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
|
|
|
// Code generated by protoc-gen-go.
|
|
|
|
|
// source: languages.proto
|
|
|
|
|
// DO NOT EDIT!
|
|
|
|
|
|
|
|
|
|
package lumirpc
|
|
|
|
|
|
|
|
|
|
import proto "github.com/golang/protobuf/proto"
|
|
|
|
|
import fmt "fmt"
|
|
|
|
|
import math "math"
|
|
|
|
|
import google_protobuf "github.com/golang/protobuf/ptypes/struct"
|
|
|
|
|
|
|
|
|
|
import (
|
|
|
|
|
context "golang.org/x/net/context"
|
|
|
|
|
grpc "google.golang.org/grpc"
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
// Reference imports to suppress errors if they are not otherwise used.
|
|
|
|
|
var _ = proto.Marshal
|
|
|
|
|
var _ = fmt.Errorf
|
|
|
|
|
var _ = math.Inf
|
|
|
|
|
|
|
|
|
|
// RunRequest asks the interpreter to execute a program.
|
|
|
|
|
type RunRequest struct {
|
|
|
|
|
Pwd string `protobuf:"bytes,1,opt,name=pwd" json:"pwd,omitempty"`
|
|
|
|
|
Program string `protobuf:"bytes,2,opt,name=program" json:"program,omitempty"`
|
|
|
|
|
Args []string `protobuf:"bytes,3,rep,name=args" json:"args,omitempty"`
|
|
|
|
|
Config map[string]string `protobuf:"bytes,4,rep,name=config" json:"config,omitempty" protobuf_key:"bytes,1,opt,name=key" protobuf_val:"bytes,2,opt,name=value"`
|
2017-08-30 03:24:12 +02:00
|
|
|
DryRun bool `protobuf:"varint,5,opt,name=dryRun" json:"dryRun,omitempty"`
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *RunRequest) Reset() { *m = RunRequest{} }
|
|
|
|
|
func (m *RunRequest) String() string { return proto.CompactTextString(m) }
|
|
|
|
|
func (*RunRequest) ProtoMessage() {}
|
|
|
|
|
func (*RunRequest) Descriptor() ([]byte, []int) { return fileDescriptor2, []int{0} }
|
|
|
|
|
|
|
|
|
|
func (m *RunRequest) GetPwd() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Pwd
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *RunRequest) GetProgram() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Program
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *RunRequest) GetArgs() []string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Args
|
|
|
|
|
}
|
|
|
|
|
return nil
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *RunRequest) GetConfig() map[string]string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Config
|
|
|
|
|
}
|
|
|
|
|
return nil
|
|
|
|
|
}
|
|
|
|
|
|
2017-08-30 03:24:12 +02:00
|
|
|
func (m *RunRequest) GetDryRun() bool {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.DryRun
|
|
|
|
|
}
|
|
|
|
|
return false
|
|
|
|
|
}
|
|
|
|
|
|
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
|
|
|
// RunResponse is the response back from the interpreter/source back to the monitor.
|
|
|
|
|
type RunResponse struct {
|
|
|
|
|
Error string `protobuf:"bytes,1,opt,name=error" json:"error,omitempty"`
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *RunResponse) Reset() { *m = RunResponse{} }
|
|
|
|
|
func (m *RunResponse) String() string { return proto.CompactTextString(m) }
|
|
|
|
|
func (*RunResponse) ProtoMessage() {}
|
|
|
|
|
func (*RunResponse) Descriptor() ([]byte, []int) { return fileDescriptor2, []int{1} }
|
|
|
|
|
|
|
|
|
|
func (m *RunResponse) GetError() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Error
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// NewResourceRequest contains information about a resource object that was newly allocated.
