pulumi/sdk/nodejs/resource.ts

605 lines
27 KiB
TypeScript
Raw Normal View History

2018-05-22 21:43:36 +02:00
// Copyright 2016-2018, Pulumi Corporation.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
Implement initial Lumi-as-a-library This is the initial step towards redefining Lumi as a library that runs atop vanilla Node.js/V8, rather than as its own runtime. This change is woefully incomplete but this includes some of the more stable pieces of my current work-in-progress. The new structure is that within the sdk/ directory we will have a client library per language. This client library contains the object model for Lumi (resources, properties, assets, config, etc), in addition to the "language runtime host" components required to interoperate with the Lumi resource monitor. This resource monitor is effectively what we call "Lumi" today, in that it's the thing orchestrating plans and deployments. Inside the sdk/ directory, you will find nodejs/, the Node.js client library, alongside proto/, the definitions for RPC interop between the different pieces of the system. This includes existing RPC definitions for resource providers, etc., in addition to the new ones for hosting different language runtimes from within Lumi. These new interfaces are surprisingly simple. There is effectively a bidirectional RPC channel between the Lumi resource monitor, represented by the lumirpc.ResourceMonitor interface, and each language runtime, represented by the lumirpc.LanguageRuntime interface. The overall orchestration goes as follows: 1) Lumi decides it needs to run a program written in language X, so it dynamically loads the language runtime plugin for language X. 2) Lumi passes that runtime a loopback address to its ResourceMonitor service, while language X will publish a connection back to its LanguageRuntime service, which Lumi will talk to. 3) Lumi then invokes LanguageRuntime.Run, passing information like the desired working directory, program name, arguments, and optional configuration variables to make available to the program. 4) The language X runtime receives this, unpacks it and sets up the necessary context, and then invokes the program. The program then calls into Lumi object model abstractions that internally communicate back to Lumi using the ResourceMonitor interface. 5) The key here is ResourceMonitor.NewResource, which Lumi uses to serialize state about newly allocated resources. Lumi receives these and registers them as part of the plan, doing the usual diffing, etc., to decide how to proceed. This interface is perhaps one of the most subtle parts of the new design, as it necessitates the use of promises internally to allow parallel evaluation of the resource plan, letting dataflow determine the available concurrency. 6) The program exits, and Lumi continues on its merry way. If the program fails, the RunResponse will include information about the failure. Due to (5), all properties on resources are now instances of a new Property<T> type. A Property<T> is just a thin wrapper over a T, but it encodes the special properties of Lumi resource properties. Namely, it is possible to create one out of a T, other Property<T>, Promise<T>, or to freshly allocate one. In all cases, the Property<T> does not "settle" until its final state is known. This cannot occur before the deployment actually completes, and so in general it's not safe to depend on concrete resolutions of values (unlike ordinary Promise<T>s which are usually expected to resolve). As a result, all derived computations are meant to use the `then` function (as in `someValue.then(v => v+x)`). Although this change includes tests that may be run in isolation to test the various RPC interactions, we are nowhere near finished. The remaining work primarily boils down to three things: 1) Wiring all of this up to the Lumi code. 2) Fixing the handful of known loose ends required to make this work, primarily around the serialization of properties (waiting on unresolved ones, serializing assets properly, etc). 3) Implementing lambda closure serialization as a native extension. This ongoing work is part of pulumi/pulumi-fabric#311.
2017-08-26 21:07:54 +02:00
import { RunError } from "./errors";
import * as runtime from "./runtime";
import { readResource, registerResource, registerResourceOutputs } from "./runtime/resource";
2017-11-26 19:04:15 +01:00
import { getRootResource } from "./runtime/settings";
Implement initial Lumi-as-a-library This is the initial step towards redefining Lumi as a library that runs atop vanilla Node.js/V8, rather than as its own runtime. This change is woefully incomplete but this includes some of the more stable pieces of my current work-in-progress. The new structure is that within the sdk/ directory we will have a client library per language. This client library contains the object model for Lumi (resources, properties, assets, config, etc), in addition to the "language runtime host" components required to interoperate with the Lumi resource monitor. This resource monitor is effectively what we call "Lumi" today, in that it's the thing orchestrating plans and deployments. Inside the sdk/ directory, you will find nodejs/, the Node.js client library, alongside proto/, the definitions for RPC interop between the different pieces of the system. This includes existing RPC definitions for resource providers, etc., in addition to the new ones for hosting different language runtimes from within Lumi. These new interfaces are surprisingly simple. There is effectively a bidirectional RPC channel between the Lumi resource monitor, represented by the lumirpc.ResourceMonitor interface, and each language runtime, represented by the lumirpc.LanguageRuntime interface. The overall orchestration goes as follows: 1) Lumi decides it needs to run a program written in language X, so it dynamically loads the language runtime plugin for language X. 2) Lumi passes that runtime a loopback address to its ResourceMonitor service, while language X will publish a connection back to its LanguageRuntime service, which Lumi will talk to. 3) Lumi then invokes LanguageRuntime.Run, passing information like the desired working directory, program name, arguments, and optional configuration variables to make available to the program. 4) The language X runtime receives this, unpacks it and sets up the necessary context, and then invokes the program. The program then calls into Lumi object model abstractions that internally communicate back to Lumi using the ResourceMonitor interface. 5) The key here is ResourceMonitor.NewResource, which Lumi uses to serialize state about newly allocated resources. Lumi receives these and registers them as part of the plan, doing the usual diffing, etc., to decide how to proceed. This interface is perhaps one of the most subtle parts of the new design, as it necessitates the use of promises internally to allow parallel evaluation of the resource plan, letting dataflow determine the available concurrency. 6) The program exits, and Lumi continues on its merry way. If the program fails, the RunResponse will include information about the failure. Due to (5), all properties on resources are now instances of a new Property<T> type. A Property<T> is just a thin wrapper over a T, but it encodes the special properties of Lumi resource properties. Namely, it is possible to create one out of a T, other Property<T>, Promise<T>, or to freshly allocate one. In all cases, the Property<T> does not "settle" until its final state is known. This cannot occur before the deployment actually completes, and so in general it's not safe to depend on concrete resolutions of values (unlike ordinary Promise<T>s which are usually expected to resolve). As a result, all derived computations are meant to use the `then` function (as in `someValue.then(v => v+x)`). Although this change includes tests that may be run in isolation to test the various RPC interactions, we are nowhere near finished. The remaining work primarily boils down to three things: 1) Wiring all of this up to the Lumi code. 2) Fixing the handful of known loose ends required to make this work, primarily around the serialization of properties (waiting on unresolved ones, serializing assets properly, etc). 3) Implementing lambda closure serialization as a native extension. This ongoing work is part of pulumi/pulumi-fabric#311.
2017-08-26 21:07:54 +02:00
export type ID = string; // a provider-assigned ID.
export type URN = string; // an automatically generated logical URN, used to stably identify resources.
/**
* Resource represents a class whose CRUD operations are implemented by a provider plugin.
*/
Implement initial Lumi-as-a-library This is the initial step towards redefining Lumi as a library that runs atop vanilla Node.js/V8, rather than as its own runtime. This change is woefully incomplete but this includes some of the more stable pieces of my current work-in-progress. The new structure is that within the sdk/ directory we will have a client library per language. This client library contains the object model for Lumi (resources, properties, assets, config, etc), in addition to the "language runtime host" components required to interoperate with the Lumi resource monitor. This resource monitor is effectively what we call "Lumi" today, in that it's the thing orchestrating plans and deployments. Inside the sdk/ directory, you will find nodejs/, the Node.js client library, alongside proto/, the definitions for RPC interop between the different pieces of the system. This includes existing RPC definitions for resource providers, etc., in addition to the new ones for hosting different language runtimes from within Lumi. These new interfaces are surprisingly simple. There is effectively a bidirectional RPC channel between the Lumi resource monitor, represented by the lumirpc.ResourceMonitor interface, and each language runtime, represented by the lumirpc.LanguageRuntime interface. The overall orchestration goes as follows: 1) Lumi decides it needs to run a program written in language X, so it dynamically loads the language runtime plugin for language X. 2) Lumi passes that runtime a loopback address to its ResourceMonitor service, while language X will publish a connection back to its LanguageRuntime service, which Lumi will talk to. 3) Lumi then invokes LanguageRuntime.Run, passing information like the desired working directory, program name, arguments, and optional configuration variables to make available to the program. 4) The language X runtime receives this, unpacks it and sets up the necessary context, and then invokes the program. The program then calls into Lumi object model abstractions that internally communicate back to Lumi using the ResourceMonitor interface. 5) The key here is ResourceMonitor.NewResource, which Lumi uses to serialize state about newly allocated resources. Lumi receives these and registers them as part of the plan, doing the usual diffing, etc., to decide how to proceed. This interface is perhaps one of the most subtle parts of the new design, as it necessitates the use of promises internally to allow parallel evaluation of the resource plan, letting dataflow determine the available concurrency. 6) The program exits, and Lumi continues on its merry way. If the program fails, the RunResponse will include information about the failure. Due to (5), all properties on resources are now instances of a new Property<T> type. A Property<T> is just a thin wrapper over a T, but it encodes the special properties of Lumi resource properties. Namely, it is possible to create one out of a T, other Property<T>, Promise<T>, or to freshly allocate one. In all cases, the Property<T> does not "settle" until its final state is known. This cannot occur before the deployment actually completes, and so in general it's not safe to depend on concrete resolutions of values (unlike ordinary Promise<T>s which are usually expected to resolve). As a result, all derived computations are meant to use the `then` function (as in `someValue.then(v => v+x)`). Although this change includes tests that may be run in isolation to test the various RPC interactions, we are nowhere near finished. The remaining work primarily boils down to three things: 1) Wiring all of this up to the Lumi code. 2) Fixing the handful of known loose ends required to make this work, primarily around the serialization of properties (waiting on unresolved ones, serializing assets properly, etc). 3) Implementing lambda closure serialization as a native extension. This ongoing work is part of pulumi/pulumi-fabric#311.
