pulumi/pkg/engine/update.go

244 lines
7.4 KiB
Go
Raw Normal View History

// Copyright 2018, Pulumi Corporation. All rights reserved.
package engine
import (
"time"
"github.com/pulumi/pulumi/pkg/diag"
"github.com/pulumi/pulumi/pkg/resource"
"github.com/pulumi/pulumi/pkg/resource/deploy"
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
"github.com/pulumi/pulumi/pkg/resource/plugin"
"github.com/pulumi/pulumi/pkg/util/contract"
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
"github.com/pulumi/pulumi/pkg/workspace"
)
// UpdateOptions contains all the settings for customizing how an update (deploy, preview, or destroy) is performed.
type UpdateOptions struct {
// an optional set of analyzers to run as part of this deployment.
Analyzers []string
// the degree of parallelism for resource operations (<=1 for serial).
Parallel int
// true if debugging output it enabled
Debug bool
}
// ResourceChanges contains the aggregate resource changes by operation type.
type ResourceChanges map[deploy.StepOp]int
// HasChanges returns true if there are any non-same changes in the resulting summary.
func (changes ResourceChanges) HasChanges() bool {
var c int
for op, count := range changes {
if op != deploy.OpSame {
c += count
}
}
return c > 0
}
func Update(u UpdateInfo, ctx *Context, opts UpdateOptions, dryRun bool) (ResourceChanges, error) {
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
contract.Require(u != nil, "update")
contract.Require(ctx != nil, "ctx")
defer func() { ctx.Events <- cancelEvent() }()
2017-10-23 00:52:00 +02:00
info, err := newPlanContext(u, "update", ctx.ParentSpan)
if err != nil {
return nil, err
}
defer info.Close()
emitter := makeEventEmitter(ctx.Events, u)
return update(ctx, info, planOptions{
UpdateOptions: opts,
SourceFunc: newUpdateSource,
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
Events: emitter,
Diag: newEventSink(emitter),
PluginEvents: &pluginActions{ctx},
}, dryRun)
}
func newUpdateSource(
opts planOptions, proj *workspace.Project, pwd, main string,
target *deploy.Target, plugctx *plugin.Context, dryRun bool) (deploy.Source, error) {
// Figure out which plugins to load by inspecting the program contents.
plugins, err := plugctx.Host.GetRequiredPlugins(plugin.ProgInfo{
Proj: proj,
Pwd: pwd,
Program: main,
})
if err != nil {
return nil, err
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
}
// Now ensure that we have loaded up any plugins that the program will need in advance.
if err = plugctx.Host.EnsurePlugins(plugins); err != nil {
return nil, err
}
// If that succeeded, create a new source that will perform interpretation of the compiled program.
// TODO[pulumi/pulumi#88]: we are passing `nil` as the arguments map; we need to allow a way to pass these.
return deploy.NewEvalSource(plugctx, &deploy.EvalRunInfo{
Proj: proj,
Pwd: pwd,
Program: main,
Target: target,
}, dryRun), nil
}
func update(ctx *Context, info *planContext, opts planOptions, dryRun bool) (ResourceChanges, error) {
result, err := plan(info, opts, dryRun)
if err != nil {
return nil, err
}
var resourceChanges ResourceChanges
if result != nil {
defer contract.IgnoreClose(result)
// Make the current working directory the same as the program's, and restore it upon exit.
done, err := result.Chdir()
if err != nil {
return nil, err
}
defer done()
if dryRun {
// If a dry run, just print the plan, don't actually carry out the deployment.
resourceChanges, err = printPlan(ctx, result, dryRun)
if err != nil {
return resourceChanges, err
}
} else {
// Otherwise, we will actually deploy the latest bits.
opts.Events.preludeEvent(dryRun, result.Ctx.Update.GetTarget().Config)
// Walk the plan, reporting progress and executing the actual operations as we go.
start := time.Now()
actions := newUpdateActions(ctx, info.Update, opts)
summary, step, _, err := result.Walk(ctx, actions, false)
2017-08-27 09:38:17 +02:00
if err != nil && summary == nil {
// Something went wrong, and no changes were made.
return resourceChanges, err
2017-08-27 09:38:17 +02:00
}
contract.Assert(summary != nil)
if err != nil {
var failedUrn resource.URN
if step != nil {
failedUrn = step.URN()
}
opts.Diag.Errorf(diag.Message(failedUrn, err.Error()))
}
// Print out the total number of steps performed (and their kinds), the duration, and any summary info.
resourceChanges = ResourceChanges(actions.Ops)
opts.Events.updateSummaryEvent(actions.MaybeCorrupt, time.Since(start), resourceChanges)
if err != nil {
return resourceChanges, err
}
}
}
return resourceChanges, nil
}
// pluginActions listens for plugin events and persists the set of loaded plugins
// to the snapshot.
type pluginActions struct {
Context *Context
}
func (p *pluginActions) OnPluginLoad(loadedPlug workspace.PluginInfo) error {
return p.Context.SnapshotManager.RecordPlugin(loadedPlug)
}
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
// updateActions pretty-prints the plan application process as it goes.
