1
0
mirror of https://github.com/golang/go synced 2024-10-01 16:18:32 -06:00
go/internal/lsp/fake/editor.go

660 lines
18 KiB
Go
Raw Normal View History

internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
// Copyright 2020 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package fake
import (
"bufio"
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
"context"
"errors"
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
"fmt"
"regexp"
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
"strings"
"sync"
"golang.org/x/tools/internal/jsonrpc2"
"golang.org/x/tools/internal/lsp/protocol"
)
// Editor is a fake editor client. It keeps track of client state and can be
// used for writing LSP tests.
type Editor struct {
Config EditorConfig
// server, client, and sandbox are concurrency safe and written only
// at construction time, so do not require synchronization.
server protocol.Server
client *Client
sandbox *Sandbox
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
// Since this editor is intended just for testing, we use very coarse
// locking.
mu sync.Mutex
// Editor state.
buffers map[string]buffer
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
// Capabilities / Options
serverCapabilities protocol.ServerCapabilities
}
type buffer struct {
version int
path string
content []string
}
func (b buffer) text() string {
return strings.Join(b.content, "\n")
}
// EditorConfig configures the editor's LSP session. This is similar to
// source.UserOptions, but we use a separate type here so that we expose only
// that configuration which we support.
//
// The zero value for EditorConfig should correspond to its defaults.
type EditorConfig struct {
Env []string
// CodeLens is a map defining whether codelens are enabled, keyed by the
// codeLens command. CodeLens which are not present in this map are left in
// their default state.
CodeLens map[string]bool
}
// NewEditor Creates a new Editor.
func NewEditor(ws *Sandbox, config EditorConfig) *Editor {
return &Editor{
buffers: make(map[string]buffer),
sandbox: ws,
Config: config,
}
}
// Connect configures the editor to communicate with an LSP server on conn. It
// is not concurrency safe, and should be called at most once, before using the
// editor.
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
//
// It returns the editor, so that it may be called as follows:
// editor, err := NewEditor(s).Connect(ctx, conn)
func (e *Editor) Connect(ctx context.Context, conn *jsonrpc2.Conn, hooks ClientHooks) (*Editor, error) {
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
e.server = protocol.ServerDispatcher(conn)
e.client = &Client{editor: e, hooks: hooks}
go conn.Run(ctx,
protocol.Handlers(
protocol.ClientHandler(e.client,
jsonrpc2.MethodNotFound)))
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
if err := e.initialize(ctx); err != nil {
return nil, err
}
e.sandbox.Workdir.AddWatcher(e.onFileChanges)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
return e, nil
}
// Shutdown issues the 'shutdown' LSP notification.
func (e *Editor) Shutdown(ctx context.Context) error {
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
if e.server != nil {
if err := e.server.Shutdown(ctx); err != nil {
return fmt.Errorf("Shutdown: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
return nil
}
// Exit issues the 'exit' LSP notification.
func (e *Editor) Exit(ctx context.Context) error {
if e.server != nil {
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
// Not all LSP clients issue the exit RPC, but we do so here to ensure that
// we gracefully handle it on multi-session servers.
if err := e.server.Exit(ctx); err != nil {
return fmt.Errorf("Exit: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
return nil
}
// Client returns the LSP client for this editor.
func (e *Editor) Client() *Client {
return e.client
}
func (e *Editor) configuration() map[string]interface{} {
config := map[string]interface{}{
"verboseWorkDoneProgress": true,
}
envvars := e.sandbox.GoEnv()
envvars = append(envvars, e.Config.Env...)
env := map[string]interface{}{}
for _, value := range envvars {
kv := strings.SplitN(value, "=", 2)
env[kv[0]] = kv[1]
}
config["env"] = env
if e.Config.CodeLens != nil {
config["codelens"] = e.Config.CodeLens
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
return config
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
func (e *Editor) initialize(ctx context.Context) error {
params := &protocol.ParamInitialize{}
params.ClientInfo.Name = "fakeclient"
params.ClientInfo.Version = "v1.0.0"
params.RootURI = e.sandbox.Workdir.RootURI()
params.Capabilities.Workspace.Configuration = true
params.Capabilities.Window.WorkDoneProgress = true
// TODO: set client capabilities
params.InitializationOptions = e.configuration()
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
params.Trace = "messages"
// TODO: support workspace folders.
if e.server != nil {
resp, err := e.server.Initialize(ctx, params)
if err != nil {
return fmt.Errorf("initialize: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
e.mu.Lock()
e.serverCapabilities = resp.Capabilities
e.mu.Unlock()
if err := e.server.Initialized(ctx, &protocol.InitializedParams{}); err != nil {
return fmt.Errorf("initialized: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
// TODO: await initial configuration here, or expect gopls to manage that?