2017-08-26 21:07:54 +02:00
export abstract class Resource {
/**
* A private field to help with RTTI that works in SxS scenarios.
*/
// tslint:disable-next-line:variable-name
/* @internal */ private readonly __pulumiResource: boolean = true;
/**
2018-01-26 03:50:58 +01:00
* urn is the stable logical URN used to distinctly address a resource, both before and after
* deployments.
*/
2018-02-06 03:37:10 +01:00
public readonly urn: Output<URN>;
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
/**
* When set to true, protect ensures this resource cannot be deleted.
*/
// tslint:disable-next-line:variable-name
/* @internal */ private readonly __protect: boolean;
/**
* The set of providers to use for child resources. Keyed by package name (e.g. "aws").
*/
// tslint:disable-next-line:variable-name
/* @internal */ private readonly __providers: Record<string, ProviderResource>;
public static isInstance(obj: any): obj is Resource {
return obj && obj.__pulumiResource;
}
// getProvider fetches the provider for the given module member, if any.
public getProvider(moduleMember: string): ProviderResource | undefined {
const memComponents = moduleMember.split(":");
if (memComponents.length !== 3) {
return undefined;
}
const pkg = memComponents[0];
return this.__providers[pkg];
}
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
/**
2018-01-26 03:50:58 +01:00
* Creates and registers a new resource object. t is the fully qualified type token and name is
* the "name" part to use in creating a stable and globally unique URN for the object.
* dependsOn is an optional list of other resources that this resource depends on, controlling
* the order in which we perform resource operations.
*
* @param t The type of the resource.
2018-02-08 00:01:55 +01:00
* @param name The _unique_ name of the resource.
* @param custom True to indicate that this is a custom resource, managed by a plugin.
* @param props The arguments to use to populate the new resource.
* @param opts A bag of options that control this resource's behavior.
*/
constructor(t: string, name: string, custom: boolean, props: Inputs = {}, opts: ResourceOptions = {}) {
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
if (!t) {
throw new RunError("Missing resource type argument");
Implement initial Lumi-as-a-library This is the initial step towards redefining Lumi as a library that runs atop vanilla Node.js/V8, rather than as its own runtime. This change is woefully incomplete but this includes some of the more stable pieces of my current work-in-progress. The new structure is that within the sdk/ directory we will have a client library per language. This client library contains the object model for Lumi (resources, properties, assets, config, etc), in addition to the "language runtime host" components required to interoperate with the Lumi resource monitor. This resource monitor is effectively what we call "Lumi" today, in that it's the thing orchestrating plans and deployments. Inside the sdk/ directory, you will find nodejs/, the Node.js client library, alongside proto/, the definitions for RPC interop between the different pieces of the system. This includes existing RPC definitions for resource providers, etc., in addition to the new ones for hosting different language runtimes from within Lumi. These new interfaces are surprisingly simple. There is effectively a bidirectional RPC channel between the Lumi resource monitor, represented by the lumirpc.ResourceMonitor interface, and each language runtime, represented by the lumirpc.LanguageRuntime interface. The overall orchestration goes as follows: 1) Lumi decides it needs to run a program written in language X, so it dynamically loads the language runtime plugin for language X. 2) Lumi passes that runtime a loopback address to its ResourceMonitor service, while language X will publish a connection back to its LanguageRuntime service, which Lumi will talk to. 3) Lumi then invokes LanguageRuntime.Run, passing information like the desired working directory, program name, arguments, and optional configuration variables to make available to the program. 4) The language X runtime receives this, unpacks it and sets up the necessary context, and then invokes the program. The program then calls into Lumi object model abstractions that internally communicate back to Lumi using the ResourceMonitor interface. 5) The key here is ResourceMonitor.NewResource, which Lumi uses to serialize state about newly allocated resources. Lumi receives these and registers them as part of the plan, doing the usual diffing, etc., to decide how to proceed. This interface is perhaps one of the most subtle parts of the new design, as it necessitates the use of promises internally to allow parallel evaluation of the resource plan, letting dataflow determine the available concurrency. 6) The program exits, and Lumi continues on its merry way. If the program fails, the RunResponse will include information about the failure. Due to (5), all properties on resources are now instances of a new Property<T> type. A Property<T> is just a thin wrapper over a T, but it encodes the special properties of Lumi resource properties. Namely, it is possible to create one out of a T, other Property<T>, Promise<T>, or to freshly allocate one. In all cases, the Property<T> does not "settle" until its final state is known. This cannot occur before the deployment actually completes, and so in general it's not safe to depend on concrete resolutions of values (unlike ordinary Promise<T>s which are usually expected to resolve). As a result, all derived computations are meant to use the `then` function (as in `someValue.then(v => v+x)`). Although this change includes tests that may be run in isolation to test the various RPC interactions, we are nowhere near finished. The remaining work primarily boils down to three things: 1) Wiring all of this up to the Lumi code. 2) Fixing the handful of known loose ends required to make this work, primarily around the serialization of properties (waiting on unresolved ones, serializing assets properly, etc). 3) Implementing lambda closure serialization as a native extension. This ongoing work is part of pulumi/pulumi-fabric#311.
2017-08-26 21:07:54 +02:00
}
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
if (!name) {
throw new RunError("Missing resource name argument (for URN creation)");
Implement initial Lumi-as-a-library This is the initial step towards redefining Lumi as a library that runs atop vanilla Node.js/V8, rather than as its own runtime. This change is woefully incomplete but this includes some of the more stable pieces of my current work-in-progress. The new structure is that within the sdk/ directory we will have a client library per language. This client library contains the object model for Lumi (resources, properties, assets, config, etc), in addition to the "language runtime host" components required to interoperate with the Lumi resource monitor. This resource monitor is effectively what we call "Lumi" today, in that it's the thing orchestrating plans and deployments. Inside the sdk/ directory, you will find nodejs/, the Node.js client library, alongside proto/, the definitions for RPC interop between the different pieces of the system. This includes existing RPC definitions for resource providers, etc., in addition to the new ones for hosting different language runtimes from within Lumi. These new interfaces are surprisingly simple. There is effectively a bidirectional RPC channel between the Lumi resource monitor, represented by the lumirpc.ResourceMonitor interface, and each language runtime, represented by the lumirpc.LanguageRuntime interface. The overall orchestration goes as follows: 1) Lumi decides it needs to run a program written in language X, so it dynamically loads the language runtime plugin for language X. 2) Lumi passes that runtime a loopback address to its ResourceMonitor service, while language X will publish a connection back to its LanguageRuntime service, which Lumi will talk to. 3) Lumi then invokes LanguageRuntime.Run, passing information like the desired working directory, program name, arguments, and optional configuration variables to make available to the program. 4) The language X runtime receives this, unpacks it and sets up the necessary context, and then invokes the program. The program then calls into Lumi object model abstractions that internally communicate back to Lumi using the ResourceMonitor interface. 5) The key here is ResourceMonitor.NewResource, which Lumi uses to serialize state about newly allocated resources. Lumi receives these and registers them as part of the plan, doing the usual diffing, etc., to decide how to proceed. This interface is perhaps one of the most subtle parts of the new design, as it necessitates the use of promises internally to allow parallel evaluation of the resource plan, letting dataflow determine the available concurrency. 6) The program exits, and Lumi continues on its merry way. If the program fails, the RunResponse will include information about the failure. Due to (5), all properties on resources are now instances of a new Property<T> type. A Property<T> is just a thin wrapper over a T, but it encodes the special properties of Lumi resource properties. Namely, it is possible to create one out of a T, other Property<T>, Promise<T>, or to freshly allocate one. In all cases, the Property<T> does not "settle" until its final state is known. This cannot occur before the deployment actually completes, and so in general it's not safe to depend on concrete resolutions of values (unlike ordinary Promise<T>s which are usually expected to resolve). As a result, all derived computations are meant to use the `then` function (as in `someValue.then(v => v+x)`). Although this change includes tests that may be run in isolation to test the various RPC interactions, we are nowhere near finished. The remaining work primarily boils down to three things: 1) Wiring all of this up to the Lumi code. 2) Fixing the handful of known loose ends required to make this work, primarily around the serialization of properties (waiting on unresolved ones, serializing assets properly, etc). 3) Implementing lambda closure serialization as a native extension. This ongoing work is part of pulumi/pulumi-fabric#311.