|
|
|
|
|
type NewResourceRequest struct {
|
|
|
|
|
Type string `protobuf:"bytes,1,opt,name=type" json:"type,omitempty"`
|
|
|
|
|
Name string `protobuf:"bytes,2,opt,name=name" json:"name,omitempty"`
|
|
|
|
|
Object *google_protobuf.Struct `protobuf:"bytes,3,opt,name=object" json:"object,omitempty"`
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceRequest) Reset() { *m = NewResourceRequest{} }
|
|
|
|
|
func (m *NewResourceRequest) String() string { return proto.CompactTextString(m) }
|
|
|
|
|
func (*NewResourceRequest) ProtoMessage() {}
|
|
|
|
|
func (*NewResourceRequest) Descriptor() ([]byte, []int) { return fileDescriptor2, []int{2} }
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceRequest) GetType() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Type
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceRequest) GetName() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Name
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceRequest) GetObject() *google_protobuf.Struct {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Object
|
|
|
|
|
}
|
|
|
|
|
return nil
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// NewResourceResponse reflects back the properties initialized during creation, if applicable.
|
|
|
|
|
type NewResourceResponse struct {
|
|
|
|
|
Id string `protobuf:"bytes,1,opt,name=id" json:"id,omitempty"`
|
|
|
|
|
Urn string `protobuf:"bytes,2,opt,name=urn" json:"urn,omitempty"`
|
|
|
|
|
Object *google_protobuf.Struct `protobuf:"bytes,3,opt,name=object" json:"object,omitempty"`
|
2017-09-05 19:01:00 +02:00
|
|
|
Stable bool `protobuf:"varint,4,opt,name=stable" json:"stable,omitempty"`
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceResponse) Reset() { *m = NewResourceResponse{} }
|
|
|
|
|
func (m *NewResourceResponse) String() string { return proto.CompactTextString(m) }
|
|
|
|
|
func (*NewResourceResponse) ProtoMessage() {}
|
|
|
|
|
func (*NewResourceResponse) Descriptor() ([]byte, []int) { return fileDescriptor2, []int{3} }
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceResponse) GetId() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Id
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceResponse) GetUrn() string {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Urn
|
|
|
|
|
}
|
|
|
|
|
return ""
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (m *NewResourceResponse) GetObject() *google_protobuf.Struct {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Object
|
|
|
|
|
}
|
|
|
|
|
return nil
|
|
|
|
|
}
|
|
|
|
|
|
2017-09-05 19:01:00 +02:00
|
|
|
func (m *NewResourceResponse) GetStable() bool {
|
|
|
|
|
if m != nil {
|
|
|
|
|
return m.Stable
|
|
|
|
|
}
|
|
|
|
|
return false
|
|
|
|
|
}
|
|
|
|
|
|
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
|
|
|
func init() {
|
|
|
|
|
proto.RegisterType((*RunRequest)(nil), "lumirpc.RunRequest")
|
|
|
|
|
proto.RegisterType((*RunResponse)(nil), "lumirpc.RunResponse")
|
|
|
|
|
proto.RegisterType((*NewResourceRequest)(nil), "lumirpc.NewResourceRequest")
|
|
|
|
|
proto.RegisterType((*NewResourceResponse)(nil), "lumirpc.NewResourceResponse")
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Reference imports to suppress errors if they are not otherwise used.
|
|
|
|
|
var _ context.Context
|
|
|
|
|
var _ grpc.ClientConn
|
|
|
|
|
|
|
|
|
|
// This is a compile-time assertion to ensure that this generated file
|
|
|
|
|
// is compatible with the grpc package it is being compiled against.