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
return nil
}
func (e *Editor) onFileChanges(ctx context.Context, evts []FileEvent) {
if e.server == nil {
return
}
var lspevts []protocol.FileEvent
for _, evt := range evts {
lspevts = append(lspevts, evt.ProtocolEvent)
}
e.server.DidChangeWatchedFiles(ctx, &protocol.DidChangeWatchedFilesParams{
Changes: lspevts,
})
}
// OpenFile creates a buffer for the given workdir-relative file.
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
func (e *Editor) OpenFile(ctx context.Context, path string) error {
content, err := e.sandbox.Workdir.ReadFile(path)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
if err != nil {
return err
}
buf := newBuffer(path, content)
e.mu.Lock()
e.buffers[path] = buf
item := textDocumentItem(e.sandbox.Workdir, buf)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
e.mu.Unlock()
if e.server != nil {
if err := e.server.DidOpen(ctx, &protocol.DidOpenTextDocumentParams{
TextDocument: item,
}); err != nil {
return fmt.Errorf("DidOpen: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
return nil
}
func newBuffer(path, content string) buffer {
return buffer{
version: 1,
path: path,
content: strings.Split(content, "\n"),
}
}
func textDocumentItem(wd *Workdir, buf buffer) protocol.TextDocumentItem {
uri := wd.URI(buf.path)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
languageID := ""
if strings.HasSuffix(buf.path, ".go") {
// TODO: what about go.mod files? What is their language ID?
languageID = "go"
}
return protocol.TextDocumentItem{
URI: uri,
LanguageID: languageID,
Version: float64(buf.version),
Text: buf.text(),
}
}
// CreateBuffer creates a new unsaved buffer corresponding to the workdir path,
// containing the given textual content.
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
func (e *Editor) CreateBuffer(ctx context.Context, path, content string) error {
buf := newBuffer(path, content)
e.mu.Lock()
e.buffers[path] = buf
item := textDocumentItem(e.sandbox.Workdir, buf)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
e.mu.Unlock()
if e.server != nil {
if err := e.server.DidOpen(ctx, &protocol.DidOpenTextDocumentParams{
TextDocument: item,
}); err != nil {
return fmt.Errorf("DidOpen: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
return nil
}
// CloseBuffer removes the current buffer (regardless of whether it is saved).
func (e *Editor) CloseBuffer(ctx context.Context, path string) error {
e.mu.Lock()
_, ok := e.buffers[path]
if !ok {
e.mu.Unlock()
return ErrUnknownBuffer
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
delete(e.buffers, path)
e.mu.Unlock()
if e.server != nil {
if err := e.server.DidClose(ctx, &protocol.DidCloseTextDocumentParams{
TextDocument: e.textDocumentIdentifier(path),
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}); err != nil {
return fmt.Errorf("DidClose: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
return nil
}
func (e *Editor) textDocumentIdentifier(path string) protocol.TextDocumentIdentifier {
return protocol.TextDocumentIdentifier{
URI: e.sandbox.Workdir.URI(path),
}
}
// SaveBuffer writes the content of the buffer specified by the given path to
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
// the filesystem.
func (e *Editor) SaveBuffer(ctx context.Context, path string) error {
if err := e.OrganizeImports(ctx, path); err != nil {
return fmt.Errorf("organizing imports before save: %w", err)
}
if err := e.FormatBuffer(ctx, path); err != nil {
return fmt.Errorf("formatting before save: %w", err)
}
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
e.mu.Lock()
buf, ok := e.buffers[path]
if !ok {
e.mu.Unlock()
return fmt.Errorf(fmt.Sprintf("unknown buffer: %q", path))
}
content := buf.text()
includeText := false
syncOptions, ok := e.serverCapabilities.TextDocumentSync.(protocol.TextDocumentSyncOptions)
if ok {
includeText = syncOptions.Save.IncludeText
}
e.mu.Unlock()
docID := e.textDocumentIdentifier(buf.path)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
if e.server != nil {
if err := e.server.WillSave(ctx, &protocol.WillSaveTextDocumentParams{
TextDocument: docID,
Reason: protocol.Manual,
}); err != nil {
return fmt.Errorf("WillSave: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
if err := e.sandbox.Workdir.WriteFile(ctx, path, content); err != nil {
return fmt.Errorf("writing %q: %w", path, err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
if e.server != nil {
params := &protocol.DidSaveTextDocumentParams{
TextDocument: protocol.VersionedTextDocumentIdentifier{
Version: float64(buf.version),
TextDocumentIdentifier: docID,
},
}
if includeText {
params.Text = &content
}
if err := e.server.DidSave(ctx, params); err != nil {
return fmt.Errorf("DidSave: %w", err)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
}
return nil
}
// contentPosition returns the (Line, Column) position corresponding to offset
// in the buffer referenced by path.