2017-08-26 21:07:54 +02:00
}
// Check the parent type if one exists and fill in any default options.
this.__providers = {};
if (opts.parent) {
if (!Resource.isInstance(opts.parent)) {
throw new RunError(`Resource parent is not a valid Resource: ${opts.parent}`);
}
if (opts.protect === undefined) {
opts.protect = opts.parent.__protect;
}
this.__providers = opts.parent.__providers;
if (custom) {
const provider = (<CustomResourceOptions>opts).provider;
if (provider === undefined) {
(<CustomResourceOptions>opts).provider = opts.parent.getProvider(t);
}
}
}
if (!custom) {
const providers = (<ComponentResourceOptions>opts).providers;
if (providers) {
this.__providers = { ...this.__providers, ...providers };
}
}
this.__protect = !!opts.protect;
if (opts.id) {
// If this resource already exists, read its state rather than registering it anew.
if (!custom) {
throw new RunError("Cannot read an existing resource unless it has a custom provider");
}
readResource(this, t, name, props, opts);
} else {
// Kick off the resource registration. If we are actually performing a deployment, this
// resource's properties will be resolved asynchronously after the operation completes, so
// that dependent computations resolve normally. If we are just planning, on the other
// hand, values will never resolve.
registerResource(this, t, name, custom, props, opts);
}
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
}
}
(<any>Resource).doNotCapture = true;
/**
* ResourceOptions is a bag of optional settings that control a resource's behavior.
*/
export interface ResourceOptions {
/**
* An optional existing ID to load, rather than create.
*/
id?: Input<ID>;
/**
* An optional parent resource to which this resource belongs.
*/
parent?: Resource;
/**
* An optional additional explicit dependencies on other resources.
*/
dependsOn?: Resource[] | Resource;
/**
* When set to true, protect ensures this resource cannot be deleted.
*/
protect?: boolean;
}
/**
* CustomResourceOptions is a bag of optional settings that control a custom resource's behavior.
*/
export interface CustomResourceOptions extends ResourceOptions {
Implement first-class providers. (#1695) ### First-Class Providers These changes implement support for first-class providers. First-class providers are provider plugins that are exposed as resources via the Pulumi programming model so that they may be explicitly and multiply instantiated. Each instance of a provider resource may be configured differently, and configuration parameters may be source from the outputs of other resources. ### Provider Plugin Changes In order to accommodate the need to verify and diff provider configuration and configure providers without complete configuration information, these changes adjust the high-level provider plugin interface. Two new methods for validating a provider's configuration and diffing changes to the same have been added (`CheckConfig` and `DiffConfig`, respectively), and the type of the configuration bag accepted by `Configure` has been changed to a `PropertyMap`. These changes have not yet been reflected in the provider plugin gRPC interface. We will do this in a set of follow-up changes. Until then, these methods are implemented by adapters: - `CheckConfig` validates that all configuration parameters are string or unknown properties. This is necessary because existing plugins only accept string-typed configuration values. - `DiffConfig` either returns "never replace" if all configuration values are known or "must replace" if any configuration value is unknown. The justification for this behavior is given [here](https://github.com/pulumi/pulumi/pull/1695/files#diff-a6cd5c7f337665f5bb22e92ca5f07537R106) - `Configure` converts the config bag to a legacy config map and configures the provider plugin if all config values are known. If any config value is unknown, the underlying plugin is not configured and the provider may only perform `Check`, `Read`, and `Invoke`, all of which return empty results. We justify this behavior becuase it is only possible during a preview and provides the best experience we can manage with the existing gRPC interface. ### Resource Model Changes Providers are now exposed as resources that participate in a stack's dependency graph. Like other resources, they are explicitly created, may have multiple instances, and may have dependencies on other resources. Providers are referred to using provider references, which are a combination of the provider's URN and its ID. This design addresses the need during a preview to refer to providers that have not yet been physically created and therefore have no ID. All custom resources that are not themselves providers must specify a single provider via a provider reference. The named provider will be used to manage that resource's CRUD operations. If a resource's provider reference changes, the resource must be replaced. Though its URN is not present in the resource's dependency list, the provider should be treated as a dependency of the resource when topologically sorting the dependency graph. Finally, `Invoke` operations must now specify a provider to use for the invocation via a provider reference. ### Engine Changes First-class providers support requires a few changes to the engine: - The engine must have some way to map from provider references to provider plugins. It must be possible to add providers from a stack's checkpoint to this map and to register new/updated providers during the execution of a plan in response to CRUD operations on provider resources. - In order to support updating existing stacks using existing Pulumi programs that may not explicitly instantiate providers, the engine must be able to manage the "default" providers for each package referenced by a checkpoint or Pulumi program. The configuration for a "default" provider is taken from the stack's configuration data. The former need is addressed by adding a provider registry type that is responsible for managing all of the plugins required by a plan. In addition to loading plugins froma checkpoint and providing the ability to map from a provider reference to a provider plugin, this type serves as the provider plugin for providers themselves (i.e. it is the "provider provider"). The latter need is solved via two relatively self-contained changes to plan setup and the eval source. During plan setup, the old checkpoint is scanned for custom resources that do not have a provider reference in order to compute the set of packages that require a default provider. Once this set has been computed, the required default provider definitions are conjured and prepended to the checkpoint's resource list. Each resource that requires a default provider is then updated to refer to the default provider for its package. While an eval source is running, each custom resource registration, resource read, and invoke that does not name a provider is trapped before being returned by the source iterator. If no default provider for the appropriate package has been registered, the eval source synthesizes an appropriate registration, waits for it to complete, and records the registered provider's reference. This reference is injected into the original request, which is then processed as usual. If a default provider was already registered, the recorded reference is used and no new registration occurs. ### SDK Changes These changes only expose first-class providers from the Node.JS SDK. - A new abstract class, `ProviderResource`, can be subclassed and used to instantiate first-class providers. - A new field in `ResourceOptions`, `provider`, can be used to supply a particular provider instance to manage a `CustomResource`'s CRUD operations. - A new type, `InvokeOptions`, can be used to specify options that control the behavior of a call to `pulumi.runtime.invoke`. This type includes a `provider` field that is analogous to `ResourceOptions.provider`.
2018-08-07 02:50:29 +02:00
/**
* An optional provider to use for this resource's CRUD operations. If no provider is supplied, the default
* provider for the resource's package will be used. The default provider is pulled from the parent's
* provider bag (see also ComponentResourceOptions.providers).