|
|
|
|
|
const _ = grpc.SupportPackageIsVersion4
|
|
|
|
|
|
|
|
|
|
// Client API for LanguageRuntime service
|
|
|
|
|
|
|
|
|
|
type LanguageRuntimeClient interface {
|
2017-08-30 03:24:12 +02:00
|
|
|
Run(ctx context.Context, in *RunRequest, opts ...grpc.CallOption) (*RunResponse, error)
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
type languageRuntimeClient struct {
|
|
|
|
|
cc *grpc.ClientConn
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func NewLanguageRuntimeClient(cc *grpc.ClientConn) LanguageRuntimeClient {
|
|
|
|
|
return &languageRuntimeClient{cc}
|
|
|
|
|
}
|
|
|
|
|
|
2017-08-30 03:24:12 +02:00
|
|
|
func (c *languageRuntimeClient) Run(ctx context.Context, in *RunRequest, opts ...grpc.CallOption) (*RunResponse, error) {
|
2017-08-28 18:18:44 +02:00
|
|
|
out := new(RunResponse)
|
2017-08-30 03:24:12 +02:00
|
|
|
err := grpc.Invoke(ctx, "/lumirpc.LanguageRuntime/Run", in, out, c.cc, 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
|
|
|
if err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
|
|
|
|
return out, nil
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Server API for LanguageRuntime service
|
|
|
|
|
|
|
|
|
|
type LanguageRuntimeServer interface {
|
2017-08-30 03:24:12 +02:00
|
|
|
Run(context.Context, *RunRequest) (*RunResponse, error)
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func RegisterLanguageRuntimeServer(s *grpc.Server, srv LanguageRuntimeServer) {
|
|
|
|
|
s.RegisterService(&_LanguageRuntime_serviceDesc, srv)
|
|
|
|
|
}
|
|
|
|
|
|
2017-08-30 03:24:12 +02:00
|
|
|
func _LanguageRuntime_Run_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
|
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
|
|
|
in := new(RunRequest)
|
|
|
|
|
if err := dec(in); err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
|
|
|
|
if interceptor == nil {
|
2017-08-30 03:24:12 +02:00
|
|
|
return srv.(LanguageRuntimeServer).Run(ctx, in)
|
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
|
|
|
}
|
|
|
|
|
info := &grpc.UnaryServerInfo{
|
|
|
|
|
Server: srv,
|
2017-08-30 03:24:12 +02:00
|
|
|
FullMethod: "/lumirpc.LanguageRuntime/Run",
|
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
|
|
|
}
|
|
|
|
|
handler := func(ctx context.Context, req interface{}) (interface{}, error) {
|
2017-08-30 03:24:12 +02:00
|
|
|
return srv.(LanguageRuntimeServer).Run(ctx, req.(*RunRequest))
|
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
|
|
|
}
|
|
|
|
|
return interceptor(ctx, in, info, handler)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
var _LanguageRuntime_serviceDesc = grpc.ServiceDesc{
|
|
|
|
|
ServiceName: "lumirpc.LanguageRuntime",
|
|
|
|
|
HandlerType: (*LanguageRuntimeServer)(nil),
|
|
|
|
|
Methods: []grpc.MethodDesc{
|
|
|
|
|
{
|
2017-08-30 03:24:12 +02:00
|
|
|
MethodName: "Run",
|
|
|
|
|
Handler: _LanguageRuntime_Run_Handler,
|
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
|
|
|
},
|
|
|
|
|
},
|
|
|
|
|
Streams: []grpc.StreamDesc{},
|
|
|
|
|
Metadata: "languages.proto",
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Client API for ResourceMonitor service
|
|
|
|
|
|
|
|
|
|
type ResourceMonitorClient interface {
|
|
|
|
|
NewResource(ctx context.Context, in *NewResourceRequest, opts ...grpc.CallOption) (*NewResourceResponse, error)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
type resourceMonitorClient struct {
|
|
|
|
|
cc *grpc.ClientConn
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func NewResourceMonitorClient(cc *grpc.ClientConn) ResourceMonitorClient {
|
|
|
|
|
return &resourceMonitorClient{cc}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func (c *resourceMonitorClient) NewResource(ctx context.Context, in *NewResourceRequest, opts ...grpc.CallOption) (*NewResourceResponse, error) {
|
|
|
|
|
out := new(NewResourceResponse)
|
|
|
|
|
err := grpc.Invoke(ctx, "/lumirpc.ResourceMonitor/NewResource", in, out, c.cc, opts...)