func contentPosition(content string, offset int) (Pos, error) {
scanner := bufio.NewScanner(strings.NewReader(content))
start := 0
line := 0
for scanner.Scan() {
end := start + len([]rune(scanner.Text())) + 1
if offset < end {
return Pos{Line: line, Column: offset - start}, nil
}
start = end
line++
}
if err := scanner.Err(); err != nil {
return Pos{}, fmt.Errorf("scanning content: %w", err)
}
// Scan() will drop the last line if it is empty. Correct for this.
if strings.HasSuffix(content, "\n") && offset == start {
return Pos{Line: line, Column: 0}, nil
}
return Pos{}, fmt.Errorf("position %d out of bounds in %q (line = %d, start = %d)", offset, content, line, start)
}
// ErrNoMatch is returned if a regexp search fails.
var (
ErrNoMatch = errors.New("no match")
ErrUnknownBuffer = errors.New("unknown buffer")
)
// regexpRange returns the start and end of the first occurrence of either re
// or its singular subgroup. It returns ErrNoMatch if the regexp doesn't match.
func regexpRange(content, re string) (Pos, Pos, error) {
var start, end int
rec, err := regexp.Compile(re)
if err != nil {
return Pos{}, Pos{}, err
}
indexes := rec.FindStringSubmatchIndex(content)
if indexes == nil {
return Pos{}, Pos{}, ErrNoMatch
}
switch len(indexes) {
case 2:
// no subgroups: return the range of the regexp expression
start, end = indexes[0], indexes[1]
case 4:
// one subgroup: return its range
start, end = indexes[2], indexes[3]
default:
return Pos{}, Pos{}, fmt.Errorf("invalid search regexp %q: expect either 0 or 1 subgroups, got %d", re, len(indexes)/2-1)
}
startPos, err := contentPosition(content, start)
if err != nil {
return Pos{}, Pos{}, err
}
endPos, err := contentPosition(content, end)
if err != nil {
return Pos{}, Pos{}, err
}
return startPos, endPos, nil
}
// RegexpSearch returns the position of the first match for re in the buffer
// bufName. For convenience, RegexpSearch supports the following two modes:
// 1. If re has no subgroups, return the position of the match for re itself.
// 2. If re has one subgroup, return the position of the first subgroup.
// It returns an error re is invalid, has more than one subgroup, or doesn't
// match the buffer.
func (e *Editor) RegexpSearch(bufName, re string) (Pos, error) {
e.mu.Lock()
defer e.mu.Unlock()
buf, ok := e.buffers[bufName]
if !ok {
return Pos{}, ErrUnknownBuffer
}
start, _, err := regexpRange(buf.text(), re)
return start, err
}
// RegexpReplace edits the buffer corresponding to path by replacing the first
// instance of re, or its first subgroup, with the replace text. See
// RegexpSearch for more explanation of these two modes.
// It returns an error if re is invalid, has more than one subgroup, or doesn't
// match the buffer.
func (e *Editor) RegexpReplace(ctx context.Context, path, re, replace string) error {
e.mu.Lock()
defer e.mu.Unlock()
buf, ok := e.buffers[path]
if !ok {
return ErrUnknownBuffer
}
content := buf.text()
start, end, err := regexpRange(content, re)
if err != nil {
return err
}
return e.editBufferLocked(ctx, path, []Edit{{
Start: start,
End: end,
Text: replace,
}})
}
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
// EditBuffer applies the given test edits to the buffer identified by path.
func (e *Editor) EditBuffer(ctx context.Context, path string, edits []Edit) error {
e.mu.Lock()
defer e.mu.Unlock()
return e.editBufferLocked(ctx, path, edits)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
// BufferText returns the content of the buffer with the given name.