Implement first-class providers. (#1695) ### First-Class Providers These changes implement support for first-class providers. First-class providers are provider plugins that are exposed as resources via the Pulumi programming model so that they may be explicitly and multiply instantiated. Each instance of a provider resource may be configured differently, and configuration parameters may be source from the outputs of other resources. ### Provider Plugin Changes In order to accommodate the need to verify and diff provider configuration and configure providers without complete configuration information, these changes adjust the high-level provider plugin interface. Two new methods for validating a provider's configuration and diffing changes to the same have been added (`CheckConfig` and `DiffConfig`, respectively), and the type of the configuration bag accepted by `Configure` has been changed to a `PropertyMap`. These changes have not yet been reflected in the provider plugin gRPC interface. We will do this in a set of follow-up changes. Until then, these methods are implemented by adapters: - `CheckConfig` validates that all configuration parameters are string or unknown properties. This is necessary because existing plugins only accept string-typed configuration values. - `DiffConfig` either returns "never replace" if all configuration values are known or "must replace" if any configuration value is unknown. The justification for this behavior is given [here](https://github.com/pulumi/pulumi/pull/1695/files#diff-a6cd5c7f337665f5bb22e92ca5f07537R106) - `Configure` converts the config bag to a legacy config map and configures the provider plugin if all config values are known. If any config value is unknown, the underlying plugin is not configured and the provider may only perform `Check`, `Read`, and `Invoke`, all of which return empty results. We justify this behavior becuase it is only possible during a preview and provides the best experience we can manage with the existing gRPC interface. ### Resource Model Changes Providers are now exposed as resources that participate in a stack's dependency graph. Like other resources, they are explicitly created, may have multiple instances, and may have dependencies on other resources. Providers are referred to using provider references, which are a combination of the provider's URN and its ID. This design addresses the need during a preview to refer to providers that have not yet been physically created and therefore have no ID. All custom resources that are not themselves providers must specify a single provider via a provider reference. The named provider will be used to manage that resource's CRUD operations. If a resource's provider reference changes, the resource must be replaced. Though its URN is not present in the resource's dependency list, the provider should be treated as a dependency of the resource when topologically sorting the dependency graph. Finally, `Invoke` operations must now specify a provider to use for the invocation via a provider reference. ### Engine Changes First-class providers support requires a few changes to the engine: - The engine must have some way to map from provider references to provider plugins. It must be possible to add providers from a stack's checkpoint to this map and to register new/updated providers during the execution of a plan in response to CRUD operations on provider resources. - In order to support updating existing stacks using existing Pulumi programs that may not explicitly instantiate providers, the engine must be able to manage the "default" providers for each package referenced by a checkpoint or Pulumi program. The configuration for a "default" provider is taken from the stack's configuration data. The former need is addressed by adding a provider registry type that is responsible for managing all of the plugins required by a plan. In addition to loading plugins froma checkpoint and providing the ability to map from a provider reference to a provider plugin, this type serves as the provider plugin for providers themselves (i.e. it is the "provider provider"). The latter need is solved via two relatively self-contained changes to plan setup and the eval source. During plan setup, the old checkpoint is scanned for custom resources that do not have a provider reference in order to compute the set of packages that require a default provider. Once this set has been computed, the required default provider definitions are conjured and prepended to the checkpoint's resource list. Each resource that requires a default provider is then updated to refer to the default provider for its package. While an eval source is running, each custom resource registration, resource read, and invoke that does not name a provider is trapped before being returned by the source iterator. If no default provider for the appropriate package has been registered, the eval source synthesizes an appropriate registration, waits for it to complete, and records the registered provider's reference. This reference is injected into the original request, which is then processed as usual. If a default provider was already registered, the recorded reference is used and no new registration occurs. ### SDK Changes These changes only expose first-class providers from the Node.JS SDK. - A new abstract class, `ProviderResource`, can be subclassed and used to instantiate first-class providers. - A new field in `ResourceOptions`, `provider`, can be used to supply a particular provider instance to manage a `CustomResource`'s CRUD operations. - A new type, `InvokeOptions`, can be used to specify options that control the behavior of a call to `pulumi.runtime.invoke`. This type includes a `provider` field that is analogous to `ResourceOptions.provider`.
2018-08-07 02:50:29 +02:00
*/
provider?: ProviderResource;
}
/**
* ComponentResourceOptions is a bag of optional settings that control a component resource's behavior.
*/
export interface ComponentResourceOptions extends ResourceOptions {
/**
* An optional set of providers to use for child resources. Keyed by package name (e.g. "aws")
*/
providers?: Record<string, ProviderResource>;
}
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
/**
2018-01-26 03:50:58 +01:00
* CustomResource is a resource whose create, read, update, and delete (CRUD) operations are managed
* by performing external operations on some physical entity. The engine understands how to diff
* and perform partial updates of them, and these CRUD operations are implemented in a dynamically
* loaded plugin for the defining package.
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
*/
export abstract class CustomResource extends Resource {
/**
* A private field to help with RTTI that works in SxS scenarios.
*/
// tslint:disable-next-line:variable-name
/* @internal */ private readonly __pulumiCustomResource: boolean = true;
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
/**
2018-01-26 03:50:58 +01:00
* id is the provider-assigned unique ID for this managed resource. It is set during
* deployments and may be missing (undefined) during planning phases.
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
*/
2018-02-06 03:37:10 +01:00
public readonly id: Output<ID>;
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
/**
* Returns true if the given object is an instance of CustomResource. This is designed to work even when
* multiple copies of the Pulumi SDK have been loaded into the same process.
*/
public static isInstance(obj: any): obj is CustomResource {
return obj && obj.__pulumiCustomResource;
}
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
/**
2018-01-26 03:50:58 +01:00
* Creates and registers a new managed resource. t is the fully qualified type token and name
* is the "name" part to use in creating a stable and globally unique URN for the object.
* dependsOn is an optional list of other resources that this resource depends on, controlling
* the order in which we perform resource operations. Creating an instance does not necessarily
* perform a create on the physical entity which it represents, and instead, this is dependent
* upon the diffing of the new goal state compared to the current known resource state.
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
*
* @param t The type of the resource.
2018-02-08 00:01:55 +01:00
* @param name The _unique_ name of the resource.
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
* @param props The arguments to use to populate the new resource.
* @param opts A bag of options that control this resource's behavior.
Implement components This change implements core support for "components" in the Pulumi Fabric. This work is described further in pulumi/pulumi#340, where we are still discussing some of the finer points. In a nutshell, resources no longer imply external providers. It's entirely possible to have a resource that logically represents something but without having a physical manifestation that needs to be tracked and managed by our typical CRUD operations. For example, the aws/serverless/Function helper is one such type. It aggregates Lambda-related resources and exposes a nice interface. All of the Pulumi Cloud Framework resources are also examples. To indicate that a resource does participate in the usual CRUD resource provider, it simply derives from ExternalResource instead of Resource. All resources now have the ability to adopt children. This is purely a metadata/tagging thing, and will help us roll up displays, provide attribution to the developer, and even hide aspects of the resource graph as appropriate (e.g., when they are implementation details). Our use of this capability is ultra limited right now; in fact, the only place we display children is in the CLI output. For instance: + aws:serverless:Function: (create) [urn=urn:pulumi:demo::serverless::aws:serverless:Function::mylambda] => urn:pulumi:demo::serverless::aws:iam/role:Role::mylambda-iamrole => urn:pulumi:demo::serverless::aws:iam/rolePolicyAttachment:RolePolicyAttachment::mylambda-iampolicy-0 => urn:pulumi:demo::serverless::aws:lambda/function:Function::mylambda The bit indicating whether a resource is external or not is tracked in the resulting checkpoint file, along with any of its children.
2017-10-14 23:18:43 +02:00
*/
constructor(t: string, name: string, props?: Inputs, opts?: CustomResourceOptions) {
super(t, name, true, props, opts);
Implement initial Lumi-as-a-library This is the initial step towards redefining Lumi as a library that runs atop vanilla Node.js/V8, rather than as its own runtime. This change is woefully incomplete but this includes some of the more stable pieces of my current work-in-progress. The new structure is that within the sdk/ directory we will have a client library per language. This client library contains the object model for Lumi (resources, properties, assets, config, etc), in addition to the "language runtime host" components required to interoperate with the Lumi resource monitor. This resource monitor is effectively what we call "Lumi" today, in that it's the thing orchestrating plans and deployments. Inside the sdk/ directory, you will find nodejs/, the Node.js client library, alongside proto/, the definitions for RPC interop between the different pieces of the system. This includes existing RPC definitions for resource providers, etc., in addition to the new ones for hosting different language runtimes from within Lumi. These new interfaces are surprisingly simple. There is effectively a bidirectional RPC channel between the Lumi resource monitor, represented by the lumirpc.ResourceMonitor interface, and each language runtime, represented by the lumirpc.LanguageRuntime interface. The overall orchestration goes as follows: 1) Lumi decides it needs to run a program written in language X, so it dynamically loads the language runtime plugin for language X. 2) Lumi passes that runtime a loopback address to its ResourceMonitor service, while language X will publish a connection back to its LanguageRuntime service, which Lumi will talk to. 3) Lumi then invokes LanguageRuntime.Run, passing information like the desired working directory, program name, arguments, and optional configuration variables to make available to the program. 4) The language X runtime receives this, unpacks it and sets up the necessary context, and then invokes the program. The program then calls into Lumi object model abstractions that internally communicate back to Lumi using the ResourceMonitor interface. 5) The key here is ResourceMonitor.NewResource, which Lumi uses to serialize state about newly allocated resources. Lumi receives these and registers them as part of the plan, doing the usual diffing, etc., to decide how to proceed. This interface is perhaps one of the most subtle parts of the new design, as it necessitates the use of promises internally to allow parallel evaluation of the resource plan, letting dataflow determine the available concurrency. 6) The program exits, and Lumi continues on its merry way. If the program fails, the RunResponse will include information about the failure. Due to (5), all properties on resources are now instances of a new Property<T> type. A Property<T> is just a thin wrapper over a T, but it encodes the special properties of Lumi resource properties. Namely, it is possible to create one out of a T, other Property<T>, Promise<T>, or to freshly allocate one. In all cases, the Property<T> does not "settle" until its final state is known. This cannot occur before the deployment actually completes, and so in general it's not safe to depend on concrete resolutions of values (unlike ordinary Promise<T>s which are usually expected to resolve). As a result, all derived computations are meant to use the `then` function (as in `someValue.then(v => v+x)`). Although this change includes tests that may be run in isolation to test the various RPC interactions, we are nowhere near finished. The remaining work primarily boils down to three things: 1) Wiring all of this up to the Lumi code. 2) Fixing the handful of known loose ends required to make this work, primarily around the serialization of properties (waiting on unresolved ones, serializing assets properly, etc). 3) Implementing lambda closure serialization as a native extension. This ongoing work is part of pulumi/pulumi-fabric#311.