|
|
|
|
|
if err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
|
|
|
|
return out, nil
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Server API for ResourceMonitor service
|
|
|
|
|
|
|
|
|
|
type ResourceMonitorServer interface {
|
|
|
|
|
NewResource(context.Context, *NewResourceRequest) (*NewResourceResponse, error)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func RegisterResourceMonitorServer(s *grpc.Server, srv ResourceMonitorServer) {
|
|
|
|
|
s.RegisterService(&_ResourceMonitor_serviceDesc, srv)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func _ResourceMonitor_NewResource_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
|
|
|
|
|
in := new(NewResourceRequest)
|
|
|
|
|
if err := dec(in); err != nil {
|
|
|
|
|
return nil, err
|
|
|
|
|
}
|
|
|
|
|
if interceptor == nil {
|
|
|
|
|
return srv.(ResourceMonitorServer).NewResource(ctx, in)
|
|
|
|
|
}
|
|
|
|
|
info := &grpc.UnaryServerInfo{
|
|
|
|
|
Server: srv,
|
|
|
|
|
FullMethod: "/lumirpc.ResourceMonitor/NewResource",
|
|
|
|
|
}
|
|
|
|
|
handler := func(ctx context.Context, req interface{}) (interface{}, error) {
|
|
|
|
|
return srv.(ResourceMonitorServer).NewResource(ctx, req.(*NewResourceRequest))
|
|
|
|
|
}
|
|
|
|
|
return interceptor(ctx, in, info, handler)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
var _ResourceMonitor_serviceDesc = grpc.ServiceDesc{
|
|
|
|
|
ServiceName: "lumirpc.ResourceMonitor",
|
|
|
|
|
HandlerType: (*ResourceMonitorServer)(nil),
|
|
|
|
|
Methods: []grpc.MethodDesc{
|
|
|
|
|
{
|
|
|
|
|
MethodName: "NewResource",
|
|
|
|
|
Handler: _ResourceMonitor_NewResource_Handler,
|
|
|
|
|
},
|
|
|
|
|
},
|
|
|
|
|
Streams: []grpc.StreamDesc{},
|
|
|
|
|
Metadata: "languages.proto",
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
func init() { proto.RegisterFile("languages.proto", fileDescriptor2) }
|
|
|
|
|
|
|
|
|
|
var fileDescriptor2 = []byte{
|
2017-09-05 19:01:00 +02:00
|
|
|
// 397 bytes of a gzipped FileDescriptorProto
|
|
|
|
|
0x1f, 0x8b, 0x08, 0x00, 0x00, 0x09, 0x6e, 0x88, 0x02, 0xff, 0x94, 0x52, 0x4d, 0x6f, 0x95, 0x40,
|
|
|
|
|
0x14, 0x95, 0x8f, 0x52, 0x7b, 0x49, 0x7c, 0x66, 0xda, 0xe8, 0x04, 0x9b, 0x48, 0x70, 0xc3, 0x8a,
|
|
|
|
|
0x97, 0xe0, 0xc2, 0x8f, 0xad, 0xe9, 0xc6, 0xa8, 0x8b, 0x71, 0xed, 0x02, 0x78, 0xb7, 0x04, 0x0b,
|
|
|
|
|