func (e *Editor) BufferText(name string) string {
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
e.mu.Lock()
defer e.mu.Unlock()
return e.buffers[name].text()
}
internal/lsp/regtest: add functions to make diagnostic assertions easier One of the tricky things about asserting on conditions in regtests is the asynchronous nature of LSP. For example, as the LSP client we cannot be sure when we've received all diagnostics for a given file. Currently, regtests are implemented by awaiting specific diagnostic expectations. This means that if gopls generates diagnostics that do not match those expectations, we can only time out the test. Ideally, we would want to know that gopls is done generating all diagnostics for the current file state. This is not possible without knowing the status of diagnostics for. Barring this, we would want to know that diagnostics are done for the current file version. Unfortunately, that also is not possible, because a new version of file B can affect diagnostics in file A. So in lieu of this information, this CL exposes a few tools that can be used to improve the experience of writing new regtests. - A new expectation is added: AnyDiagnosticAtCurrentVersion, that is satisfied if any diagnostics have been received for the current buffer version. - ExpectDiagnostics is added to Env, to help check whether the current diagnostics matches expectations. Updates golang/go#38113 Change-Id: I48d2c3db87c13ac3ab424d01d9444cbc285af9e1 Reviewed-on: https://go-review.googlesource.com/c/tools/+/226842 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-04-01 12:56:48 -06:00
// BufferVersion returns the current version of the buffer corresponding to
// name (or 0 if it is not being edited).
func (e *Editor) BufferVersion(name string) int {
e.mu.Lock()
defer e.mu.Unlock()
return e.buffers[name].version
}
func (e *Editor) editBufferLocked(ctx context.Context, path string, edits []Edit) error {
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
buf, ok := e.buffers[path]
if !ok {
return fmt.Errorf("unknown buffer %q", path)
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
var (
content = make([]string, len(buf.content))
err error
evts []protocol.TextDocumentContentChangeEvent
)
copy(content, buf.content)
content, err = editContent(content, edits)
if err != nil {
return err
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
buf.content = content
buf.version++
e.buffers[path] = buf
// A simple heuristic: if there is only one edit, send it incrementally.
// Otherwise, send the entire content.
if len(edits) == 1 {
evts = append(evts, edits[0].toProtocolChangeEvent())
} else {
evts = append(evts, protocol.TextDocumentContentChangeEvent{
Text: buf.text(),
})
}
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
params := &protocol.DidChangeTextDocumentParams{
TextDocument: protocol.VersionedTextDocumentIdentifier{
Version: float64(buf.version),
TextDocumentIdentifier: e.textDocumentIdentifier(buf.path),
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
},
ContentChanges: evts,
}
if e.server != nil {
if err := e.server.DidChange(ctx, params); err != nil {
return fmt.Errorf("DidChange: %w", err)
}
}
return nil
internal/lsp/fake: add fakes for testing editor interaction A lot of bug reports originating from LSP clients are related to either the timing or sequence of editor interactions with gopls (or at least they're originally reported this way). For example: "when I open a package and then create a new file, I lose diagnostics for existing files". These conditions are often hard to reproduce, and to isolate as either a gopls bug or a bug in the editor. Right now we're relying on govim integration tests to catch these regressions, but it's important to also have a testing framework that can exercise this functionality in-process. As a starting point this CL adds test fakes that implement a high level API for scripting editor interactions. A fake workspace can be used to sandbox file operations; a fake editor provides an interface for text editing operations; a fake LSP client can be used to connect the fake editor to a gopls instance. Some tests are added to the lsprpc package to demonstrate the API. The primary goal of these fakes should be to simulate an client that complies to the LSP spec. Put another way: if we have a bug report that we can't reproduce with our regression tests, it should either be a bug in our test fakes or a bug in the LSP client originating the report. I did my best to comply with the spec in this implementation, but it will certainly develop as we write more tests. We will also need to add to the editor API in the future for testing more language features. Updates golang/go#36879 Updates golang/go#34111 Change-Id: Ib81188683a7066184b8a254275ed5525191a2d68 Reviewed-on: https://go-review.googlesource.com/c/tools/+/217598 Run-TryBot: Robert Findley <rfindley@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rebecca Stambler <rstambler@golang.org>
2020-02-02 10:53:30 -07:00
}
// GoToDefinition jumps to the definition of the symbol at the given position
// in an open buffer.
func (e *Editor) GoToDefinition(ctx context.Context, path string, pos Pos) (string, Pos, error) {
if err := e.checkBufferPosition(path, pos); err != nil {
return "", Pos{}, err
}
params := &protocol.DefinitionParams{}
params.TextDocument.URI = e.sandbox.Workdir.URI(path)
params.Position = pos.toProtocolPosition()
resp, err := e.server.Definition(ctx, params)
if err != nil {
return "", Pos{}, fmt.Errorf("definition: %w", err)
}
if len(resp) == 0 {
return "", Pos{}, nil
}
newPath := e.sandbox.Workdir.URIToPath(resp[0].URI)
newPos := fromProtocolPosition(resp[0].Range.Start)
if err := e.OpenFile(ctx, newPath); err != nil {
return "", Pos{}, fmt.Errorf("OpenFile: %w", err)
}
return newPath, newPos, nil
}
// OrganizeImports requests and performs the source.organizeImports codeAction.