2017-08-26 21:07:54 +02:00
}
}
(<any>CustomResource).doNotCapture = true;
Implement first-class providers. (#1695) ### First-Class Providers These changes implement support for first-class providers. First-class providers are provider plugins that are exposed as resources via the Pulumi programming model so that they may be explicitly and multiply instantiated. Each instance of a provider resource may be configured differently, and configuration parameters may be source from the outputs of other resources. ### Provider Plugin Changes In order to accommodate the need to verify and diff provider configuration and configure providers without complete configuration information, these changes adjust the high-level provider plugin interface. Two new methods for validating a provider's configuration and diffing changes to the same have been added (`CheckConfig` and `DiffConfig`, respectively), and the type of the configuration bag accepted by `Configure` has been changed to a `PropertyMap`. These changes have not yet been reflected in the provider plugin gRPC interface. We will do this in a set of follow-up changes. Until then, these methods are implemented by adapters: - `CheckConfig` validates that all configuration parameters are string or unknown properties. This is necessary because existing plugins only accept string-typed configuration values. - `DiffConfig` either returns "never replace" if all configuration values are known or "must replace" if any configuration value is unknown. The justification for this behavior is given [here](https://github.com/pulumi/pulumi/pull/1695/files#diff-a6cd5c7f337665f5bb22e92ca5f07537R106) - `Configure` converts the config bag to a legacy config map and configures the provider plugin if all config values are known. If any config value is unknown, the underlying plugin is not configured and the provider may only perform `Check`, `Read`, and `Invoke`, all of which return empty results. We justify this behavior becuase it is only possible during a preview and provides the best experience we can manage with the existing gRPC interface. ### Resource Model Changes Providers are now exposed as resources that participate in a stack's dependency graph. Like other resources, they are explicitly created, may have multiple instances, and may have dependencies on other resources. Providers are referred to using provider references, which are a combination of the provider's URN and its ID. This design addresses the need during a preview to refer to providers that have not yet been physically created and therefore have no ID. All custom resources that are not themselves providers must specify a single provider via a provider reference. The named provider will be used to manage that resource's CRUD operations. If a resource's provider reference changes, the resource must be replaced. Though its URN is not present in the resource's dependency list, the provider should be treated as a dependency of the resource when topologically sorting the dependency graph. Finally, `Invoke` operations must now specify a provider to use for the invocation via a provider reference. ### Engine Changes First-class providers support requires a few changes to the engine: - The engine must have some way to map from provider references to provider plugins. It must be possible to add providers from a stack's checkpoint to this map and to register new/updated providers during the execution of a plan in response to CRUD operations on provider resources. - In order to support updating existing stacks using existing Pulumi programs that may not explicitly instantiate providers, the engine must be able to manage the "default" providers for each package referenced by a checkpoint or Pulumi program. The configuration for a "default" provider is taken from the stack's configuration data. The former need is addressed by adding a provider registry type that is responsible for managing all of the plugins required by a plan. In addition to loading plugins froma checkpoint and providing the ability to map from a provider reference to a provider plugin, this type serves as the provider plugin for providers themselves (i.e. it is the "provider provider"). The latter need is solved via two relatively self-contained changes to plan setup and the eval source. During plan setup, the old checkpoint is scanned for custom resources that do not have a provider reference in order to compute the set of packages that require a default provider. Once this set has been computed, the required default provider definitions are conjured and prepended to the checkpoint's resource list. Each resource that requires a default provider is then updated to refer to the default provider for its package. While an eval source is running, each custom resource registration, resource read, and invoke that does not name a provider is trapped before being returned by the source iterator. If no default provider for the appropriate package has been registered, the eval source synthesizes an appropriate registration, waits for it to complete, and records the registered provider's reference. This reference is injected into the original request, which is then processed as usual. If a default provider was already registered, the recorded reference is used and no new registration occurs. ### SDK Changes These changes only expose first-class providers from the Node.JS SDK. - A new abstract class, `ProviderResource`, can be subclassed and used to instantiate first-class providers. - A new field in `ResourceOptions`, `provider`, can be used to supply a particular provider instance to manage a `CustomResource`'s CRUD operations. - A new type, `InvokeOptions`, can be used to specify options that control the behavior of a call to `pulumi.runtime.invoke`. This type includes a `provider` field that is analogous to `ResourceOptions.provider`.
2018-08-07 02:50:29 +02:00
/**
* ProviderResource is a resource that implements CRUD operations for other custom resources. These resources are
* managed similarly to other resources, including the usual diffing and update semantics.
*/
export abstract class ProviderResource extends CustomResource {
/**
* Creates and registers a new provider resource for a particular package.
*
* @param pkg The package associated with this provider.
* @param name The _unique_ name of the provider.
* @param props The configuration to use for this provider.
* @param opts A bag of options that control this provider's behavior.
*/
constructor(pkg: string, name: string, props?: Inputs, opts?: ResourceOptions) {
if (opts && (<any>opts).provider !== undefined) {
Implement first-class providers. (#1695) ### First-Class Providers These changes implement support for first-class providers. First-class providers are provider plugins that are exposed as resources via the Pulumi programming model so that they may be explicitly and multiply instantiated. Each instance of a provider resource may be configured differently, and configuration parameters may be source from the outputs of other resources. ### Provider Plugin Changes In order to accommodate the need to verify and diff provider configuration and configure providers without complete configuration information, these changes adjust the high-level provider plugin interface. Two new methods for validating a provider's configuration and diffing changes to the same have been added (`CheckConfig` and `DiffConfig`, respectively), and the type of the configuration bag accepted by `Configure` has been changed to a `PropertyMap`. These changes have not yet been reflected in the provider plugin gRPC interface. We will do this in a set of follow-up changes. Until then, these methods are implemented by adapters: - `CheckConfig` validates that all configuration parameters are string or unknown properties. This is necessary because existing plugins only accept string-typed configuration values. - `DiffConfig` either returns "never replace" if all configuration values are known or "must replace" if any configuration value is unknown. The justification for this behavior is given [here](https://github.com/pulumi/pulumi/pull/1695/files#diff-a6cd5c7f337665f5bb22e92ca5f07537R106) - `Configure` converts the config bag to a legacy config map and configures the provider plugin if all config values are known. If any config value is unknown, the underlying plugin is not configured and the provider may only perform `Check`, `Read`, and `Invoke`, all of which return empty results. We justify this behavior becuase it is only possible during a preview and provides the best experience we can manage with the existing gRPC interface. ### Resource Model Changes Providers are now exposed as resources that participate in a stack's dependency graph. Like other resources, they are explicitly created, may have multiple instances, and may have dependencies on other resources. Providers are referred to using provider references, which are a combination of the provider's URN and its ID. This design addresses the need during a preview to refer to providers that have not yet been physically created and therefore have no ID. All custom resources that are not themselves providers must specify a single provider via a provider reference. The named provider will be used to manage that resource's CRUD operations. If a resource's provider reference changes, the resource must be replaced. Though its URN is not present in the resource's dependency list, the provider should be treated as a dependency of the resource when topologically sorting the dependency graph. Finally, `Invoke` operations must now specify a provider to use for the invocation via a provider reference. ### Engine Changes First-class providers support requires a few changes to the engine: - The engine must have some way to map from provider references to provider plugins. It must be possible to add providers from a stack's checkpoint to this map and to register new/updated providers during the execution of a plan in response to CRUD operations on provider resources. - In order to support updating existing stacks using existing Pulumi programs that may not explicitly instantiate providers, the engine must be able to manage the "default" providers for each package referenced by a checkpoint or Pulumi program. The configuration for a "default" provider is taken from the stack's configuration data. The former need is addressed by adding a provider registry type that is responsible for managing all of the plugins required by a plan. In addition to loading plugins froma checkpoint and providing the ability to map from a provider reference to a provider plugin, this type serves as the provider plugin for providers themselves (i.e. it is the "provider provider"). The latter need is solved via two relatively self-contained changes to plan setup and the eval source. During plan setup, the old checkpoint is scanned for custom resources that do not have a provider reference in order to compute the set of packages that require a default provider. Once this set has been computed, the required default provider definitions are conjured and prepended to the checkpoint's resource list. Each resource that requires a default provider is then updated to refer to the default provider for its package. While an eval source is running, each custom resource registration, resource read, and invoke that does not name a provider is trapped before being returned by the source iterator. If no default provider for the appropriate package has been registered, the eval source synthesizes an appropriate registration, waits for it to complete, and records the registered provider's reference. This reference is injected into the original request, which is then processed as usual. If a default provider was already registered, the recorded reference is used and no new registration occurs. ### SDK Changes These changes only expose first-class providers from the Node.JS SDK. - A new abstract class, `ProviderResource`, can be subclassed and used to instantiate first-class providers. - A new field in `ResourceOptions`, `provider`, can be used to supply a particular provider instance to manage a `CustomResource`'s CRUD operations. - A new type, `InvokeOptions`, can be used to specify options that control the behavior of a call to `pulumi.runtime.invoke`. This type includes a `provider` field that is analogous to `ResourceOptions.provider`.