0x33, 0x38, 0x1f, 0x36, 0xec, 0xfc, 0x99, 0xfe, 0x1c, 0xc3, 0x30, 0xa3, 0x6d, 0x7c, 0x9b, 0xee,
|
|
|
|
|
0xce, 0xb9, 0x9c, 0xcb, 0x39, 0xf7, 0x00, 0xec, 0xc6, 0x86, 0xf7, 0xa6, 0xe9, 0x51, 0x55, 0xb3,
|
|
|
|
|
0x14, 0x5a, 0x90, 0xd3, 0xd1, 0x4c, 0x83, 0x9c, 0xbb, 0xec, 0xb2, 0x17, 0xa2, 0x1f, 0x71, 0x6f,
|
|
|
|
|
0xc7, 0xad, 0xb9, 0xde, 0x2b, 0x2d, 0x4d, 0xa7, 0x37, 0x59, 0xf1, 0x3b, 0x00, 0x60, 0x86, 0x33,
|
|
|
|
|
0xfc, 0x61, 0x50, 0x69, 0xf2, 0x14, 0xa2, 0xf9, 0xf6, 0x40, 0x83, 0x3c, 0x28, 0xcf, 0xd8, 0x0a,
|
|
|
|
|
0x09, 0x85, 0xd3, 0x59, 0x8a, 0x5e, 0x36, 0x13, 0x0d, 0xed, 0xd4, 0x53, 0x42, 0x20, 0x6e, 0x64,
|
|
|
|
|
0xaf, 0x68, 0x94, 0x47, 0xe5, 0x19, 0xb3, 0x98, 0xbc, 0x81, 0xa4, 0x13, 0xfc, 0x7a, 0xe8, 0x69,
|
|
|
|
|
0x9c, 0x47, 0x65, 0x5a, 0xbf, 0xac, 0x5c, 0x8c, 0xea, 0x9f, 0x49, 0xf5, 0xc1, 0x2a, 0xae, 0xb8,
|
|
|
|
|
0x96, 0x0b, 0x73, 0x72, 0xf2, 0x0c, 0x92, 0x83, 0x5c, 0x98, 0xe1, 0xf4, 0x24, 0x0f, 0xca, 0xc7,
|
|
|
|
|
0xcc, 0xb1, 0xec, 0x1d, 0xa4, 0x77, 0xe4, 0x6b, 0xbe, 0x1b, 0x5c, 0x7c, 0xbe, 0x1b, 0x5c, 0xc8,
|
|
|
|
|
0x05, 0x9c, 0xfc, 0x6c, 0x46, 0x83, 0x2e, 0xdd, 0x46, 0xde, 0x87, 0x6f, 0x83, 0xe2, 0x15, 0xa4,
|
|
|
|
|
0xd6, 0x54, 0xcd, 0x82, 0x2b, 0x5c, 0x85, 0x28, 0xa5, 0x90, 0x6e, 0x79, 0x23, 0xc5, 0x04, 0xe4,
|
|
|
|
|
0x0b, 0xde, 0x32, 0x54, 0xc2, 0xc8, 0x0e, 0x7d, 0x0d, 0x04, 0x62, 0xbd, 0xcc, 0xe8, 0xa4, 0x16,
|
|
|
|
|
0xaf, 0x33, 0xde, 0x4c, 0xde, 0xc7, 0x62, 0xb2, 0x87, 0x44, 0xb4, 0xdf, 0xb1, 0xd3, 0x34, 0xca,
|
|
|
|
|
0x83, 0x32, 0xad, 0x9f, 0x57, 0x5b, 0xd9, 0x95, 0x2f, 0xbb, 0xfa, 0x6a, 0xcb, 0x66, 0x4e, 0x56,
|
|
|
|
|
0xfc, 0x0a, 0xe0, 0xfc, 0x9e, 0x9f, 0x0b, 0xf7, 0x04, 0xc2, 0xc1, 0xd7, 0x1e, 0x0e, 0x87, 0xf5,
|
|
|
|
|
0x4e, 0x23, 0xb9, 0xf3, 0x5a, 0xe1, 0x83, 0xad, 0xd6, 0x46, 0x95, 0x6e, 0xda, 0x11, 0x69, 0xbc,
|
|
|
|
|
0x35, 0xba, 0xb1, 0xfa, 0x0a, 0x76, 0x9f, 0xdc, 0xbf, 0xc2, 0x0c, 0xd7, 0xc3, 0x84, 0xa4, 0x86,
|
|
|
|
|
0x88, 0x19, 0x4e, 0xce, 0x8f, 0x7c, 0xac, 0xec, 0xe2, 0xfe, 0x70, 0xcb, 0x5b, 0x3c, 0xaa, 0xbf,
|
|
|
|
|
0xc1, 0xce, 0x5f, 0xf1, 0x59, 0xf0, 0x41, 0x0b, 0x49, 0x3e, 0x42, 0x7a, 0xe7, 0x36, 0xf2, 0xe2,
|
|
|
|
|
0xef, 0xe6, 0xff, 0x0d, 0x67, 0x97, 0xc7, 0x1f, 0xfa, 0xd7, 0xb7, 0x89, 0x3d, 0xeb, 0xf5, 0x9f,
|
|
|
|
|
0x00, 0x00, 0x00, 0xff, 0xff, 0xaf, 0x0e, 0x79, 0x3c, 0xd8, 0x02, 0x00, 0x00,
|
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
|
|
|
}
|