phraser: the reminder summary is cut in runes, and silence is an error (V-620)

Three defects in internal/phraser, all of the shape "reports done when
nothing happened".

The reminder summary was cut in bytes: `len(s) > 60` and `s[:57]`, in two
copies (Stub.PhraseReminder and LLMPhraser.PhraseReminder). On Russian a
letter is two bytes, so the cut fell at about 28 letters instead of 60 and
landed inside a letter about half the time. Sendable.Summary is what
voicesink hands to piper and what the telegram sink posts, so the half rune
was spoken and sent. One rune-counting helper now, shared by both. The test
that covered this was ASCII, which is what let the arithmetic stand.

The evidence branch of PhraseQuery and the bare-prose tail of PhraseChat
both returned ("", nil) when the server answered and the model wrote no
tokens. The knowledge branch has guarded that with errEmptyResponse since it
was written; these two did not. The daemon's callers check for the empty
string and paper over it, so the visible cost was the eval, which scored a
silent model as bad phrasing rather than as a failure, and a log line that
never appeared.

Stub.PhraseReminder set no Mood. Its sibling PhraseNudge sets "neutral" and
says in a comment why: the Stub is a production fallback and owes the output
contract a value. The zero value is not one of the five moods.

No prompt and no spoken wording changed, so the phrasing eval is unmoved.

internal/phraser/llmphraser.go

@@ -299,6 +299,16 @@ func (p *LLMPhraser) PhraseQuery(ctx context.Context, utterance string, notes []
 	if text != "" {
 		return text, nil
 	}
+	if raw == "" {
+		// Same guard the knowledge branch above has had since it was written,
+		// and this branch did not: the server answered and the model wrote
+		// nothing, which returned ("", nil) — an empty answer reported as a
+		// successful phrasing. The daemon's callers happen to check for the
+		// empty string, so it read as a silent fallback there; the eval scored
+		// it as bad phrasing rather than as the failure it is, and nothing on
+		// either path logged that the model had produced no tokens.
+		return SourcesFallback(strings.Join(notes, "; ")), errEmptyResponse
+	}
 	return raw, nil
 }
 
@@ -375,7 +385,13 @@ func (p *LLMPhraser) PhraseChat(ctx context.Context, utterance string, history [
 	if i := strings.IndexByte(resp, '\n'); i >= 0 {
 		resp = resp[:i]
 	}
-	return strings.TrimSpace(resp), nil
+	if resp = strings.TrimSpace(resp); resp == "" {
+		// The model was up and wrote nothing. Same rule as PhraseQuery: the
+		// fallback keeps the turn alive and the failure stays visible, rather
+		// than ("", nil) telling the caller the chat path succeeded.
+		return ChatFallback(), fmt.Errorf("phrase chat: %w", errEmptyResponse)
+	}
+	return resp, nil
 }
 
 func (p *LLMPhraser) PhraseReminder(ctx context.Context, d loop.ReminderDecision) (delivery.PhrasedReminder, error) {
@@ -414,11 +430,9 @@ func (p *LLMPhraser) PhraseReminder(ctx context.Context, d loop.ReminderDecision
 	if mood == "" {
 		mood = "neutral"
 	}
-	summary := body
-	if len(summary) > 60 {
-		summary = summary[:57] + "..."
-	}
-	return delivery.PhrasedReminder{Decision: d, Body: body, Summary: summary, Mood: mood}, nil
+	return delivery.PhrasedReminder{
+		Decision: d, Body: body, Summary: reminderSummary(body), Mood: mood,
+	}, nil
 }
 
 func (p *LLMPhraser) chat(ctx context.Context, userPrompt string) (string, error) {