2018-08-07 02:50:29 +02:00
throw new RunError("Explicit providers may not be used with provider resources");
}
super(`pulumi:providers:${pkg}`, name, props, opts);
}
}
/**
2018-01-26 03:50:58 +01:00
* ComponentResource is a resource that aggregates one or more other child resources into a higher
* level abstraction. The component resource itself is a resource, but does not require custom CRUD
* operations for provisioning.
*/
export class ComponentResource extends Resource {
/**
2018-01-26 03:50:58 +01:00
* Creates and registers a new component resource. t is the fully qualified type token and name
* is the "name" part to use in creating a stable and globally unique URN for the object. parent
* is the optional parent for this component, and dependsOn is an optional list of other
* resources that this resource depends on, controlling the order in which we perform resource
* operations.
*
* @param t The type of the resource.
2018-02-08 00:01:55 +01:00
* @param name The _unique_ name of the resource.
* @param props The arguments to use to populate the new resource.
* @param opts A bag of options that control this resource's behavior.
*/
constructor(t: string, name: string, props?: Inputs, opts?: ComponentResourceOptions) {
if (opts && (<any>opts).provider !== undefined) {
Implement first-class providers. (#1695) ### First-Class Providers These changes implement support for first-class providers. First-class providers are provider plugins that are exposed as resources via the Pulumi programming model so that they may be explicitly and multiply instantiated. Each instance of a provider resource may be configured differently, and configuration parameters may be source from the outputs of other resources. ### Provider Plugin Changes In order to accommodate the need to verify and diff provider configuration and configure providers without complete configuration information, these changes adjust the high-level provider plugin interface. Two new methods for validating a provider's configuration and diffing changes to the same have been added (`CheckConfig` and `DiffConfig`, respectively), and the type of the configuration bag accepted by `Configure` has been changed to a `PropertyMap`. These changes have not yet been reflected in the provider plugin gRPC interface. We will do this in a set of follow-up changes. Until then, these methods are implemented by adapters: - `CheckConfig` validates that all configuration parameters are string or unknown properties. This is necessary because existing plugins only accept string-typed configuration values. - `DiffConfig` either returns "never replace" if all configuration values are known or "must replace" if any configuration value is unknown. The justification for this behavior is given [here](https://github.com/pulumi/pulumi/pull/1695/files#diff-a6cd5c7f337665f5bb22e92ca5f07537R106) - `Configure` converts the config bag to a legacy config map and configures the provider plugin if all config values are known. If any config value is unknown, the underlying plugin is not configured and the provider may only perform `Check`, `Read`, and `Invoke`, all of which return empty results. We justify this behavior becuase it is only possible during a preview and provides the best experience we can manage with the existing gRPC interface. ### Resource Model Changes Providers are now exposed as resources that participate in a stack's dependency graph. Like other resources, they are explicitly created, may have multiple instances, and may have dependencies on other resources. Providers are referred to using provider references, which are a combination of the provider's URN and its ID. This design addresses the need during a preview to refer to providers that have not yet been physically created and therefore have no ID. All custom resources that are not themselves providers must specify a single provider via a provider reference. The named provider will be used to manage that resource's CRUD operations. If a resource's provider reference changes, the resource must be replaced. Though its URN is not present in the resource's dependency list, the provider should be treated as a dependency of the resource when topologically sorting the dependency graph. Finally, `Invoke` operations must now specify a provider to use for the invocation via a provider reference. ### Engine Changes First-class providers support requires a few changes to the engine: - The engine must have some way to map from provider references to provider plugins. It must be possible to add providers from a stack's checkpoint to this map and to register new/updated providers during the execution of a plan in response to CRUD operations on provider resources. - In order to support updating existing stacks using existing Pulumi programs that may not explicitly instantiate providers, the engine must be able to manage the "default" providers for each package referenced by a checkpoint or Pulumi program. The configuration for a "default" provider is taken from the stack's configuration data. The former need is addressed by adding a provider registry type that is responsible for managing all of the plugins required by a plan. In addition to loading plugins froma checkpoint and providing the ability to map from a provider reference to a provider plugin, this type serves as the provider plugin for providers themselves (i.e. it is the "provider provider"). The latter need is solved via two relatively self-contained changes to plan setup and the eval source. During plan setup, the old checkpoint is scanned for custom resources that do not have a provider reference in order to compute the set of packages that require a default provider. Once this set has been computed, the required default provider definitions are conjured and prepended to the checkpoint's resource list. Each resource that requires a default provider is then updated to refer to the default provider for its package. While an eval source is running, each custom resource registration, resource read, and invoke that does not name a provider is trapped before being returned by the source iterator. If no default provider for the appropriate package has been registered, the eval source synthesizes an appropriate registration, waits for it to complete, and records the registered provider's reference. This reference is injected into the original request, which is then processed as usual. If a default provider was already registered, the recorded reference is used and no new registration occurs. ### SDK Changes These changes only expose first-class providers from the Node.JS SDK. - A new abstract class, `ProviderResource`, can be subclassed and used to instantiate first-class providers. - A new field in `ResourceOptions`, `provider`, can be used to supply a particular provider instance to manage a `CustomResource`'s CRUD operations. - A new type, `InvokeOptions`, can be used to specify options that control the behavior of a call to `pulumi.runtime.invoke`. This type includes a `provider` field that is analogous to `ResourceOptions.provider`.
2018-08-07 02:50:29 +02:00
throw new RunError("Explicit providers may not be used with component resources");
}
super(t, name, false, props, opts);
}
// registerOutputs registers synthetic outputs that a component has initialized, usually by allocating
// other child sub-resources and propagating their resulting property values.
protected registerOutputs(outputs: Inputs | Promise<Inputs> | Output<Inputs> | undefined): void {
if (outputs) {
registerResourceOutputs(this, outputs);
}
}
}
(<any>ComponentResource).doNotCapture = true;
(<any>ComponentResource.prototype).registerOutputs.doNotCapture = true;
/**
* Output helps encode the relationship between Resources in a Pulumi application. Specifically an
* Output holds onto a piece of Data and the Resource it was generated from. An Output value can
* then be provided when constructing new Resources, allowing that new Resource to know both the
* value as well as the Resource the value came from. This allows for a precise 'Resource
2018-02-08 00:01:55 +01:00
* dependency graph' to be created, which properly tracks the relationship between resources.
*/
export class Output<T> {
/**
* A private field to help with RTTI that works in SxS scenarios.
*
* This is internal instead of being truly private, to support mixins and our serialization model.
*/
// tslint:disable-next-line:variable-name
/* @internal */ public readonly __pulumiOutput?: boolean = true;
/**
* Whether or not this 'Output' should actually perform .apply calls. During a preview,
* an Output value may not be known (because it would have to actually be computed by doing an
* 'update'). In that case, we don't want to perform any .apply calls as the callbacks
* may not expect an undefined value. So, instead, we just transition to another Output
* value that itself knows it should not perform .apply calls.
*/
Discriminate unknown values in the JS runtime. (#1414) These changes add support for distinguishing an output property with an unknown value from an output property with a known value that is undefined. In a broad sense, the Pulumi property type system is just JSON with the addition of unknown values. Notably absent, however, are undefined values. As it stands, our marshalers between JavaScript and Pulumi property values treat all undefined JavaScript values as unknown Pulumi values. Unfortunately, this conflates two very different concepts: unknown Pulumi values are intended to represent values of output properties that are unknown at time of preview, _not_ values that are known but undefined. This results in difficulty reasoning about when transforms are run on output properties as well as confusing output in the `diff` view of Pulumi preview (user-specifed undefined values are rendered as unknown values). As it turns out, we already have a way to decide whether or not an Output value is known or not: Output.performApply. These changes rename this property to `isKnown`, clarify its meaning, and take advantage of the result to decide whether or not an Output value should marshal as an unknown Pulumi value. This also allowed these changes to improve the serialization of undefined object keys and array elements s.t. we better match JavaScript to JSON serialization behavior (undefined object keys are omitted; undefined array elements are marshaled as `null`). Fixes https://github.com/pulumi/pulumi-cloud/issues/483.
2018-05-23 23:47:40 +02:00
/* @internal */ public isKnown: Promise<boolean>;
/**
* Method that actually produces the concrete value of this output, as well as the total
* deployment-time set of resources this output depends on.
*
* Only callable on the outside.
*/
/* @internal */ public readonly promise: () => Promise<T>;
/**
* The list of resource that this output value depends on.
*
* Only callable on the outside.
*/
/* @internal */ public readonly resources: () => Set<Resource>;
/**
* Transforms the data of the output with the provided func. The result remains a
* Output so that dependent resources can be properly tracked.