func (e *Editor) OrganizeImports(ctx context.Context, path string) error {
return e.codeAction(ctx, path, nil, protocol.SourceOrganizeImports)
}
// ApplyQuickFixes requests and performs the quickfix codeAction.
func (e *Editor) ApplyQuickFixes(ctx context.Context, path string, diagnostics []protocol.Diagnostic) error {
return e.codeAction(ctx, path, diagnostics, protocol.QuickFix, protocol.SourceFixAll)
}
func (e *Editor) codeAction(ctx context.Context, path string, diagnostics []protocol.Diagnostic, only ...protocol.CodeActionKind) error {
if e.server == nil {
return nil
}
params := &protocol.CodeActionParams{}
params.TextDocument.URI = e.sandbox.Workdir.URI(path)
params.Context.Only = only
if diagnostics != nil {
params.Context.Diagnostics = diagnostics
}
actions, err := e.server.CodeAction(ctx, params)
if err != nil {
return fmt.Errorf("textDocument/codeAction: %w", err)
}
e.mu.Lock()
defer e.mu.Unlock()
for _, action := range actions {
var match bool
for _, o := range only {
if action.Kind == o {
match = true
break
}
}
if !match {
continue
}
for _, change := range action.Edit.DocumentChanges {
path := e.sandbox.Workdir.URIToPath(change.TextDocument.URI)
if float64(e.buffers[path].version) != change.TextDocument.Version {
// Skip edits for old versions.
continue
}
edits := convertEdits(change.Edits)
if err := e.editBufferLocked(ctx, path, edits); err != nil {
return fmt.Errorf("editing buffer %q: %w", path, err)
}
}
}
return nil
}
func convertEdits(protocolEdits []protocol.TextEdit) []Edit {
var edits []Edit
for _, lspEdit := range protocolEdits {
edits = append(edits, fromProtocolTextEdit(lspEdit))
}
return edits
}
// FormatBuffer gofmts a Go file.
func (e *Editor) FormatBuffer(ctx context.Context, path string) error {
if e.server == nil {
return nil
}
e.mu.Lock()
version := e.buffers[path].version
e.mu.Unlock()
params := &protocol.DocumentFormattingParams{}
params.TextDocument.URI = e.sandbox.Workdir.URI(path)
resp, err := e.server.Formatting(ctx, params)
if err != nil {
return fmt.Errorf("textDocument/formatting: %w", err)
}
e.mu.Lock()
defer e.mu.Unlock()
if versionAfter := e.buffers[path].version; versionAfter != version {
return fmt.Errorf("before receipt of formatting edits, buffer version changed from %d to %d", version, versionAfter)
}
edits := convertEdits(resp)
return e.editBufferLocked(ctx, path, edits)
}
func (e *Editor) checkBufferPosition(path string, pos Pos) error {
e.mu.Lock()
defer e.mu.Unlock()
buf, ok := e.buffers[path]
if !ok {
return fmt.Errorf("buffer %q is not open", path)
}
if !inText(pos, buf.content) {
return fmt.Errorf("position %v is invalid in buffer %q", pos, path)
}
return nil
}
// RunGenerate runs `go generate` non-recursively in the workdir-relative dir
// path. It does not report any resulting file changes as a watched file
// change, so must be followed by a call to Workdir.CheckForFileChanges once
// the generate command has completed.
func (e *Editor) RunGenerate(ctx context.Context, dir string) error {
if e.server == nil {
return nil
}
absDir := e.sandbox.Workdir.filePath(dir)
params := &protocol.ExecuteCommandParams{
Command: "generate",
Arguments: []interface{}{absDir, false},
}
if _, err := e.server.ExecuteCommand(ctx, params); err != nil {
return fmt.Errorf("running generate: %v", err)
}
// Unfortunately we can't simply poll the workdir for file changes here,
// because server-side command may not have completed. In regtests, we can
// Await this state change, but here we must delegate that responsibility to
// the caller.
return nil
}
// CodeLens execute a codelens request on the server.
func (e *Editor) CodeLens(ctx context.Context, path string) ([]protocol.CodeLens, error) {
if e.server == nil {
return nil, nil
}
e.mu.Lock()
_, ok := e.buffers[path]
e.mu.Unlock()
if !ok {
return nil, fmt.Errorf("buffer %q is not open", path)
}
params := &protocol.CodeLensParams{
TextDocument: e.textDocumentIdentifier(path),
}
lens, err := e.server.CodeLens(ctx, params)
if err != nil {
return nil, err
}
return lens, nil
}