrevRunState):
// If there are no changes planned but we detected some
// outside-Terraform changes while refreshing then we consider
// that applyable in isolation only if this was a refresh-only
// plan where we expect updating the state to include these
// changes was the intended goal.
//
// (We don't treat a "refresh only" plan as applyable in normal
// planning mode because historically the refresh result wasn't
// considered part of a plan at all, and so it would be
// a disruptive breaking change if refreshing alone suddenly
// became applyable in the normal case and an existing configuration
// was relying on ignore_changes in order to be convergent in spite
// of intentional out-of-band operations.)
return p.UIMode == RefreshOnlyMode
default:
// Otherwise, there are either no changes to apply or they are changes
// our cases above don't consider as worthy of applying in isolation.
return false
}
}
// ProviderAddrs returns a list of all of the provider configuration addresses
// referenced throughout the receiving plan.
//
// The result is de-duplicated so that each distinct address appears only once.
func (p *Plan) ProviderAddrs() []addrs.AbsProviderConfig {
if p == nil || p.Changes == nil {
return nil
}
m := map[string]addrs.AbsProviderConfig{}
for _, rc := range p.Changes.Resources {
m[rc.ProviderAddr.String()] = rc.ProviderAddr
}
if len(m) == 0 {
return nil
}
// This is mainly just so we'll get stable results for testing purposes.
keys := make([]string, 0, len(m))
for k := range m {
keys = append(keys, k)
}
sort.Strings(keys)
ret := make([]addrs.AbsProviderConfig, len(keys))
for i, key := range keys {
ret[i] = m[key]
}
return ret
}
// Backend represents the backend-related configuration and other data as it
// existed when a plan was created.
type Backend struct {
// Type is the type of backend that the plan will apply against.
Type string
// Config is the configuration of the backend, whose schema is decided by
// the backend Type.
Config DynamicValue
// Workspace is the name of the workspace that was active when the plan
// was created. It is illegal to apply a plan created for one workspace
// to the state of another workspace.
// (This constraint is already enforced by the statefile lineage mechanism,
// but storing this explicitly allows us to return a better error message
// in the situation where the user has the wrong workspace selected.)
Workspace string
}
func NewBackend(typeName string, config cty.Value, configSchema *configschema.Block, workspaceName string) (*Backend, error) {
dv, err := NewDynamicValue(config, configSchema.ImpliedType())
if err != nil {
return nil, err
}
return &Backend{
Type: typeName,
Config: dv,
Workspace: workspaceName,
}, nil
}