*
* 'func' is not allowed to make resources.
*
* 'func' can return other Outputs. This can be handy if you have a Output<SomeVal>
* and you want to get a transitive dependency of it. i.e.
*
* ```ts
* var d1: Output<SomeVal>;
* var d2 = d1.apply(v => v.x.y.OtherOutput); // getting an output off of 'v'.
* ```
*
* In this example, taking a dependency on d2 means a resource will depend on all the resources
* of d1. It will *not* depend on the resources of v.x.y.OtherDep.
*
* Importantly, the Resources that d2 feels like it will depend on are the same resources as d1.
* If you need have multiple Outputs and a single Output is needed that combines both
* set of resources, then 'pulumi.all' should be used instead.
*
* This function will only be called execution of a 'pulumi update' request. It will not run
* during 'pulumi preview' (as the values of resources are of course not known then). It is not
* available for functions that end up executing in the cloud during runtime. To get the value
2018-02-08 00:01:55 +01:00
* of the Output during cloud runtime execution, use `get()`.
*/
public readonly apply: <U>(func: (t: T) => Input<U>) => Output<U>;
/**
* Retrieves the underlying value of this dependency.
*
* This function is only callable in code that runs in the cloud post-deployment. At this
* point all Output values will be known and can be safely retrieved. During pulumi deployment
* or preview execution this must not be called (and will throw). This is because doing so
* would allow Output values to flow into Resources while losing the data that would allow
* the dependency graph to be changed.
*/
public readonly get: () => T;
// Statics
/**
* Returns true if the given object is an instance of Output<T>. This is designed to work even when
* multiple copies of the Pulumi SDK have been loaded into the same process.
*/
public static isInstance<T>(obj: any): obj is Output<T> {
return obj && obj.__pulumiOutput;
}
/* @internal */ public static create<T>(
Discriminate unknown values in the JS runtime. (#1414) These changes add support for distinguishing an output property with an unknown value from an output property with a known value that is undefined. In a broad sense, the Pulumi property type system is just JSON with the addition of unknown values. Notably absent, however, are undefined values. As it stands, our marshalers between JavaScript and Pulumi property values treat all undefined JavaScript values as unknown Pulumi values. Unfortunately, this conflates two very different concepts: unknown Pulumi values are intended to represent values of output properties that are unknown at time of preview, _not_ values that are known but undefined. This results in difficulty reasoning about when transforms are run on output properties as well as confusing output in the `diff` view of Pulumi preview (user-specifed undefined values are rendered as unknown values). As it turns out, we already have a way to decide whether or not an Output value is known or not: Output.performApply. These changes rename this property to `isKnown`, clarify its meaning, and take advantage of the result to decide whether or not an Output value should marshal as an unknown Pulumi value. This also allowed these changes to improve the serialization of undefined object keys and array elements s.t. we better match JavaScript to JSON serialization behavior (undefined object keys are omitted; undefined array elements are marshaled as `null`). Fixes https://github.com/pulumi/pulumi-cloud/issues/483.
2018-05-23 23:47:40 +02:00
resource: Resource, promise: Promise<T>, isKnown: Promise<boolean>): Output<T> {
return new Output<T>(new Set<Resource>([resource]), promise, isKnown);
}
/* @internal */ public constructor(
Discriminate unknown values in the JS runtime. (#1414) These changes add support for distinguishing an output property with an unknown value from an output property with a known value that is undefined. In a broad sense, the Pulumi property type system is just JSON with the addition of unknown values. Notably absent, however, are undefined values. As it stands, our marshalers between JavaScript and Pulumi property values treat all undefined JavaScript values as unknown Pulumi values. Unfortunately, this conflates two very different concepts: unknown Pulumi values are intended to represent values of output properties that are unknown at time of preview, _not_ values that are known but undefined. This results in difficulty reasoning about when transforms are run on output properties as well as confusing output in the `diff` view of Pulumi preview (user-specifed undefined values are rendered as unknown values). As it turns out, we already have a way to decide whether or not an Output value is known or not: Output.performApply. These changes rename this property to `isKnown`, clarify its meaning, and take advantage of the result to decide whether or not an Output value should marshal as an unknown Pulumi value. This also allowed these changes to improve the serialization of undefined object keys and array elements s.t. we better match JavaScript to JSON serialization behavior (undefined object keys are omitted; undefined array elements are marshaled as `null`). Fixes https://github.com/pulumi/pulumi-cloud/issues/483.
2018-05-23 23:47:40 +02:00
resources: Set<Resource>, promise: Promise<T>, isKnown: Promise<boolean>) {
this.isKnown = isKnown;
// Always create a copy so that no one accidentally modifies our Resource list.
this.resources = () => new Set<Resource>(resources);
this.promise = () => promise;
this.apply = <U>(func: (t: T) => Input<U>) => {
return new Output<U>(resources, promise.then(async v => {
// During previews do not perform the apply if the engine was not able to
// give us an actual value for this Output.
Discriminate unknown values in the JS runtime. (#1414) These changes add support for distinguishing an output property with an unknown value from an output property with a known value that is undefined. In a broad sense, the Pulumi property type system is just JSON with the addition of unknown values. Notably absent, however, are undefined values. As it stands, our marshalers between JavaScript and Pulumi property values treat all undefined JavaScript values as unknown Pulumi values. Unfortunately, this conflates two very different concepts: unknown Pulumi values are intended to represent values of output properties that are unknown at time of preview, _not_ values that are known but undefined. This results in difficulty reasoning about when transforms are run on output properties as well as confusing output in the `diff` view of Pulumi preview (user-specifed undefined values are rendered as unknown values). As it turns out, we already have a way to decide whether or not an Output value is known or not: Output.performApply. These changes rename this property to `isKnown`, clarify its meaning, and take advantage of the result to decide whether or not an Output value should marshal as an unknown Pulumi value. This also allowed these changes to improve the serialization of undefined object keys and array elements s.t. we better match JavaScript to JSON serialization behavior (undefined object keys are omitted; undefined array elements are marshaled as `null`). Fixes https://github.com/pulumi/pulumi-cloud/issues/483.
2018-05-23 23:47:40 +02:00
const perform = await isKnown;
if (runtime.isDryRun() && !perform) {
return <U><any>undefined;
}
const transformed = await func(v);
if (Output.isInstance(transformed)) {
// Note: if the func returned a Output, we unwrap that to get the inner value
// returned by that Output. Note that we are *not* capturing the Resources of
// this inner Output. That's intentional. As the Output returned is only
// supposed to be related this *this* Output object, those resources should
// already be in our transitively reachable resource graph.
return await transformed.promise();
} else {
return transformed;
}
Discriminate unknown values in the JS runtime. (#1414) These changes add support for distinguishing an output property with an unknown value from an output property with a known value that is undefined. In a broad sense, the Pulumi property type system is just JSON with the addition of unknown values. Notably absent, however, are undefined values. As it stands, our marshalers between JavaScript and Pulumi property values treat all undefined JavaScript values as unknown Pulumi values. Unfortunately, this conflates two very different concepts: unknown Pulumi values are intended to represent values of output properties that are unknown at time of preview, _not_ values that are known but undefined. This results in difficulty reasoning about when transforms are run on output properties as well as confusing output in the `diff` view of Pulumi preview (user-specifed undefined values are rendered as unknown values). As it turns out, we already have a way to decide whether or not an Output value is known or not: Output.performApply. These changes rename this property to `isKnown`, clarify its meaning, and take advantage of the result to decide whether or not an Output value should marshal as an unknown Pulumi value. This also allowed these changes to improve the serialization of undefined object keys and array elements s.t. we better match JavaScript to JSON serialization behavior (undefined object keys are omitted; undefined array elements are marshaled as `null`). Fixes https://github.com/pulumi/pulumi-cloud/issues/483.
2018-05-23 23:47:40 +02:00
}), isKnown);
};
this.get = () => {
2018-08-01 06:09:47 +02:00
throw new RunError(`Cannot call '.get' during update or preview.
To manipulate the value of this Output, use '.apply' instead.`);
};
}
}
/**
2018-09-12 06:17:52 +02:00
* [output] takes any Input value and converts it into an Output, deeply unwrapping nested Input
* values as necessary".
*
2018-09-12 06:17:52 +02:00
* The expected way to use this function is like so:
*
* ```ts
2018-09-12 06:17:52 +02:00
* var transformed = pulumi.output(someVal).apply(unwrapped => {
* // Do whatever you want now. 'unwrapped' will contain no outputs/promises inside
* // here, so you can easily do whatever sort of transformation is most convenient.
* });
*
2018-09-12 06:17:52 +02:00
* // the result can be passed to another Resource. The dependency information will be
* // properly maintained.
