Refactor lsp.Server God Object: extract chat and completion subsystems

The Server type accumulated two large, tangled feature subsystems (in-editor
chat and code completion) alongside its core LSP dispatch role. Pull them into
cohesive types that own their state and logic while delegating shared
infrastructure back to Server via a back-reference.

- Add completionService (completion_service.go): owns the completion
  cache/throttle state (completionState) and all completion request-handling
  logic (handleCompletion, plan/jobs/execute, provider-native path,
  post-processing, message building, prefix heuristics). Methods moved off
  Server in handlers_completion.go to completionService.
- Add chatService (chat_service.go + chat_handlers.go): owns the input-activity
  clock (lastInput, with its own mutex instead of Server.mu) and all in-editor
  chat logic (detection, transcript history, message building, edit
  application, inline prompts) plus the slash-command handlers
  (chat_commands.go).
- Server now holds chat/completion fields, wires them in NewServer, and
  delegates (dispatch table, didChange, debounce gate) to them. Thin Server
  shims preserve the existing completion-state API for callers/tests.

Pure refactor: no behavior or LSP protocol changes. Tests adjusted only to
reach the relocated methods/state via the services. All tests pass with -race;
coverage 86.1%.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

1 files changed, 6 insertions, 6 deletions

diff --git a/internal/lsp/debounce_throttle_test.go b/internal/lsp/debounce_throttle_test.go
index 7efd439..d02bdb8 100644
--- a/internal/lsp/debounce_throttle_test.go
+++ b/internal/lsp/debounce_throttle_test.go
@@ -21,7 +21,7 @@ func (t *timeLLM) DefaultModel() string { return "m" }
 
 func TestCompletionDebounce_WaitsUntilQuiet(t *testing.T) {
 	s := newTestServer()
-	s.compCache = make(map[string]string)
+	s.completion.compCache = make(map[string]string)
 	cfg := s.cfg
 	cfg.TriggerCharacters = []string{".", ":", "/", "_"}
 	cfg.MaxTokens = 32
@@ -37,7 +37,7 @@ func TestCompletionDebounce_WaitsUntilQuiet(t *testing.T) {
 	p.Context = json.RawMessage([]byte(`{"triggerKind":1}`))
 
 	start := time.Now()
-	_, ok, _ := s.tryLLMCompletion(p, "", line, "", "", "", false, "")
+	_, ok, _ := s.completion.tryLLMCompletion(p, "", line, "", "", "", false, "")
 	if !ok {
 		t.Fatalf("expected ok=true")
 	}
@@ -51,7 +51,7 @@ func TestCompletionDebounce_WaitsUntilQuiet(t *testing.T) {
 
 func TestCompletionThrottle_SerializesCalls(t *testing.T) {
 	s := newTestServer()
-	s.compCache = make(map[string]string)
+	s.completion.compCache = make(map[string]string)
 	cfg := s.cfg
 	cfg.TriggerCharacters = []string{".", ":", "/", "_"}
 	cfg.MaxTokens = 32
@@ -65,7 +65,7 @@ func TestCompletionThrottle_SerializesCalls(t *testing.T) {
 	p := CompletionParams{Position: Position{Line: 0, Character: len(line)}, TextDocument: TextDocumentIdentifier{URI: "file://throttle.go"}}
 	p.Context = json.RawMessage([]byte(`{"triggerKind":1}`))
 	start := time.Now()
-	if _, ok, _ := s.tryLLMCompletion(p, "", line, "", "", "", false, ""); !ok {
+	if _, ok, _ := s.completion.tryLLMCompletion(p, "", line, "", "", "", false, ""); !ok {
 		t.Fatalf("first call expected ok=true")
 	}
 	if f1.t.IsZero() {
@@ -74,10 +74,10 @@ func TestCompletionThrottle_SerializesCalls(t *testing.T) {
 
 	// second call immediately after; should be delayed by ~interval.
 	// Clear cache to ensure we actually call the LLM again.
-	s.compCache = make(map[string]string)
+	s.completion.compCache = make(map[string]string)
 	f2 := &timeLLM{}
 	s.llmClient = f2
-	if _, ok, _ := s.tryLLMCompletion(p, "", line, "", "", "", false, ""); !ok {
+	if _, ok, _ := s.completion.tryLLMCompletion(p, "", line, "", "", "", false, ""); !ok {
 		t.Fatalf("second call expected ok=true")
 	}
 	if f2.t.IsZero() {