type updateActions struct {
Context *Context
Steps int
Ops map[deploy.StepOp]int
Seen map[resource.URN]deploy.Step
MaybeCorrupt bool
Update UpdateInfo
Opts planOptions
}
func newUpdateActions(context *Context, u UpdateInfo, opts planOptions) *updateActions {
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
return &updateActions{
Context: context,
Ops: make(map[deploy.StepOp]int),
Seen: make(map[resource.URN]deploy.Step),
Update: u,
Opts: opts,
Switch to parent pointers; display components nicely This change switches from child lists to parent pointers, in the way resource ancestries are represented. This cleans up a fair bit of the old parenting logic, including all notion of ambient parent scopes (and will notably address pulumi/pulumi#435). This lets us show a more parent/child display in the output when doing planning and updating. For instance, here is an update of a lambda's text, which is logically part of a cloud timer: * cloud:timer:Timer: (same) [urn=urn:pulumi:malta::lm-cloud::cloud:timer:Timer::lm-cts-malta-job-CleanSnapshots] * cloud:function:Function: (same) [urn=urn:pulumi:malta::lm-cloud::cloud:function:Function::lm-cts-malta-job-CleanSnapshots] * aws:serverless:Function: (same) [urn=urn:pulumi:malta::lm-cloud::aws:serverless:Function::lm-cts-malta-job-CleanSnapshots] ~ aws:lambda/function:Function: (modify) [id=lm-cts-malta-job-CleanSnapshots-fee4f3bf41280741] [urn=urn:pulumi:malta::lm-cloud::aws:lambda/function:Function::lm-cts-malta-job-CleanSnapshots] - code : archive(assets:2092f44) { // etc etc etc Note that we still get walls of text, but this will be actually quite nice when combined with pulumi/pulumi#454. I've also suppressed printing properties that didn't change during updates when --detailed was not passed, and also suppressed empty strings and zero-length arrays (since TF uses these as defaults in many places and it just makes creation and deletion quite verbose). Note that this is a far cry from everything we can possibly do here as part of pulumi/pulumi#340 (and even pulumi/pulumi#417). But it's a good start towards taming some of our output spew.
2017-11-17 03:21:41 +01:00
}
}
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
func (acts *updateActions) OnResourceStepPre(step deploy.Step) (interface{}, error) {
// Ensure we've marked this step as observed.
acts.Seen[step.URN()] = step
acts.Opts.Events.resourcePreEvent(step, false /*planning*/, acts.Opts.Debug)
// Inform the snapshot service that we are about to perform a step.
return acts.Context.SnapshotManager.BeginMutation(step)
}
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
func (acts *updateActions) OnResourceStepPost(ctx interface{},
step deploy.Step, status resource.Status, err error) error {
assertSeen(acts.Seen, step)
// If we've already been terminated, exit without writing the checkpoint. We explicitly want to leave the
// checkpoint in an inconsistent state in this event.
if acts.Context.Cancel.TerminateErr() != nil {
return nil
}
// Report the result of the step.
stepop := step.Op()
if err != nil {
if status == resource.StatusUnknown {
acts.MaybeCorrupt = true
}
// Issue a true, bonafide error.
acts.Opts.Diag.Errorf(diag.GetPlanApplyFailedError(step.URN()), err)
acts.Opts.Events.resourceOperationFailedEvent(step, status, acts.Steps, acts.Opts.Debug)
} else {
if step.Logical() {
// Increment the counters.
acts.Steps++
acts.Ops[stepop]++
}
2018-04-20 20:52:33 +02:00
// Also show outputs here for custom resources, since there might be some from the initial registration. We do
// not show outputs for component resources at this point: any that exist must be from a previous execution of
// the Pulumi program, as component resources only report outputs via calls to RegisterResourceOutputs.
if step.Res().Custom {
acts.Opts.Events.resourceOutputsEvent(step, false /*planning*/, acts.Opts.Debug)
}
}
// Write out the current snapshot. Note that even if a failure has occurred, we should still have a
// safe checkpoint. Note that any error that occurs when writing the checkpoint trumps the error
// reported above.
return ctx.(SnapshotMutation).End(step, err == nil)
}
General prep work for refresh This change includes a bunch of refactorings I made in prep for doing refresh (first, the command, see pulumi/pulumi#1081): * The primary change is to change the way the engine's core update functionality works with respect to deploy.Source. This is the way we can plug in new sources of resource information during planning (and, soon, diffing). The way I intend to model refresh is by having a new kind of source, deploy.RefreshSource, which will let us do virtually everything about an update/diff the same way with refreshes, which avoid otherwise duplicative effort. This includes changing the planOptions (nee deployOptions) to take a new SourceFunc callback, which is responsible for creating a source specific to the kind of plan being requested. Preview, Update, and Destroy now are primarily differentiated by the kind of deploy.Source that they return, rather than sprinkling things like `if Destroying` throughout. This tidies up some logic and, more importantly, gives us precisely the refresh hook we need. * Originally, we used the deploy.NullSource for Destroy operations. This simply returns nothing, which is how Destroy works. For some reason, we were no longer doing this, and instead had some `if Destroying` cases sprinkled throughout the deploy.EvalSource. I think this is a vestige of some old way we did configuration, at least judging by a comment, which is apparently no longer relevant. * Move diff and diff-printing logic within the engine into its own pkg/engine/diff.go file, to prepare for upcoming work. * I keep noticing benign diffs anytime I regenerate protobufs. I suspect this is because we're also on different versions. I changed generate.sh to also dump the version into grpc_version.txt. At least we can understand where the diffs are coming from, decide whether to take them (i.e., a newer version), and ensure that as a team we are monotonically increasing, and not going backwards. * I also tidied up some tiny things I noticed while in there, like comments, incorrect types, lint suppressions, and so on.
2018-03-28 16:45:23 +02:00
func (acts *updateActions) OnResourceOutputs(step deploy.Step) error {
assertSeen(acts.Seen, step)
acts.Opts.Events.resourceOutputsEvent(step, false /*planning*/, acts.Opts.Debug)
// There's a chance there are new outputs that weren't written out last time.
// We need to perform another snapshot write to ensure they get written out.
return acts.Context.SnapshotManager.RegisterResourceOutputs(step)
}