* var someResource = new SomeResource(name, { data: transformed ... });
* ```
*/
export function output<T>(val: Input<T>): Output<Unwrap<T>>;
export function output<T>(val: Input<T> | undefined): Output<Unwrap<T | undefined>>;
export function output<T>(val: Input<T | undefined>): Output<Unwrap<T | undefined>> {
if (val === null || typeof val !== "object") {
// strings, numbers, booleans, functions, symbols, undefineds, nulls are all returned as
// themselves. They are always 'known' (i.e. we can safely 'apply' off of them even during
// preview).
return createSimpleOutput(val);
}
else if (Resource.isInstance(val)) {
// Don't unwrap Resources, there are existing codepaths that return Resources through
// Outputs and we want to preserve them as is when flattening.
return createSimpleOutput(val);
}
else if (val instanceof Promise) {
// For a promise, we can just treat the same as an output that points to that resource. So
// we just create an Output around the Promise, and immediately apply the unwrap function on
// it to transform the value it points at.
return <any>new Output(new Set(), val, /*isKnown*/ Promise.resolve(true)).apply(output);
}
else if (Output.isInstance(val)) {
return <any>val.apply(output);
}
else if (val instanceof Array) {
return <any>all(val.map(output));
}
else {
const unwrappedObject: any = {};
Object.keys(val).forEach(k => {
unwrappedObject[k] = output((<any>val)[k]);
});
return <any>all(unwrappedObject);
}
}
function createSimpleOutput(val: any) {
return new Output(new Set(), Promise.resolve(val), /*isKnown*/ Promise.resolve(true));
}
/**
* Allows for multiple Output objects to be combined into a single Output object. The single Output
* will depend on the union of Resources that the individual dependencies depend on.
*
* This can be used in the following manner:
*
* ```ts
* var d1: Output<string>;
* var d2: Output<number>;
*
* var d3: Output<ResultType> = Output.all([d1, d2]).apply(([s, n]) => ...);
* ```
*
* In this example, taking a dependency on d3 means a resource will depend on all the resources of
* d1 and d2.
*/
// tslint:disable:max-line-length
export function all<T>(val: Record<string, Input<T>>): Output<Record<string, Unwrap<T>>>;
export function all<T1, T2, T3, T4, T5, T6, T7, T8>(values: [Input<T1> | undefined, Input<T2> | undefined, Input<T3> | undefined, Input<T4> | undefined, Input<T5> | undefined, Input<T6> | undefined, Input<T7> | undefined, Input<T8> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>, Unwrap<T3>, Unwrap<T4>, Unwrap<T5>, Unwrap<T6>, Unwrap<T7>, Unwrap<T8>]>;
export function all<T1, T2, T3, T4, T5, T6, T7>(values: [Input<T1> | undefined, Input<T2> | undefined, Input<T3> | undefined, Input<T4> | undefined, Input<T5> | undefined, Input<T6> | undefined, Input<T7> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>, Unwrap<T3>, Unwrap<T4>, Unwrap<T5>, Unwrap<T6>, Unwrap<T7>]>;
export function all<T1, T2, T3, T4, T5, T6>(values: [Input<T1> | undefined, Input<T2> | undefined, Input<T3> | undefined, Input<T4> | undefined, Input<T5> | undefined, Input<T6> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>, Unwrap<T3>, Unwrap<T4>, Unwrap<T5>, Unwrap<T6>]>;
export function all<T1, T2, T3, T4, T5>(values: [Input<T1> | undefined, Input<T2> | undefined, Input<T3> | undefined, Input<T4> | undefined, Input<T5> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>, Unwrap<T3>, Unwrap<T4>, Unwrap<T5>]>;
export function all<T1, T2, T3, T4>(values: [Input<T1> | undefined, Input<T2> | undefined, Input<T3> | undefined, Input<T4> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>, Unwrap<T3>, Unwrap<T4>]>;
export function all<T1, T2, T3>(values: [Input<T1> | undefined, Input<T2> | undefined, Input<T3> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>, Unwrap<T3>]>;
export function all<T1, T2>(values: [Input<T1> | undefined, Input<T2> | undefined]): Output<[Unwrap<T1>, Unwrap<T2>]>;
export function all<T>(ds: (Input<T> | undefined)[]): Output<Unwrap<T>[]>;
export function all<T>(val: Input<T>[] | Record<string, Input<T>>): Output<any> {
if (val instanceof Array) {
const allOutputs = val.map(v => output(v));
const [resources, isKnown] = getResourcesAndIsKnown(allOutputs);
const promisedArray = Promise.all(allOutputs.map(o => o.promise()));
return new Output<Unwrap<T>[]>(new Set<Resource>(resources), promisedArray, isKnown);
} else {
const keysAndOutputs = Object.keys(val).map(key => ({ key, value: output(val[key]) }));
const allOutputs = keysAndOutputs.map(kvp => kvp.value);
const [resources, isKnown] = getResourcesAndIsKnown(allOutputs);
const promisedObject = getPromisedObject(keysAndOutputs);
return new Output<Record<string, Unwrap<T>>>(new Set<Resource>(resources), promisedObject, isKnown);
}
}
async function getPromisedObject<T>(keysAndOutputs: { key: string, value: Output<Unwrap<T>> }[]): Promise<Record<string, Unwrap<T>>> {
const result: Record<string, Unwrap<T>> = {};
for (const kvp of keysAndOutputs) {
result[kvp.key] = await kvp.value.promise();
}
return result;
}
function getResourcesAndIsKnown<T>(allOutputs: Output<Unwrap<T>>[]): [Resource[], Promise<boolean>] {
const allResources = allOutputs.reduce<Resource[]>((arr, o) => (arr.push(...o.resources()), arr), []);
// A merged output is known if all of its inputs are known.
const isKnown = Promise.all(allOutputs.map(o => o.isKnown)).then(ps => ps.every(b => b));
return [allResources, isKnown];
}
/**
2018-02-06 03:37:10 +01:00
* Input is a property input for a resource. It may be a promptly available T, a promise
* for one, or the output from a existing Resource.
*/
export type Input<T> = T | Promise<T> | Output<T>;
/**
2018-02-06 03:37:10 +01:00
* Inputs is a map of property name to property input, one for each resource
2018-01-26 03:50:58 +01:00
* property value.
*/
2018-02-06 03:37:10 +01:00
export type Inputs = Record<string, Input<any>>;
/**
* The 'Unwrap' type allows us to express the operation of taking a type, with potentially deeply
* nested Promises and Outputs and to then get that same type with all the Promises and Outputs
* replaced with their wrapped type. Note that this Unwrapping is 'deep'. So if you had:
*
* `type X = { A: Promise<{ B: Output<{ c: Input<boolean> }> }> }`
*
* Then `Unwrap<X>` would be equivalent to:
*
* `... = { A: { B: { C: boolean } } }`
*
* Unwrapping sees through Promises, Outputs, Arrays and Objects.
*
* Note: due to TypeScript limitations there are some things that cannot be expressed. Specifically,
* if you had a `Promise<Output<T>>` then the Unwrap type would not be able to undo both of those
* wraps. In practice that should be ok. Values in an object graph should not wrap Outputs in
* Promises. Instead, any code that needs to work Outputs and also be async should either create
* the Output with the Promise (which will collapse into just an Output). Or, it should start with
* an Output and call [apply] on it, passing in an async function. This will also collapse and just
* produce an Output.
*
* In other words, this should not be used as the shape of an object: `{ a: Promise<Output<...>> }`.
* It should always either be `{ a: Promise<NonOutput> }` or just `{ a: Output<...> }`.
*/
type Unwrap<T> =
// 1. If we have a promise, just get the type it itself is wrapping and recursively unwrap that.
// 2. Otherwise, if we have an output, do the same as a promise and just unwrap the inner type.
// 3. Otherwise, we have a basic type. Just unwrap that.
2018-09-15 11:59:01 +02:00
T extends Promise<infer U1> ? UnwrapSimple<U1> :
T extends Output<infer U2> ? UnwrapSimple<U2> :
UnwrapSimple<T>;
type primitive = string | number | boolean | undefined | null;
/**
* Handles encountering basic types when unwrapping.
*/
type UnwrapSimple<T> =
// 1. Any of the primitive types just unwrap to themselves.
// 2. An array of some types unwraps to an array of that type itself unwrapped. Note, due to a
// TS limitation we cannot express that as Array<Unwrap<U>> due to how it handles recursive
// types. We work around that by introducing an structurally equivalent interface that then
// helps make typescript defer type-evaluation instead of doing it eagerly.
// 3. An object unwraps to an object with properties of the same name, but where the property
// types have been unwrapped.
// 4. return 'never' at the end so that if we've missed something we'll discover it.
T extends primitive ? T :
T extends Resource ? T :
T extends Array<infer U> ? UnwrappedArray<U> :
T extends object ? UnwrappedObject<T> :
never;
interface UnwrappedArray<T> extends Array<Unwrap<T>> {}
type UnwrappedObject<T> = {
[P in keyof T]: Unwrap<T[P]>;
};