Telegram formatting

[dzarlax]

Note

Medium Risk
Moderate risk because it changes core Telegram response delivery logic (HTML conversion, splitting, fallbacks) and could affect message ordering/formatting or introduce truncation edge cases for long outputs.

Overview
Improves Telegram output formatting by converting Markdown checklist items in bullet lists into Unicode checkboxes (☐/☑) and adding tests for the new rendering.

Reworks sendResponse to prefer HTML messages, then split and send long responses in multiple HTML/plain chunks (with retry fallbacks), only falling back to sending a .md document when chunked delivery fails. Adds splitMessage (paragraph/line-aware) plus unit tests to ensure chunks respect Telegram’s 4096-byte limit.

^{Written by Cursor Bugbot for commit 3da2235. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Autofix prepared a fix for the issue found in the latest run.

• ✅ Fixed: Hard break in splitMessage may split multi-byte UTF-8 characters
  • Added UTF-8 boundary check that backs up from maxLen until finding a valid start byte before splitting, ensuring all chunks contain valid UTF-8.

Or push these changes by commenting:

@cursor push 8ee6ffb5b0

Preview (8ee6ffb5b0)

diff --git a/internal/telegram/handler.go b/internal/telegram/handler.go
--- a/internal/telegram/handler.go
+++ b/internal/telegram/handler.go
@@ -12,6 +12,7 @@
 	"strings"
 	"sync"
 	"time"
+	"unicode/utf8"
 
 	tgbotapi "github.com/go-telegram-bot-api/telegram-bot-api/v5"
 
@@ -849,9 +850,16 @@
 			continue
 		}
 
-		// Hard break at maxLen.
-		parts = append(parts, chunk)
-		text = text[maxLen:]
+		// Hard break at maxLen, but respect UTF-8 character boundaries.
+		breakAt := maxLen
+		for breakAt > 0 && !utf8.RuneStart(text[breakAt]) {
+			breakAt--
+		}
+		if breakAt == 0 {
+			breakAt = maxLen // Avoid empty chunk; shouldn't happen with valid UTF-8.
+		}
+		parts = append(parts, text[:breakAt])
+		text = text[breakAt:]
 	}
 	return parts
 }

diff --git a/internal/telegram/markdown_test.go b/internal/telegram/markdown_test.go
--- a/internal/telegram/markdown_test.go
+++ b/internal/telegram/markdown_test.go
@@ -2,6 +2,7 @@
 
 import (
 	"testing"
+	"unicode/utf8"
 )
 
 func TestMarkdown_PlainText(t *testing.T) {
@@ -205,6 +206,29 @@
 	}
 }
 
+func TestSplitMessage_UTF8Boundary(t *testing.T) {
+	// "日本語" is 9 bytes (3 chars × 3 bytes each). No newlines, so hard break is needed.
+	// Create a string that requires hard break in the middle of a multi-byte character.
+	text := "日本語日本語日本語" // 27 bytes, 9 characters
+	parts := splitMessage(text, 10)
+	if len(parts) < 2 {
+		t.Errorf("expected multiple parts, got %d: %v", len(parts), parts)
+	}
+	for i, p := range parts {
+		if !utf8.ValidString(p) {
+			t.Errorf("part %d is invalid UTF-8: %q", i, p)
+		}
+	}
+	// Verify all parts joined equal the original
+	joined := ""
+	for _, p := range parts {
+		joined += p
+	}
+	if joined != text {
+		t.Errorf("parts don't reassemble to original:\n got: %q\nwant: %q", joined, text)
+	}
+}
+
 // --- helpers ---
 
 func assertMD(t *testing.T, input, want string) {

_{This Bugbot Autofix run was free. To enable autofix for future PRs, go to the Cursor dashboard.}

[cursor]

Hard break in splitMessage may split multi-byte UTF-8 characters

Low Severity

The hard-break fallback in splitMessage slices at a raw byte offset (text[:maxLen]), which can split a multi-byte UTF-8 character in half. This produces invalid UTF-8 in the resulting chunk, potentially causing garbled text in the Telegram message or an API rejection. The test suite already uses Cyrillic text, indicating multi-byte input is expected. Long paragraphs without newlines in CJK or other non-ASCII languages could trigger this path.